payment policymedicare...march 2003 report to the congress medicare payment policy 601 new jersey...

REPORT TO THE CONGRESS

MedicarePayment Policy

M A R C H 2 0 0 3

601 New Jersey Avenue, NW • Suite 9000 • Washington, DC 20001(202) 220-3700 • Fax: (202) 220-3759 • www.medpac.gov

Re

po

rt to

the

Co

ng

re

ss

: Me

dic

ar

e P

ay

me

nt P

olic

y

|M

ar

ch

20

03

The Medicare Payment Advisory Commission (MedPAC) is an independent federal body

established by the Balanced Budget Act of 1997 (P.L. 105–33) to advise the U.S. Congress

on issues affecting the Medicare program. In addition to advising the Congress on payments

to health plans participating in the Medicare�Choice program and providers in Medicare’s

traditional fee-for-service program, MedPAC is also tasked with analyzing access to care,

quality of care, and other issues affecting Medicare.

The Commission’s 17 members bring diverse expertise in the financing and delivery of

health care services. Commissioners are appointed to three-year terms (subject to renewal) by

the Comptroller General and serve part time. Appointments are staggered; the terms of five

or six Commissioners expire each year. The Commission is supported by an executive

director and a staff of analysts, who typically have backgrounds in economics, health policy,

public health, or medicine.

MedPAC meets publicly to discuss policy issues and formulate its recommendations to the

Congress. In the course of these meetings, Commissioners consider the results of staff

research, presentations by policy experts, and comments from interested parties. (Meeting

transcripts are available at www.medpac.gov.) Commission members and staff also seek

input on Medicare issues through frequent meetings with individuals interested in the

program, including staff from congressional committees and the Centers for Medicare &

Medicaid Services (CMS), health care researchers, health care providers, and beneficiary

advocates.

Two reports—issued in March and June each year—are the primary outlet for Commission

recommendations. This volume fulfills MedPAC’s requirement to submit an annual report on

Medicare payment policy. In addition to annual reports and occasional reports on subjects

requested by the Congress, MedPAC advises the Congress through other avenues, including

comments on reports and proposed regulations issued by the Secretary of the Department of

Health and Human Services, testimony, and briefings for congressional staff.

M A R C H 2 0 0 3




•

•

This report was prepared with the assistance of many people. Their support was key asthe Commission considered policy issues and worked toward consensus on itsrecommendations.

Despite a heavy workload, staff members of the Centers for Medicare & MedicaidServices were particularly helpful during preparation of the report. We thank: BradyAugustine, Stuart Barranco, Bob Cereghino, Christy Cornell, Jack Fletcher, MarkFreeland, Tom Gustafson, Marc Hartstein, Stephen Heffler, Amy Heller, Russell Hendel,Kimberlie Jasmin, Mary Kapp, John Klemm, Kim Kotova, Nora Kraemer, SheilaLambowitz, Barry Levi, Susan Levy, Shannon Martin, Ann Meadow, Renee Mentnech,Parasher Patel, Cindy Read, Chester Robinson, Joan Sanow, John Shatto, Darryl Simms,Michelle Slowikowski, Tom Talbott, Sean Tunis, Meredith Walz, Carter Warfield, LauraWarfield, Laurence Wilson, Dave Zanardelli, and Carlos Zarabozo.

The Commission also received valuable insights and assistance from others—ingovernment, industry, and the research community—who generously offered their timeand knowledge. They include: Bob Berenson, Amy Bernstein, Cindy Brown, JennyBryant, Kathy Bryant, Catharine Burt, Nathan Childs, Geri Dallek, William Dombi,Joyce DuBow, Lois Ehle, Randolph Fenninger, David Gifford, Kurt Gillis, GeorgeGreenberg, Christopher Hogan, Ellen Kramarow, Jo Ann Lamphere, Kevin McAnaney,Sharon McIlrath, Marianne Miller, Penny Mohr, Margaret Nowak, Julia Paradise, DaePark, Mark Peavy, Susan Polniaszek, Ellen Pryga, Michael Romansky, Sarah Silberstein,Libby Solon, Elise Smith, Kathleen Smith, Elaine Swift, Ashley Thompson, ChrisTopoleski, Sally Trude, Amanda Twiss, Bob Wardwell, Eric Wedum, Robin Weinick,Chapin White, and Eric Zimmerman.

Once again, the programmers at Social and Scientific Systems provided highly capableassistance to Commission staff. In particular, we appreciate the hard work of: NancyAllen, Valerie Aschenbach, Tom Bell, Christine Brady, Danita Byrd-Holt, DakshaDamera, Deborah Johnson, John May, Jeff McCartney, Carl Newman, Margie Odle,Anthony Pridgen, Charles Thomson, and Arlene Turner.

Finally, the Commission wishes to thank Carol Adams, Kerry Kemp, Cynthia Pratt, andRebecca Weber for their help editing and producing this report. �

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 v

Acknowledgments

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v

Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi

Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xv

Executive summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xix

Chapters

1 Context for Medicare spending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3

Medicare spending trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4

Medicare spending compared with other indicators of health spending . . . . . . . . .7

Implications of Medicare spending given limited resources . . . . . . . . . . . . . . . .13

Spending and other implications of MedPAC’s recommendations . . . . . . . . . . .17

2 Assessing payment adequacy and updating payments in fee-for-service Medicare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27

2A Hospital inpatient and outpatient services . . . . . . . . . . . . . . .35

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36

Assessing payment adequacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37

Policies affecting the distribution of payments . . . . . . . . . . . . . . . . . . . . . .43

Update for inpatient services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62

Update for outpatient services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63

2B Physician services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71


Accounting for cost changes in the coming year . . . . . . . . . . . . . . . . . . . . .79

Update recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .80

2C Skilled nursing facility services . . . . . . . . . . . . . . . . . . . . . . . . . . .85

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86


Accounting for cost changes in the coming year . . . . . . . . . . . . . . . . . . . . .95

Update recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .95

2D Home health services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104

Assessing payment adequacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .106

Accounting for providers’ cost changes in the coming year . . . . . . . . . . .114

Update recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 vii

Table of contents

2E Outpatient dialysis services . . . . . . . . . . . . . . . . . . . . . . . . . . . . .121


Accounting for cost changes in the coming year . . . . . . . . . . . . . . . . . . . .131


2F Ambulatory surgical center services . . . . . . . . . . . . . . . . . . . .135

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .136

Collecting recent ASC cost data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .140


Accounting for cost changes in the coming year . . . . . . . . . . . . . . . . . . . .141


Variations in payment for ambulatory surgical procedures by setting . . .143

3 Access to care in the Medicare program . . . . . . . . . . . . . . . . . . . . .153

Evaluating access to care: an overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .154

The capacity of the health system to meet beneficiaries’ needs . . . . . . . . . . . .155

Beneficiaries’ ability to obtain care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .163

Beneficiaries’ ability to obtain appropriate care . . . . . . . . . . . . . . . . . . . . . . . .168

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170

4 Payment for new technologies in Medicare’s prospectivepayment systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .177

Payment for new technologies in prospective payment systems: an overview .178

Hospital inpatient services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180

Hospital outpatient services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181

Lessons from other health care purchasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . .184

5 Health insurance choices for Medicare beneficiaries . . . . . . . .193

What health insurance options do Medicare beneficiaries have? . . . . . . . . . . .194

Medicare beneficiaries and health plans in the marketplace . . . . . . . . . . . . . . .204

When supply and demand meet in the marketplace . . . . . . . . . . . . . . . . . . . . . .210

viii Tab l e o f c on t e n t s

Appendixes

A How Medicare pays for services: an overview . . . . . . . . . . . . . .219

Key structural elements of Medicare’s prospective payment systems . . . . . . .220

Acute inpatient services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .222

Ambulatory care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .228

Post-acute care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .233

Services for special populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .238

Other services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .239

Medicare�Choice plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .242

B An introduction to how Medicare makes coverage decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .245

The national coverage determination process . . . . . . . . . . . . . . . . . . . . . . . . . .246

Local medical review process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .247

Coverage policies implemented in program manuals . . . . . . . . . . . . . . . . . . . .248

Medicare’s coding process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .248

Appeals process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .248

C Inpatient payments for rural hospitals . . . . . . . . . . . . . . . . . . . . . .253

Implementing a low-volume adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253

Reevaluating the labor share used in geographic adjustment . . . . . . . . . . . . . .255

Eliminating the base rate differential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .256

Raising the cap on disproportionate share payments . . . . . . . . . . . . . . . . . . . . .257

Impact of recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .259

D A data book on hospital financial performance . . . . . . . . . . . . .265

E Commissioners’ voting on recommendations . . . . . . . . . . . . . . .283

Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .289

More about MedPAC

Commission members . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .295

Commissioners’ biographies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .297

Commission staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .303

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 ix


1-1 Annual change in Medicare spending, selected settings, 1993–2002 . . . . . . . . . . .6

1-2 Medicare HI trust fund solvency projections . . . . . . . . . . . . . . . . . . . . . . . . . . . .15

1-3 Dollar ranges for one- and five-year spending estimates . . . . . . . . . . . . . . . . . . .18

2 Assessing payment adequacy and updating payments in fee-for-service Medicare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27

2A Hospital inpatient and outpatient services . . . . . . . . . . . . . .35

2A-1 Trend in hospital total margin, 1998–2002 . . . . . . . . . . . . . . . . . . . . . . .38

2A-2 Overall Medicare margin and margin by major service line, 1999–2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40

2A-3 Overall Medicare margin by hospital group, 1999–2003 . . . . . . . . . . . .41

2A-4 Hospital cases under Medicare’s inpatient transfer payment policy currently and under proposed expansion . . . . . . . . . . . . . . . . . . . .45

2A-5 Medicare hospital discharges to post-acute care providers, 1991–2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47

2A-6 Use of post-acute care providers and cases affected by expansion of transfer policy to all DRGs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49

2A-7 Share of cases discharged to post-acute settings for selected DRGs, by hospital group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50

2A-8 Use of post-acute care, transfers, and length of stay by region, 2001 . . .51

2A-9 Use of post-acute care, Medicare inpatient operating margins, and length of stay, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51

2A-10 Distribution of hospital cases under the transfer policy, 2001 . . . . . . . . .52

2A-11 Characteristics of hospital cases under the transfer policy, 2001 . . . . . .53

2A-12 Change in inpatient payments from expanded transfer policy, 2001 . . . .54

2A-13 Percent increase in inpatient payment rates under alternative levels of the indirect medical education adjustment . . . . . . . . . . . . . . . . . . . . . .56

2A-14 Medicare inpatient margin in fiscal year 2000 and at alternative indirect medical education adjustment levels . . . . . . . . . . . . . . . . . . . . . .57

2A-15 Overall Medicare margin in fiscal year 2000 and at alternative indirect medical education adjustment levels . . . . . . . . . . . . . . . . . . . . . .58

2A-16 One- and two-year impacts on Medicare inpatient payments ofrecommendations to improve payments to rural hospitals . . . . . . . . . . . .62

2A-17 Impact on Medicare inpatient payments of update and distributionalrecommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64

2B Physician services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71

2B-1 Physicians billing traditional Medicare, 1995–2001 . . . . . . . . . . . . . . . .73

2B-2 Change in per capita use of physician services by beneficiaries intraditional Medicare, by selected type of service, 1999–2002 . . . . . . . . .78

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 xi

Tables

2B-3 Medicare Economic Index weights and forecast of input price changes for 2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .79

2C Skilled nursing facility services . . . . . . . . . . . . . . . . . . . . . . . . .85

2C-1 Medicare margins for skilled nursing facilities, 2000 and estimated 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87

2C-2 Change in the number of certified skilled nursing facilities, by type,1998–2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92

2C-3 Payment and use of skilled nursing facilities, 1996–2000 . . . . . . . . . . . .92

2D Home health services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103

2D-1 Freestanding home health Medicare margin, by agency group, 2001 and estimated 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107

2D-2 Ratio of payments to charges, by type of home health episode,2001 and 2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .108

2D-3 Use of home health after the PPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109

2D-4 Share of visits per home health episode, by type of visit . . . . . . . . . . .110

2E Outpatient dialysis services . . . . . . . . . . . . . . . . . . . . . . . . . . . .121

2E-1 Payment-to-cost ratios for composite rate services and separately billable drugs for freestanding dialysis facilities, 1996 and 2001 . . . . .124

2E-2 Indicators to assess changes in services furnished during in-centerhemodialysis treatments, 1997 and 2001 . . . . . . . . . . . . . . . . . . . . . . . .127

2E-3 Characteristics of dialysis facilities, 1993–2001 . . . . . . . . . . . . . . . . . .129

2E-4 Clinical performance indicators, 1994–2000 . . . . . . . . . . . . . . . . . . . . .130

2F Ambulatory surgical center services . . . . . . . . . . . . . . . . . . .135

2F-1 Characteristics of Medicare-certified ambulatory surgical centers,1991–2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .136

2F-2 Most common categories of procedures provided to Medicare beneficiaries in ASCs, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .138

2F-3 Hospital outpatient department and ASC payment rates for ambulatory surgery services, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . .143

2F-4 Average risk scores for Medicare beneficiaries receiving surgicalprocedures in ASCs and outpatient departments, 1999 . . . . . . . . . . . . .145

2F-5 Average total Medicare payments for all services for beneficiariesreceiving surgical procedures in ASCs and outpatient departments, 1999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .146

2F-6 Estimated impact of limiting ASC payment rates to hospital outpatient rates, by procedure category . . . . . . . . . . . . . . . . . . . . . . . . .147

2F-7 Distribution of ambulatory surgical centers by specialty type, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .148


3-1 Share of physicians accepting all new Medicare patients, 1997–2002 . . . . . . .157

xii Tab l e s

3-2 Acceptance of all or some new patients, by type of patient: MedPAC physician survey results, 1999 and 2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .157

3-3 Characteristics of the noninstitutionalized aged Medicare population, 1996–1999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .165

3-4 Aged Medicare population reporting access to care problems, by beneficiary characteristics, 1996–1999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .166

3-5 Share of beneficiaries using selected clinically necessary services, by supplemental coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .169

4 Payment for new technologies in Medicare’s prospectivepayment systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .177

4-1 Design of the hospital inpatient and outpatient new technology paymentmechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181

4-2 Outpatient PPS payments by service type and hospital group, 2001 . . . . . . . . .185

5 Health insurance choices for Medicare beneficiaries . . . . . . . .193

5-1 Availability of alternatives to the traditional Medicare fee-for-service program, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .198

5-2 Benefits, enrollment, and average premiums in standardized Medigap plans, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .201

5-3 Sources of additional coverage by selected beneficiary characteristics, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .206

5-4 Consumer Assessment of Health Plans ratings of Medicare FFS andMedicare�Choice plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .208

A How Medicare pays for services: an overview . . . . . . . . . . . . . .219

A-1 Summary of Medicare’s current payment systems by setting . . . . . . . . . . . . . .222


C-1 Medicare inpatient margin, by discharge volume, 1999 . . . . . . . . . . . . . . . . . .254

C-2 Components of national labor share for inpatient care . . . . . . . . . . . . . . . . . . .255

C-3 Medicare inpatient margin, by location, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . .257

C-4 One-year impact on Medicare inpatient payments of four recommendations to improve payments for rural hospitals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .259

C-5 Two-year impact on Medicare inpatient payments of four recommendations to improve payments for rural hospitals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .260

D A data book on hospital financial performance . . . . . . . . . . . . .265

D-1 Change in hospital payment, cost, and length of stay indicators, 1991–2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .266

D-2 Change in Medicare inpatient costs per discharge, 1991–2000 . . . . . . . . . . . . .267

D-3 Change in Medicare inpatient length of stay, 1991–2000 . . . . . . . . . . . . . . . . .268

D-4 Medicare inpatient margins excluding payments for direct graduate medical education, by hospital group, 1991–2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .269

D-5 Distribution of Medicare inpatient margins excluding payments for direct graduate medical education, by hospital group, 2000 . . . . . . . . . . . . . . . . . . . .270

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 xiii

D-6 Medicare outpatient margins excluding payments for direct graduate medicaleducation, by hospital group, 1996–2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .271

D-7 Distribution of Medicare outpatient margins excluding payments for directgraduate medical education, by hospital group, 2000 . . . . . . . . . . . . . . . . . . . .272

D-8 Hospital-based Medicare skilled nursing facility margins excluding graduate medical education, by hospital group, 1996–2000 . . . . . . . . . . . . . . .273

D-9 Hospital-based Medicare home health agency margins excluding graduate medical education, by hospital group, 1996–2000 . . . . . . . . . . . . . . . . . . . . . . .274

D-10 Hospital Medicare PPS-exempt unit margins excluding graduate medicaleducation, by hospital group, 1996–2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .275

D-11 Overall Medicare margins including payments for direct graduate medicaleducation, by hospital group, 1996–2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .276

D-12 Distribution of overall Medicare margins including payments for direct graduate medical education, by hospital group, 2000 . . . . . . . . . . . . . . . . . . . .277

D-13 Hospital payment-to-cost ratios, by source of revenue, 1991–2001 . . . . . . . . .278

D-14 Gains or losses as a percent of total hospital costs, by source of revenue,1991–2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .278

D-15 Hospital total margins, by hospital group, 1991–2000 . . . . . . . . . . . . . . . . . . .279

D-16 Distribution of hospital total margins, by hospital group, 2000 . . . . . . . . . . . .280

xiv Tab l e s


1-1 Change in distribution of Medicare spending by setting, fiscal years 1992–2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5

1-2 Distribution of Medicare spending among beneficiaries, 1997 . . . . . . . . . . . . . . . .6

1-3 Total Medicare spending, 1980–2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

1-4 National spending for personal health care, by payment source, 2001 . . . . . . . . . .8

1-5 National spending for personal health care, by setting and payment source, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9

1-6 Medicare share of national spending for personal health care, 1980–2011 . . . . .10

1-7 Real change in spending per enrollee, Medicare and PHI, 1968–2000 . . . . . . . .11

1-8 Change in spending per enrollee, Medicare and other purchasers, selected years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

1-9 Medicare spending as share of the federal budget, 1980–2012 . . . . . . . . . . . . . .14

1-10 Medicare spending as share of GDP, 1980–2030 . . . . . . . . . . . . . . . . . . . . . . . . .16

1-11 Composition of out-of-pocket spending, by out-of-pocket spending level, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18

2 Assessing payment adequacy and updating payments in fee-for-service Medicare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27

2-1 Approach for assessing payment adequacy and updating payment rates . . . . . . .28

2-2 Steps and factors in assessing payment adequacy . . . . . . . . . . . . . . . . . . . . . . . . .29

2A Hospital inpatient and outpatient services . . . . . . . . . . . . . .35

2A-1 Medicare hospital payments by major service line, 2000 . . . . . . . . . . . .36

2A-2 Payments to Medicare providers for all hospital inpatient and outpatient services, fiscal years 1992–2001 . . . . . . . . . . . . . . . . . . . . . . .37

2A-3 Hospital payment-to-cost ratios for Medicare, Medicaid, and privatepayers, 1991–2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39

2A-4 Trend in hospital total margin, 1990–2000 . . . . . . . . . . . . . . . . . . . . . . .39

2A-5 Trend in Medicare inpatient and overall Medicare margin, 1990–2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41

2A-6 Length of stay distribution for stroke (DRG 14) . . . . . . . . . . . . . . . . . . .47

2A-7 Payment-to-cost ratios for transfer cases before and after transfer policy, respiratory infections (DRG 79) . . . . . . . . . . . . . . . . . . . . . . . . . .48

2A-8 Change in Medicare inpatient margins, by teaching status, 1990–2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58

2A-9 Uncompensated care costs as a percentage of total hospital costs, byhospital group, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59

2A-10 Spending on all hospital outpatient services, 1991–2001 . . . . . . . . . . . .65

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 xv

Figures

2B Physician services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71

2B-1 Physician services program spending and payment updates, 2001–2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73

2B-2 Physician participation rates, 1997–2002 . . . . . . . . . . . . . . . . . . . . . . . . .74

2B-3 Quarterly changes in professional liability insurance premiums, 1990–2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .80

2C Skilled nursing facility services . . . . . . . . . . . . . . . . . . . . . . . . .85

2C-1 Medicare spending for skilled nursing facility services, 1992–2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87

2C-2 Skilled nursing facility payments and minutes of therapy for rehabilitation patients, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91

2D Home health services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103

2D-1 Estimated spending for home health, 1992–2002 . . . . . . . . . . . . . . . . .106

2D-2 Certified home health agencies, 1996–2002 . . . . . . . . . . . . . . . . . . . . .112

2E Outpatient dialysis services . . . . . . . . . . . . . . . . . . . . . . . . . . . .121

2E-1 Medicare spending for outpatient dialysis services furnished byfreestanding dialysis facilities, 1991–2001 . . . . . . . . . . . . . . . . . . . . . .122

2E-2 Aggregate payment-to-cost ratios for dialysis services, adjusted and unadjusted, 1997–2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .125

2F Ambulatory surgical center services . . . . . . . . . . . . . . . . . . .135

2F-1 States with the most Medicare-certified ASCs, 2001 . . . . . . . . . . . . . .137

2F-2 Growth in total Medicare payments for ASC services, 1991–2001 . . .139

2F-3 Growth in the number of ASCs and volume of procedures provided to Medicare beneficiaries in ASCs, 1996–2001 . . . . . . . . . . . . . . . . . .142


3-1 Supply and demand projections for full-time equivalent registered nurses,2000–2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .160

3-2 Per capita Medicare spending for beneficiaries ages 65–69 and 85 and over,selected services, 1998 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .161

B An introduction to how Medicare makes coverage decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .245

B-1 Medicare’s process for appeals of Part A and Part B claims . . . . . . . . . . . . . . .249


C-1 Hospital discharge volume and hospital cost per case, 1997 . . . . . . . . . . . . . . .254

C-2 Rural hospitals’ shares of low-income patient costs and disproportionate share payments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .258

xvi F i g u r e s

Executive summary

The Congress has charged the Medicare Payment Advisory Commission with reviewingand making recommendations concerning Medicare payment policies. The Commission’srecommendations aim to ensure that Medicare’s payment systems set rates that cover thecosts efficient providers would incur in furnishing care to beneficiaries. If payments areset too low, providers may not want to participate in the program and Medicarebeneficiaries may not have access to quality care. If payments are set too high, taxpayersand beneficiaries will bear too large a burden.

In this report, we review Medicare prospective payment systems (PPSs) for seven sectors:hospital inpatient, hospital outpatient, physician, skilled nursing, home health, outpatientdialysis, and ambulatory surgical center services. We also discuss several broader issuesrelated to Medicare payments:

• considering the context for Medicare payment recommendations (e.g. how does thegrowth of Medicare expenditures compare to that of the economy, the federalbudget, and the amount paid by other payers; how to characterize the spendingimpact of our recommendations);

• assessing Medicare beneficiaries’ access to care;

• deciding how Medicare should deal with payments for new technologies; and

• examining what health insurance choices are available to Medicare beneficiaries andwhat characteristics of insurance markets determine those choices.

At the beginning of each chapter, we list the recommendations contained in that chapter.In Appendix E, we present a list of all recommendations and the votes byCommissioners.

ContextUnderstanding the overall context for Medicare payment policies is important forpolicymakers. Therefore, we have included in Chapter 1 spending trends not just forMedicare but also for private sector payers and other federal health care programs. Overthe long term, the rate of increase in per capita spending for Medicare beneficiaries hasbeen similar to that for members of private sector health insurance plans and severalgovernment-sponsored plans. Year to year, there are different patterns and fluctuations,but the factors driving health care costs appear to operate similarly for all payers. We alsoreport trends in Medicare’s share of health care spending in the United States and of thefederal budget, and the share overall health care spending represents of gross domesticproduct (GDP). Over the next few decades Medicare will constitute a greater proportionof economic output. Similarly, it will create greater pressure within the federal budgetand increased cost sharing may stress beneficiary resources. For these reasons, pressuresto restrain Medicare’s rate of spending growth will likely increase.

When considering a policy direction, policymakers need a clear understanding of howrecommendations will affect spending. Therefore, we introduce a taxonomy forestimating the fiscal implications of each of our recommendations. Specifically, estimatesof spending changes are presented as ranges over one- and five-year periods; theimplications for beneficiaries and providers are highlighted. These spending estimatescannot simply be added together to compute an overall estimate. Unlike official budgetestimates, they do not take into account the complete package of policyrecommendations, the interactions among them, or assumptions about changes inprovider behavior.

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 xix

Executive summary

Assessing payment adequacy and updating paymentsIn Chapter 2 we recommend payment adjustments for seven different Medicareprospective payment systems. For each system, we assess whether payments are adequateto cover the cost of efficient providers by using indicators such as providers’ financialperformance under Medicare, changes in the volume of services, the quality of and accessto care, providers’ access to capital, and market entry or exit. We then address the likelychange in efficient providers’ costs in 2004. We estimate input price inflation (asmeasured by a “market basket” index for each sector), allow for technological changesthat both improve quality and significantly increase costs, and determine a reasonableexpectation for productivity gains. For expected productivity gains, we use the 10-yearaverage change in multifactor productivity in the general economy. Our updaterecommendations combine these judgments for each payment system. When appropriate,we also make recommendations to improve the distribution of payments among providerswithin each payment system.

Hospital inpatient and outpatient servicesIn the hospital sector we make both update and distributional recommendations. Theserecommendations should be considered jointly as a package because they are so closelyinterrelated and because some distributional recommendations would help certainhospitals—such as some rural hospitals—that are particularly vulnerable.

We find that overall Medicare payments for hospital services are adequate as of fiscalyear 2003. Using a margin calculation that encompasses nearly all Medicare payments tothe hospitals, and thus is not influenced by cost accounting differences, we estimate amargin for hospital services in 2003 of 3.9 percent (adjusted for changes legislated forfiscal year 2004 that will reduce payments). Other broad indicators, such as trends involume and access to capital, are also generally consistent with a conclusion of adequatepayments. This conclusion, together with consideration of other factors that are likely toaffect costs in the coming year—including input price inflation, technological advances,and productivity—support an update for 2004 of market basket minus 0.4 percent forinpatient services. Because significant technological advances that affect outpatientservices are accounted for through new technology provisions in that payment system, werecommend an outpatient update of market basket minus 0.9 percent for productivityimprovement.

In addition, five policy changes are needed to improve the distribution of inpatientpayments:

• expanding the current transfer policy for patients in certain diagnosis related groups(DRGs) who are discharged to post-acute settings;

• implementing a low-volume adjustment;

• reevaluating the labor share used for geographic adjustment of rates;

• eliminating the differential in base rates for hospitals in rural and small urban areas;and

• increasing the cap on disproportionate share payments that applies to most ruralhospitals.

We recommend expanding the post acute care transfer policy to additional DRGs tobetter allow payments to follow patient care and to prevent hospitals that cannotdischarge patients to post-acute care from being disadvantaged. We have recommendedthe other four policy changes in previous reports and reiterate them now as part of thecomprehensive package that, taken together with the update recommendation, will help

xx Exe cu t i v e s umma r y

maintain the financial viability of the hospital sector. A final important issue is the currentindirect medical education adjustment to inpatient payments. That adjustment providespayments above the level justified by the empirical evidence on the relation betweenteaching activity and hospitals’ Medicare costs. The Commission is not satisfied with thecurrent policy, because there is no accountability for the use of the payments above theempirical level. We will explore ways to better target those payments to advance specificMedicare policy objectives through increased accountability.

Physician servicesMedicare payment rates for physician services are based on a fee schedule and areupdated annually based on the so-called sustainable growth rate system, which tiesupdates to growth in the national economy and other factors. Under this system, theupdate for 2003 is a minus 4.4 percent.

In assessing payment adequacy we find a mixed picture. The number of physiciansbilling Medicare has increased and national indicators of access are still good. There are,however, anecdotal reports of access problems in some geographic markets andspecialities. A national survey of physicians suggests that physicians are becoming moreselective about accepting new Medicare patients—but that is true for private HMO andMedicaid patients as well. Finally, Medicare payment rates have fallen somewhat relativeto payment rates in the private sector, although they are still above levels seen in the mid-1990s.

From this assessment, the Commission concludes that payments would be adequate thisyear if the Congress were to change current law and require a modest, positive update for2003 instead of the 4.4 percent payment reduction. Therefore, if the Congress acts, werecommend an update for 2004 that equals the estimated change in input prices forphysician services less an adjustment for productivity growth. If the Congress does notrequire a positive update for 2003, a higher update will be necessary in 2004.

Skilled nursing facility servicesAggregate Medicare payments for skilled nursing facilities (SNFs) are at least adequatefor fiscal year 2003. For freestanding SNFs—about 90 percent of providers in thissector—we estimate aggregate Medicare margins to be 11 percent in 2003. Including the10 percent of SNFs that are hospital-based brings the aggregate SNF margin to about 5percent. The high margin for freestanding SNFs reflects a decline in costs in recent yearsin response to incentives in the skilled nursing facility prospective payment systemfollowing high cost growth prior to its introduction. Preliminary evidence indicates thatthe decline in costs has not resulted in a lower quality of care. Because the prospectivepayment system for skilled nursing facilities is still relatively new, we expect this costtrend to continue into 2004, offsetting increases in input prices and other factors.Therefore, we recommend that the Congress not update payment rates for SNFs for fiscalyear 2004.

Because of weaknesses in the current classification system for care in SNFs, however,payments are not distributed appropriately to account for the expected resource needs ofdifferent types of Medicare beneficiaries. Resources should be reallocated until theclassification system is improved or replaced. As a start, we recommend that theCongress give the Secretary authority to reallocate money currently used as a paymentadd-on for rehabilitation classification groups to other classification groups so thatpayment more closely follows patient costs. This reallocation will benefit hospital-basedSNFs to the extent that they serve patients with conditions more complex than those ofpatients in freestanding SNFs; therefore, no separate update for hospital-based SNFs isrecommended. However, if this reallocation does not occur in a timely manner, theCongress should provide a market basket update less productivity adjustment of 0.9percent for hospital-based SNFs only.

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 xxi

Home health servicesCurrent aggregate Medicare payments for home health services are more than adequaterelative to costs. For the first time, we now have cost data showing how home healthagencies are performing under the PPS. We estimate that the Medicare margin for homehealth services in fiscal year 2003 will be 23.3 percent, even after accounting for the so-called 15 percent payment reduction and the expiration of the current 10 percent ruraladd-on. Providers have responded to the new PPS by changing the home health productand the cost of providing an episode of home health services is lower as a result. Otherbroad indicators also suggest that payments are adequate: access to care is generallygood, the rate of decline in the number of users has decreased, and the entry and exit ofagencies has remained stable for the third year in a row.

In the past, we have recommended updates that emphasized stability for this sectorbecause we lacked data on agencies’ financial performance and also wanted to giveproviders time to adapt to the new payment system. Home health agencies have adapted,and we expect them to continue to adapt during the coming year, further reducing thecosts of providing an episode of care. Therefore, we recommend that the Congress notupdate payment rates for home health services for fiscal year 2004. Because of potentialchallenges that providers may face in rural areas, we also recommend that the Congressextend for one year, at a rate of 5 percent, add-on payments for home health servicesprovided to Medicare beneficiaries who live in rural areas.

Outpatient dialysis services Current aggregate Medicare payments for outpatient dialysis services for beneficiarieswith end-stage renal disease appear to be adequate. Together, payments for compositerate services and injectable drugs—the two main components of payment to providers ofoutpatient dialysis services—exceeded providers’ costs by about four percent in 2001. Inaddition, other indicators—such as continued entry of for-profit freestanding providers,increases in the volume of services provided, lack of evidence of beneficiaries facingsystematic problems in accessing care, continued improvements in the quality of dialysiscare, and providers enjoying adequate access to capital—together support the conclusionthat Medicare’s outpatient dialysis payments are adequate relative to efficient providers’costs. To account for changes in providers’ costs in the coming year, the Congress shouldupdate the composite rate for outpatient dialysis services for calendar year 2004 by thechange in input prices less a 0.9 percent adjustment for productivity gains.

Ambulatory surgical center servicesAn ambulatory surgical center (ASC) is a distinct entity that exclusively furnishesoutpatient surgical services. The current payment rates for ASC services are based on acost survey conducted in 1986. Because of the age of the data, our first recommendationin this sector is that the Secretary expedite the collection of recent ASC charge and costdata for the purpose of analyzing and revising the ASC payment system. Because thereare no recent data on the cost of providing ASC services to Medicare beneficiaries, welooked at market factors and concluded that current payments for ASC services are morethan adequate. There has been rapid growth in the number of ASCs; between 1997 and2001, the number of Medicare-certified ASCs more than doubled. The volume ofprocedures provided by ASCs to beneficiaries increased by over 60 percent between 1997and 2001. In addition, as indicated by their rapid growth, ASCs have sufficient access tocapital. Current Medicare payments for ASC services are at least adequate to cover nextyear’s expected increase in ASCs’ costs. Therefore, we recommend that the Congress notupdate the payment rates for ASC services for fiscal year 2004.

xxii Exe cu t i v e s umma r y

In addition, although costs in ASCs should be lower than in hospital outpatientdepartments because ASCs have less regulatory burden and serve less medically complexpatients, the ASC rate is currently higher than the outpatient hospital rate for several high-volume procedures. Therefore, we recommend the Congress should ensure that paymentrates for ASC procedures do not exceed hospital outpatient PPS rates for those sameprocedures after accounting for differences in the bundle of services covered.

Access to careA basic goal of Medicare is to ensure that elderly and disabled Americans have access toappropriate, quality health care. Therefore, we plan each year to monitor beneficiaries’access to Medicare-covered services along three dimensions: (1) the health system’scapacity; (2) beneficiaries’ ability to obtain care; and (3) access to appropriate care. InChapter 3, we present our analysis for this year and do not find widespread problems inbeneficiaries’ access to care. Although more selective about accepting patients from anumber of payers than in the past, the vast majority of physicians are accepting at leastsome new Medicare beneficiaries. Post-acute services are generally available, although ithas become more difficult to place the most complex patients in skilled nursing facilities.Nonetheless, some issues will require careful monitoring. As in other populations, certainbeneficiaries—those in poor health, with low incomes, and without supplementalinsurance—report more difficulty than others in accessing appropriate services. Otherbeneficiaries, even though reporting good access, may not be receiving appropriateservices. In addition, shortages of nurses could affect the availability or timeliness ofcertain services, and demographic trends raise concerns about the capacity of the healthsystem over time.

Payment for new technologies Medicare has the dual responsibility to pay enough for beneficial new technologies toensure beneficiaries’ access to care, while also being a prudent purchaser of newtechnologies. In Chapter 4 we examine how this dual role is addressed in the inpatientand outpatient prospective payment systems and how those systems might be improved.The incentives built into prospective payment systems promote the use of newtechnologies that reduce costs, but they may also slow adoption of new technologies thatincrease costs. To offset that tendency, the inpatient and outpatient prospective paymentsystems currently incorporate the costs of new technologies through special paymentmechanisms for specific new technologies as well as through an annual review ofpayment rates. To ensure fair treatment across technologies and payment systems,MedPAC recommends that the clinical criteria currently applied to all new technologyapplicants under the inpatient PPS, and to new medical device applicants under theoutpatient PPS, be extended to new drugs and biologicals applicants under the outpatientPPS.

Health insurance choices for Medicare beneficiariesDepending on where they live, Medicare beneficiaries may have a wide array ofinsurance options beyond traditional fee-for-service Medicare available to them. Thoseoptions may include Medicare+Choice comprehensive care plans and private fee-for-service plans, cost contract plans, preferred provider plans, and varying forms ofsupplemental coverage. What options are available, and how and when beneficiarieschoose among them, depends on specific market conditions and the circumstances ofindividual beneficiaries. The determinants of market conditions are both local andnational. Although Medicare is a national program, it is only at the local level thatmedical care is delivered, beneficiaries choose insurance options and delivery systems,and insurers make decisions to enter the insurance market. In Chapter 5 we review theentire spectrum of insurance choices, as a first step in MedPAC’s effort to betterunderstand beneficiaries’ choices and market conditions. �

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 xxiii




















advocates.







Context forMedicare spending

C H A P T E R1

edicare spending increased by an annual average of 9.6

percent per beneficiary between 1968 and 2000.

Although slightly lower than the growth rate of health

care spending by private insurers, increases of this mag-

nitude have unique implications given limited federal budget, trust fund, and

beneficiary resources. Moreover, because the growth in Medicare spending

has exceeded growth of the gross domestic product—as has all health care

spending—an increasing portion of the nation’s economic resources are devoted

to health care services. Medicare’s spending growth is a concern because it re-

quires policymakers to weigh competing priorities and ultimately to make trade-

offs in allocating limited resources.

This chapter explores trends in Medicare spending, compares Medicare growth

to that of other health spending indicators, and examines the implications of

spending increases given limited resources.

M

C H A P T E R

Context for Medicarespending

1In this chapter

• Medicare spending trends

• Medicare spending comparedwith other indicators of healthspending

• Implications of Medicarespending given limitedresources

• Spending and otherimplications of MedPAC’srecommendations

Repo r t t o t h e Cong r e s s : Med i ca r e Paymen t P o l i c y | Ma r ch 2003 3

The Congress has charged MedPAC withassessing the design and implementationof Medicare payment policy and makingrecommendations to the Congress and theSecretary of the Department of Health andHuman Services to address any problemsidentified. In carrying out theseresponsibilities, MedPAC examineswhether Medicare’s payment policysupports the ultimate goal of the program:ensuring that its beneficiaries have accessto medically necessary acute care of highquality in the most appropriate clinicalsetting, without imposing undue financialburdens on beneficiaries and taxpayers.This examination requires that weevaluate not only the technical aspects ofpayment policy as they affect access tocare, but also the implications forbeneficiaries and taxpayers of risingMedicare and health care spending.

This chapter shows that after a fewanomalous years of low rates of growth,Medicare spending has resumed its moretypical trajectory, growing an average of7.7 percent between 2001 and 2002. Toprovide a context in which to assess thisgrowth rate, the chapter comparesMedicare’s growth to that of other typesof national health care spending. The datasuggest that while growth rates diverge atcertain points, over the long runMedicare’s growth is roughly comparableto that of other purchasers.

The chapter also identifies resourceconstraints that ought to be consideredwhen evaluating both the short-termpayment policy recommendations in thisreport and the need for longer-rangeMedicare reforms. Medicare is absorbinga growing proportion of the nation’sbudget and economic resources; theMedicare Hospital Insurance trust fundinsolvency date looms; and beneficiariesare spending a growing percentage oftheir resources on health care, which forsome means painful trade-offs betweengetting medical care and purchasing otheressentials of living. This chapter does notgo so far as to recommend solutions to

these problems, but MedPAC will analyzeand report on innovations in health carefinancing and delivery that may holdpromise for addressing them.

Given Medicare’s limited resources,MedPAC makes its recommendationswith—and policymakers should considerthem with—an understanding of theirconsequences on spending as well as onbeneficiaries and providers. To further thisgoal, MedPAC is making the implicationsof its recommendations more explicit bysummarizing the implications below eachrecommendation and providing anestimate of the change in spending, whenpossible.

This chapter first presents backgroundinformation on Medicare spending trends.Then it discusses overall national healthspending and other health care spendingthat may serve as a benchmark againstwhich to assess Medicare’s scope andgrowth. Third, the chapter identifies theresource constraints associated with thefederal budget, Medicare trust funds, theeconomy, and beneficiaries. Finally, giventhese trends and constraints, the chapterdiscusses how MedPAC assesses andpresents the implications of itsrecommendations.

Medicare spending trends

Understanding how much Medicarespends for which services and for whichbeneficiaries, and also how fast thisspending is expected to grow, is essentialto assessing the performance and financialsustainability of the program. Informationon spending trends lays the foundation forcomparing Medicare’s spending growthwith that of other payers and forconsidering various spending constraints,such as the federal budget and Medicaretrust funds. In addition, this informationprovides a sense of scale for assessing theimpact of various policy options. Forexample, an option that is estimated to

increase hospital payments by 1 percent isfar more costly than an option increasinghospice payments by 1 percent.

Spending levels anddistribution The amount of Medicare spending can beexpressed in many different ways that areuseful for different purposes. For ageneral understanding, perhaps the bestway to consider Medicare spending is toinclude all the money the Medicareprogram pays for benefits. In 2002,Medicare spent about $250 billion, or$6,200 per enrollee.1 In the same yearbeneficiaries, often through asupplemental insurer, also paid anadditional $38 billion in Medicarecoinsurance and deductibles to theirproviders.

Medicare spending is concentrated oncertain services, beneficiaries, andgeographic areas. Inpatient hospitalservices were by far the largest spendingcategory (40 percent), followed byphysicians (17 percent), skilled nursingfacilities (6 percent), and home health (5percent). Spending for beneficiariesenrolled in the Medicare�Choiceprogram accounted for 15 percent of thetotal. This distribution has changed overtime, particularly as enrollment in theMedicare�Choice program has fluctuatedand major changes in payment policyhave affected spending levels ofindividual sectors. For example, althoughinpatient hospital spending has grown 53percent from 1992 to 2002, it has shrunkas a percentage of Medicare’s spending,falling from 51 percent to 40 percent(Figure 1-1).

Like private insurance spending, Medicarespending is concentrated in a smallpercentage of beneficiaries. In 1997, halfof Medicare spending was for the costliest5 percent of beneficiaries, and 90 percentwas for the costliest 25 percent ofbeneficiaries. By contrast, the least costly50 percent of beneficiaries consumed only2 percent of all Medicare spending in

4 Con t e x t f o r Med i ca r e s pend i ng

1 For the purposes of this chapter, unless otherwise noted, spending numbers are presented as gross outlays, meaning that they include spending financed by beneficiarypremiums but do not include spending by beneficiaries (or on their behalf) for cost-sharing associated with Medicare-covered services. In general, they are reported on afiscal year, incurred basis and do not include spending on program administration.

Change in distribution of Medicare spendingby setting, fiscal years 1992–2002

FIGURE1-1

Hospital inpatient150%

Managed care5%Other fee-for-service

settings2

9%Ambulatory surgical center0.3%

Free-standing dialysis2%

Home health6%

Skilled nursing facility3%

Physician20%

Hospital outpatient5%

Total spending 1992 = $130 billion

Total spending 2002 = $252 billion

Hospital inpatient1

40%

Managed care15%

Other fee-for-servicesettings2

10%

Ambulatory surgicalcenter1%

Free-standing dialysis2%

Home health5%

Skilled nursing facility6%

Physician17%

Hospital outpatient4%

Note: Includes program outlays only. Totals may not add to 100 due to rounding.1Includes all hospitals, those paid under the prospective payment system (PPS), and PPS-exempt hospitals.2Includes hospice; outpatient laboratory; durable medical equipment; Part B drugs, ambulance services, and supplies; and Rural Health Clinics, Federally Qualified Health Centers, and outpatient rehabilitation facilities.

Source: CMS, Office of the Actuary, 2002.

1997 (Figure 1-2, p. 6).2 When examinedover a five-year period, the concentrationis less dramatic: roughly 75 percent ofspending between 1993 and 1997 was forthe costliest 25 percent of beneficiaries.

Focusing on the characteristics of costlybeneficiaries is illuminating, but theimplications of these characteristics mustbe considered carefully. Costlybeneficiaries in one year are more likelythan other beneficiaries to have high costsin the following years. Of the high-costbeneficiaries who were alive at the end of1993, over half remained in the highestquartile of spending in the next calenderyear—a rate twice as high as would beexpected by chance (Crippen 2002a).

Costly beneficiaries are also likely to havemultiple chronic conditions. One analysisfound that beneficiaries with three or moreconditions (46 percent of beneficiaries)account for almost 90 percent of totalMedicare spending, while those with nochronic conditions account for less than 1percent (Anderson 2002). Because thisanalysis measured all spending for eachtype of beneficiary regardless of whetherthe spending was associated with thebeneficiaries’ chronic conditions, it isunclear to what extent the costly acute-care episodes were attributable to chronicconditions. It is known, however, thatcostly beneficiaries tend to use a lot ofinpatient hospital care. More than half ofMedicare spending on the most expensive5 percent of beneficiaries was for inpatienthospital services in 1997 (Crippen 2002a).

Costly beneficiaries often include those inthe last year of life. About 25 percent ofMedicare outlays are spent on the last yearof life for the 4.7 percent of beneficiarieswho die each year. It is important toremember, however, that because the yearor time of death is not predictable, thisfigure shows the cost of caring forseverely ill individuals with unknown lifeexpectancy, not the cost of care deliveredin anticipation of impending death(MedPAC 2000).


2 This data is based on a Congressional Budget Office analysis of claims data for fee-for-service beneficiaries. The five-year analysis includes only beneficiaries enrolled inMedicare since 1993.

In addition, beneficiaries in some areas ofthe country are more costly, on average,than beneficiaries in other areas of thecountry. Some of this variation is due todeliberate payment adjustments to reflectdifferences in input prices, such as wagesand rent, and to support other missions,such as payments for medical educationand provision of uncompensated care. Anadditional part of the geographic variationis due to beneficiaries receiving differentamounts of medical services, which isinfluenced by differences in providers’practice patterns and beneficiaries’propensity to seek care, which in turn areinfluenced by factors such as their healthstatus, income, culture, and presence ofsupplemental coverage.

Spending growthPrior to 1997, Medicare spending hadbeen increasing rapidly, averaging 11.1percent annually between 1981 and 1997.3

This rate of increase declined sharply

between 1998 and 2000 to 1.7 percent, asthe effects of provider payment reductionsin the Balanced Budget Act of 1997(BBA) and enhanced efforts to deter fraudand abuse were felt. For 2001 and 2002,however, the rates of increase in spendingresumed more typical trajectories of about9 percent and 5.6 percent, respectively.

This general growth pattern was observedin virtually every service sector, butseveral specific trends are worthhighlighting (Table 1-1):

• Leading up to the passage of theBBA, home health and skillednursing facility (SNF) spending weregrowing at double-digit rates,peaking at 34 percent and 43 percent,respectively. Between 1997 and2000, however, home health and SNFspending levels decreased. By 2001and 2002, annual growth rates foreach sector were again positive, andin the double digits.

• Inpatient hospital growth rates havenot shown the same volatility as thosefor post-acute care, but becauseinpatient hospital care represents alarge portion of Medicare spending,its growth greatly influencesMedicare’s overall growth. Between1993 and 1997, inpatient hospitalspending grew 6.1 percent annually,on average. Growth dipped to just 0.1percent between 1998 and 2000 (afterthe BBA), before resuming a 6.7percent annual growth rate between2001 and 2002.

• Managed care spending grew nearly30 percent annually, on average,between 1993 and 1997, asenrollment more than doubled. After


Distribution of Medicare spending amongbeneficiaries, 1997

FIGURE1-2

Source: Congressional Budget Office, 2002.

70%

60%

50%

100%

90%

80%

40%

30%

20%

10%

0%Fee-for-service beneficiaries Total fee-for-service spending

5%

20%

25%

50%

47%

42%

2%9%

Annual change in Medicare spending, selectedsettings, 1993–2002

Setting 1993–1997 1998–2000 2001–2002

Hospital inpatient 6.1% 0.1% 6.7%Physician 4.3 4.7 8.6Skilled nursing facility 27.6 �4.8 15.8Home health 19.7 �21.8 22.8

Source: CMS, Office of the Actuary, 2002. Hospital inpatient includes all hospitals, those under the prospectivepayment system (PPS), and PPS-exempt hospitals. Includes program outlays only, gross mandatory, fiscal year,incurred.

T A B L E1-1

3 In calculating average annual growth rates over a span of years, growth for the first year is calculated as the difference in spending from the prior year (1980, in thiscase) to spending in the year noted (1981, in this case). This convention is followed throughout the report.

the passage of the BBA, a number ofplans withdrew from the program orreduced their service area andenrollment declined, resulting inannual growth rates that averaged16.4 percent between 1998–2000 and–4.2 percent between 2001–2002.

Projections of future growth suggest thatMedicare will continue to grow at about 6percent annually, on average, until theretirement of the baby boom generation,when growth will accelerate significantly.Forecasts of future Medicare spending areinherently uncertain but need to beconsidered in order to evaluate whetherthe program is financially sustainable.Several entities project future Medicarespending, including the CongressionalBudget Office (CBO), the Office ofManagement and Budget (OMB), and theMedicare Trustees (Figure 1-3). Amongthe factors contributing to the uncertaintyof their estimates is that they assume nochange in current law, despite the fact thatCongress regularly intervenes to adjustpayment policies and occasionallychanges coverage policies. Another sourceof uncertainty is difficulty predictingchanges in the volume and intensity ofservices to be delivered to Medicarebeneficiaries and, in particular, how newtechnology will influence these factors.

With these caveats in mind, we note thatCBO projects that mandatory spending forMedicare will grow at an annual averagerate of 6.5 percent over the 2003–2012period (3.9 percent real growth). CBO’sestimate of cumulative spending over thefirst 5 years of the projection window is7.7 percent higher than the estimates ofthe Office of Management and Budget; itis 10.2 percent higher than OMB’sestimates over the 10-year window.4 TheMedicare Trustees’ intermediateprojection for 2003–2011 assumes 6.1percent average annual growth (3.5percent real growth).

Medicare spendingcompared with otherindicators of healthspending

As policymakers debate how to improveMedicare’s ability to be a prudentpurchaser and whether policy changes areneeded to change the projected trajectoryof Medicare spending, it may be helpful tocompare Medicare spending with totalhealth spending and spending by otherpayers. This comparison provides abenchmark, albeit an imperfect one, thathelps policymakers understand the size ofMedicare in the marketplace and, in turn,its potential influence in the market.

To give a better sense of how Medicarespending compares with other health carespending, this section first discusses thecomparative scope of Medicare, thencompares Medicare’s growth rates tothose of other private and public healthspending, and finally explores the factorsdriving growth in health care spending.This discussion draws heavily from thenational health expenditure (NHE) datacompiled by the CMS Office of theActuary, which disaggregates totalspending by source of funding andservice.5

Comparative scope ofMedicareIn 2001, the Medicare program spent $235billion (about $5,900 per beneficiary) andaccounted for 19 percent of total national


1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012

Trustees low

OMB

CBO

intermediate

Trustees high

600

500

400

300

200

100

0

Total Medicare spending, 1980–2012FIGURE1-3

Note: CBO (Congressional Budget Office), OMB (Office of Management and Budget), Trustees (2002annual report of the Boards of Trustees of the Medicare trust funds). All data are nominal, gross

mandatory program outlays. Trustees and OMB projections include administrative spending,and Trustees projections are presented on a calendar year basis ending in year 2011.

Source: CMS, Office of the Actuary 2002 (historical spending). 2002 annual report of the Boards of Trustees of theMedicare trust funds, CBO 2002, OMB 2002 (projections).

Spen

din

g (

dolla

rs in

bill

ions)

Fiscal year

Historical Projected

Trustees

4 The differences between CBO’s and OMB’s estimates are attributable to different assumptions about annual updates for provider payment rates, administrative actionson outpatient drug payment, managed care enrollment, and the rate of increase in the volume and mix of services in the fee-for-service sector (Crippen 2002b).

5 NHE’s Medicare estimates are derived from the Medicare Trustees reports. Its latest year of actual data is 2001.

spending on personal health careservices.6 As such, Medicare is the singlelargest payer for health services in themarketplace. Of the $1.24 trillion (about$4,400 per person) spent on personalhealth care services in the United States in2001, about 35 percent was privateinsurance payments from a wide array ofpayers and 17 percent was consumer out-of-pocket spending. Medicare, Medicaid,the State Children’s Health InsuranceProgram (SCHIP), and all other publicspending combined accounted for about43 percent (Figure 1-4).

The level and distribution of Medicarespending differ somewhat from those ofother payers largely because Medicarecovers an older, sicker population anddoes not cover most prescription drugs or

dental care. Accordingly, a greaterpercentage of Medicare’s total spending isdevoted to hospital and home healthservices compared with that of privateinsurers. Medicare is the single largestpurchaser of these services. In 2001, itpaid for 30 percent of both hospital andhome health services. However, Medicarepaid for only 2 percent of prescriptiondrugs and 12 percent of nursing home care(Figure 1-5). For some types of providers,including certain hospitals and physicianspecialties, Medicare accounts for morethan half of their revenue. As such,Medicare’s payment and coveragepolicies can be a strong influence on thehealth care delivery system.

Comparing growth in spending

In this section, we compare the growth inMedicare spending with total spending onpersonal health care, private insurancespending on benefits, and premiumgrowth of other government insuranceprograms. Although comparingMedicare’s per enrollee growth rate withother payers’ growth rates may beinformative, it must be undertaken with anappreciation for the limits of thecomparison. First, Medicare and otherpurchasers do not buy the same mix ofservices. So, for example, Medicare islargely unaffected by the rapid growth inspending for outpatient prescription drugs,one of the main drivers of otherpurchasers’ spending increases. Inaddition, Medicare covers an older


National spending for personal health care, by payment source, 2001

FIGURE1-4

Other public2

7%

Other private1

5%

Medicaid and allSCHIP17%

Medicare19%

PHI35%

Out-of-pocket17%

Total = $1.24 trillion

Note: PHI (Private Health Insurance), SCHIP (State Children's Health Insurance Program). Out-of-pocket spending includes cost-sharing for both privately and publicly insuredindividuals. Personal health spending includes spending for clinical and professional services received by patients. It excludes administrative costs and profits.1Includes industrial in-plant, privately funded construction, and nonpatient revenues including philanthropy.2Includes programs such as workers' compensation, public health activity, Department of Defense, Department of Veterans Affairs, Indian Health Service, and state and local government hospital subsidies and school health.

Source: CMS, Office of the Actuary, National Health Accounts, 2003.

6 Medicare spending does not include beneficiary spending on cost sharing for Medicare benefits. Personal health care spending excludes spending for such categoriesas research, construction, public health, and administrative costs.

population that tends to be more costlyand may use expensive technology at afaster pace than younger people (Moon1999). This comparison is alsocomplicated because the NHE includes inits private insurance spendingsupplemental insurers’ spending forMedicare beneficiaries.

Another concern about comparing privatepayers’ spending with Medicare spendingis that these measures do not isolatechanges in cost-sharing for coveredservices. Because changes in the level of

enrollee cost-sharing can either increase ordecrease spending by payers, examiningchanges in the spending by payers can bemisleading about their ability to containoverall health care costs. In previousdecades, private insurers tended to reducecost-sharing. Recently, however, evidencefrom employer surveys and focus groupssuggests that enrollees are facing highercost-sharing as private-sector purchasersseek to inject greater cost-consciousnessamong enrollees and slow the growth inthe use of health care services (Robinson

2002).7 This shift of health care costs fromthe premium to cost-sharing may beequivalent to a 2 to 3 percent increase inpremiums (Strunk 2002).8

Comparing personal health carespending and Medicare spending

To see how Medicare’s growth compareswith growth in national spending onhealth care services, we examined NHEmeasurements of personal healthspending, which include consumer out-of-pocket spending as well as spending by a


Total spending = $1.24 trillion

Service type

MedicareOther* Medicaid and all SCHIP

National spending for personal health care, by setting and payment source, 2001FIGURE1-5

Share

of

spen

din

g

60%

100%

80%

40%

20%

0%hospital physician and

clinical serviceshome health nursing home prescription

drugsdurable medical

equipment

Note: SCHIP (State Children's Health Insurance Program). Personal health spending includes spending for clinical and professional services received bypatients. It excludes administrative costs and profits.*Other includes private health insurance, out-of-pocket, and other private and public spending.

Source: CMS, Office of the Actuary, National Health Accounts, 2003.

7 One survey found that, between 2001 and 2002, preferred provider organizations (PPOs) increased their deductibles 37 percent and that the percentage of workers inhealth maintenance organizations (HMOs) facing a $20 copayment for outpatient physician services rose from 2 percent to 11 percent (Kaiser-HRET 2002).

8 Ideally, our analysis would tease out this shifting of costs between insurers’ spending and beneficiary cost-sharing to ensure the most accurate comparison. However, thedata on out-of-pocket spending do not specify the extent to which such spending has been associated with benefits covered by Medicare as opposed to privateinsurance, or the extent to which spending has been related to uncovered services.

multitude of payers, including Medicare,insurance companies, and employers.Between 1991 and 1997, Medicare’sspending growth generally outpaced theaverage growth of all other components ofpersonal health spending (e.g., privateinsurance, Medicaid, and out-of-pocketspending) combined. However,Medicare’s growth slowed dramaticallyafter 1997, while other components ofpersonal health care spending continuedgrowing at a faster rate. Medicarerepresented 19 percent of personal healthcare spending in 2001, down from 21percent in 1997. The actuaries whodevelop the NHE data project that thisproportion will decline further to 18percent by 2003 and remain relativelysteady through the remainder of the

projection window (Figure 1-6), whichends just before the retirement of the babyboom generation.

Comparing Medicare spendingand private insurance spending

Two of the major sources of personalhealth care spending are Medicare andprivate insurance.9 Over a 33-year period,despite some fluctuation, the per enrolleeaverage growth rates in Medicare andprivate insurance have been roughlycomparable, with Medicare growingslightly more slowly (see Figure 1-7).After adjusting spending levels fordifferences in age and gender,unpublished CMS data show that real perenrollee Medicare growth over this periodwas 3.1 percent compared to 4.4 percentfor private health insurance. Whenestimated spending on outpatient

prescription drugs is subtracted fromprivate health insurance and Medicarespending, the growth rates of Medicareand private health insurance are evenmore comparable (3.1 percent forMedicare vs. 4.0 percent for private healthinsurance). Over shorter periods withinthis time frame, the growth rates of thetwo sectors have diverged as each trieddifferent cost-containment strategies(Figure 1-8, p. 12).

Projections of future growth rates arehighly uncertain and usually fail toanticipate the timing of peaks and valleysin spending growth rates. Nevertheless,they are useful for gaining a sense of thelikely direction of the spending trajectoryand the relationship between payers.Assuming current law, Medicare perenrollee spending is expected to grow


Medicare share of national spending for personal health care, 1980–2011FIGURE1-6

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

15%

10%

5%

0%

20%

25%

Med

icare

share

Note: Personal health spending includes spending for clinical and professional services received by patients. It excludes spending on research, construction, public health andadministrative costs.


9 Recall that private insurance includes spending by private insurers for Medicare beneficiaries, so these measures are not entirely independent.

more slowly than private health insurancespending through 2011. However, ifCongress intervenes and raises paymentrates to Medicare providers, the slowerout-year growth may not be realistic.

The accuracy of the estimates for near-term private insurance growth (10.4percent increase in premiums for 2002) isalso uncertain. Surveys suggest higherprivate premium increases in the short-term—between 12 and 16 percent in2002, and more than 15 or 16 percent in2003 (Mercer 2002, Kaiser FamilyFoundation 2002, Hewitt Associates 2002,Towers Perrin 2002).

Comparing Medicare to othergovernment health purchasers

Comparing Medicare’s growth to that ofother large public purchasers, each of whichhas a different approach to containing costs,

tells a similar story: While growth ratesdiffer over selected periods, over the long-term they tend to be similar.

The Federal Employees Health BenefitsProgram (FEHBP) and California PublicEmployees’ Retirement System(CalPERS) are two examples of publicentities that use a market-orientedapproach to contract with privateinsurance plans for employee healthcoverage. While the strategies these publicentities use to contain costs offer someinsight into potential payment alternativesfor Medicare, policymakers mustrecognize important differences betweenthese purchasers and Medicare. Forexample, in contrast to Medicare, bothFEHBP and CalPERS serve currentworkers as well as retirees; CalPERSenrollees are concentrated in Californiaand FEHBP annuitants are largelyconcentrated in urban areas, which

enables greater competition amongcontracting plans; and both programs havefar fewer beneficiaries than Medicaredoes. Also, CalPERS and FEHBP providecoverage for outpatient drugs, whereas, asmentioned above, Medicare does not.

• FEHBP is the health benefit programrun by the federal government for itscivilian employees. It contracts with188 plans each year to cover about 9million lives, of which approximately31 percent are annuitants (Quayle2003). FEHBP requires annual bidsubmissions from plans andnegotiates with plans to determinepremiums and benefit packages. Overthe last 10 years, FEHBP’s averagegrowth was slightly higher thanMedicare’s, although for differentperiods within that time frame,growth rates differed (Figure 1-8,next page).10


1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998

Per

enro

llee

change

Real change in spending per enrollee, Medicare and PHI, 1968–2000FIGURE1-7

Note: PHI (private health insurance). Age and gender adjusted. Private health insurance spending includes spending for clinical and professional services received by patients.It excludes administrative costs and profits.


20%

5%

0%

�5%

10%

15%

Medicare PHI

10 FEHBP annual increases are a weighted average of the premiums of all individual and family contracts (including of both active workers and annuitants) calculated atthe end of the annual open season.

• CalPERS is a public agency thatcontracts annually for health benefitscoverage on behalf of 1,100 memberstate and local public agencies inCalifornia. Many public agencies inlower cost markets choose not to joinCalPERS. Approximately 1 millionCalifornia public employees, retirees,and dependents were in CalPERSplans in 1997, 20 percent of themretirees. The rate of growth ofCalPERS’ premiums was lower thanMedicare’s over the last 10 years buthigher over the last 12 years.11

A comparison between Medicare andMedicaid growth is of limited utility giventhe myriad eligibility and payment policyissues that are unique to Medicaid andhave greatly influenced its growth rate.For example, Medicaid’s growth has beeninfluenced by increases in enrollmentacross all eligibility categories in the early1990s; state use of financing mechanisms,such as provider taxes anddisproportionate share payments;escalating prescription drug costs; andfluctuations in the economy that affecteligibility. In addition, there is widevariation in the amount of resources usedby Medicaid enrollees, depending on age

and eligibility category. On a per-enrolleebasis, Medicaid spending grew at roughlythe same pace as Medicare between 1987and 2001, and has grown at a slower pacethan Medicare recently.

Comparisons with the Department ofVeterans Affairs (VA) and theDepartment of Defense (DoD), significantpublic purchasers of health care services,are also not particularly apt. The VAdiffers from Medicare in that it owns andmanages its own hospitals and clinics andoperates within a capped budget. DoDalso owns and operates some facilities,although it relies increasingly onTRICARE—a managed care entity that


Change

6%

14%

8%

4%

2%

0%1992–1996 1997–2000 2001–2002

10%

12%

CalPERSMedicare PHI MedicaidFEHBP

Change in spending per enrollee, Medicare and other purchasers, selected yearsFIGURE1-8

Note: PHI (private health insurance), FEHBP (Federal Employees Health Benefits Program), CalPERS (California Public Employees' Retirement System). Changes in spendingare nominal. Private health insurance spending excludes profits and spending on administration.

Source: CMS, Office of the Actuary 2002, Medicare, PHI, and Medicaid (not including SCHIP) spending; FEHBP 2003, FEHBP premium increases;CalPERS 2002, CalPERS premium increases.

11 CalPERS’ increases are a weighted average of the premiums of all individual and family policies calculated at the beginning of the annual open enrollment period forall enrollees except Medicare beneficiaries. CalPERS has a separate benefit design and associated premium for its retirees who are eligible for Medicare.

employs private-sector contractors—todeliver care to its Medicare beneficiaries,and operates within a capped budget (seeChapter 5 for further discussion).

Factors affecting Medicare andother health spending growth

Growth in aggregate public and privatespending for health care are influenced bymany of the same underlying factors, butsome dynamics affect one sectordifferently than the other. The mostsignificant underlying factors that the twosectors share are inflation and increases inthe volume and intensity of servicesdelivered. Increases in volume andintensity (that is, shifts in the compositionof services toward those that are moreresource intensive) are due totechnological developments and consumerdemand, among other factors.

New technologies tend to increase costs,on balance, because they often mean thatmore services can be performed and morepeople can benefit from them.12 As aresult, total spending increases eventhough the unit cost of services maydecline. New technologies may alsoreplace less expensive technologies.Because these costlier technologies mayoffer only marginal improvement inpatient outcomes, the increased spendingis not necessarily offset by reducedspending on subsequent care. Of course,some new technologies may yield somesavings. In particular, some suggest thatnew technologies that improve the processof health care delivery, such as electronicmedical records and physician order entrytechnology, are likely to result in savings.However, because they have start-up costsand have not been widely implemented,their savings potential has not been fullytested.

Increases in consumer demand forservices also lead to increases in volumeand intensity. Because individuals are

shielded from much of the cost of theircare, they tend to use more than theywould otherwise. Similarly, physicians,who often direct beneficiaries’ care, maybe insensitive to costs when makingtreatment decisions. Second, increases inincome, as experienced in the 1990s, tendto increase demand for health careservices. A third factor is beneficiaries’changing expectations about their healthstatus as they age. Beneficiaries do notview illness and debilitation as anecessary part of the aging processanymore. Instead, beneficiaries expect thatmedical services should enable them toretain their health and mobility, and evenagility, as they age (Alliance for AgingResearch 2001).

The aging of the population and impact ofincreased managed care enrollment areexamples of dynamics that can affect thetwo sectors differently. While growth inthe nation’s population has been a steadyand comparatively small factor drivingoverall health care spending for thepopulation under 65 years of age(Ginsburg 2002), the looming retirementof the baby boom generation is certain todramatically affect Medicare’s spending.Medicare spending is greatly influencedby both the number of people over 65 andthe increased longevity of those people.Accordingly, with the leading edge of thebaby boom generation becoming eligiblefor Medicare in 2011 and life expectancyat age 65 projected to increase by 20–25percent between now and 2075, Medicarespending is expected to increasesignificantly over the long term. In fact, asa result of these demographic shifts, theproportion of the nation’s population over65 is expected to nearly double by 2075(from 12 percent to 23 percent by 2075)(CBO 2002b).

Throughout the 1990s, the private sector(and other public purchasers) turned tomanaged care as a way of controlling

spending growth. In a marketcharacterized by excess capacity amongproviders, managed care plans were ableto negotiate lower prices per service and,to a lesser extent, reduce the number ofservices provided. In contrast, Medicare’spayment method for managed careservices prevented the Medicare programfrom capturing any direct savings frommanaged care.13 In fact, increases inmanaged care enrollment led to increasedMedicare spending because of Medicare’sinability to appropriately adjust paymentsto reflect the relative health status ofmanaged care enrollees.

Implications of Medicarespending given limitedresources

Assessing the implications of spendinggrowth requires an understanding of thenature of resource constraints and ofaccompanying pressures on policymakersto make choices in allocating resources.Among the resource constraints affectingMedicare spending are the federal budget,the Medicare trust funds, the size of theeconomy, and beneficiaries’ ability toafford to pay the costs of their care.

The federal budget Medicare is an increasingly large portionof the federal budget, leaving fewerresources available for other spendingpriorities. Current and anticipated annualbudget deficits tend to increase pressureon policymakers to make choices aboutspending and find sources of budgetsavings. Because Medicare is such a largepart of the budget, policymakers oftenlook to savings from Medicare to reducebudget deficits.

Throughout the 1980s, Medicare programoutlays accounted for between 6 and 8percent of total federal spending. Over the


12 For example, cataract procedures rose from 334,000 performed on an impatient basis in 1980 to 1,487,000 performed either in hospital outpatient departments or inambulatory surgical centers in 1996 (Moon 1999).

13 Indirect fee-for-service savings may have resulted from increased managed care enrollment overall to the extent managed care plans induced providers to adopt moreconservative practice styles in caring for all their patients, including Medicare beneficiaries. This indirect savings is called the “spillover effect” (Baker 1997).

course of the 1990s, Medicare’s shareincreased sharply to 13 percent in 1997,dipping 1 percent in the period followingthe BBA, then returning to 13 percent by2001 (Figure 1-9).

According to the CBO, Medicarespending is projected to remain at about13 percent of federal spending until 2007,when it is expected to grow faster thanoverall spending, reaching 16 percent oftotal spending by 2012. While projectionsof Medicare spending as a percentage oftotal federal spending provide a sense ofthe direction of the trend, they areinherently uncertain and may change ifcurrent law changes.

The Medicare trust fundsThe Medicare program is financedthrough two trust funds: the HospitalInsurance (HI) trust fund for Part Aservices and the Supplementary MedicalInsurance (SMI) trust fund for Part Bservices. Unlike the SMI trust fund, the HItrust fund can be exhausted if spendingexceeds revenue plus reserves.14 Once theHI trust fund is exhausted, Medicare stopspaying its bills for Part A services. Thepending insolvency date therefore exertspressure on policymakers to balance trustfund revenue and spending to ensurecontinued operation of much of theprogram.

In recognition of the uncertainty ofprojections, the Medicare Trustees, whoare responsible for reporting on the statusof the Medicare trust funds, make a low-cost, high-cost, and intermediateprojection. Solvency dates are reassessedannually and are subject to substantivechange from year to year. Economic andlegislative changes can quickly alterprojections of solvency, in much the sameway that they alter total annual federalbudgetary surplus or deficit projections.

The HI fund is projected to becomeinsolvent in 2030 under the Trustee’sintermediate estimate. Costs are projectedto begin exceeding tax revenues in 2016,


14 The HI fund’s receipts come primarily from current payroll taxes (87 percent in 2001) and interest earnings on assets held by the trust fund (8 percent in 2001), with theremainder from beneficiary premiums, income taxes on Social Security benefits, and other sources (approximately 5 percent in 2001).

Medicare spending as share of the federal budget, 1980–2012FIGURE1-9

Note: Federal budget includes spending on Social Security.

Source: Congressional Budget Office, 2002.

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012

8%

6%

2%

4%

0%

14%

12%

10%

18%

16%

Med

icare

share


Fiscal year

requiring the fund to use interest incometo pay some costs. In 2021, projectedcosts would exceed all HI income, so trustfund assets would need to be spent tomeet costs. Finally, the HI fund assets areprojected to be exhausted in 2030. Underthe Trustees’ low estimate, the HI fundwould remain solvent throughout the75-year projection window (ending in2076). Under the high-cost estimate,however, it would be exhausted in theyear 2018 (Table 1-2).

In contrast to the HI fund, the SMI fund,financed primarily by federal generalrevenues and beneficiary premiums, isdesigned to remain solvent indefinitely.Current law automatically sets annualfinancing to cover SMI’s expected costsfor the upcoming year plus a “contingencyreserve.” However, as Medicare’sbeneficiary population grows with theretirement of the baby boom generation,and as health care costs continue to rise,the SMI fund is expected to requireincreasing amounts of general revenueand substantial increases in beneficiarypremiums.

In addition, the trust fund financingstructure affects the distributional impactof any policy and may encourage certaintypes of policy decisions. For example, ifextending the solvency date of the HI trustfund is paramount, either spendingreductions on Part A services or changesin the 2.9 percent payroll tax on workerwages (half of which is paid by employersand half of which is paid by employees)that finances the HI trust fund must bepursued. On the other hand, if the goal isto reduce beneficiary premiums, changesin Part B spending are needed. From abudgetary perspective, changes to Part Bresult in relatively smaller changes to thebudget, because 25 percent of the changewould be offset by premium changes.

The economyMedicare spending is growing as apercentage of the nation’s economy, asmeasured by the gross domestic product(GDP). Depending on one’s point of view,

Medicare’s growth may signal thenation’s collective preferences, a programgrowing out of control, or something inbetween. Regardless of one’s point ofview, however, this growing portionhighlights the need to improve the valuegained from increased spending.

For the historical period 1980 to 2001,Medicare’s share of GDP rose from 1.2percent in 1980 to a high of 2.5 percent in1997 (Figure 1-10, p. 16). As a result ofspending reduction provisions in theBBA, increased fraud and abuse scrutiny,and strong economic growth, Medicarespending declined slightly as a share ofGDP to 2.2 percent in 2000. However,after passage of legislation that temperedpreviously enacted payment reductions, ithas since risen to 2.4 percent in 2001 andis projected to increase steadily to 2.8percent by 2012. It is estimated that by2030 Medicare will climb to 5.4 percentof GDP. When the three big entitlementprograms—Medicare, Social Security,and Medicaid—are taken as a whole, theywill account for 14.7 percent of GDP by2030 (Crippen 2002b). Because thesefigures exclude spending by beneficiaries,or on behalf of beneficiaries by Medicaidor private insurers, for coinsurance anddeductibles associated with the Medicarebenefit package, the total share of GDPrelated to Medicare-covered serviceswould be even higher.

Medicare growth of this magnitude raisesquestions about how these costs will beborne by taxpayers and beneficiaries in

the future. If Medicare’s spending werefinanced by raising taxes or increasingbeneficiary contributions, less disposableincome would be available forconsumption or investment. Raisingpayroll taxes affects all workers, butparticularly affects low-income workersbecause the payroll tax is not graduated;raising income taxes would likely affectincome groups more progressivelybecause income taxes are calculated as agraduated percentage; and raisingpremiums affects beneficiaries exclusivelyand would have different distributionaleffects depending on whether the increasewere adjusted by income. Alternatively,Medicare’s growth could be financed bymore borrowing. In that case, morecapital would be invested in government-issued debt and less would be availablefor private investment, which in turncould slow economic growth.

Beneficiaries’ ability toabsorb health care costsLike other people, many beneficiarieshave limited ability to absorb rising healthcare costs. Although beneficiaries 65years of age and older have lower povertyrates than younger people, most elderlyhouseholds—56 percent in 1999—haveincomes below $20,000. On average,these households spend 25 percent of theirincome on health care (CMS 2002).Beneficiary out-of-pocket spending onhealth care includes direct spending onuncovered services, cost-sharing for


Medicare HI trust fund solvency projections

Year outgo exceeds income Year HI trust fundfrom payroll taxes assets exhausted

EstimateHigh 2008 2018Intermediate 2016 2030Low * *

Note: HI (Hospital Insurance). *Not exhausted within the 75-year projection period (ending 2076).

Source: 2002 annual report of the Boards of Trustees of the Medicare trust funds; CMS, Office of the Actuary, 2002.

T A B L E1-2

Medicare-covered services, payments forMedicare Part B premiums, and paymentsfor supplemental insurance premiums.Because there is a potential for high out-of-pocket spending, the vast majority ofbeneficiaries have supplemental insurancecoverage (see Chapter 5 for furtherdiscussion).

Beneficiaries’ resource constraints areimportant to keep in mind when assessingthe level and distribution of out-of-pocketspending and evaluating policy options.Changes in the scope of Medicare’scoverage and levels of cost-sharing affectbeneficiaries’ out-of-pocket spending. Inaddition, beneficiaries’ out-of-pocketspending is directly affected by changes inpayment for Part B services becausecoinsurance for Part B services is

calculated, in general, as 20 percent ofpayment and Part B premiums arecalculated as 25 percent of total Part Bspending.

Extent of Medicare coverage

Medicare provides considerable financialprotection to its enrollees, butbeneficiaries are at risk for substantial out-of-pocket costs. For all beneficiaries,including the institutionalized and those inMedicare�Choice (M�C), Medicarecovered 52 percent of total costs, or$9,573, in 2000. On average, beneficiarieswho were in the traditional fee-for-serviceprogram and living in the communityconsumed $8,200 in health care servicesin 2000, of which Medicare covered 57percent.

While the proportion of beneficiaries’health care costs covered by Medicare hasremained largely unchanged since 1993for institutionalized beneficiaries andthose in managed care, the proportion forfee-for-service beneficiaries living in thecommunity has declined from 63.2percent in 1993. This decline may resultfrom several factors, including an increasein the working aged, for whom Medicareis the secondary insurer, and an increasein the proportion of the disabled, forwhom Medicare pays a smaller proportionof total costs than for the aged. However,much of this change is attributable togrowth in out-of-pocket spending onprescription drugs—a trend that can beexpected to continue absent legislativechange. CBO estimates that spending per


Medicare spending as share of GDP, 1980–2030FIGURE1-10

Note: GDP (gross domestic product).

Source: Congressional Budget Office, 2002; years 2020 to 2030 are from 2002 annual report of the Boards of Trustees of the Medicare trust funds, and are presentedon a calendar year basis.

1980 1984 1988 1992 1996 2000 2004 2008 2016 2020 2024 20282012

3.0%

3.5%

2.0%

2.5%

1.0%

1.5%

0.0%

0.5%

4.0%

5.0%

4.5%

Med

icare

share

Fiscal year


Medicare beneficiary for prescriptiondrugs will increase from $2,439 in 2003 to$5,816 in 2012, an average annual changeof 10.1 percent (CBO 2002a).

According to a MedPAC analysis ofMedicare Current Beneficiary Survey(MCBS) data, growth in out-of-pocketcosts for fee-for-service beneficiariesliving in the community has outpacedgrowth in their income and the largestsource of out-of-pocket growth has beenfor noncovered services. Between 1993and 2000, growth in beneficiaries’ out-of-pocket spending was slightly faster (5.4percent on average) than their growth inincome (3.8 percent on average). Morethan three-quarters of growth in out-of-pocket spending in this time period wasdue to increased spending on noncoveredservices and supplemental insurancepremiums.

On average, beneficiaries spend about 20percent of their income on health careservices, but it is perhaps more useful toconsider the distribution of spending byincome. Households with incomes lessthan $10,000 in 2000 spent 29 percent oftheir income on health care, andhouseholds with incomes between$10,000 and $19,000 spent 22 percent oftheir income on health. In contrast,households with incomes greater than$70,000 spent 5 percent of their incomeon health care (CMS 2002).

Entities that subsidize supplementalcoverage also find it difficult to keep upwith rapidly growing health care costs.Medicaid provides assistance to certainlow-income beneficiaries by providingcoverage for benefits that Medicare doesnot cover and paying for beneficiaries’Medicare premiums and/or cost-sharingfor Medicare-covered benefits, dependingon beneficiaries’ income and stateeligibility income thresholds. Growth inthese costs has contributed to recent statebudget strains and deficits. Employers arealso affected to the extent that they offersupplemental coverage for their retirees.Recent surveys indicate that they areconsidering reducing this coverage oreliminating it for new employees (KaiserFamily Foundation 2002).

Assessing the implications

Because beneficiaries differ in their use ofservices, access to supplemental insurancecoverage, and ability to afford their care,the current burden of out-of-pocketliability and spending varies. Any policychanges would have different implicationsfor different types of beneficiaries. Toassess the distributional implications ofgrowth in beneficiary out-of-pocketspending, policymakers must considerthese characteristics and theirinterrelationships.

Out-of-pocket spending is concentratedamong a minority of beneficiaries, thoughless so than Medicare spending. In 2000, 5percent of all beneficiaries account for 20percent of total out-of-pocket spending.The highest levels of out-of-pocketspending are related to higher levels ofspending for noncovered services.Spending for noncovered servicesaccounted for nearly 46 percent of out-of-pocket spending for beneficiaries in thehighest quartile, while out-of-pocketspending for noncovered services hoveredaround 30–35 percent of total out-of-pocket spending for all other beneficiaries(Figure 1-11, p. 18).

In general, MCBS data show thatMedicare beneficiaries who have low out-of-pocket spending fit one of two profiles.The first group includes relatively youngand healthy people, between ages 65 and74, for instance, and disabled beneficiarieswho have stable conditions and use fewservices. Within this group are peoplewho have only Medicare coverage andthose who have additional coverage butdo not pay the associated premiums. Thesecond group includes people withcomprehensive supplemental coverage,including beneficiaries eligible forMedicaid, and relatively high-incomepeople with good employer-sponsoredcoverage who pay a small or no portion ofthe premium.

In contrast, people who have high out-of-pocket spending pay more forsupplemental coverage and noncoveredservices. They tend to be older, use manyservices, and have relatively high

incomes, and they are more likely to havesupplemental coverage, primarilyMedigap that does not pay much of theirnoncovered services. Accordingly, to theextent that employers reduce thesupplemental coverage they offer, affectedbeneficiaries may buy Medigap coverage,but between higher premiums and lesscomprehensive coverage, they will paymore out-of-pocket.

Spending and otherimplications of MedPAC’srecommendations

Given limited budgetary, economic, andbeneficiary resources, MedPAC’srecommendations should be made andconsidered by policymakers with anunderstanding of their consequences forspending as well as for beneficiaries andproviders. Accordingly, a few changesfrom previous MedPAC reports will beevident in the pages that follow. First, inthis report, we will make the implicationsof MedPAC’s policy recommendationsprominent in the text.

Second, where applicable, MedPAC willprovide one- and five-year estimates ofspending change for its recommendations,expressed as being within one of severalpredetermined dollar ranges (Table 1-3, p. 18). In the past, our estimates ofspending impact were often expressed as apercentage increase in baseline spendingor were discussed in general terms. Thisnew approach is intended to give readers abetter and more direct sense of thepotential spending impact of a givenpolicy recommendation.

MedPAC recognizes that otherorganizations, including CBO, CMS’Office of the Actuary (OACT), OMB, andthe Medicare Trustees, specialize in andhave a legislated role in forecastingMedicare spending and estimating theimpact of policy options. MedPAC’sestimates are intended only to aid readersin considering the implications and scaleof a given recommendation. They are notformal budget or trust fund estimates.MedPAC will consult, or work in tandem,


with CBO and the OACT to inform theestimates and reduce the likelihood ofwidely different estimates for the samepolicies. Nevertheless, separatelyproduced CBO or OACT estimates are inno way constrained by MedPACestimates.

We have elected to express our estimatesof spending changes in terms of a one-and five-year dollar range. One-yearestimates may be particularly relevant forour payment update recommendations,where we fully expect to revisit therecommendations for the following year.Five-year estimates may be more helpfulfor more long-term policyrecommendations, particularly those that

include a phased-in approach that delaysrealization of the full spending impactbeyond the first year.

We are presenting a range for eachestimate, rather than a point estimate, forseveral reasons. First, because MedPAC’sestimates are intended to give readers onlya sense of scale, ranges are more realisticindications of impact than point estimates(see text box). Second, many of ourrecommendations are not sufficientlydetailed to produce a point estimate.Third, we hope that by presenting a range,we reduce any possible confusion betweenour estimates and those of CBO or theOACT.


Per

capita o

ut-

of-

pock

et s

pen

din

g (

dolla

rs)

Groups of beneficiaries ranked by out-of-pocket spending (percentile ranges)

3,000

6,000

4,000

2,000

1,000

0� 25 75 to 10025 to 50 50 to 75

5,000

Part B premiums

Supplemental premiums Noncovered services

Cost sharing

Composition of out-of-pocket spending, by out-of-pocket spending level, 2000FIGURE1-11

Note: Sample of 9,577 includes community-dwelling beneficiaries who participated in traditional Medicare in 2000. Out-of pocket spending includesbeneficiaries' direct spending in four categories: the Part B premium, cost sharing for covered services, supplemental premiums, and noncovered services.The vertical bars represent per enrollee out-of-pocket spending, divided into the four categories, for each group. For example, the � 25 group illustratesper enrollee out-of-pocket spending for beneficiaries with the 25 percent smallest values (the lowest quartile). Likewise, the 75 to 100 group illustratesper enrollee out-of-pocket spending for beneficiaries with the 25 percent largest values (the highest quartile).

Source: MedPAC analysis of Medicare Current Beneficiary Survey, Cost and Use file, 2000.

Dollar ranges for one- and five-year spending estimates

1-year estimates 5-year estimates

No spending No spending� $50 million � $250 million$50–$200 million $250 million–$1 billion$200–$600 million $1 billion–$5 billion$600–$1.5 billion $5 billion–$10 billionOver $1.5 billion over $10 billion

T A B L E1-3

Three other caveats should also beconsidered. First, the spendingimplications for each recommendationhave been developed as if the policy werethe sole change. If other policy changeswere to be made simultaneously, therecould be interactions that would influencethe spending implications. Accordingly,we caution against attempts to add up thespending implications acrossrecommendations. Second, our estimatesdo not reflect the impact on spending forother programs, such as Medicaid, VA, orDoD, and as such do not approximateformal budget estimates. Third,differences may arise between what isintuitively thought to affect spending andwhat is considered “scorable” forpurposes of budget laws. For example,CBO generally scores changes in law, notchanges in administrative policy. �


Why spending estimates may change

Spending estimates may changebecause of the considerableuncertainty in projecting future

spending and the complex technicalaspects of such projections. For thisreason, the Commission is providingsome background information onestimating Medicare spending.

First, spending estimates depend onassumptions about Medicare spendingabsent any new policy changes. Thesespending assumptions define thecurrent law baseline (also referred tosimply as the baseline). Threeversions of the current law baselineare produced separately by CBO, theOffice of the Actuary for OMB, andthe Medicare Trustees, and each isupdated at least once a year toincorporate new assumptions aboutspending or the impact of recentlegislative or regulatory changes tothe program.

Sometimes the baseline will changesignificantly based on new

information about the use and/or mixof services or the prices paid forservices. Accordingly, an estimatethat was based on a baseline includingone set of assumptions may be verydifferent if the underlyingassumptions change. For example,baseline assumptions about M�Cenrollment have changeddramatically. A policy change toM�C payments will have differentimplications now than it would havein 1998 when enrollment in M�Cwas higher than its current level.

Second, estimating the behavioralresponse of providers andbeneficiaries to a policy proposal ishighly imprecise. Different estimatesare likely based on differentassumptions about whether the policywill, for example, increase or decreasethe volume of services delivered.Differences in these assumptions canresult in major changes in thespending estimate for the policy. �

References

Alliance for Aging Research. Great expectations: Americans’ views on aging.Washington (DC), Alliance for Aging Research. May 2001. Results of a national surveyon aging.

Andersen G, Robert Wood Johnson Foundation. Partnership for solutions: better lives forpeople with chronic conditions. Written testimony before the Subcommittee on Health,U.S. House of Representatives Committee on Ways and Means. April 16, 2002.

Baker LC. The effect of HMOs on fee-for-service health expenditures: evidence fromMedicare, Journal of Health Economics. 1997, Vol. 16, No. 4, p. 453–481.

Centers for Medicare & Medicaid Services. Chartbook. Baltimore (MD), CMS. June2002, Section III.B.5., p.10.

Congressional Budget Office. Issues in designing a prescription drug benefit forMedicare: a CBO study. Washington (DC), CBO. October 2002a, p. 4.

Congressional Budget Office. The looming budgetary impact of society’s aging, long-range fiscal policy brief. Washington (DC), CBO. July 3, 2002b.

Crippen DL, Congressional Budget Office. Disease management in Medicare: dataanalysis and benefit design issues. Written testimony before the U.S. Senate SpecialCommittee on Aging. September 19, 2002a.

Crippen DL, Congressional Budget Office. Projections of Medicare and prescription drugspending. Written testimony before the U.S. Senate Committee on Finance. March 7,2002b.

Ginsburg PB, Center for Studying Health System Change. Looking behind the numbers:what’s driving health care costs. Written testimony before the Subcommittee onEmployer-Employee Relations, U.S. House of Representatives Committee on Educationand the Workforce. June 18, 2002.

Hewitt Associates. Health care cost increases expected to continue double-digit pace in2003. Lincolnshire (IL), press release. October 14, 2002.

Kaiser Family Foundation and Health Research and Educational Trust. Employer healthbenefits: 2002 summary of findings. Washington (DC), Kaiser Family Foundation andHRET. September 2002, available at http://www.kff.org/content/2002/20020905a/3252a.pdf.

Medicare Payment Advisory Commission. Medicare beneficiaries’ costs and use of carein the last year of life. Washington (DC), MedPAC. May 2000.

Mercer WM. Mercer/Foster Higgins national survey of employer-sponsored health plans2000. New York (NY), William M. Mercer, Inc. 2002.

Moon M. Beneath the averages: an analysis of Medicare and private expenditures.Washington (DC), Henry J. Kaiser Family Foundation. September 1999, p. iii and 11.

Quayle, James R, Office of the Actuary. Electronic communication with the Office ofPersonnel Management. Washington (DC). January 10, 2003.


Robinson J. Renewed emphasis on consumer cost sharing in health insurance benefitdesign, Health Affairs. March 20, 2002.

Strunk BC, Ginsburg PB, Gabel JR. Tracking health care costs: growth accelerates againin 2001, Health Affairs. September 25, 2002, p. W307.

Towers Perrin. Towers Perrin forecasts 15 percent increase in health care costs. NewYork (NY), press release. October 2002.


Assessing payment adequacyand updating payments infee-for-service Medicare

C H A P T E R2

R E C O M M E N D A T I O N S

Section A: Hospital inpatient and outpatient services

2A-1 The Secretary should add 13 DRGs to the post-acute transfer policy in fiscal year 2004and then evaluate the effects on hospitals and beneficiaries before proposing furtherexpansions.

*YES: 15 • NO: 1 • NOT VOTING: 1 • ABSENT: 0

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-2 The Congress should enact a low-volume adjustment to the rates used in the inpatientPPS. This adjustment should apply only to hospitals that are more than 15 miles fromanother facility offering acute inpatient care.

YES: 17 • NO: 0 • NOT VOTING: 0 • ABSENT: 0

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-3 The Secretary should reevaluate the labor share used in the wage index system thatgeographically adjusts rates in the inpatient PPS, with any resulting change phased inover two years.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-4 The Congress should raise the inpatient base rate for hospitals in rural and other urbanareas to the level of the rate for those in large urban areas, phased in over two years.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-5 The Congress should raise the cap on the disproportionate share add-on a hospital canreceive in the inpatient PPS from 5.25 percent to 10 percent, phased in over two years.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-6 The Congress should increase payment rates for the inpatient PPS by the rate of increasein the hospital market basket, less 0.4 percent, for fiscal year 2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-7 The Congress should increase payment rates for the outpatient PPS by the rate of increasein the hospital market basket, less 0.9 percent, for calendar year 2004.


Section B: Physician services

2B The Congress should update payments for physician services by the projected change ininput prices, less an adjustment for productivity growth of 0.9 percent, for 2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Section C: Skilled nursing facility services

2C-1 The Secretary should continue a series of nationally representative studies on access toskilled nursing facility services (similar to studies previously conducted by theDepartment of Health and Human Services’ Office of Inspector General).


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2C-2 The Congress should eliminate the update to payment rates for skilled nursing facilityservices for fiscal year 2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2C-3A Consistent with previous MedPAC recommendations, the Secretary should develop anew classification system for care in skilled nursing facilities.

Because it may take time to develop this system, the Secretary should draw on new andexisting research to reallocate payments to achieve a better balance of available resourcesbetween the rehabilitation and nonrehabilitation groups.

To allow for immediate reallocation of resources, the Congress should give the Secretarythe authority to:� remove some or all of the 6.7 percent payment add-on currently applied to the

rehabilitation RUG–III groups.� reallocate money to the nonrehabilitation RUG–III groups to achieve a better balance

of resources among all of the RUG–III groups.

2C-3B If necessary action does not occur within a timely manner, the Congress should providefor a market basket update, less an adjustment for productivity growth of 0.9 percent, forhospital-based skilled nursing facilities to be effective October 1, 2003.


*COMMISSIONERS’ VOTING RESULTS

Section D: Home health services

2D-1 The Secretary should continue a series of nationally representative studies on access tohome health services (similar to studies previously conducted by the Department ofHealth and Human Services’ Office of Inspector General).


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2D-2 The Congress should extend for one year add-on payments at 5 percent for home healthservices provided to Medicare beneficiaries who live in rural areas.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2D-3 The Congress should eliminate the update to payment rates for home health services forfiscal year 2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Section E: Outpatient dialysis services

2E The Congress should update the composite rate payment by the projected change in inputprices, less 0.9 percent, for calendar year 2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Section F: Ambulatory surgical center services

2F-1 The Secretary should expedite collection of recent ASC charge and cost data for thepurpose of analyzing and revising the ASC payment system.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2F-2 The Congress should eliminate the update to payment rates for ASC services for fiscalyear 2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2F-3 Until the Secretary implements a revised ASC payment system, the Congress shouldensure that payment rates for ASC procedures do not exceed hospital outpatient PPS ratesfor those procedures, after accounting for differences in the bundle of services covered.



he law requires MedPAC to develop payment update recom-

mendations for each major service sector in fee-for-service

Medicare. While the process of setting updates is inherently im-

perfect, we have developed a framework to help us formulate our

recommendations in the most thoughtful and consistent way possible. Our model

breaks the process into two parts: assessing the adequacy of current Medicare

payments and accounting for the increase in efficient providers’ costs in the com-

ing year. We also take current law into account. We applied our updating model

to services in seven sectors: hospital inpatient, hospital outpatient, physician,

skilled nursing facility, home health, outpatient dialysis, and, for the first time,

ambulatory surgical center. Generally we found that current payments are at least

adequate—and in some cases more than adequate—in these sectors. For physi-

cian payments, however, our finding of adequate payments is linked to

Congressional action to provide a modest increase in payments for 2003.

T

C H A P T E R

Assessing payment adequacyand updating payments infee-for-service Medicare

2In this chapter

• Hospital inpatient andoutpatient services

• Physician services

• Skilled nursing facilityservices

• Home health services

• Outpatient dialysis services

• Ambulatory surgical centerservices


The goal of Medicare payment policy is toalign payments with efficient providers’marginal costs of furnishing health care,and in so doing to help ensurebeneficiaries’ access to high-qualityservices. Achieving this goal involvessetting the base payment rate (for servicesof average complexity) at the right level,developing payment adjustments toaccurately reflect cost differences amongtypes of services and for varying marketconditions, and then annually consideringthe need for a payment update.

MedPAC’s general approach to paymentpolicy attempts to:

• make enough funding available forpaying providers to preserveMedicare beneficiaries’ access tohigh-quality care, and

• distribute payments accurately acrossservices and among providers indifferent health care markets.

The Commission’s annual updatedecisions address the first of theseobjectives. Other recommendationsaddress distributional issues. Often thesewill coincide with the updating processbecause policy changes affecting thedistribution of payments can also affectthe overall amount of payments.

In practice, we have no way of measuringproviders’ marginal costs or determiningthe costs associated with efficientoperation. But the law nonethelessrequires MedPAC to develop paymentupdate recommendations for each majorservice sector in fee-for-service Medicare.Consequently, we have developed aframework to guide our update decisionmaking, so as to carry out this inherentlyimperfect process in the most thoughtfuland consistent way possible.

In our model, we sequentially address twoquestions that together determine theappropriate level of aggregate funding fora given payment system:

• Is the current base payment rate toohigh or too low?

• How much will efficient providers’costs change in the next paymentyear?

As shown in Figure 2-1, if the currentbase rate is too high or too low, we willrecommend a compensating percentagechange factor, and we recommend asecond percentage change factor toaccount for cost changes expected duringthe forthcoming year. The two are thensummed to produce our recommendedupdate. As a practical matter, theCommission may not publish thesepercentage factors separately, but weconsider both questions in arriving at ourfinal update recommendation.

This section of the chapter begins byreviewing the basics of our two-partsystem and then discusses two specialissues in updating payments:

• taking current law into account, and

• considering the impact of newtechnology pass-through payments.

The chapter then proceeds through theCommission’s analysis of paymentadequacy and development of update andother recommendations for hospitalinpatient and outpatient, physician, skillednursing facility, home health, outpatientdialysis, and ambulatory surgery services.

Model for assessingpayment adequacy andupdating payments

Our model attempts to separate assessingthe adequacy of current payments fromprojecting likely changes in efficientproviders’ costs for the coming yearbecause commingling these processes hascaused confusion in the past. For example,one of the factors the Commissionbelieved was responsible for hospitalpayments being too high in the 1990s wasunbundling of the payment unit. Hospitalsshifted care at the end of patients’ acuteinpatient stays to other settings, such asrehabilitation or skilled nursing facilities,which reduced hospitals’ costs. TheCommission’s decision to recommendreduced updates in response to thisphenomenon brought charges that theupdates would not adequately coverhospital cost inflation. Publishing thereduction as a response to currentpayments being too high—separate froman allowance for cost growth in thecoming year—might have presented aclearer picture of the rationale for ourrecommendation.

Multiple factors can contribute to a gapbetween current payments and costs,including errors in past forecasts of inputprice inflation, changes in codingpractices, unbundling of the payment unit,

28 As s e s s i ng paymen t adequacy and upda t i ng paymen t s i n f e e - f o r - s e r v i c e Med i ca r e

Approach for assessing payment adequacy andupdating payment rates

FIGURE2-1

How much will efficientproviders' costs changein next payment year?

Is current base paymenttoo high or too low?

Updaterecommendation

Addcomponents

Percentagechangeneeded

Percentagechangeneeded


or other changes in product. The mostimportant issue for our attention iswhether payments are too high or too low,as opposed to how they became so. Butwhen we believe that a specific factor mayhave played a major role in makingpayments too high or too low—particularly in the most recent year—developing an estimate of the effect ofthat factor may help in deciding whetherand how much to adjust for the adequacyof current payments.

Part one: assessing payment adequacy The first part of MedPAC’s approach todeveloping payment updates is to assessthe adequacy of current payments. In mostcases, we address payments for theservices covered by a single paymentsystem (for example, home health orphysician services). When a singleorganization provides services acrossmultiple payment systems, however,cross-subsidization and inaccurateallocation of costs among services maydistort our measures of payments andcosts. The prime examples of thisphenomenon are hospitals (that provideacute inpatient, outpatient, home health,skilled nursing, and inpatientrehabilitation and psychiatric services)and dialysis facilities (that provide dialysistreatments and furnish separately billablemedications to dialysis patients).

In these instances, we assess the adequacyof payments for all the Medicare servicesthat one type of provider furnishes. If wedecide that payments in aggregate are toohigh or too low, we must then also decidehow to distribute the resulting changeamong services. We would do this byadjusting one or more of the applicablebase rates.

As discussed below, MedPAC’s approachto assessing the adequacy of currentMedicare payments includes three steps:

• estimating current payments andcosts,

• assessing the adequacy of currentpayments relative to costs, and

• adjusting current payments via anupdate or distributional change(Figure 2-2).

Estimating current payments and costs

We begin our assessment by estimatingtotal Medicare payments nationally, alongwith the corresponding provider costs oftreating Medicare beneficiaries. Therelationship between costs and paymentsis typically expressed as a margin.1 Thebase margin estimate covers the yearpreceding the one to which our updaterecommendation will apply. In this report,we are estimating payments and costs forfiscal year 2003 to inform our updaterecommendations for fiscal year 2004.

Unfortunately, because of processingdelays caused by changes in the format ofMedicare cost reports, the latest dataavailable to us from providers’ costreports are from fiscal year 2000.Consequently, we have estimated thechanges in both Medicare’s payments andproviders’ costs (assuming a constantvolume of service) from 2000 to 2003.

On the payment side, we first apply theannual payment updates specified in lawfor 2001 through 2003 to our 2000 basenumbers. We then model the effects ofother policy changes that will affect thelevel of payments during this three-yearperiod. For changes other than updates,we also include provisions scheduled togo into effect in the decision year (fiscalor calendar year 2004).2 This allows us toconsider whether current payments wouldbe adequate under all applicableprovisions of current law. Thus, we endup with estimates of what payments in

Steps and factors in assessing payment adequacyFIGURE2-2

For appropriatenessof current costs:

Estimate: Assess: Adjust:

For relationship ofpayments to costs:

• target relationship of payments to efficient providers' costs

Market factors Policy factor

• current Medicare costs

• current Medicare payments

• relationship of payments to appropriate costs

• appropriateness of current costs

• through a distributional change

(if applicable)• through the update

• changes in quality

• beneficiaries' access to care

• providers' access to capital

• entry and exit of providers

• changes in volume of services

• changes in product

• changes in unit costs

1 A margin is calculated as payments less costs divided by payments. Alternatively, the data can be expressed as a ratio of payments to costs.

2 An example of a payment policy scheduled to go into effect in 2004 is eliminating the hold-harmless provision for small rural hospitals under the outpatient prospectivepayment system.

fiscal year 2003 would have been, hadfiscal year 2004 payment rules been ineffect.

On the cost side, we estimate the increasesin costs per unit of output over the samethree-year period—a difficult task giventhat fiscal year 2003 had just started whenwe had to make our decisions. Generallywe assume that cost per unit of output hasincreased at the rate of input priceinflation, as measured by the applicablemarket basket index from the Centers forMedicare & Medicaid Services (CMS),adjusted downward slightly in anticipationof productivity improvments.3 In somecases, however, more recent estimates ofcost growth are available through claimsanalysis and alternative data sources suchas the National Hospital IndicatorsSurvey, which CMS and MedPACcosponsor.

Assessing the adequacy ofcurrent payments relative tocosts

The next step in assessing paymentadequacy involves two interrelated issues:

• the appropriateness of providers’costs (that is, whether actual costsprovide a reasonable representationof the costs of efficient providers),and

• the relationship of payments to anappropriate cost base.

In examining the cost base (aggregatecurrent costs), we generally treat thevolume of services as given. At a certainvolume, providers’ total costs are drivenby the average cost per unit of output,which then becomes the focal point of ouranalysis. If this unit cost is consideredappropriate, we then proceed to thequestion of whether payments areadequate to cover costs and to providesufficient funds for keeping plant andequipment up to date. However, if costsare too high (implying some degree ofinefficiency) or too low (implying that

additional spending is needed to ensureappropriate quality and access to care),then an adjustment to actual costs may beneeded before we decide whetherpayments are adequate in relation to thosecosts.

The tasks of assessing the appropriatenessof the cost base and the adequacy ofpayments inevitably require Commissionjudgment. Available information isinvariably limited. Nonetheless, severaltypes of data about the market conditionsthat providers face may provide usefulclues (Figure 2-2).

Market factors Two market factorsrelate primarily to the appropriateness ofcurrent costs:

• the trend in average cost per unit ofoutput, and

• evidence of change in the productbeing furnished.

Although it is nearly impossible to knowwhether costs are “efficient” in theabsolute, the rate of change in unit costs atleast provides evidence of whether theinitial level of appropriateness has beenmaintained. Other things being equal, wewould generally expect average growth inunit costs to approximate the rate ofincrease in the applicable market basketindex, or be slightly below the marketbasket increase with productivityimprovements. Changes in product canhave a major effect on unit costs,however. For example, substantialreductions in the length or visit content ofhome health episodes would be expectedto reduce the growth in provider costs(inflation adjusted).

Changes in several other market factorsmay suggest that payments are too high ortoo low relative to costs, even in theabsence of any direct evidence as towhether the cost base is appropriate(Figure 2-2):

• changes in the volume of services,

• entry or exit of providers,

• changes in the quality of care,

• changes in beneficiaries’ access tocare, and

• changes in providers’ access tocapital.

Reductions in the volume of servicesfurnished or in the number of providersoffering services to Medicare beneficiariesmay indicate that revenue flows areinadequate for providers to continueoperating or to provide the same level ofservices. Facilities closing is the extremeoutcome, although it can be difficult todistinguish between closures that haveserious implications for access to care in acommunity and those that have resultedfrom excess capacity. Evidence that moreprivately practicing physicians arerefusing to accept new Medicare patientsis another example. By the same token,substantial increases in volume or thenumber of providers may indicate thatpayments are more than sufficient tocover providers’ financial needs,potentially leading to unnecessary servicesbeing provided.4

Although difficult to measure,deteriorating quality or access to care mayindicate that revenues (either specific toMedicare or across all payers) areinadequate. It is unlikely, however, thatquality measures alone would everprovide the basis for concluding thatMedicare payments are too high. Changesin bond ratings may indicate thatproviders’ access to needed capital hasdeteriorated or improved, although thedata are difficult to interpret becauseaccess to capital depends on more thanjust bond ratings. The industry’s volumeof borrowing and overall level of capitalexpenditures may provide indirectevidence of access to capital.


3 Actual changes in the market basket index were used for 2002 together with CMS’s forecasts of the market basket for 2003.

4 Changes in the volume of physician services must be interpreted cautiously because in this case there is some evidence to suggest that volume goes up when paymentrates go down—the so-called “volume offset.”

Policy factor Apart from market factors,there is a policy factor to be consideredwhen assessing the adequacy of currentMedicare payments—namely, the desiredrelationship between payments and costs(Figure 2-2). Given a judgment that thecurrent level of costs is appropriate, atarget ratio of payment to costs couldsimplify MedPAC’s assessment ofpayment adequacy—if our projection ofcurrent year payments and costs produceda margin above the target, then we wouldrecommend a downward adjustment, andvice versa.

The appropriate margin of payments overappropriate costs—which could be anarrow range, rather than a specificpoint—is difficult to discern. Difficultyarises for several reasons: the degree ofrisk among specific providers variesdepending on their size, the actions ofother payers, their exposure to nonpayingpatients, and other factors. Even onaverage across all providers, however, riskcould vary by sector and over time for agiven sector. Moreover, even if we couldidentify a target aggregate margin, itwould still be only one element of acomposite picture that is also informed bythe other factors described above (theeffects of changes in product, quality ofcare, access to care, and so forth).

In sum, our deliberations have suggestedthat it will not be possible to develop astandard relationship between paymentsand appropriate costs. Thus, theCommission will still need to think aboutan appropriate range for this relationshipeach year, one sector at a time.

Adjusting current payments viaan update or distributionalchange

A finding that current Medicare paymentsare too high or too low will lead to anadjustment to the payment update thatotherwise would apply. If the adjustmentis large, the Commission typicallyrecommends phasing it in over two ormore years. Sometimes, however, we mayfind it appropriate to increase or decreasethe amount of money in the system in away that simultaneously redistributes

payments. In this case, we would intendthe combined impact of the distributionalchanges and the update itself to providefor an appropriate level of payments in thepolicy year.

It may be useful to quantify a percentageadjustment factor when we find thatcurrent payments are too high or too low.Often, however, the Commission simplymakes clear that current payments are toohigh or too low and then considers thatfinding together with the expected costchange in the coming year (as discussedbelow) in developing its updaterecommendation.

Part two: accounting forproviders’ cost changes inthe coming year The second part of MedPAC’s approachto developing payment updaterecommendations is to account forexpected cost changes in the next paymentyear. This involves reviewing evidenceabout the likelihood and extent of changesin factors that are expected to affectproviders’ costs. One major factor ischange in input prices, as measured by theapplicable CMS price index. Forinstitutional providers, we use theforecasted increase in an industry-specificindex of national input prices, called amarket basket index. For physicianservices, we use a similar index known asthe Medicare Economic Index. Theseindexes approximate how muchproviders’ costs would rise in the comingyear if the quality and mix of inputs theyuse to furnish care remain constant.Several other factors may also affectproviders’ costs in the coming year:

• Scientific and technologicaladvances—Many improvements inmedical science and technologyenhance quality and reduceproviders’ costs (or leave costsunchanged). No increase inMedicare’s payment rates is neededto accommodate these changesbecause providers have a financialincentive to adopt them. But weshould consider the effects of

technological advances that improvequality of care and also increasecosts, when these effects aresubstantial and the technologies arebroadly disseminated. TheCommission monitors industry trendsand has informal discussions withindustry representatives in eachservice area. When evidence suggeststhat one or more technologicaladvances in a specific area areplaying an unusually large role inincreasing providers’ costs, we mayattempt to estimate the cost impact ofthese advances.

• Improvements in productivity—TheCommission believes that providersshould be able to reduce the quantityof inputs required to produce a unit ofservice by at least a modest amounteach year while maintaining servicequality. Productivity gains are oftenachieved by adopting newtechnology. We have adopted thelong-term growth rate forproductivity in the general economyas our standard of expectedproductivity improvement.Specifically, we use the 10-yearaverage annual change in total-factorproductivity as published by the U.S.Bureau of Labor Statistics, which iscurrently estimated at 0.9 percent.

• One-time factors—On occasion, werecommend an adjustment to theupdate to reflect a one-time factorthat has a systematic and substantialeffect on costs and will improve carefor beneficiaries or is necessary foranother reason (such as a legalmandate). Examples of one-timefactors the Commission has takeninto account in the past includeMedicare’s share of the 2000computer problem and the cost ofcomplying with the Health InsurancePortability and Accountability Act of1996.

We generally consider the estimate ofinput price inflation as the most importantfactor influencing providers’ costs,particularly since the costs of


technological advances and improvementsin productivity at least partially offseteach other. This focus on inflation alsoreflects the reality that the costs of newtechnology and productivity gains aredifficult to measure. To the extent thatimportant changes do not get addressedwhen we update payments in a given year,their effects can be considered in ouranalysis of payment adequacy in the nextcycle.

Special issues in updatingpayments

This section addresses two special issuesthat have arisen this year for assessingpayment adequacy and updatingpayments: considering the budgetimplications of potential changes tocurrent law and considering the impact oftechnology pass-through payments.

Budget implications The Commission is aware of—and wedocument in our report—how spendingunder our recommendation wouldcompare to that under current law. Webegin by developing a list of current lawprovisions and changes scheduled to gointo effect in the coming year, by sector,to illustrate any differences betweenMedPAC recommendations and presentpolicy. We also develop rough estimatesof the impact of recommendations relativeto the current budget baseline, placingeach recommendation into one of severalcategories. (Our method of documenting

the budget implications ofrecommendations is discussed in greaterdetail in Chapter 1.)

Considering the impact oftechnology pass-throughpayments For hospital outpatient and inpatientpayments, Medicare makes additionalpayments for specific new technologiesthat have a substantial impact on providercosts. These payments are intended to betemporary, to ensure that Medicare canpay for a new or substantially improvedtechnology during its initial diffusionperiod and until its effects on providers’costs can be reflected in the paymentweights for the affected groups of patientsor procedures.5 After two to three years,during which necessary coding changesare implemented and charge data arecollected from providers, permanentadjustments will be made to the relativepayment weights and the temporarypayment adjustments stopped.

It may be necessary to take technologypass-through payments into account in thesecond part of our update framework—theallowance for expected increases inefficient providers’ costs. However, theimpact of pass throughs on the overalllevel of payments will depend on whetherthey have been implemented in a budgetneutral fashion.

If the payment adjustments are not budgetneutral, which was the case initially withthe outpatient pass-through payments,then they will augment the payment

increase provided by the update. Thismeans that any allowance fortechnological advancement in our updateneed only consider major technologicalcost impacts that are outside the scope ofthe pass-through system. The effect willbe greatest in the first years after pass-through payments are implemented, whennew technologies are approved forpayment adjustments and there are notexisting pass-through technologies readyto be folded into the prospective paymentsystem rates. In later years, the impact onaggregate payments each year will be thenet of new adjustments added and currentadjustments eliminated.

If payments are made budget neutrally,which is the case now for both theoutpatient and inpatient pass-throughpayments, then the net increase in costsresulting from the technologies should beconsidered in developing paymentupdates—but only if they are substantialand systematic. The data from the pass-through payments (utilization andpayment rate for each technology) mayprovide useful input into the decision onhow the impact of cost-increasing newtechnologies compares to expectedproductivity improvement. However,there are several limitations on how wellaggregate pass-through payments willrepresent the overall impact of cost-increasing new technology, such that thedata must be used guardedly. A detaileddiscussion of the treatment of newtechnology in Medicare’s paymentsystems is presented in Chapter 4 �.


5 These are ambulatory payment classification (APC) groups for outpatient payments and diagnosis related groups (DRGs) for inpatient payments.

2AAssessing payment adequacy and updating payments for

hospital inpatient and outpatient services

S E C T I O N


2A-1 The Secretary should add 13 DRGs to the post-acute transfer policy in fiscal year 2004 andthen evaluate the effects on hospitals and beneficiaries before proposing further expansions.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-2 The Congress should enact a low-volume adjustment to the rates used in the inpatient PPS.This adjustment should apply only to hospitals that are more than 15 miles from anotherfacility offering acute inpatient care.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-3 The Secretary should reevaluate the labor share used in the wage index system thatgeographically adjusts rates in the inpatient PPS, with any resulting change phased in overtwo years.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-4 The Congress should raise the inpatient base rate for hospitals in rural and other urban areasto the level of the rate for those in large urban areas, phased in over two years.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-5 The Congress should raise the cap on the disproportionate share add-on a hospital canreceive in the inpatient PPS from 5.25 percent to 10 percent, phased in over two years.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-6 The Congress should increase payment rates for the inpatient PPS by the rate of increase inthe hospital market basket, less 0.4 percent, for fiscal year 2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2A-7 The Congress should increase payment rates for the outpatient PPS by the rate of increasein the hospital market basket, less 0.9 percent, for calendar year 2004.



*

Section 2A: Assessing paymentadequacy and updating payments forhospital inpatient and outpatient services

The Commission finds that Medicare payments for all hospital services are at

least adequate as of fiscal year 2003, even after accounting for legislated changes.

Our conclusion is based on an estimated overall Medicare margin of 3.9 percent

for 2003; broad indicators such as access to capital; and factors affecting costs in

the coming year such as inflation and technological advances. We recommend an

update of market basket minus 0.4 percent for inpatient services, but because

technological advances affecting outpatient services are frequently handled

through new technology provisions, we recommend a lower outpatient update—

market basket minus 0.9 percent. We view our inpatient update as part of a pack-

age that includes five other policy changes aimed at appropriately distributing

payments: extending the post-acute transfer policy; implementing a low-volume

adjustment; reevaluating the labor share used with Medicare’s wage index; elim-

inating the differential in base rates for hospitals in rural and small urban areas;

and increasing the cap on disproportionate share payments. In addition, we are

not satisfied with the current indirect medical education adjustment because it

provides payments well above the empirically justified level without account-

ability, and we will explore ways to target these payments to advance specific

Medicare policy objectives.

2AIn this section

• Assessing payment adequacy

• Policies affecting thedistribution of payments

• Update for inpatient services

• Update for outpatient services


S E C T I O N

In this section of Chapter 2, we presentthe Commission’s analysis of Medicarepayments for hospital services, togetherwith seven recommendations on inpatientand outpatient payments. As background,we begin with an overview of the serviceshospitals provide to Medicarebeneficiaries and of Medicare spending onthese services. We also describeMedicare’s inpatient and outpatientprospective payment systems (PPSs),which account for the bulk of Medicarespending on hospital services.

Next, we analyze the adequacy of Medicarepayments for all hospital services—inpatient,outpatient, and other services—in fiscal year2003. We then discuss the Commission’sfindings and recommendations for Medicarepayments to hospitals under the inpatientPPS for patients transferred from inpatienthospital to post-acute settings, the indirectmedical education (IME) adjustment for thecosts of teaching hospitals, and payments torural hospitals. Finally, we presentMedPAC’s recommendations for updates toMedicare’s hospital inpatient and outpatientPPS payment rates.

Background

Hospitals provide a variety of services toMedicare beneficiaries, but the bulk ofMedicare spending on hospitals is forinpatient and outpatient care. Each year,approximately one-fifth of Medicarebeneficiaries receive hospital inpatientcare, and one-half receive care in hospitaloutpatient departments. Medicarepurchases these and other services fromover 4,800 short-term general hospitalsthat meet its conditions of participationand agree to accept the program’spayment rates as full payment.

The services hospitalsprovide Short-term general hospitals provideMedicare beneficiaries with inpatient carefor the diagnosis and treatment of acuteconditions and manifestations of chronicconditions. They also provide ambulatorycare through outpatient departments andemergency rooms. Many hospitalsprovide home health, skilled nursing

facility (SNF), and rehabilitation servicesfollowing surgery or an inpatient stay formedical care, and many also furnishpsychiatric care.

Medicare spending onhospitals In 2000, about three-fourths of Medicarepayments to hospitals were for inpatientcare and about one-seventh was foroutpatient care, including emergencyroom services (Figure 2A-1). Most of theremaining Medicare payments went forhome health care, care provided by SNFs,and care provided by hospital unitsexempt from the inpatient PPS.

Total hospital spending grew 8.3 percent in2001 after increasing 5.8 percent in 2000.CMS estimates that hospital inflationincreased 3.2 percent in 2001 after growthof 2.6 percent in 2000 (Levit et al. 2003).Total Medicare spending for inpatient andoutpatient care increased from about $83billion in 1992 to $119 billion in 2001(Figure 2A-2). These expendituresincreased 4.2 percent per year over theperiod, growing at annual rates of 4.9percent from 1992 to 1998 and 2.7 percent

from 1998 to 2001. Medicare spent $86billion on services paid under the inpatientprospective payment system in fiscal year2001. The Congressional Budget Office(CBO) projects that PPS inpatient spendingwill increase at an average annual rate of6.2 percent from 2002 to 2007.

Medicare’s payment systems forhospital inpatient and outpatientservicesMedicare has used prospective payment forinpatient services since 1984. Medicareintroduced prospective payment for hospitaloutpatient department services (includingemergency room services) in 2000.

Medicare’s hospital inpatient PPSMedicare’s hospital inpatient PPS payshospitals predetermined amounts perdischarge based primarily on the patient’scondition and market conditions in thehospitals’ location. Medicare assignsdischarges to diagnosis related groups(DRGs), which group patients withsimilar clinical problems that are expectedto require similar amounts of hospitalresources. Separate DRG-based paymentsapply for operating and capital costs.

36 As s e s s i ng paymen t adequacy and upda t i ng paymen t s f o r ho sp i t a l i n pa t i e n t a nd ou t pa t i e n t s e r v i c e s

Medicare hospital payments by majorservice line, 2000

FIGURE2A–1

Note: PPS (prospective payment system), SNF (skilled nursing facility). PPS exempt units include inpatient psychiatric and rehabilitation services. Data are imputed for hospitals whose cost reports were not available (about 27 percent of observations). Excludes critical access hospitals.

Source: MedPAC analysis of Medicare cost report data from CMS.

77%

14%

5%2% 2%

Inpatient

HomehealthSNFPPS

exempt

Outpatient

CMS sets relative weights for 508 DRGs;weights are intended to measure theexpected relative costliness for a patient ineach DRG compared with costs for theaverage Medicare patient.1 The basepayment rate reflects the averagecostliness of Medicare inpatient casesnationwide. The labor share of the basepayment amount (71 percent) is adjustedby a wage index to reflect the relativelevel of input prices in the hospital’s localarea. The product of the hospital’s basepayment rate and the relative weight ofthe DRG to which a patient is assigned isthe hospital’s DRG payment rate.

The inpatient PPS makes special paymentsfor unusual cases and to hospitals withspecific characteristics. These additionaladjustments are intended to recognizedifferences in patient treatment costs or toaccomplish a policy goal. Extremely costlycases qualify for outlier payments inaddition to the regular DRG payment. Anindirect medical education (IME)adjustment accounts for the higher patientcare costs of teaching hospitals. Thedisproportionate share (DSH) adjustment

provides additional payment for hospitalsthat treat an unusually large share of low-income patients. DRG payments arereduced when a patient is transferred toanother PPS hospital, or in some instancesto a post-acute care setting. Specialpayments are made to rural hospitals thatqualify as sole community providers,referral centers, or Medicare-dependenthospitals. Additional payments are madefor new technologies when PPS paymentrates for specific DRGs or cases withinDRGs are inadequate.

Medicare’s hospital outpatient PPS By contrast with the inpatient PPS, theoutpatient PPS pays hospitals apredetermined amount per service. Ahospital receives payment for each inpatientdischarge but a separate payment for eachservice provided during an outpatientencounter. Each service provided to abeneficiary is assigned to 1 ofapproximately 570 ambulatory paymentclassification (APC) groups. The APCsgroup procedures, evaluation andmanagement services, and some drugs usedin hospital outpatient departments. Each

APC has a relative weight based on themedian cost of the services grouped in theAPC. A conversion factor translates relativeweights into dollar payment amounts; theoutpatient payment rate equals the relativeweight for the APC times the conversionfactor. The labor portion of the conversionfactor (60 percent) is adjusted by thehospital wage index to reflect differences inlocal input prices.

The outpatient PPS includes five paymentadjustments. Pass-through payments fornew technologies supplement paymentsfor individual services. These technologiesinclude drugs, biologicals, and medicaldevices used in the delivery of services.Outlier payments are made for individualservices or procedures with extraordinarilyhigh costs relative to the payment rate forthe APC. In addition, certain services areassigned to new technology APCs. Hold-harmless payments are made to cancer,children’s, and small rural hospitals if theiroutpatient PPS payments are lower thanthey would have received under priorpolicy. Hold-harmless payments to smallrural hospitals end in calendar year 2003.Transitional corridor payments, which arealso made through 2003, are intended topartially compensate all hospitals for thedifference between PPS payments andpayments they would have received underprevious policy.

Assessing paymentadequacy

Each year, MedPAC makes paymentupdate recommendations for hospitalinpatient and outpatient services for thecoming fiscal year. To inform ourrecommendations, we consider multiplefactors, including the relationship ofMedicare’s current payments to providers’costs, the appropriateness of providers’current costs, and various marketindicators of payment adequacy. MedPACanalysis finds that aggregate Medicarepayments for all hospital servicesprovided to beneficiaries are at leastadequate as of fiscal year 2003.


1 There are 527 DRGs in 2003, but 19 of these are no longer used for Medicare payment, leaving 508 DRGs in use.

Payments to Medicare providers for all hospitalinpatient and outpatient services, fiscal years

1992–2001

FIGURE2A-2

Note: Includes inpatient services covered by prospective payment (PPS); PPS-exempt inpatient services (psychiatric,rehabilitation, long-term care, cancer, and children's hospitals and units); outpatient services covered byprospective payment; and other outpatient services. Payments include both program outlays and cost-sharingincurred by beneficiaries.


Dolla

rs in

bill

ions

80

100

140

120

40

60

20

019941992 1993 1995 1996 1997 1998 1999 2000 2001

Financial status of hospitals


Hospitals’ general financialhealth is of concern to theCommission because a severe

decline could affect the ability ofhospitals to provide high-quality care toMedicare beneficiaries. For that reason,it is important to monitor the impact onhospitals of the payment policies ofprivate and public purchasers ofhospital care, as well as measures ofhospitals’ general financial status.

During the 1990s, increased pressurefrom private payers on hospitals’revenues was generally credited withproducing low growth in hospital costs.In 1998 and 1999, both private payerand Medicare payments fell relative tocosts, providing additional pressure tocontrol costs in those years (Figure 2A-3). The situation changed in 2000 whenprivate payments increased relative tocosts, while the decline in Medicarepayments relative to costs slowed.

The increase in the private-sectorpayment-to-cost ratio reflects moreaggressive negotiations by providers aswell as shifts by payers and consumersto less-intrusive approaches to caremanagement. Less-restrictive forms ofmanaged care such as preferredprovider organizations (PPOs) havedisplaced health maintenanceorganizations (HMOs) as the dominantprivate insurance models. Plans haveresponded to consumer demand byestablishing broader provider networks.These changes have weakened thebargaining power of plans in dealingwith providers—hospitals have beenwilling to cease contracting withspecific plans to avoid priceconcessions (Strunk 2001).

The total margin for all payers—Medicare, Medicaid, and privatepayers—reflects the relationship of allhospital revenues to all hospital costsincluding inpatient, outpatient, post-acute, and nonpatient services. Thetotal margin does not provide ameasure of the adequacy of Medicarepayments, but it is certainly the mostcomprehensive measure of hospitals’general financial performance.1 Datafrom Medicare cost reports show thatthe average total margin for the periodfrom 1990 through 2000 was 4.6percent. After reaching a high of 6.1percent in fiscal year 1996, the totalmargin fell to 3.4 percent in fiscal year2000 (Figure 2A-4).

The decline in total margins appears tohave halted in 2002. MedPACexamined data from the AmericanHospital Association (AHA) ondevelopments since 2000. The AHAannual survey indicates that the totalmargin fell in 2001 from 4.6 to 4.2percent (Table 2A-1). However, thenational hospital indicators survey,conducted by AHA with funding fromthe Centers for Medicare & MedicaidServices (CMS) and MedPAC,indicates that this decline had stoppedin the first nine months of fiscal 2002.This data source yields a 4.5 percentmargin for fiscal years 2001 and2002. �

1 There is substantial variation in financial reporting among hospitals and between the Medicare cost report and audited financial statements of individualhospitals. These considerations suggest that comparisons of total margins among hospitals and across data sources should be treated with caution (Kane2001).

Trend in hospital total margin, 1998–2002

National hospitalFiscal year Medicare cost report AHA annual survey indicators survey

1998 4.3% 5.8% 4.3%1999 3.8 4.7 2.72000 3.4 4.6 4.72001 N/A 4.2 4.52002 N/A N/A 4.5

Note: AHA (American Hospital Association). Medicare cost report margins are imputed for hospitals whose 2000cost reports were not available (about 27 percent of observations) and exclude critical access hospitals. The2002 value for the national hospital indicators survey is based on three quarters of data and is seasonallyadjusted.

Source: MedPAC analysis of cost report data from CMS, AHA annual survey of hospitals, and the national hospitalindicators survey (sponsored by CMS and MedPAC, conducted by AHA).

T A B L E2A-1


Hospital payment-to-cost ratios for Medicare, Medicaid, and private payers,1991–2001

FIGURE2A-3

Note: Payment-to-cost ratios indicate the relative degree to which payments from each payer cover the costs of treating that payer's patients. Data are for community hospitals andcover all hospital services. Imputed values were used for missing data (about 35 percent of observations). Most Medicare and Medicaid managed care patients areincluded in the private payers category.

Source: Medicare analysis of data from the American Hospital Association's annual survey of hospitals.

1991 1992 1993 1994 1995 1996 1997 1999 200120001998

120

130

100

110

80

90

70

140

Per

cent

Private payersMedicare Medicaid

Trend in hospital total margin, 1990–2000FIGURE2A-4

Note: Data are imputed for hospitals whose 2000 cost reports were not available (about 27 percent of observations). Excludes critical access hospitals.


Per

cent 6

10

8

2

019921990

3.6

1991

4.4

1993

4.4 4.45.0

5.8

1994 1995 1996

6.1

1997 1998 2000

4

1999

6.0

4.3

3.43.8

Current payments and costs One factor the Commission uses to informits Medicare payment updaterecommendations for fiscal year 2004 isthe estimated relationship betweenMedicare payments and costs (margins)for fiscal year 2003. Separate margins canbe calculated for inpatient, outpatient, andall services provided to Medicarebeneficiaries. We use the latest cost reportdata available (fiscal year 2000) fromhospitals as the cost and payment base,and then estimate margins in fiscal year2003 by projecting cost and paymentincreases from fiscal years 2000 to 2003.This involves applying payment updatesin current law and modeling paymentpolicy changes, including those scheduledto take effect in fiscal year 2004. Wecompare payments to costs in fiscal year2003 assuming that all provisions ofcurrent law for fiscal year 2004 are ineffect except the inpatient and outpatientupdates.

Inpatient and outpatientMedicare margins The inpatient and outpatient marginsreflect payments and costs for servicescovered under Medicare’s hospitalinpatient PPS and all outpatient services,respectively.2 The inpatient margin isoverstated and the outpatient marginunderstated because of the way hospitalsallocate their costs between these twosettings. This variation results fromaccounting practices introduced whenMedicare paid prospectively determinedpayments for inpatient services but paidfor outpatient and other services at cost.Research for the Health Care FinancingAdministration (HCFA, now CMS) foundthat outpatient costs might be overstatedby 15 to 20 percent (CHPS Consulting1994).3 This implied that inpatient costswere understated by 3 to 4 percent. Costsfor the other components of hospitalservices are overstated for similar reasons,

implying a small additionalunderstatement on the inpatient side.

From 1999 to 2000, hospital inpatientmargins declined from 12.3 to 10.8percent, and outpatient margins increasedfrom –16.4 to –13.7 percent (Table 2A-2).These changes were accompanied byincreases in the PPS-exempt and homehealth margins and almost no change inthe skilled nursing facility margin.

Overall Medicare margin The overall Medicare margin incorporatespayments and costs for inpatient,outpatient, skilled nursing, home health,psychiatric, and rehabilitative services forMedicare beneficiaries, as well asgraduate medical education and Medicarebad debts.4

The overall margin is available since 1996and the inpatient margin since 1984.Inpatient payments compriseapproximately three-fourths of totalMedicare payments to PPS hospitals. As aresult, the overall margin follows a trendsimilar to that of the inpatient margin(Figure 2A-5). The inpatient marginincreased steadily from 1991 through1997. The overall margin increased aswell, reaching a high point of 10.4 percentin 1997.

The overall Medicare margin was 5.1percent in 1999 and 5.0 percent in 2000.We estimate that the overall Medicaremargin will be 3.9 percent in 2003 (Table2A-3). The overall margin of majorteaching hospitals increased between 1999and 2000 but is expected to decline in 2003,largely because of the scheduled reductionin IME payments. The overall margin ofrural hospitals declined from 1999 to 2000.It is expected to increase by 2003, in partbecause of the increase in disproportionateshare payments implemented in 2001through the Medicare, Medicaid, andSCHIP Benefits Improvement andProtection Act of 2000 (BIPA).

Appropriateness of current costs In general, we find that the hospital costbase as of fiscal year 2003 is appropriate.A number of factors put downwardpressure on costs in the late 1990s and2000, so that hospital costs wereconstrained. Large reductions in length ofstay occurred in the mid-1990s, andrevenue pressure from both private andpublic payers increased in 1998 and 1999.Declining interest rates reduced costs andimproved hospital access to capital.However, as length-of-stay declineslowed, revenue pressure moderated, andwage pressures emerged, Medicare cost-per-case growth increased in 2001.

The most direct indicator of theappropriateness of the hospital cost base isgrowth in Medicare inpatient cost per case


2 Outpatient margins encompass all outpatient services, not just outpatient PPS services. This approach provides consistency over time and reflects the fact that costreporting periods for some hospitals span the implementation of the new payment system in August 2000.

3 The final report of HCFA’s study contains a series of DRG-specific values, rather than a national figure for outpatient cost overstatement. However, the study’s principalinvestigator has estimated that the national figure is between 15 and 20 percent.

4 Because of data limitations, small amounts spent on certain other services, such as ambulance and hospice, are not reflected in the overall Medicare margin.

T A B L E2A-2 Overall Medicare

margin and marginby major service line,

1999–2000

Service 1999 2000

Inpatient 12.3% 10.8%Outpatient –16.4 –13.7PPS-exempt –1.8 0.6Skilled nursing facility –55.9 –57.4Home health –13.1 –9.9Overall 5.1 5.0

Note: PPS (prospective payment system). Data arebased on Medicare-allowable costs. Marginsare imputed for hospitals whose 2000 costreports were not available (about 27 percent ofobservations). Excludes critical access hospitals.PPS-exempt includes inpatient psychiatric andrehabilitation services. Payments and costs forgraduate medical education are included in theoverall Medicare margin but not in the othermargins.

Source: MedPAC analysis of Medicare cost report datafrom CMS.

over time. Growth in hospitals’ Medicarecost per case was modest—less than theincrease in the hospital market basket—from 1993 through 1998. From 1994 to1996, cost-per-case growth was negative.This is partly because from 1990 to 1999hospitals reduced Medicare length of stayabout 33 percent, resulting in lowerresource use. In an earlier study, MedPACfound that during the period of the largestlength-of-stay reductions, each percentagepoint drop in length of stay resulted in acorresponding 0.8 percent drop in realcosts per case (Ashby et al. 2000).

Cost-per-case growth began to accelerateat the end of the 1990s as the decline inlength of stay slowed. Length of stay forall hospital inpatients has continued tofall, though at a more moderate rate, withdeclines of 1.8 percent in fiscal year 1999,1.9 percent in 2000, and 1.3 percent in2001. Similarly, the reductions inMedicare length of stay of 1.3 percent in1999 and 1.9 percent in 2000 trailedannual declines exceeding 5.5 percentfrom 1993 through 1996.

Wages are the largest component of thehospital market basket. As a result, recentwage growth has contributed significantlyto higher overall cost growth. Shortages ofspecific occupational groups, such asnurses, pharmacists, and therapists, havecontributed to this greater wage pressure.Hospital industry wages rose more rapidlythan wages in the general economy in2001 and 2002, reversing a trend ofslower hospital wage growth from 1994 to2000. The employment cost index (ECI)for wages and salaries of hospital workersincreased 5.4 percent compared with anincrease of 3.6 percent for all workers infiscal year 2001, and continued at 4.4percent for hospital workers and 3.2percent for all workers in fiscal year2002.5

Efforts by private payers to exact priceconcessions from hospitals havemoderated as the expansion of less-intrusive forms of managed care hasincreased the bargaining power of


Trend in Medicare inpatient and overall Medicare margin,1990–2000

Note: Data are based on Medicare-allowable costs and imputed for hospitals whose 2000 cost reports were not available (about 27 percent of observations). Excludes critical access hospitals. Overall Medicare margin covers the costs and payments of hospital inpatient, outpatient, psychiatric and rehabilitation (prospective payment system exempt), skilled nursing facility, and home health services, as well as graduate medical education. Data on overall Medicare margin are unavailable before 1996.


1990 1991 1992 1993 1994 1995 1996 1997 1999 20001998

10

15

0

5

�5

20

Per

cent

Medicare inpatient

Overall Medicare

FIGURE2A-5

5 Growth in the ECI for wages and salaries of hospital workers is reflected in the market basket, which leads to higher payments under both the inpatient and outpatientPPSs.

T A B L E2A-3 Overall Medicare margin by

hospital group, 1999–2003

EstimatedHospital group 1999 2000 2003

All hospitals 5.1% 5.0% 3.9%

Large urban 8.4 8.4 6.9Other urban 3.3 2.9 1.7Rural –2.5 –2.9 –1.9

Major teaching 13.7 14.9 12.7Other teaching 5.7 5.0 3.8Nonteaching 0.1 –0.2 –0.6

Note: Data are based on Medicare-allowable costs. Margins are imputed for hospitals whose 2000 cost reportswere not available (about 27 percent of observations). Excludes critical access hospitals. Projections for 2003reflect the effects of all policy changes implemented between 2000 and 2003, plus policy changes otherthan updates that are scheduled under current law to go into effect in 2004.

Overall Medicare margin covers the costs and payments of hospital inpatient, outpatient, psychiatric,rehabilitation, skilled nursing facility, and home health services, as well as graduate medical education.

Source: MedPAC analysis of Medicare cost report and market basket data from CMS, American Hospital Association(AHA) annual survey of hospitals, and the national hospital indicators survey (sponsored by CMS andMedPAC, and conducted by AHA).

providers in their dealings with insurers(Levit et al. 2003).

One measure of hospital cost growth thatis available for 2001 is the change in costper adjusted admission from the AmericanHospital Association’s annual survey ofhospitals. It measures costs for allinpatient and outpatient services for allpayers. Cost per adjusted admission rose4.7 percent in 2001—the most rapidincrease since 1992—reflecting thestabilizing length of stay and greater wagepressures discussed above.

Relationship of paymentsand costs We next assess the relationship betweenpayments and an appropriate cost base. Indoing this, we consider measures of thevolume of hospital services, entry and exitof providers, access to capital, andbeneficiary access to care. We concludethat current payments are at leastadequate.

Changes in volume Substantial increases in volume couldindicate that payment rates are too high,and decreases that payment rates are toolow. The trend in hospital volume also hasimplications for the appropriateness ofcosts. If volume increases, hospitalsshould have more cases over which tospread fixed costs, which will reduce per-unit cost.

We measure hospital volume by totaladmissions, total days of care, and totaloutpatient visits. The volume of hospitalservices has grown strongly in recentyears. The total number of hospitaladmissions grew a cumulative 6.1 percentfrom 1990 through 2000 despite fallingfrom 1990 to 1994. According to theAmerican Hospital Association annualsurvey, total admissions grew by 1.7percent in 1999, 2.3 percent in 2000, and2.2 percent in 2001. Medicare dischargesgrew even more rapidly, by 1.9 percent in

1999, 4.2 percent in 2000, and 3.2 percentin 2001.

Large declines in length of stay andmodest admissions growth combined toreduce total inpatient days at communityhospitals by 15 percent from 1990 to1998. Stabilizing length of stay and fasteradmission growth have since turned thisreduction around, with an increase of 1.4percent from 1998 to 2001 to about 195million days.

Total outpatient visits have increasedsteadily over 20 years. Total outpatientvisits to community hospitals, includingemergency visits, increased 73 percentfrom 1990 to 2000. Growth continued in2001, with an increase of 3.3 percent over2000 to almost 540 million visits.

Entry and exit of providers Significant changes in the number ofproviders can indicate the relative healthof the hospital market. If payments are toolow, some providers may be forced toclose; if payments are too high, moreproviders than are necessary for accessmay enter the field. Because Medicare issuch a large purchaser of hospitalservices, entry and exit could beinfluenced by Medicare payment policy.

As the volume of patient days declinedthrough the 1990s, a small number ofhospitals closed each year. From 1990 to2000, there was a net reduction of 469community hospitals across the country.This reduced the total supply of beds byabout 10 percent. Closed hospitals tend tobe in areas with low levels of demand forhospital services. At the same time, asmaller number of hospitals opened eachyear in areas with excess demand. In 1999through 2001, the number of closuresaveraged 56 per year, with an average of21 openings or reopenings (OIG 2002,AHA 2002). In 2002, 52 short stayhospitals ended their participation inMedicare while 42 were accepted in theprogram.6

Hospitals that closed in 2000 had lowoccupancy rates. Closed urban hospitalswere smaller in size than urban hospitalsnationally and had lower occupancy.Closed rural hospitals were the sameaverage size as all rural hospitals withmodestly lower occupancy rates.

The Office of Inspector General (OIG) ofthe Department of Health and HumanServices found that hospital closures in2000 generally had modest effects onaccess to care. Rural hospitals that closedhad an average of 23 patients each day inthe year before closure, whereas closingurban hospitals had an average of 70patients. Inpatient care was available within20 miles of 86 percent of rural hospitalsthat closed and all urban hospitals thatclosed. While 24-hour emergency servicessometimes disappeared when hospitalsclosed, patients in 73 percent of ruralclosures and 91 percent of urban closuresstill had emergency services within 10miles of the closed facility.

Despite hospital closures, increasedvolume of hospital services—in bothadmissions and total days—supports theconclusion that the capacity of the hospitalindustry remains adequate.

Access to capital Access to capital is necessary for hospitalsto maintain and modernize their facilitiesand capabilities for patient care. Aninability to access capital that waswidespread throughout the sector mightindicate inadequate payments. Borrowingby hospitals was strong in 2002,indicating good access to capital. Long-term borrowing by acute care hospitalsreached $20.0 billion in 2002, an increaseof 7.3 percent over 2001.7 Because about85 percent of nongovernment short-termacute care hospitals are nonprofit, thelevel of borrowing is a strong indicator ofaccess to capital for the hospital industry.

Hospitals obtain capital through equitymarkets (in the case of for-profithospitals), bond markets, bank lending,receivables financing, and cash flow. The


6 MedPAC analysis of CMS data on number of new participating hospitals, January 2002 to December 2002.

7 MedPAC analysis of data on general acute care long-term bond issuance, Thomson Corporation, January 28, 2003.

outlook for the health care sectorremained favorable in 2002. Stock ofinvestor-owned hospital firms generallyperformed better than the overall stockmarket in both 2001 and 2002. Wall Streetanalysts predict continuing financialhealth for the investor-owned firmsattributable to healthy volume growth andhigh private sector payments (MerrillLynch 2003).

For those hospitals that are able to borrowon the bond market, access to capital isgood. The ability of hospitals to borrow isstrongly affected by the bond ratings theyreceive from credit rating agencies. Betterratings reduce the interest expenseshospitals must incur to raise capital whilelower ratings increase them. Hospitalbond downgrades exceeded upgrades eachyear from 1999 through 2001. In 2002,downgrades exceeded upgrades by lessthan in 2001, indicating easier access tocapital for hospitals. Among bond issuesrated by Moody’s, there were 1.9downgrades of health care bonds for everyupgrade in 2002. This was animprovement compared to 2.5downgrades for every upgrade in 2001.Fitch Ratings reported a 4:1 ratio ofdowngrades to upgrades in 2002 for theacute-care bond issues it rated. This wasan improvement over 2001, when therewere nine downgrades for every upgrade(Fitch 2003). However, these ratios maskthe dollar volume of these ratings, whichshows the opposite picture. The dollarvolume of upgrades exceeded downgradesby 80 percent in 2002, while the volumeof downgrades exceeded upgrades by 80percent in 2001 (Moody’s 2003).

Two events in 2002 led to investorconcern about hospital finances. Thesewere the difficulties of Tenet Healthcareand the bankruptcy of National CenturyFinancial Enterprises (NCFE). In October,Tenet disclosed unusually large Medicarepayments for outliers under the inpatientPPS. The Department of Health andHuman Services announced an audit ofoutlier payments for all hospitals.

NCFE is a privately held company thatprovided financing to a variety of healthcare providers in exchange for theproviders’ receivables. It packaged thereceivables and sold bonds based on themto raise capital and pay for morereceivables. The company haltedpayments to its health care providerclients in October and filed for bankruptcyprotection in November.

Although these events received significantpublicity and had major repercussions onselected stock prices and bond ratings,neither one was expected by Wall Streetto overshadow access to capital for thehospital sector as a whole for long. WallStreet analysts see factors such asadmissions growth and pricing as the mostimportant determinants of the financialstatus of investor-owned hospitals.

Expansion of for-profit chains in rural orsmall urban areas may indicate goodaccess to capital. These firms haveexpanded by acquiring nonprofit hospitalsthat reportedly have not been able to makethemselves attractive to patients. This maybe a symptom of inability of these smallhospitals to obtain capital, suggesting thataccess is constrained. Conversely, theability of for-profit chains to acquire thesehospitals hinges on their ability to enterthe capital markets, suggesting that accessis good. This contrast illustrates that thecapital markets make distinctions amonghospitals regarding their financial viabilityas one would expect in a properlyfunctioning market.

Overall, the trends in the equity and bondmarkets indicate that both for-profit andnonprofit hospitals have sufficient accessto capital if they are financially viable.

Beneficiary access to care If payments for Medicare services are toolow and providers are forced to exit themarket, some Medicare beneficiaries mayexperience problems with access to care.Access to hospital services does notappear to be a problem for most Medicarebeneficiaries.

Most hospitals typically have lowoccupancy rates—the national averageoccupancy rate was 64.4 percent in 2001.This suggests that hospitals have thecapacity to treat Medicare patients.Reports of diversions of ambulances fromoverburdened emergency rooms andevidence of shortage of emergencydepartment capacity in some areas,however, suggest that in a few instanceshospitals may be unable to provide someservices sought by beneficiaries. But it isnot clear that these problems are related tothe level of Medicare payments. Anycapacity problems that exist may beaggravated by shortages of nurses andother health care personnel as well asrising malpractice premiums (seeChapter 3).

Medicare beneficiaries in rural areas facechallenges with access to hospital servicesbecause of longer distances betweenhospitals, but MedPAC found that use ofhealth services is not lower in rural areasthan in urban areas (MedPAC 2001a).Medicare has addressed issues of supplyof inpatient care in rural areas with thecritical access hospital (CAH) program.CAHs are paid their current Medicareallowable costs for inpatient andoutpatient services. The CAH programrequires a hospital to be located in anisolated area without another nearbyprovider, or to be designated as anessential provider in a state health plan.The program has grown rapidly from 375hospitals in April 2001 to 725 in January2003.8

Policies affecting thedistribution of payments

In this section, we discuss six componentsof the inpatient PPS that we believe couldbe modified to improve the distribution ofpayments. For the first component, werecommend extending the post acutetransfer policy to 13 additional DRGs(Recommendation 2A-1). For the secondone, we discuss the need to explore ways


8 MedPAC analysis of data on critical access hospitals, January 30, 2003.

to target the portion of indirect medicaleducation payments above the empiricallyjustified level to advance specific policyobjectives within the Medicare program.Finally, we reissue four previousrecommendations designed to improvepayments to rural hospitals(Recommendations 2A-2 though 2A-5).These would implement a low-volumeadjustment, reevaluate the labor shareused in adjusting inpatient payments forgeographic differences in input prices,eliminate the differential in base paymentrates for hospitals in rural and small urbanareas, and raise the cap ondisproportionate share payments for ruralhospitals.

Inpatient payments for casestransferred to other settings When hospitals discharge patients toanother care setting, some of the carefurnished in the other setting maysubstitute for services that otherwisewould have been provided during thehospital inpatient stay; thus the hospital isfurnishing a product that does not includethe full course of care implied by thediagnosis related group (DRG) payment.Under the inpatient PPS, Medicare treatsall cases discharged from the hospital toanother PPS hospital with shorter thanaverage stays as partial cases, paying a perdiem rate for each day, up to the full DRGamount. Starting in 1999, policymakersexpanded the inpatient transfer paymentpolicy to include cases in 10 DRGs thatare discharged to post-acute care settingsafter shorter than average inpatient stays.

The incentives created by Medicare’sinpatient transfer payment policy areconsistent with the goal of paying efficientproviders’ costs, and the Commission haspreviously stated that this policy should bemaintained (MedPAC 2000a). Thetransfer payment policy helps to link acuteand post-acute payment systems byadjusting inpatient payments when a

portion of care is shifted to another settingwhere Medicare also pays for thebeneficiaries’ care. This policy alsoimproves hospitals’ financial incentives toprovide quality care. By matchingpayments more closely to the incrementalcosts of each day of care, the transferpolicy helps to diminish the influence offinancial considerations on hospitals’clinical decision-making. The transferpolicy also adjusts payments to reflect thecircumstances faced by individualhospitals, recognizing that hospitals mayhave different access to post-acute careservices; thus, payment reductions aretargeted only to hospitals that dischargepatients to post-acute care with shortstays.

The Balanced Budget Act of 1997 (BBA)expanded Medicare’s transfer paymentpolicy. Before 1997, Medicare considereda case to be a transfer only if an inpatientwas discharged from one PPS hospital andimmediately admitted to another PPShospital.9 Under the BBA, transfers alsoinclude inpatients in selected DRGs whoare discharged from a PPS hospital eitherto a skilled nursing facility (SNF), to aPPS-exempt hospital or unit (i.e., arehabilitation hospital or unit, psychiatrichospital or unit, long-term care hospital,cancer hospital, or children’s hospital), orwith a written plan for home health carethat starts within three days of dischargeand is related to the condition or diagnosisthat accounted for the inpatient stay.10

A number of factors probably entered intothe Congress’ decision in 1997 to expandMedicare’s inpatient transfer paymentpolicy to include discharges from PPShospitals to PPS-exempt hospitals andother post-acute care settings. At the timethe Congress was considering this policy,data showed that Medicare beneficiaries’average inpatient length of stay haddropped substantially—22 percentbetween 1990 and 1995 (ProspectivePayment Assessment Commission

[ProPAC] 1997a). In addition, the drop inbeneficiaries’ average length of stay wasaccompanied by dramatic growth in use ofand spending for post-acute care byMedicare beneficiaries (ProPAC 1997b).Furthermore, hospitals’ Medicareinpatient margins had risen to recordlevels. The conference reportaccompanying the BBA noted that theconferees were concerned that Medicaremight in some cases be overpayinghospitals for patients who were transferredto a post-acute setting after a very shortacute care hospital stay (U.S. House ofRepresentatives 1997).

Analysis by MedPAC and by ProPAC, itspredecessor, showed that declines ininpatient lengths of stay were greatest forDRGs in which post-acute care use wasmost prevalent (MedPAC 1998).Furthermore, hospitals operating post-acute care facilities discharged theirpatients one day sooner on average thanhospitals without such facilities, and theirpatients used post-acute care about 10percent more frequently than patients ofhospitals without such facilities (ProPAC1996).

These trends in inpatient length of staywere consistent with the financialincentives of Medicare’s hospital inpatientPPS. When the inpatient PPS began in1984, relatively few patients weredischarged to post-acute care. Prospectivepayments provided hospitals with a strongfinancial incentive to shorten the length ofhospital stays; and the growth in theavailability and capabilities of post-acutecare providers allowed hospitals to shiftsome of the care once provided duringinpatient hospital stays to SNFs and otherpost-acute care settings.

CMS’s implementation of Medicare’sinpatient transfer policy following theBBA, the current operation of the transferpolicy, and a proposal for CMS to applythe policy more broadly are discussed


9 Discharges to hospitals excluded from Medicare’s inpatient PPS because they participated in a statewide cost control program were also considered transfers. Recently,this policy has affected only discharges from PPS hospitals to acute-care hospitals located in Maryland.

10 Discharges to hospital swing beds, which are designated beds in small rural acute care hospitals that can be used for acute or skilled care, are not counted astransfers. CMS considered treating discharges to swing beds as transfers in the proposed rule for implementing the expanded transfer policy, but withdrew thisproposal in the final rule in response to comments.

below. Although Medicare beneficiaries’post-acute care use has grown relativelylittle in recent years, the shift of servicesfrom inpatient to post-acute settingscontinues. Medicare’s inpatient transferpolicy is intended to adjust payments tohospitals for inpatient services to reflectthis shift of services from inpatient topost-acute settings. So far, however, thepolicy has been implemented for only 10

DRGs. Consequently, Medicare stilloverpays hospitals for inpatient serviceswhen they discharge patients in otherDRGs to post-acute care after very shortstays. This is particularly true for the 13DRGs with high use of post-acute careservices. For that reason, the Commissionrecommends extending the policy to these13 additional DRGs next year (Table2A-4).

Implementation of Medicare’sinpatient transfer policy CMS implemented Medicare’s expandedinpatient transfer payment policy in fiscalyear 1999. To comply with the law, theSecretary selected 10 DRGs in whichinpatient cases would be consideredtransfers if they met certain statutorycriteria. The Secretary was authorized bythe BBA to expand the list of DRGs


Hospital cases under Medicare’s inpatient transfer payment policy currently and under proposed expansion

Percent of Percent of allcases Percent of cases in DRG that

Number of discharged to transfer cases are short-stayDRG Title cases post-acute care with short stays transfers

DRGs under current policy14 Specific cerebrovascular disorders except transient ischemic attack 302,095 51.8% 22.1% 11.5%

113 Amputation for circulatory system disorders, except upper limb and toe 39,267 71.4 45.2 32.3209 Major joint and limb reattachment procedure of lower extremity 356,891 76.7 28.7 22.0210 Hip and femur procedures except major joint age � 17 with CC 115,722 81.9 28.1 23.0211 Hip and femur procedures except major joint age � 17 without CC 30,572 79.8 21.7 17.3236 Fractures of hip and pelvis 37,919 67.3 12.3 8.3263 Skin graft and/or debridement for skin ulcer or cellulitis with CC 22,919 61.9 40.8 25.3264 Skin graft and/or debridement for skin ulcer or cellulitis without CC 3,711 50.5 37.0 18.7429 Organic disturbances and mental retardation 25,373 58.4 31.4 18.3483 Tracheostomy except for face, mouth, and neck diagnoses 40,954 52.5 47.5 24.9

Total for 10 DRGs 975,423 67.2 27.8 18.7

DRGs under proposed expansion12 Degenerative nervous system disorders 47,929 56.2 30.9 17.479 Respiratory infections and inflammations age � 17 with CC 158,062 49.6 29.8 14.880 Respiratory infections and inflammations age � 17 without CC 8,019 42.7 25.2 10.8

107 Coronary bypass with cardiac catheterization 79,444 42.7 35.8 15.3109 Coronary bypass without PTCA or cardiac catheterization 54,830 38.9 25.5 9.9148 Major small and large bowel procedures with CC 123,995 39.2 31.3 12.3149 Major small and large bowel procedures without CC 18,498 15.4 22.3 3.4239 Pathological fractures and musculoskeletal and CT malignancy 45,479 53.6 25.2 13.5243 Medical back problems 88,618 40.9 9.7 4.0320 Kidney and urinary tract infections age � 17 with CC 184,099 44.0 26.9 11.9321 Kidney and urinary tract infections age � 17 without CC 29,862 29.0 17.4 5.0415 OR procedure for infectious and parasitic diseases 37,974 53.6 36.6 19.6468 Extensive OR procedure unrelated to principal diagnosis 57,861 43.6 27.5 12.0

Total for 13 additional DRGs 934,670 44.0 27.5 12.1

All other DRGs 8,902,789 25.1 17.7 4.4

All DRGs 10,812,882 30.5 20.9 6.4

Note: DRG (diagnosis related group), CC (complication or comorbidity), PTCA (percutaneous transluminal coronary angioplasty), CT (connective tissue), OR (operating room).Short stays are those that are more than one day less than the geometric mean length of stay for the DRG. Percentages may be inexact because of rounding.

Source: MedPAC analysis of 2001 MedPAR data from CMS.

T A B L E2A-4

beginning in fiscal year 2001, but, inconjunction with the Balanced BudgetRefinement Act of 1999 (BBRA), shedecided to delay any expansion by at leasttwo years.

The Secretary considered expanding thetransfer policy to encompass DRGsbeyond the original 10 in the proposedrule on Medicare’s inpatient prospectivepayments for fiscal year 2003 (CMS2002b). In the final rule the Secretarydecided to defer a decision aboutexpanding the policy until 2004.Commenters raised many issues regardingthe impact of expanding the policy thatneed to be considered carefully beforeproceeding, and the Secretary stated thatthe limited time between the close of thecomment period and the requiredpublication date for the final rule was notsufficient for analyzing and responding toall the points raised. CMS plans tocontinue research to assess whetherexpansion of the policy to additionalDRGs is warranted for fiscal year 2004 orsubsequent years (CMS 2002a).

Medicare’s current inpatienttransfer payment policy For transfer cases with hospital lengths ofstay substantially shorter than the nationalaverage for the DRG, Medicare payshospitals a per diem rate up to the fullDRG payment—which is reached whenthe length of stay is one day less than thegeometric mean length of stay for theDRG.11 The per diem amount equals thefull per discharge payment for the DRGdivided by its national geometric meanlength of stay. Hospitals receive twice theper diem amount for the first day of careand the per diem amount for allsubsequent days up to the full DRGpayment. Very expensive cases mayqualify for outlier payments as well.

The Secretary may provide a modified perdiem payment for DRGs in which asubstantial portion of the cost of care isincurred in the early days of the stay. Thisensures that the transfer payment will

cover the full cost of care for these cases.By law, the modified payment may be nomore than the average of the paymentunder the basic transfer policy and the fullDRG payment. Currently, this modifiedtransfer payment is provided in 3 of the 10DRGs affected by the transfer policy; all 3are surgical DRGs. In these instances,hospitals receive half the full DRGpayment plus a single per diem paymentfor the first day of care. They then receivehalf of a per diem payment for allsubsequent days of care up to the fullDRG payment for the case.

In selecting the 10 DRGs originallyauthorized for the expanded transferpolicy, the Secretary chose DRGs with alarge number of discharges to post-acutecare and a high rate of post-acute care use(Table 2A-4). More than half the cases ineach of these DRGs were discharged topost-acute care settings. In these 10DRGs, as in most DRGs, the patients whouse post-acute care tend to have longerthan average inpatient stays. For example,patients who are transferred to post-acutecare settings in DRG 14 (strokes) haveaverage acute inpatient stays of 6.8 days,which is 2.1 days higher than the nationalgeometric mean length of stay for allcases in this DRG (Figure 2A-6). SinceMedicare’s transfer policy applies only totransfer cases with lengths of stay that aremore than one day shorter than thenational geometric mean for the DRG,patients transferred to post-acute caresettings are affected only when their staysare several days shorter than those typicalof post-acute care users.

Proposals to expand Medicare’sinpatient transfer paymentpolicy CMS, as part of the proposed rule forMedicare’s hospital inpatient PPS forfiscal year 2003, considered two optionsfor expanding the transfer policy toencompass additional DRGs:

• Expansion of the policy to someadditional DRGs. Under one option

CMS would apply the inpatienttransfer payment policy to 13additional DRGs with high rates oftransfers to post-acute care settings(similar to the initial group of 10DRGs).

• Expansion of the policy to all DRGs.Under the second option, CMS wouldapply the inpatient transfer policy toall DRGs.

As discussed below, we believe that theweight of evidence supports expandingMedicare’s inpatient transfer paymentpolicy beyond the original 10 DRGs.Because expanding the policy to all DRGsmight reduce PPS payments to somehospitals by as much as 4 percent, werecommend that the policy be extendedinitially to 13 additional DRGs, with theeffects of this expansion evaluated beforeextending the policy to more DRGs.

Why the expanded transfer policy isneeded Medicare’s inpatient PPS isintended to encourage providers to seekmore efficient ways to furnish high-quality care to its beneficiaries. In manyinstances, substituting less costly post-acute care services for more expensiveinpatient care may provide a moreefficient overall episode of care ofcomparable quality. As long as this is trueand Medicare’s payment policies adaptappropriately, it is in everyone’s interestto promote such changes in the quantityand mix of services furnished across caresettings.

There would be no need to broadenMedicare’s inpatient transfer paymentpolicy if all of the observed increases inthe use of post-acute care representedadditional care, or if the substitution ofpost-acute care for inpatient care wereoccurring roughly similarly across allDRGs and hospitals. In the latter case, thepayments could adjust to reflect thisthrough the update. Available evidence,however, strongly suggests that observedincreases in the use of post-acute carereflect the substitution of post-acute care


11 The geometric mean length of stay provides a more representative measure of the usual length of stay than the arithmetic average when the distribution includes manycases with extremely long stays.

for some inpatient care services, and thatthe substitution of services differs amongDRGs and hospitals. Furthermore,expanding the transfer policy wouldimprove the incentives for providingquality care by reducing the strongfinancial incentives Medicare’s inpatientPPS gives providers to discharge patientsto post-acute settings as quickly aspossible. Expanding the policy would alsoimprove payment equity among hospitalsby accounting for differences in hospitals’short stay post-acute transfer rates.

Shifts in services from inpatient to post-acute settings In recent years, Medicarebeneficiaries’ average inpatient length ofstay has declined, while the number ofdischarges from inpatient to post-acutecare settings has increased. Between 1991and 2001, Medicare length of stay fell 34percent and Medicare discharges fromPPS hospitals to post-acute care settingsincreased 49 percent. During this period,Medicare discharges from acute carehospitals to hospitals and units exemptfrom the inpatient PPS doubled;discharges from acute care hospitals toskilled nursing facilities climbed 65percent; and discharges from acute carehospitals to home health care increased by14 percent (Table 2A-5).

Analysis of Medicare hospital inpatientclaims also suggests that some of theincrease in post-acute care has substitutedfor inpatient days and related services.12

One analysis, for example, has shown thatlength of stay declines were greater inDRGs with high rates of post-acute careuse (MedPAC 1998). Another studyshowed that length-of-stay declinesbetween 1991 and 1998 were greater forpost-acute users compared withnonusers—4.5 days and 2.4 days,respectively (Gillman et al. 2000).Further, average length of stay in DRGswith high use of post-acute care dropped7.1 days for users versus 5.6 days fornonusers.

Analysis of inpatient payments andestimated costs for patients discharged topost-acute care suggests that hospitalsincur much lower costs for post-acute careusers with very short stays. In the absenceof an expanded transfer policy, this canresult in large financial gains to hospitalsthat discharge many patients to post-acutecare settings. Although hospitals receive alower payment for short-stay cases under

the transfer policy, per diem transferpayments still exceed the cost of caringfor these cases on average. In DRG 79(respiratory infections), for example, perdiem transfer payments prior to reachingthe full DRG payment would exceedestimated daily costs by about 50 percent(Figure 2A-7, p. 48). Almost all otherDRGs would have similar outcomes underan expanded transfer policy. In the few


Length of stay distribution for stroke (DRG 14)FIGURE2A-6

1 2 3 4 5 6 7 8 9 10 11

Num

ber

of

case

s

30,000

15,000

10,000

0

5,000

20,000

25,000 Geometric meanLOS 4.7 days

Length of stay (days)

Nontransfer casesTransfer cases

Casesaffected by

transferpolicy

Note: DRG (diagnosis related group), LOS (length of stay).


Medicare hospital discharges to post-acute care providers, 1991–2001Percent of hospital cases

1991 1998 2001

PPS exempt hospital or unit 2.7% 4.7% 5.5%Skilled nursing facility 9.3 15.8 15.3Home health 8.5 9.7 9.7Total 20.5 30.2 30.5

Note: PPS (prospective payment system).

Source: Gillman et al. (2000) and MedPAC analysis of 2001 MedPAR data from CMS.

T A B L E2A-5

12 Although Medicare beneficiaries’ use of post-acute care has not been growing rapidly since 1998, the substitution of post-acute services for hospital inpatient care hasnot yet been fully addressed in the inpatient PPS.

instances where per diem payments wouldnot cover per diem costs (DRG 107, forexample), applying the modified transferpayment method would provide per diempayments well above per diem costs forcases affected by the policy.

Thus there is strong evidence that thecurrent payment system overpayshospitals that discharge patients to post-acute settings with shorter than averagestays in DRGs where the expandedtransfer policy does not apply. Theexpanded transfer policy provides amechanism for linking acute and post-acute payments by reducing theoverpayment when a portion of the care isshifted to a post-acute setting whereMedicare also pays for services. Thepolicy, however, also is necessary toaccount for differences in hospitals’circumstances, recognizing that access topost-acute services can vary, contributingto different financial outcomes amonghospitals.

Post-acute care use patterns amongDRGs and hospitals The percentage ofcases discharged to post-acute care

settings varies widely among DRGs. Overhalf of all cases are discharged to a post-acute care provider in 41 DRGs thataccount for about 12 percent of Medicaredischarges. Between 25 and 50 percent ofall cases are discharged to post-acute carein another 177 DRGs. In all, more thanthree-fifths of all Medicare discharges arein DRGs in which at least 25 percent of allcases go on to post-acute care. At theother end of the spectrum, 12 percent ofall discharges are in 108 DRGs where lessthan 10 percent of cases go on to post-acute care.

Transfers to post-acute care are similarlyuneven across hospitals. For instance,urban hospitals on average transfer alarger proportion of Medicare cases topost-acute care providers than do ruralhospitals—32 percent and 26 percent,respectively (Table 2A-6). Finerbreakdowns of urban and rural hospitalsshow even greater differences in use ofpost-acute care. Hospitals located in largeurban areas (metropolitan statistical areaswith over 1 million people) transferred 34percent of their cases to post-acute care,compared with only 20 percent for rural

hospitals with less than 50 beds that donot currently receive special treatmentunder Medicare.

Regional disparities in the use of post-acute care are even larger. For example,post-acute care transfer rates are twice ashigh in New England (46 percent) as inthe West South Central census division(23 percent). Most other differencesamong hospital groups (e.g., ownership orteaching status) tend to be small, however.These small differences reflect the widevariability in post-acute care transfer ratesamong the hospitals within most hospitalgroups.

Individual hospitals’ transfer rates doappear to be strongly consistent acrossDRGs (Table 2A-7, p. 50). When wegrouped hospitals by their overallpercentage of Medicare cases transferredto post-acute care, it turned out that thosewith low overall transfer rates also hadlow transfer rates in each of the DRGs weexamined. Similarly, those with highoverall transfer rates also had high rates ineach of the DRGs. These findings suggestthat hospitals’ transfers to post-acute careare driven more by their specificcircumstances than by any shared hospitalcharacteristic. Short-stay transfers to post-acute care—those that have inpatient staysprior to transfer that are more than oneday less than the national geometric meanlength of stay for the DRG—show similarpatterns, with strong hospital-specificdifferences and relatively smalldifferences among hospital groups.

Average length of stay for Medicarebeneficiaries varies across regions,although the differences are much lessthan they were 10 years ago (Table 2A-8,p. 51). Some observers have suggestedthat expansion of the transfer policywould penalize hospitals in regions withshort stays. This concern is only valid to apoint, however, because the relationshipbetween regional average Medicarelengths of stay and the proportion of casesaffected by the policy is fairly weak. Twofactors influence the proportion of casesaffected by the policy. One is the share ofcases discharged to post-acute care and


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Paym

ent

to c

ost

ratio

Payment-to-cost ratios for transfer cases before and after transfer policy, respiratory infections (DRG 79)

Note: DRG (diagnosis related group), LOS (length of stay).


6

3

2

0

1

4

5

Geometric meanLOS 6.6 days

Length of stay (days)

Before policy

After policy

FIGURE2A-7

the other is the share of these cases thathave short stays. Hospitals in regions withrelatively short stays tend to have a higherproportion of their transfer casesdischarged after a short stay, but there isno relationship between length of stay andthe proportion of cases discharged to post-acute care.

Further, although hospitals located inshort-stay regions may have more transfercases affected by the policy, they benefitfinancially from their short stay pattern ofcare on all their other cases. The perdischarge payment rates under theinpatient PPS reflect national average carepatterns. Other things being equal,

however, cases in relatively short-stayregions tend to have lower than averagecosts.

Hospitals with high short-stay transferrates would lose some payments under thepost-acute transfer policy, but they benefitfrom having relatively short stays and low


Use of post-acute care providers and cases affected by expansion of transfer policy to all DRGs

Percent of allPercent of cases discharged

Percent of cases transfer to post-acuteNumber of Number of discharged to cases with care with

Hospital group hospitals cases post-acute care short stays short stays

Total 4,613 10,812,882 30.5% 20.9% 6.4%

Urban 2,632 8,646,905 31.7 20.5 6.5Rural 1,656 2,103,922 25.9 22.7 5.9

Large urban 1,537 4,902,476 33.6 20.3 6.8Other urban 1,095 3,744,429 29.2 20.8 6.1Rural referral 248 833,371 28.3 20.7 5.8Sole community 521 476,975 23.6 24.4 5.8Small rural Medicare dependent 241 183,454 25.0 24.1 6.0Other rural � 50 beds 313 161,707 19.9 27.6 5.5Other rural �50 beds 333 448,415 26.3 23.5 6.2

Major teaching 298 1,493,872 32.3 21.5 6.9Other teaching 824 3,620,550 31.6 20.3 6.4Nonteaching 3,166 5,636,405 29.4 21.1 6.2

New England 183 537,570 46.2 24.6 11.4Middle Atlantic 474 1,605,852 37.2 17.0 6.3South Atlantic 687 2,084,098 29.3 20.0 5.8East North Central 696 1,914,994 32.3 22.1 7.1East South Central 406 958,806 26.1 19.9 5.2West North Central 571 872,834 28.6 22.7 6.5West South Central 648 1,260,795 22.8 20.4 4.7Mountain 328 467,115 27.9 26.1 7.3Pacific 564 972,134 30.8 22.8 7.0

Voluntary 2,596 7,900,024 32.0 20.8 6.6Proprietary 650 1,239,981 27.8 20.3 5.6Urban government 368 985,048 26.9 21.8 5.8Rural government 669 625,657 23.4 21.9 5.1

Note: DRGs (diagnosis related groups). Short stays are those that are more than one day less than the geometric mean length of stay for the DRG.


T A B L E2A-6

discharging patients to post-acute caresettings but that they substitute post-acuteservices for acute care. Nevertheless, theexpanded transfer policy provides a betterset of incentives to protect beneficiariesfrom potential premature discharge topost-acute care. When hospitals are paidless for short stays and more for longstays, the decision to transfer will beinfluenced less by financialconsiderations. Hospitals should befinancially indifferent to the decision totransfer a patient to a post-acute setting ifthe marginal cost of care and the per diempayment amounts are close. Past researchhas shown that Medicare’s current transferpayment method provides a reasonableapproximation of marginal cost (Carterand Rumpel 1993).

HCFA (now CMS) analysis of the initial10 DRGs showed that per diem paymentswould on average more than cover thecost of care for the affected transfer cases(HCFA 1998). Consequently, hospitalsstill had a financial incentive to dischargepatients to post-acute care, and in fact thepercentage of cases in the original 10DRGs discharged to post-acute careincreased slightly after the policy wasimplemented. As discussed earlier, ouranalysis also shows that per diem transferpayments would more than cover theestimated daily cost of care for short-staycases in the original 10 DRGs and in otherDRGs to which the policy might beexpanded.

Providing a more equitable distributionof payments Another reason to expand thetransfer policy is that it would improvepayment equity across cases and hospitals.The expanded transfer policy would helpimprove payment equity in two ways.First, it would account for differencesacross providers in the availability and useof post-acute care for short-stay cases. Ingeneral, the policy would provide a


Share of cases discharged to post-acute care settingsfor selected DRGs, by hospital group

Hospital Percent ofDRGs

group hospitals 14 79 89 107 116 204 209

Group defined by percent of hospital cases discharged to post-acute care� 10 10% 11% 6% 5% 7% 2% 2% 20%� 10–20 17 36 28 17 19 5 8 52� 20–30 31 49 43 28 32 8 12 73� 30–40 28 57 55 38 52 12 17 81� 40–50 10 63 66 50 72 20 23 88� 50 4 69 74 62 82 32 31 93

Total 100 52 50 34 43 11 15 77

Group defined by percent of hospital cases discharged to post-acute care with short stays� 2 10 2 2 1 2 –* –* 3� 2–5 30 7 8 3 8 –* 1 13� 5–10 44 13 16 6 18 1 3 24� 10–15 13 19 25 12 32 2 3 34� 15 3 31 38 21 45 3 9 45

Total 100 11 15 6 15 1 2 22

Note: DRGs (diagnosis related groups), DRG 14 � stroke, DRG 79 � respiratory infections, DRG 89 �pneumonia, DRG 107 � coronary bypass with cardiac catheterization, DRG 116 � other permanentcardiac pacemaker implant, DRG 204 � disorders of the pancreas except malignancy, DRG 209 � majorjoint and limb reattachment procedures of lower extremity. Short stays are those that are more than one dayless than the geometric mean length of stay for the DRG.* Less than 0.5 percent


T A B L E2A-7

13 Medicare inpatient margins were calculated excluding disproportionate share hospital payments and IME payments above the teaching cost relationship. Theseamounts were excluded because they are unrelated to the transfer policy and they tend to obscure the relationships between average length of stay, short-stay post-acute transfers, and hospital financial performance.

14 To make this calculation, we compared each hospital’s actual average length of stay for Medicare patients with what the average would have been if its cases had thenational average length of stay in each DRG.

costs throughout their Medicare caseload,which contributes to their having higherMedicare inpatient margins (Table2A-9).13 When hospitals are grouped bytheir short-stay post-acute transfer rates,those with high proportions of short-staytransfers on average have relatively shortoverall Medicare lengths of stay. Forinstance, hospitals that had more than 15percent of their cases transferred to post-acute care settings after short inpatientstays had average actual lengths of stayabout 20 percent lower than we wouldexpect given their mix of cases among the

DRGs.14 They also had a higherproportion of cases discharged to post-acute care overall.

Improving incentives for quality care A per case payment system providesstrong financial incentives for hospitals toshorten inpatient stays. Per diempayments reduce hospitals’ incentives totransfer patients to post-acute settings bybringing payments more in line with theestimated incremental cost of providingcare. The rationale for the policy does notassume that hospitals are prematurely

payment reflecting the care providedduring the acute inpatient stay,recognizing that use of post-acute care canbegin at different points in similarpatients’ care. Hospitals that have theirown post-acute care units, for example,may be able to move patients safely to apost-acute care unit earlier than hospitalswhere patients would need to betransported for post-acute care. In

addition, the timing of discharge or use ofpost-acute care may be affected by theavailability of open beds in facilities thatare able to handle patients’ specifictreatment needs. The transfer policymatches payments to the localcircumstances, rather than applying thesame payment in widely differingcircumstances.

Second, expanding the transfer policywould improve the accuracy of the DRGweights in the affected DRGs. The DRGsnot included in the expanded transferpolicy are now affected adversely becausecases that would be treated as transfers aretreated as discharges—and not discountedin recalibration. Thus DRGs not includedin the expanded transfer policy that haveexperienced substantial declines in lengthof stay (and charges) because of increasedpost-acute transfers have likely seen theirrelative weights fall. In this situation,hospitals able to discharge patients earlylikely are paid too much while those thatare unable to do so (because of limitedaccess to post-acute services) are paid toolittle.

Tracheostomy cases provide an exampleof the potential inequities of the paymentbefore the expanded transfer policy wasput in place. Cases in DRG 483 tend tohave very long lengths of stay (thegeometric mean is 35 days) and receivevery high DRG payments (the paymentrate is more than 10 times the average forall cases). However, hospitals located inareas with facilities that can provideventilator support for these patients arepotentially able to transfer patientsrelatively early in a stay (after as few asthree days) and thus receive a full DRGpayment and a large financial gain. Underthe expanded transfer policy, cases with


Use of post-acute care, transfers, and length of stay by region, 2001

Percent of Percent of allAverage Percent of cases cases dischargedlength transfer cases discharged to to post-acute care

Region of stay with short stays post-acute care with short stays

Total 5.6 days 20.9% 30.5% 6.4%

Middle Atlantic 6.7 17.0 37.2 6.3South Atlantic 5.6 20.0 29.3 5.8East South Central 5.5 19.9 26.1 5.2West South Central 5.5 20.4 22.8 4.7New England 5.5 24.6 46.2 11.4Pacific 5.3 22.8 30.8 7.0East North Central 5.3 22.1 32.3 7.1West North Central 5.1 22.7 28.6 6.5Mountain 4.8 26.1 27.9 7.3

Note: Short stays are those that are more than one day less than the geometric mean length of stay for the DRG.


T A B L E2A-8

Use of post-acute care, Medicare inpatient operating margins, and length of stay, 2001

Hospital group defined Change inby percent of cases Medicare payments iftransferred to Percent of Ratio of actual inpatient margin transfer policypost-acute care Percent of cases discharged to expected excluding DSH and expanded towith short stays hospitals to post-acute care length of stay above cost IME* all DRGs

�2 10% 10% 115% –1.8% –0.2� 2–5 30 24 107 –0.9 –0.7� 5–10 44 33 97 1.8 –1.3� 10–15 13 42 88 7.2 –2.3� 15 3 50 80 10.4 –3.8

Note: DSH (disproportionate share), IME (indirect medical education). Short stays are those that are more than one day less than the geometric mean length of stay for the DRG.

*Portion of the IME adjustment above the relationship between teaching intensity and cost per discharge.


T A B L E2A-9

short stays receive much smaller per diempayments, and the DRG relative paymentweight is raised for the remaining cases.Even though short-stay transfer cases arepaid less than the full DRG amount,analysis shows that transfer payments forDRG 483 are still greater on average thanthe cost of care provided in the hospital(Gillman et al. 2000, HCFA 1998). Theavailability of long-term care hospitalsand SNFs with ventilator support capacityvaries tremendously, and hospitals inclose proximity to these providersbenefited relative to other hospitals beforethe expanded transfer policy was adopted.The expanded transfer policy, however,brings payments more in line with the costof providing care for all hospitals.

Criticisms of the expanded transferpolicy One of the criticisms leveledagainst the transfer policy is that in asystem based on averages, expansion ofthe transfer policy penalizes hospitals forproviding efficient care. However, ifhospitals establish true efficiency gains byreducing length of stay, the transfer policydoes not penalize them. If length-of-staydeclines result from transferring patientsto another setting, this change results in atransfer of costs to another setting, not again in efficiency by the hospital. In suchcircumstances, Medicare ends up payingtwice for the care, once through a fullDRG payment and then again in thepayments made to the post-acute careprovider. The transfer policy allowsMedicare to split the total paymentappropriately between the two providersinvolved in the episode of care. Moreover,even though payments are reduced forshort-stay transfers, they will on averagecontinue to exceed the hospital’s cost ofcare for these cases.

Critics have also argued that the currentpolicy (and its expansion to other DRGs)violates the averaging principle of PPS bytaking away the opportunity for hospitalsto balance losses associated with long staycases with gains on short-stay cases. Thisargument, however, ignores the costreducing effect of site-of-care substitution.The transfer policy treats short-stay cases

that are discharged to post-acute care aspartial cases, reflecting that part of thecare is provided in another setting. Eventhough the policy reduces payments forthese cases, our analysis shows thathospitals on average would continue to bepaid more than the cost of care for thesecases. On average, gains made on short-stay cases would continue to offset losseson high-cost longer stay cases.

Some critics of the transfer policy suggestthat it creates a disincentive to providequality care by encouraging hospitals toattain a target length of stay in each DRG.Without a transfer policy, the currentpayment system gives hospitals anincentive to discharge patients to post-acute care as quickly as possible. Thetransfer policy changes hospitals’financial incentives by setting paymentrates close to the marginal cost of care.The additional financial gains a hospitalmight achieve by keeping the patient anadditional day, however, are small. As aresult, the transfer policy provides a betterbalance between financial and clinicalconsiderations.

R E C O M M E N D A T I O N 2 A - 1

The Secretary should add 13 DRGs tothe post-acute transfer policy in fiscalyear 2004 and then evaluate theeffects on hospitals and beneficiariesbefore proposing further expansions.

I M P L I C A T I O N S 2 A - 1

Spending• This policy would reduce Medicare

payments by between $200 millionand $600 million in the first year andbetween $1 billion and $5 billionover 5 years.

Beneficiary and provider• This policy would not adversely

affect beneficiaries and would betteralign incentives for hospitals as theyconsider when to place patients inpost-acute care.

• It would reduce payments toproviders who discharge manypatients to post-acute care more thanone day before reaching the nationalgeometric mean length of stay forcases in DRGs affected by the policy.

Adding the 13 DRGs considered by theSecretary would allow the transfer policyto capture a larger share of casestransferred to post-acute care providers.With these 13 plus the original 10 DRGs,almost one-third of cases discharged topost-acute care and about two-fifths of theshort-stay transfers would be affected(Table 2A-10). The 13 DRGs have alower percentage of cases transferred topost-acute care settings compared with theinitial 10 DRGs, but a similar proportionof transfer cases with short stays (Table2A-11).


Distribution of hospital cases under the transfer policy, 2001

Share of cases Share of Share of savingsdischarged to transfer if transfer policy

Share of post-acute cases with applied toall cases care short stays all DRGs

DRGs under current policy 9% 20% 26% 34%DRGs under proposed expansion 9 12 16 20All other DRGs 82 68 57 46All DRGs 100 100 100 100

Note: DRGs (diagnosis related groups). Columns may not total to 100 percent because of rounding. Short stays arethose that are more than one day less than the geometric mean for the DRG.


T A B L E2A-10

Adding 13 DRGs to the transfer policywould decrease Medicare payments by 0.4percent, assuming hospitals’ transferringbehavior remains unchanged. Theproportion of all cases affected by thepolicy would increase by about 1percentage point. The effects on PPSpayments would be fairly uniform acrossprovider groups, although this woulddiffer substantially across regions;hospitals in New England would see thelargest decline (0.7 percent) and those inthe West South Central region thesmallest (0.2 percent) (Table 2A-12, p. 54).

Extending the policy to all DRGs wouldreduce Medicare payments by about 1.2percent. About 6 percent of Medicarecases would receive a partial DRGpayment. Despite the drop in Medicarespending, per case payments under theexpanded transfer policy on averagewould remain above the cost of care forcovered cases.

The effects of expanding the transferpolicy to all DRGs would be substantiallylarger than expansion to 13 additionalDRGs, but with similar patterns acrosshospital groups and regions. In NewEngland, which has the highest proportionof cases transferred to post-acute settings,payments would fall by about 2.4 percent,compared with 0.8 percent in the West

South Central Census division, which hasone of the lowest rates of transfer to post-acute care. Differences in the financialimpact for rural and urban hospitals aremostly in the original 10 DRGs; theimpact of expanding to all DRGs is muchmore uniform for the remaining set ofcases.

The indirect medicaleducation adjustment forinpatient payments Teaching hospitals—hospitals that trainphysicians in approved residency trainingprograms—have always had higherMedicare inpatient costs per dischargethan nonteaching hospitals. Part of thecost difference reflects the direct costs ofoperating graduate medical education(GME) programs, such as stipends forresidents, salaries for teaching physicians,and related overhead expenses. Teachinghospitals’ costs per discharge also tend tobe higher for other reasons that areassociated with teaching activity butdifficult to measure directly. Theseinclude unmeasured differences inpatients’ severity of illness, inefficienciesin service use associated with residents’learning by doing, greater use of emergingtechnologies, and so forth.

When the Congress established thehospital inpatient PPS in 1983, it

recognized teaching hospitals’ highercosts in two ways. First, it excluded directGME costs from the PPS payment rates;these costs continued to be reimbursed ona reasonable cost basis. The Congresslater established a separate prospectivepayment for direct GME costs based onhospital-specific costs per resident in 1984trended forward to account for inflation.15

Second, the Congress included an indirectmedical education (IME) adjustment tothe hospital inpatient payment rates. TheIME adjustment is a percentage add-on tothe PPS payment rates for teachinghospitals, which is based on the estimatedrelationship between their Medicare costsper discharge and their teaching intensityas measured by the ratio of residents tobeds. Because of doubts about the abilityof the PPS to fully capture differences inpatient severity and other factors thatmight account for teaching hospitals’higher costs, the Congress required theSecretary to double the empiricallyestimated IME adjustment (see the textbox on the history of the IME adjustmentfor more information on how theadjustment has changed over time, p. 55).Teaching hospitals receive IME add-onpayments for Medicare patients whosecare is paid for under the inpatient PPSand also for those whose care is paid forby a Medicare�Choice plan.


15 Teaching hospitals’ per resident amounts vary widely. In the Balanced Budget Refinement Act of 1999, the Congress established a floor per-resident payment currentlyset at 85 percent of the geographically adjusted national average per-resident amount. The Congress also reduced annual increases in per-resident payments forhospitals with very high per-resident amounts (above 130 percent of the national average, after geographic adjustment).

Characteristics of hospital cases under the transfer policy, 2001

Percent of cases Percent of all Change indischarged to Percent of cases discharged payments from

Share of post-acute transfer cases to post-acute care expandedall cases care with short stays with short stays transfer policy

DRGs under current policy 9% 67% 28% 19% –0.6%DRGs under proposed expansion 9 44 28 12 –0.4All other DRGs 82 25 18 4 –0.8All DRGs 100 31 21 6 –1.8

Note: DRGs (diagnosis related groups). Short stays are those that are more than one day less than the geometric mean length of stay for the DRG.


T A B L E2A-11

Based on current law and the most recentdata, the adjustment is still set at a levelthat is twice the estimated effect ofteaching intensity on hospitals’ costs perdischarge. The Commission haspreviously recommended that theCongress combine IME and direct GMEpayments into a single paymentadjustment that would better account forthe higher costs of inpatient care in

teaching hospitals (MedPAC 2000b). Inthe absence of congressional action,teaching hospitals continue to receiveseparate direct GME and IME payments.This section focuses on the IMEadjustment.

Current IME adjustment Medicare’s IME adjustment is based on astatutory formula that in fiscal year 2003

increases payments by about 5.5 percentfor each 10 percent increment in teachingintensity, as measured by the ratio ofresidents to hospital beds (see text box onIME adjustment formula, p. 56). Theadjustment in fiscal year 2003 is about 15percent lower than it was in fiscal year2002, when it was set at 6.5 percent.Hospitals with a higher ratio receive alarger add-on adjustment to their inpatient


Change in inpatient payments from expanded transfer policy, 2001

Change in Change in Changepayments for the payments for in payments

Number of Number of DRGs under the DRGs under if policy expandedHospital group hospitals cases current policy proposed expansion to all DRGs

Total 4,613 10,812,882 –0.6% –0.4% –1.2%

Urban 2,632 8,646,905 –0.7 –0.4 –1.2Rural 1,656 2,103,922 –0.5 –0.4 –1.2

Large urban 1,537 4,902,476 –0.7 –0.4 –1.2Other urban 1,095 3,744,429 –0.6 –0.4 –1.1Rural referral 248 833,371 –0.5 –0.3 –1.1Sole community 521 476,975 –0.4 –0.4 –1.2Small rural

Medicare dependent 241 183,454 –0.3 –0.4 –1.4Other rural � 50 beds 313 161,707 –0.3 –0.4 –1.3Other rural � 50 beds 333 448,415 –0.5 –0.4 –1.3

Major teaching 298 1,493,872 –0.7 –0.4 –1.3Other teaching 824 3,620,550 –0.6 –0.3 –1.2Nonteaching 3,166 5,636,405 –0.6 –0.4 –1.2

New England 183 537,570 –0.6 –0.7 –2.4Middle Atlantic 474 1,605,852 –0.5 –0.3 –1.1South Atlantic 687 2,084,098 –0.6 –0.3 –1.0East North Central 696 1,914,994 –0.8 –0.4 –1.4East South Central 406 958,806 –0.6 –0.3 –0.9West North Central 571 872,834 –0.6 –0.4 –1.3West South Central 648 1,260,795 –0.7 –0.2 –0.8Mountain 328 467,115 –0.8 –0.4 –1.3Pacific 564 972,134 –0.6 –0.4 –1.3

Voluntary 2,596 7,900,024 –0.6 –0.4 –1.2Proprietary 650 1,239,981 –0.7 –0.3 –1.0Urban government 368 985,048 –0.7 –0.3 –1.1Rural government 669 625,657 –0.4 –0.3 –1.1

Note: DRGs (diagnosis related groups).


T A B L E2A-12

DRG payments. A teaching hospital with400 beds and 40 residents, for example,would receive a payment add-on of 5.3percent for each Medicare discharge infiscal year 2003 compared with anadjustment of 24.1 percent per dischargefor a 400-bed hospital with 200 residents

(Table 2A-13, p. 56). The CongressionalBudget Office (CBO) estimates thatMedicare IME payments will total $5.1billion in fiscal year 2003. Thesepayments go to about 1,100 hospitals thattrain residents, or about one-fourth of allPPS hospitals.

Commission’s views onMedicare’s payments toteaching hospitals In an August 1999 report to the Congresson Medicare payment policies forgraduate medical education and teachinghospitals, the Commission concluded that


History of Medicare’s indirect medical education adjustment for inpatient payments

Medicare’s indirect medicaleducation (IME) adjustmenthas changed over time.

Setting the IME adjustment for thestart of Medicare’s inpatientprospective payment system• Regression analysis was used to

estimate indirect medical education(IME) costs—the relationshipbetween inpatient operating costsper discharge and teaching intensityas measured by the ratio of residentsper bed. This analysis (conducted in1983 using 1981 data) suggestedthat inpatient operating costsincrease by about 5.8 percent forevery 10 percent increase in theresident-to-bed ratio.

• At the start of Medicare’s inpatientprospective payment system, theCongress doubled the IMEadjustment to 11.6 percent, becauseanalyses suggested that teachinghospitals would not fare as well asother hospitals under the newpayment system.1 Doubling theadjustment was the simple, butarbitrary, way the Congress thenchose to ensure that teachinghospitals would not be harmed bythe new payment system. Because

total projected payments were heldconstant, the revenues to double theadjustment were obtained byreducing the base payment rates forall hospitals.

Modifying the IME adjustmentwhen disproportionate sharehospital payments wereintroduced• When the disproportionate share

hospital (DSH) adjustment wasintroduced in 1986, the IMEadjustment was reduced to 8.1percent to help pay for part of thecosts of the new adjustment and toreflect the impact of the DSHadjustment on the empirical level ofthe IME estimate. At this point theadjustment was still set at double therelationship between residentintensity and costs per case.

• With additional expansion of theDSH adjustment, the IMEadjustment was further reduced to7.7 percent in 1988 (1.89 times theempirical level as calculated whenthe DSH adjustment wasimplemented in 1986).

Recent legislative history• The Balanced Budget Act of 1997

(BBA) reduced the level of the IME

adjustment from 7.7 percent in fiscalyear 1997 to:

• 7.0 percent in fiscal year 1998,

• 6.5 percent in fiscal year 1999,

• 6.0 percent in fiscal year 2000,and

• 5.5 percent in fiscal year 2001 andsubsequent years.

• The Balanced Budget RefinementAct of 1999 modified the BBAreductions by holding the IMEadjustment at 6.5 percent throughfiscal year 2000, then lowering theadjustment to 6.25 percent in fiscalyear 2001, and finally reducing it to5.5 percent in fiscal year 2002 andsubsequent years.

• The Medicare, Medicaid, andSCHIP Benefits Improvement andProtection Act of 2000 (BIPA)further delayed the reduction byholding it to an average of 6.5percent in both fiscal year 2001 andfiscal year 2002, before allowing itto fall to 5.5 percent in fiscal year2003. �

1 Two factors contributed to the projected adverse effects on teaching hospitals. First, they understated the complexity of their case mix in the base year, leadingto an underestimate of the prospective payment system (PPS) payments they would receive. Second, the analysis used to estimate the relationship betweenteaching intensity and costs per case included some factors, such as number of beds, which were not a part of the new payment system, lowering the estimatedIME cost relationship. Teaching hospitals in fact did not perform poorly under PPS.

residents bear the cost of their training byreceiving lower wages than they mightotherwise earn and that Medicarepayments for direct GME costs shouldtherefore be considered patient careexpenses (MedPAC 1999). TheCommission consequently recommendedfolding costs for inpatient direct GME intoMedicare’s PPS rates for inpatientservices through a revised adjustment toteaching hospital payments (MedPAC2000b). The Commission alsorecommended that federal policiesintended to affect the number, specialtymix, and geographic distribution of healthcare professionals be implementedthrough specific targeted programs ratherthan through Medicare payment policies.

As part of the Commission’s report onteaching hospitals, we assumed that theIME adjustment would gradually phasedown to 5.5 percent as the BalancedBudget Act of 1997 instructed. Inaddition, last year’s payment updaterecommendation and evaluation ofpayment adequacy for inpatient serviceswas based on the assumption that the IMEadjustment would be set at 5.5 percent infiscal year 2003, down from 6.5 percent infiscal year 2002.

Relationship of Medicare’s IMEpayments to patient care costs Medicare’s IME payments exceed theestimated cost relationship betweenteaching intensity and costs per case. Ourmost recent analysis of the relationshipbetween teaching intensity and patientcare costs, conducted with 1999 costreport data, found that inpatient operatingcosts increase about 2.7 percent for every10 percent increase in the ratio ofresidents to hospital beds (or 2.8 percent ifcapital costs are included). Our analysis of1997 data showed that this relationshipwas 3.2 percent (or 3.1 percent if capitalcosts were included). Payments above thiscost relationship are unrelated to higherpatient care costs or to education andtraining costs of residents—which arepaid separately on a per-resident basis. Infiscal year 2003 these payments (thoseabove the cost relationship) will accountfor about 2.5 percent of Medicareinpatient operating payments.

In conducting our analysis we standardizehospitals’ inpatient costs for cost-relatedpayment factors (the area wage index,case mix, and outlier payments) to reflecthow these factors are used in the PPS.This method allows the IME adjustmentto pick up the effect of any remaining


IME adjustment formula

The current IME adjustmentis based on the followingformula that is multiplied

by hospitals’ base payment rate fora case to determine the IMEpayment:

1.35 � [(1 � number of residents/beds)0.405 � 1]

The formula essentially has threeparts:

• The 1.35 multiplier increasesthe level of the adjustment tothe target level. The 5.5 percentadjustment level for every 10percent increase in the residentto bed ratio is derived bymultiplying 1.35 by the 0.405exponent. This multiplier iswhat the Congress changedwhen it altered the level of theIME adjustment.

• The resident-to-bed ratioreflects the number of residentstraining in the hospital and thenumber of licensed inpatientbeds that a hospital is operating.The resident count used in theIME formula, however, iscapped at 1996 levels.1

• The 0.405 exponent factor wasderived from a CongressionalBudget Office analysis of 1980cost report data on therelationship between teachingintensity and costs per case andseveral other factors. �

1 The Congress capped the number ofresidents in the BBA to counter hospitals’financial incentives to increase residentsin order to raise payments.

Percent increase in inpatient payment rates under alternative levels of the indirect medical

education adjustment

Indirect medical education

Resident-to-bed ratio

adjustment percentage* .05 .10 .25 .50 .75

6.5 3.2% 6.3% 15.1% 28.6% 40.7%5.5 2.7 5.3 12.8 24.1 34.32.7 1.3 2.6 6.2 11.6 16.3

Note: The 6.5 percent indirect medical education (IME) adjustment percentage was in effect from fiscal year 2000through fiscal year 2002. The 5.5 percent IME adjustment started in fiscal year 2003. The 2.7 percentadjustment is the estimated empirical level of the IME adjustment.* Per 10 percent increment of teaching intensity, measured by the ratio of residents to beds.

Source: MedPAC analysis.

T A B L E2A-13

16 Estimated Medicare inpatient margins for major teaching hospitals remain more than 3 percentage points higher than those for nonteaching hospitals afterdisproportionate share hospital (DSH) payments and IME payments above costs are removed from the calculation, also suggesting that the IME estimate is conservative.

In fiscal year 2003, Medicare’s IMEpayments above the empirical costrelationship will total an estimated $2.6billion, accounting for a little more thanhalf of total IME payments received byteaching hospitals. Reducing the IMEadjustment to the empirically justifiedlevel would substantially lower Medicareinpatient payments to teaching hospitals;for major teaching hospitals—those with25 or more residents per 100 hospitalbeds—payments would fall by 7.2percent, and other teaching hospitals’payments would decline by 1.7 percent.Lowering the IME adjustment from 5.5 to5.0 percent would decrease IME paymentsby about 8 percent, or about 0.4 percent oftotal Medicare inpatient revenues, withpayments to major teaching hospitalsfalling 1.3 percent and payments to otherteaching hospitals dropping by 0.3percent.

Financial performance ofteaching hospitals underMedicare Teaching hospitals have substantiallyhigher Medicare margins than otherhospitals. In fiscal year 2000 (the latestdata available), the Medicare inpatientmargin for major teaching hospitals was22.9 percent (Table 2A-14). Thiscompares with 10.2 percent for other

teaching hospitals, and 4.9 percent fornonteaching hospitals. If the IMEadjustment had been set at 5.5 percent infiscal year 2000 instead of 6.5 percent, theinpatient margin would have been 20.7percent for major teaching hospitals and9.5 percent for other teaching hospitals.

The overall Medicare margin (consideringmost Medicare services furnished byhospitals) was also substantially higher formajor teaching hospitals in fiscal year2000: 14.9 percent compared with 5.0percent for other teaching and –0.2percent for nonteaching hospitals (Table2A-15, p. 58). Teaching hospitals’ overallMedicare margins still remainsubstantially higher even after accountingfor the current 5.5 percent IMEadjustment level: 13.1 percent for majorteaching hospitals and 4.5 percent forother teaching hospitals.

In 2000, the portion of the IME paymentabove the measured cost relationshipaccounted for about 10 percent of majorteaching hospitals’ Medicare inpatientpayments. If this portion of IME paymentswere removed, the net inpatient marginfor major teaching hospitals in fiscal year2000 still would have been 13.8 percent,and the overall Medicare margin 7.5percent.


variation in costs not captured in thepayment system that may be related to thelevel of teaching activity in the hospital.These methods tend to produce higherestimates of the effect of teaching onhospital patient care costs than we wouldget if we included other cost factors(patient severity within DRG, forexample) in the analysis. Thus theestimated impact of teaching on hospitalcosts would be lower (and the amount ofpayments above the cost relationshipwould be even higher) if we were tocontrol for other factors like these.16 Wedo not control for these other factors,however, because the payment systemdoes not consider them in setting paymentrates.

The empirical level of the IME adjustmenthas fallen over time, probably as a resultof two factors. One is that teachinghospitals have had lower cost growth thanother hospitals over time. The second isthat increases in the resident-to-bed ratiodo not necessarily correspond to higherpatient care costs. The resident-to-bedratio, for instance, can increase if hospitalsdecrease the number of beds without anychange in the number of residents trained.In addition, the number of residents intraining has also grown by more than 35percent since the beginning of PPS, andincreases in the number of residentstrained may cause little if any increase incosts per case (especially if residentsalaries and benefit costs are excluded andpaid separately as is the case in the currentpayment system).

The calculation of the empirical level ofthe IME adjustment is based on policyparameters at a point in time and maychange somewhat with futuremodifications in the payment system. Forexample, changes in the wage index—such as the addition of an occupationalmix adjustment—might raise the IMEestimate somewhat. On the other hand,case-mix refinements might lower theestimate because more of the difference incosts between teaching and nonteachinghospitals would be captured in measuredcase-mix differences.

Medicare inpatient margin in fiscal year 2000 and at alternative indirect

medical education adjustment levels

HospitalIndirect medical education adjustment percentage

group 6.5 5.5 2.7


Major teaching 22.9 20.7 13.8Other teaching 10.2 9.5 7.6Nonteaching 4.9 4.9 4.9

Note: The 6.5 percent indirect medical education (IME) adjustment percentage was in effect from fiscal year 2000through fiscal year 2002. The 5.5 percent IME adjustment started in fiscal year 2003. The 2.7 percentadjustment is the estimated empirical level of the IME adjustment. Margins were imputed for hospitals whose2000 cost reports were not available (about 27 percent of observations). Excludes critical access hospitals.

Source: MedPAC analysis of Medicare cost report data (fourth quarter 2002) from CMS.

T A B L E2A-14

Medicare inpatient margins grew for allhospitals in the 1990s, but the largestgrowth was for major teaching hospitals,which saw Medicare inpatient marginsclimb from 6 percent in 1990 to 26

percent in 1997 (Figure 2A-8). In contrast,inpatient margins for nonteachinghospitals rose from –5 percent in 1990 to12 percent in 1997. Recently, Medicareinpatient margins have fallen from their

1997 peak, down to 23 percent for majorteaching hospitals and to 5 percent fornonteaching hospitals.

Uncompensated care in teachinghospitals One argument against reducing theindirect medical education adjustment isthat teaching hospitals provide asubstantial amount of uncompensatedcare, which the IME payments mayoffset.17 The cost burden ofuncompensated care, however, is notuniform across teaching hospitals. AHAannual survey data show thatuncompensated care accounts for 20percent of costs in public major teachinghospitals compared with 5 percent inprivate major teaching hospitals (Figure2A-9). Private major teaching hospitalsprovide about the same amount ofuncompensated care as other privatehospitals.

IME payments are not targeted tohospitals with the most uncompensated


Overall Medicare margin in fiscal year 2000 and at alternative indirect medical

education adjustment levels

HospitalIndirect medical education adjustment percentage

group 6.5 5.5 2.7


Major teaching 14.9 13.1 7.5Other teaching 5.0 4.5 2.9Nonteaching –0.2 –0.2 –0.2

Note: The 6.5 percent indirect medical education (IME) adjustment percentage was in effect from fiscal year 2000through fiscal year 2002. The 5.5 percent IME adjustment started in fiscal year 2003. The 2.7 percentadjustment is the estimated empirical level of the IME adjustment. Margins were imputed for hospitals whose2000 cost reports were not available (about 27 percent of observations). Excludes critical access hospitals.


T A B L E2A-15

Change in Medicare inpatient margins, by teaching status, 1990–2000

Note: Major teaching hospitals have at least 25 residents per 100 hospital beds.


1990 1992 19931991 1994 1995 1996 1997 1999 20001998

20

25

10

15

0

5

�10

�5

30

Marg

in

NonteachingOther teachingMajor teaching

FIGURE2A-8

17 Uncompensated care is defined as care provided by hospitals or other providers that is not paid directly (by the patient, or by a government or private insuranceprogram). It includes charity care, which is furnished without the expectation of payment, and bad debts, for which the provider has made an unsuccessful effort tocollect payment due.

care. Only 27 percent of major teachinghospitals and 8 percent of other teachinghospitals, for instance, are public hospitalsthat tend to have higher than averagelevels of uncompensated care. Further,because Medicare accounts for only 20percent of patient care costs in publicmajor teaching hospitals, IME paymentsabove the cost relationship can have onlylimited effectiveness in helping defraythese hospitals’ uncompensated careburdens. Moreover, the variable thatdetermines IME payments, the resident-to-bed ratio, does not reflectuncompensated care costs.

Medicare’s DSH payments are explicitlydesigned to help hospitals with a highshare of low-income patients and,presumably, a high load ofuncompensated care. In 2000, teachinghospitals received about $3 billion or two-thirds of Medicare DSH payments.Teaching hospitals’ share ofuncompensated care costs was about 62percent in 2000. Teaching hospitals alsoreceived more than 80 percent of state andlocal payments for uncompensated care,the vast majority of which went to publicmajor teaching hospitals.

Hospitals can also make up for lossesfrom uncompensated care through otherpayments, such as those from privatepayers and nonpatient care revenues. Onaverage, private major teaching hospitalshave much lower payment-to-cost ratiosfrom private payers than other providers,3.4 percent compared to 12.5 percent forall providers.18 Although public majorteaching hospitals have private payerpayment-to-cost ratios that are muchhigher than average—25.8 percent—private payers account for a much smallershare of their case load. Teachinghospitals use nonpatient care revenue(e.g., endowments, parking) more tosupport operations than other facilities,accounting for 6.5 percent of total revenuein major teaching hospitals, compared to3.8 percent in nonteaching hospitals.

Conclusion The Commission believes that Medicareshould recognize the higher costs teachinghospitals incur in caring for beneficiaries.The IME adjustment currently providespayments well above an empiricallyjustified level. The Commission is notsatisfied with the current policy because itprovides payments to teaching hospitalsabove the empirically justified levelwithout accountability for their use orwithout targeting policy objectivesconsistent with Medicare’s goals.However, we were not able to reachconsensus on reducing the adjustment tothe empirical level at this time. To addressthis problem, the Commission willexplore ways to target some or all of theIME payments above the empiricallyjustified level to advance specificMedicare policy objectives such asproviding enhanced medical education tobetter prepare providers with the capacityto manage the changing needs ofMedicare beneficiaries. The Commission

believes this problem should be addressedpromptly.

Inpatient payments for rural hospitals In a Congressional report devotedexclusively to rural health care issues,MedPAC found that rural hospitals onaverage had worse financial performanceunder Medicare than their urbancounterparts (about 7 percentage pointslower on both Medicare inpatient andoverall Medicare margins). TheCommission responded to this finding byreviewing Medicare’s payment policiesand making four recommendationsdesigned to improve inpatient payments torural hospitals (MedPAC 2001a). Thenext year we issued a fifthrecommendation with a similar objective(MedPAC 2002). In each case, ourrecommendation was based on evidencethat the current payment system does notaccount for factors that systematicallyraise some providers’ unit costs beyondtheir short-term control, or that the current


Uncompensated care costs as a percentage oftotal hospital costs, by hospital group, 2000

FIGURE2A-9

Source: 2002 American Hospital Association annual survey of hospitals.

Unco

mpen

sate

d c

are

cost

s (p

erce

nt)

15

25

20

5

0Public major

teaching

20.1

10

Private majorteaching

5.3

12.7

4.4

6.2

4.6

Publicnonteaching

Privatenonteaching

Public otherteaching

Private otherteaching

18 The payment-to-cost ratio is expressed as a percentage by taking the ratio, subtracting 1.0 and multiplying by 100.

system does not treat rural and urbanhospitals equitably.

CMS has already implemented one of therecommendations we made in the ruralreport administratively. Thatrecommendation was to implementimmediately (in contrast to a three-yearphase-out) the policy of excluding thesalaries of personnel categories paid underPart B from the hospital wage index.Because these personnel—teachingphysicians, residents, and certifiedregistered nurse anesthetists—all receiverelatively high wages and are morefrequently employed by urban than ruralhospitals, excluding them in calculatingthe wage index modestly increasespayments for areas with low wage indexvalues (mostly rural) and decreasespayments for areas with high wage indexvalues (mostly urban).

The other four recommendations wouldrequire legislative changes, and althoughthe Congress has considered all four, nonehas been enacted to date. We are repeatingthese four recommendations this year. Inthis section we summarize therecommendations along with the rationalefor and impact of each. Appendix Cprovides additional background,explanation, and support for the fourrecommendations.

Implementing a low-volumeadjustment The inpatient PPS applies the same baserate to payment for hospitals of all sizes.Our analysis revealed that hospitals with asmall volume of total discharges havehigher costs per discharge than largerfacilities, after controlling for the othercost-related factors recognized in thepayment system.19 Thus, the currentsystem places smaller providers at afinancial disadvantage. The critical accesshospital, sole community hospital, andMedicare-dependent hospital programsbenefit many small and isolated hospitals,but eligibility for these programs is not

well targeted to those with low dischargevolume. Consequently, low-volumehospitals on average have much lowerMedicare inpatient margins than largerfacilities.

A low-volume adjustment is most criticalfor isolated hospitals, where the facility isimportant for maintaining beneficiaries’access to care. Adjusting payments for alow-volume hospital that is near otherfacilities offering similar services, on theother hand, is not a priority; in fact, theclose proximity of two hospitals in thesame rural market may be one of theprimary reasons for the low volume ofservice.


The Congress should enact a low-volume adjustment to the rates usedin the inpatient PPS. This adjustmentshould apply only to hospitals thatare more than 15 miles from anotherfacility offering acute inpatient care.


Spending• This policy change would be

implemented with new monieswithout a phase-in schedule, but it isexpected to increase total spendingfor PPS inpatient services by lessthan $50 million in the first year andless than $250 million over fiveyears.

Beneficiary and provider• This additional payment option

should help maintain access to basicemergency service and inpatient carein isolated rural areas by maintainingthe financial viability of small ruralhospitals. A number of theseinstitutions do not qualify forassistance under the current paymentmechanisms designed to help ruralhospitals within the PPS.

• A low-volume adjustment willprovide substantial financial relief to

small and isolated rural hospitals,enabling some to earn a margin ontheir inpatient services by remainingin the PPS rather than electing cost-based payment through the criticalaccess hospital program.

Reducing the labor share usedin geographic adjustment The labor share is an estimate of thenational average proportion of hospitals’costs associated with inputs that aredirectly or indirectly affected by localwage levels. It is used to determine theportion of the PPS base payment rate towhich the wage index is applied. Forinpatient hospital services, the labor sharecurrently is set at 71.1 percent.

Most of the inputs that CMS has includedwithin the labor share are purchased inlocal markets. However, a number ofcategories (data processing andaccounting services, for instance) appearto include some inputs that are purchasedin national markets and some in localmarkets. As a result, the national averagelabor share may be somewhat lower than71.1 percent.

Since our rural report, we have obtainedpreliminary results from an analysis of thefactors explaining variation in hospitals’costs per discharge that provide strongevidence that the current labor share is toohigh. However, the study found that,contrary to what many observers haveassumed, the labor-related share ofexpenses is lower in high-wage markets(most of which are in large urban areas)than in low-wage markets (most of whichare rural). This pattern occurs becausehospitals in major metropolitan areasgenerally provide more sophisticatedservices and treat more complex patients,which raises their costs for plant andequipment. In the coming year, MedPACwill undertake a follow-up study designedto identify the best labor share value forthe hospital industry as a whole.


19 Although Medicare payments are intended to cover the costs of Medicare patients, a hospital’s total volume of service (that is, including patients covered by all payers)determines its unit costs of production.


The Secretary should reevaluate thelabor share used in the wage indexsystem that geographically adjustsrates in the inpatient PPS, with anyresulting change phased in over twoyears.


Spending• Any change in the labor share used

for geographic adjustment of ratesshould be implemented budgetneutrally, such that it would have noimpact on aggregate spending forPPS inpatient services.

Beneficiary and provider• By better aligning payments to

efficient providers’ costs, a lowerlabor share should contribute tomaintaining access to care in low-wage communities, many of whichare in isolated rural areas.

• Depending on the exact labor sharechosen, this recommendation shouldmarginally increase payments forhospitals in areas with below-averagewage index values (mostly ruralareas) and marginally reduce them inareas with above-average values(mostly large urban areas).

Eliminating the base ratedifferential In Medicare’s inpatient PPS, the operatingbase payment rate for hospitals in largeurban areas (metropolitan areas with morethan 1 million people) is 1.6 percent abovethe payment rate for other hospitals, andthe differential is 3.0 percent for thecapital base rate (comprising about 10percent of the overall rate).

When we compared hospitals’ costs bylocation, we found no statisticallysignificant difference between the costs ofhospitals in large urban and other areasafter controlling for other cost-relatedpayment adjustments in the inpatient PPS.In addition, after removing the effects ofDSH payments and IME payments above

the measured relationship betweenteaching and unit costs, hospitals in largeurban areas still have Medicare inpatientmargins that are three percentage pointsabove those of hospitals in other urbanand rural areas.


The Congress should raise theinpatient base rate for hospitals inrural and other urban areas to thelevel of the rate for those in largeurban areas, phased in over twoyears.


Spending• Because this policy change would be

implemented with new monies, itwould raise aggregate spending forPPS inpatient payments by between$200 and $600 million in the firstyear and between $1 and $5 billionover 5 years.

Beneficiary and provider• This policy change should help to

maintain access to care in rural andless populated urban areas of thecountry by better aligning hospitals’payments to their average costs.

• The change will increase paymentsfor hospitals in rural and other urbanareas, while having no impact onhospitals located in large urban areas.

Raising the cap ondisproportionate sharepayments Medicare’s DSH adjustment for hospitalinpatient services is designed primarily tooffset the financial pressure ofuncompensated care. However, theCommission has concluded that thecurrent system has two key design flaws(MedPAC 2000b, 2001a):

• The current low-income sharemeasure (used to distribute DSHpayments) does not includeuncompensated care.

• The system has separate paymentrates for 10 hospital groups, with theleast favorable rates given to mostrural hospitals and to urban facilitieswith fewer than 100 beds.

Previous legislation mandated that CMScollect the uncompensated care dataneeded to reform the system and partiallyaddressed the unequal treatment of ruralfacilities. Since MedPAC’s completereform package probably cannot beimplemented until at least fiscal year2005, because of the time required tocollect and process uncompensated caredata, an appropriate interim step isneeded.


The Congress should raise the cap onthe disproportionate share add-on ahospital can receive in the inpatientPPS from 5.25 percent to 10 percent,phased in over two years.


Spending• This policy change would be

implemented with new monies. Dueto the 2-year phase-in schedule, thefirst-year impact on aggregatespending for PPS inpatient serviceswould fall into the $50 to $200million range.

• Over five years, it would raisespending by between $200 millionand $1 billion.

Beneficiary and provider• Because this policy change would

mitigate the effects ofuncompensated care for many ruralhospitals, it should help protectaccess to care for Medicarebeneficiaries in rural communities.

• The policy would raise payments forqualifying rural hospitals as well asurban facilities with fewer than 100beds. Other hospitals would not beaffected.


Impact of ruralrecommendations (2A-2 through 2A-5) In two instances—our recommendationscalling for Congress to implement a low-volume adjustment and to reevaluate thelabor share used in its geographicadjustment of rates—the impact woulddepend on design decisions that Congressor CMS must make. But we havesimulated sample designs to illustrate thepattern and general magnitude of impactthese policy changes would likely have.

To illustrate the financial impact of a low-volume adjustment, we simulated anadjustment that increases payments by upto 25 percent and drops to zero forhospitals with 500 or more discharges.20

Payments would rise by about 8 percentfor hospitals with fewer than 200discharges and 4 percent for those with201 to 500 discharges. Since low-volumehospitals account for a small share ofMedicare discharges, however, theaggregate increase in payments across allhospitals would be less than 0.1 percent.

We simulated an illustrative change inlabor share from 71.1 percent to 68percent. On average, this change wouldraise payments for hospitals in both ruraland small urban areas by 0.2 percentwhile reducing payments for those inlarge urban areas by the same amount. Bydesign, the change would have no effecton overall payments.

Our recommendation to eliminate thedifferential in base payment rates wouldraise payments for hospitals in rural areasby 1.2 percent. With the two year phase-inwe are recommending, a 0.3 percentincrease in funding would be needed infiscal year 2004, followed by a 0.4 percentincrease in 2005.

Implemented with new funding, ourrecommendation to raise the cap on theDSH add-on from 5.25 to 10 percentwould increase rural hospitals’ paymentsby 1.2 percent. However, because ruralhospitals account for only about 15

percent of PPS spending, the changewould increase aggregate inpatientpayments by 0.2 percent. With a two-yearphase-in schedule, an increase in fundingof 0.1 percent would be needed in each offiscal years 2004 and 2005.

As shown in Table 2A-16, the fourrecommendations combined wouldincrease rural hospitals’ payments by 1.3percent in 2004 and 2.6 percent in 2005,eliminating more than a third of thedifference in inpatient margins betweenrural and urban facilities. (The impact ofeach policy change implemented inisolation is detailed in Appendix C.)Although the policy changes affect ruralhospitals the most, hospitals in smallurban areas would receive a 1.7 percentincrease because the recommendedincrease in base rates applies to them.

Payments would decline by 0.1 percentfor hospitals in large urban areas becauseof budget neutral implementation of thereduction in the labor share. By far thelargest payment increases—over 4percent—would go to hospitals that do notbenefit from any of the existing programsaimed at helping rural hospitals. Thesefacilities currently have the lowestinpatient margins.

Update for inpatientservices

We now turn to the question of theappropriate payment update for inpatientservices in fiscal year 2004. TheCommission concluded that payments areadequate in light of current costs. The


20 The formula we used in this simulation, which derives from our multivariate cost analysis, is documented in Appendix C.

One- and two-year impacts on Medicare inpatient payments of recommendations to improve

payments to rural hospitals

Hospital BaselineChange in payments

group margin One-year Two-year


Urban 11.3 0.3 0.6Rural 3.9 1.3 2.6

Large urban 13.6 –0.1 –0.1Other urban 7.7 0.8 1.7Rural referral 3.9 1.3 2.5Sole community 4.6 0.5 1.1Small rural Medicare-dependent 7.2 1.6 3.0Other rural � 50 beds 3.7 2.3 4.4Other rural � 50 beds 2.5 2.1 4.2

Major teaching 20.7 0.1 0.1Other teaching 9.6 0.4 0.8Nonteaching 5.4 0.7 1.3

Note: Baseline margin is the actual 2000 margin adjusted to reflect the increase in disproportionate share paymentsimplemented in 2001 and the decrease in indirect medical education payments implemented in 2003.Margins were imputed for hospitals whose 2000 cost reports were not available (about 27 percent ofobservations). Excludes critical access hospitals.

Source: MedPAC analysis of cost report, MedPAR, and impact file data from CMS.

T A B L E2A-16

update must account for the expectedincrease in efficient providers’ costs.

Accounting for cost changesin the coming year After any adjustments to the update forpayment adequacy, the Commissionexamines likely changes in providers’costs in the coming year. The estimate ofchanges in the costs of efficient providersreflects expected changes in prices, theimpact of the costs of scientific andtechnological advances that improvequality but increase costs, and expectedimprovements in productivity.

Changes in input prices CMS measures price inflation for thegoods and services that hospitals use inproducing inpatient services with thehospital market basket. Separate marketbaskets measure operating and capital costchanges. CMS’s latest forecast for fiscalyear 2004 is 3.5 percent for the operatingmarket basket and 1.4 percent for thecapital market basket. Under current lawthe operating update will equal the rate ofincrease in the market basket, while thecapital update is not specified by law andis set by CMS.

Scientific and technologicaladvances Technological advances may increase thecosts hospitals incur in providing care toMedicare beneficiaries. MedPAC takesaccount of this in its updaterecommendation based on information onanticipated technological changes in thehospital industry in the coming year.Although we have not conducted acomprehensive review of new technology,we note that CMS approved only onetechnology for inpatient technology pass-through payments. Accordingly, webelieve that an allowance of 0.5 percentfor fiscal year 2004 will compensateadequately for this one majortechnological advance as well asnumerous other smaller advances.

Increases in productivity The Commission believes that hospitalsshould be able to cover the costs of

technological advances with the savingsresulting from productivity gains.Increases in productivity decrease hospitalunit costs. An index of productivitychange estimates the change in outputassociated with a given increase in inputs.MedPAC has established a standard forexpected productivity growth based on the10-year average growth rate of total factorproductivity in the general economy,which currently equals 0.9 percent.Productivity growth has been even higherthan this average in the last several years.

Update recommendation Medicare separately updates payments foroperating costs (such as labor andsupplies) and capital costs (primarilybuildings and equipment) in the PPS foracute inpatient services. The Congress setsthe update for operating payments, usuallyseveral years in advance, and CMS setsthe capital update. The Commission’sinpatient update recommendation focusessolely on the operating update becauseoperating costs account for about 92percent of total hospital costs and becausethe operating update is of primary interestto the Congress.


The Congress should increasepayment rates for the inpatient PPSby the rate of increase in the hospitalmarket basket, less 0.4 percent, forfiscal year 2004.


Spending• This recommendation would increase

payments by a smaller amount thanunder current law. Consequently, itwould result in savings of between$200 and $600 million in one year.Over 5 years, the savings would bebetween $1 and $5 billion.

Beneficiary and provider• The recommendation results in a

payment increase that should beadequate to cover increases inprovider costs for 2004. To the extentthat adequate payment allows

hospitals to meet beneficiaries’ healthcare needs, beneficiaries’ access tocare would be unchanged.

• The recommended update wouldincrease Medicare inpatient paymentsto hospitals covered by the inpatientPPS by 3.1 percent in fiscal year2004. In combination with theCommission’s recommendations onexpansion of the post-acute caretransfer policy and its ruralrecommendations the updaterecommendation would increasepayments by 3.2 percent (Table 2A-17, p. 64).

The increase in the market basket and therecommended offset for the costs oftechnological advances net of productivitychange affect recommended payments toall hospitals equally. The distributionalimpacts of the rural and transfer policyrecommendations affect hospital groupsdifferently. Together, the Commission’srecommendations lead to paymentincreases of 4.2 percent for hospitals inrural areas, 3.6 percent for hospitals inother urban areas, and 2.7 percent forhospitals in large urban areas. Paymentswould rise 3.5 percent for nonteachinghospitals, 3.2 percent for other teachinghospitals, and 2.8 percent for majorteaching hospitals.

Update for outpatientservices

At the beginning of this chapter, wereviewed the adequacy of Medicare’spayments in relationship to current costsfor most of hospitals’ services and foundthem to be at least adequate. AlthoughMedPAC considers Medicare paymentadequacy for the hospital as a whole, wemake a separate update recommendationfor hospital outpatient services covered byMedicare’s outpatient PPS.

As shown in Table 2A-2 (p. 40), theaggregate margins for Medicare hospitaloutpatient services improved between1999 (–16.4 percent) and 2000 (–13.7percent). The improved margins areconsistent with policies that added funds


to the outpatient PPS: transitional corridorpayments to limit hospitals’ losses underthe new payment system and newtechnology payments. The transitionalcorridor payments made up some of thedifference between what hospitalsreceived under the PPS and what theywould have received under previouspayment policies for hospitals thatreceived lower reimbursements under thePPS. Hospitals receiving higherreimbursements under the PPS kept thegains. The Congress authorized newmonies to fund these payments. Incontrast, the pass-through payments forcertain new technologies are budgetneutral by law.21 However, from August2000 to April 2002, CMS did not enforcethe budget-neutrality provisions due toadministrative and congressional actions.

As discussed previously, the largenegative values for the outpatient margins

appear to be the result of cost allocationdecisions by hospitals, where adisproportionate share of fixed costs seemto be allocated to outpatient servicesrather than to inpatient services.22

Consequently, the outpatient margins areunderstated and the inpatient marginsoverstated. In examining overall Medicarepayments to hospitals in relationship tocosts, the fiscal year 2000 margin is 5.0percent, with an estimated overallMedicare margin of 3.9 percent in 2003(Table 2A-3, p. 41). This and otherindicators, including volume, entry andexit, and access to capital suggest thatpayments are at least adequate.

The Congress mandated development ofthe outpatient PPS in the BBA; it wasimplemented in August 2000. Unlike thehospital inpatient PPS, the outpatient PPSoperates on a calendar year. Updates foroutpatient services were set in legislation

for calendar years 2001 and 2002. TheSecretary set the update for 2003 at theprojected rate of increase for the hospitalmarket basket. Current law also providesfor an update equal to the rate of increasein the hospital market basket for 2004.

Trends in Medicarepayments for outpatientservices Total Medicare payments for servicescovered by the outpatient PPS in calendaryear 2001 were $16.3 billion, including$9.2 billion by the program and $7.1billion in beneficiary cost-sharing.23 This$16.3 billion represents about 6 percent oftotal Medicare spending. Given that theoutpatient PPS was implemented inAugust 2000, calendar year 2001 is thefirst year in which spending data areavailable specifically for services coveredby the outpatient PPS.

In 2001, services covered under theoutpatient PPS represented about 87percent of all hospital outpatient spending.Hospital outpatient services not coveredby the outpatient PPS include those paidon a separate fee schedule (e.g.,ambulance, clinical lab services,rehabilitation and other therapies, anddurable medical equipment), as well asthose still reimbursed on a cost basis (e.g.,organ acquisition, and, beginning in 2003,some vaccines).

Information on trends in Medicarespending on outpatient services is onlyavailable for all outpatient services, notjust those covered under the outpatientPPS. Such spending has grownconsiderably over the past decade, almostdoubling in nominal dollars from calendaryear 1991 to 2001 (Figure 2A-10).


Impact on Medicare inpatient payments of update and distributional recommendations

Hospital Market Update Distributional Net changegroup basket offset changes in payments

All hospitals 3.5% –0.4% 0.1% 3.2%

Large urban 3.5 –0.4 –0.4 2.7Other urban 3.5 –0.4 0.5 3.6Rural 3.5 –0.4 1.1 4.2

Major teaching 3.5 –0.4 –0.3 2.8Other teaching 3.5 –0.4 0.1 3.2Nonteaching 3.5 –0.4 0.4 3.5

Note: Recommendations include the update; a low-volume adjustment; eliminating the base rate differential; reducingthe labor share; raising the cap on disproportionate share payments; and expanding the transfer policy.Payments are imputed for hospitals whose 2000 cost reports were not available (about 27 percent ofobservations). Excludes critical access hospitals.

Source: MedPAC analysis of Medicare cost report, MedPAR, impact file, and market basket data from CMS.

T A B L E2A-17

21 See Chapter 4 for further discussion of the pass-through payment mechanism.

22 The Health Care Financing Administration (now CMS) commissioned a study of hospitals’ cost allocation practices and found that the general pattern of over-allocationto outpatient services existed, at least in part as a response to the introduction of prospective payment for inpatient services, while outpatient services continued to bereimbursed based on reported costs (CHPS Consulting 1994).

23 Beneficiary cost-sharing for hospital outpatient services has not been based on 20 percent of total payments, as it has been for most other Part B services. Historically,the Medicare program based its payments on hospitals’ costs, whereas the beneficiary coinsurance was based on 20 percent of charges. Over time, charges increasedmore quickly than costs, resulting in beneficiaries paying a greater share of total payments, reaching 50 percent by 2000. This trend was reversed under the outpatientPPS, and beneficiary cost-sharing will slowly decline, although it will continue to be greater than 20 percent for the foreseeable future.

Growth was fastest early in the 1990s andslowed from 1997 to 2001.

Several factors contributed to the slowingof growth since the mid-1990s, includingpolicy changes such as the elimination ofinadvertent overpayments in the BBA andthe introduction of Medicare’s outpatientPPS in 2000.24 Other explanatory factorsmight be reactions to stepped-up fraudand abuse efforts and a slowing ofmedical inflation in the late 1990s.Projections by both the CMS Office of theActuary and the Congressional BudgetOffice, however, forecast future growth.Payments under the outpatient PPS areprojected to increase at an average annualrate of about 8 percent between calendaryears 2002 and 2007.

Payments for outpatient servicesaccounted for approximately 14 percent ofMedicare payments to hospitals in 2000(Figure 2A-1, p. 36).

Accounting for cost changesin the coming yearWe now turn to factors likely to affecthospitals’ unit costs for outpatient servicesin 2004, such as changes in input prices,scientific and technological advances, andincreases in productivity.

Changes in input prices The hospital market basket forecast is ourbest approximation of increases in inputprices paid by providers. The outpatientupdate will be implemented on January 1,in contrast to October 1 for the inpatientupdate. The latest forecast of the hospitalmarket basket for calendar year 2004 is3.4 percent.

Scientific and technologicaladvances Technological advances may increase ordecrease unit costs for outpatient servicesin 2004, but most new outpatienttechnologies that increase costs will be

paid for explicitly through two specialprovisions discussed below:

• new technology ambulatory paymentclassification groups; and

• transitional pass-through payments.25

Given these special mechanisms to payfor new technology, MedPAC concludesthere is no need for an addition to theoutpatient PPS update for scientific andtechnological advances in 2004.

New technology ambulatorypayment classification groups The new technology APCs pay forcompletely new services, such as apositron emission tomography (PET) scanor a new surgical procedure. Services areplaced in a new technology APC basedonly on their expected costs. New

technology APCs start at $0 to $50 andcontinue through $5,000 to $6,000, withan additional category for $19,500 to$20,500. Payment is set at the midpoint ofthe range. Currently, 75 services (asdenoted by a Healthcare CommonProcedure Coding System, or HCPCS,code) are paid for under the newtechnology APCs. In addition, CMS hasfive applications under review for servicesto be placed in new technology APCs.Technologies that fall into newtechnology APCs will generate paymentsfor each service rendered. This paymentmechanism has no budget neutralityprovision, so these payments representincreased expenditures. The costs of newtechnologies covered by the newtechnology APCs, therefore, do not needto be factored into the update. In 2001,payments for services in new technology


Spending on all hospital outpatient services,1991–2001

FIGURE2A-10

Note: Spending includes both services covered by the outpatient prospective payment system and those paid onseparate fee schedules or on a cost basis.

Source: Office of the Actuary, CMS.

Spen

din

g (

dolla

rs in

bill

ions)

14

12

10

20

18

16

8

6

4

2

019921991 1993 1994 1995 1996 1997 1998 1999 2000 2001

Beneficiaries Program

24 The BBA eliminated so-called formula-driven overpayments, which were generated by a mistake in the payment formula for some ambulatory surgery, radiology, andother diagnostic services that inadequately accounted for beneficiary copayments when setting program payments, leading to excessive total payments.

25 See Chapter 4 for a full discussion of these payment mechanisms for new technology.

APCs accounted for about 1 percent oftotal payments.26

Transitional pass-through paymentsPass-through payments cover technologiesthat are inputs to a service, such as a drugor medical device, rather than a service asa whole. Pass-through payments are madein addition to base APC payments. TheCongress required CMS to implement thepass-through payments in a budget neutralmanner, with a cap of 2.5 percent of totalpayments. If CMS estimates that pass-through payments will be above the cap,all payments should, by law, be subject toa pro rata reduction. From August 2000 toApril 2002, however, no pro ratareduction was made. Consequently, in2001, payments for pass-through itemsexceeded $1.3 billion (8 percent of totalpayments), rather than the limit of about$450 million (2.5 percent of totalpayment).27 Thus, excess payments ofabout $750 million were made. For thelast nine months of 2002, however, CMSimposed a pro rata reduction of 64 percenton pass-through payments to ensure thecap was met.

CMS estimates that pass-throughspending for calendar year 2003 will bebelow the cap. Projections by industry andCMS suggest that the same will be true in2004 (2004 marks a change in thestatutory limit for the cap from 2.5 percentto 2 percent). Currently, fewer than 10applications for new pass-throughtechnologies are pending. Therefore, thefull costs of pass-through items should becovered by the payment mechanism.

If estimated pass-through paymentsexceed the cap in 2004, requiring a prorata reduction, some might argue that thereductions in payments represent coststhat are not covered by the paymentsystem that should be factored into theupdate. If this situation arises, however, ajudgment would be needed to determinewhether the reduced payments actually

cover hospitals’ costs for these items. Theestimated payments are based on theexisting payment mechanisms, which theCommission has previously stated couldresult in overpayments (MedPAC 2002)and likely exceed providers’ costs.Payments for pass-through drugs equal 95percent of average wholesale price,generally considered to be well aboveproviders’ acquisition costs. Payments fordevices equal hospitals’ charges reducedto costs using a cost-to-charge ratio. Thispayment mechanism provides hospitals anincentive to increase charges to increasepayments.

Increases in productivity Whereas technological advances mayincrease or decrease the unit costs ofproviding services, increases inproductivity decrease unit costs. Last year,MedPAC conservatively assumed that theincreases in unit costs from newtechnologies were offset by improvedproductivity. We acknowledged that thisassumption was likely to benefit hospitals,given the limited number of pass-throughtechnologies expected to be approved in2003. The decision hinged on the newnessof the payment system and the uncertaintyover the flow of pass-through items. Theexperience in setting rates for 2003,however, has confirmed that fewertechnologies are currently flowingthrough the pass-through mechanism.Consequently, this year we conclude thatmost increases in costs of technology arealready reflected in the payments and donot offset productivity gains.

Given that prospective payment systemsare designed to provide incentives forefficiency, hospitals should be expected toimprove productivity at a rate that isconsistent with multifactor productivityimprovement in the economy as a whole.The latest estimate of the 10-year movingaverage of multifactor productivity in the

economy as a whole is 0.9 percent. Thisestimate averages lower productivitygrowth in the past with larger increases inmore recent years.

Update recommendation After reviewing the adequacy of currentpayment and costs, as well as the factorslikely to affect hospitals’ costs in calendaryear 2004, we make the followingrecommendation:


The Congress should increasepayment rates for the outpatient PPSby the rate of increase in the hospitalmarket basket, less 0.9 percent, forcalendar year 2004.



payments by a smaller amount thanunder current law. Consequently, itwould result in savings of between$50 and $200 million in one year.Over 5 years, the savings would bebetween $250 million and $1 billion.

Beneficiary and provider• Although it is below the update

established in current law, thisrecommendation would result in apayment increase that is adequate tocover increases in provider costs foroutpatient services for 2004.Hospitals should be able to realizeproductivity gains to partially offsetthe increases in input prices reflectedin the hospital market basket.

• To the extent that adequate paymentallows hospitals to meetbeneficiaries’ health care needs,beneficiaries’ access to care would beunchanged. �


26 Based on MedPAC analysis of 2001 outpatient PPS claims from CMS.

27 Based on MedPAC analysis of 2001 outpatient PPS claims from CMS.

References

American Hospital Association. Closed hospitals–2001. Washington (DC), AHA. 2002.

Ashby J, Guterman S, Greene T. An analysis of hospital productivity and product change,Health Affairs. September/October 2000, Vol. 19, No. 5, p. 197–205.

Carter G, Rumpel D. Report to the Health Care Financing Administration under contractNo. 500–92–0023, an evaluation of Medicare payment for transfer cases. Santa Monica(CA), RAND. 1993.

Center for Health Policy Studies (CHPS Consulting). Final report: outpatient resourcecosting study. Columbia (MD), CHPS. December 1994.

Centers for Medicare & Medicaid Services, Department of Health and Human Services.Medicare program; changes to the hospital inpatient prospective payment systems andfiscal year 2003 rates; final rule, Federal Register. August 1, 2002a, Vol. 67, No. 148, p. 49982–50289.

Centers for Medicare & Medicaid Services, Department of Health and Human Services.Medicare program; changes to the hospital inpatient prospective payment systems andfiscal year 2003 rates; proposed rule, Federal Register. May 9, 2002b, Vol. 67, No. 90, p. 31404–31689.

Fitch Ratings. Health care special report: health care ratings actions for the 12 monthsended December 31, 2002. January 6, 2003. Available at http://www.fitchratings.com/.

Gillman BH, Cromwell J, Adamache KW, Donoghue S. Report to the Health CareFinancing Administration under contract No. 500–95–0006. Study of the effect ofimplementing the postacute care transfer policy under the inpatient prospective paymentsystem. Waltham (MA), Health Economics Research, Inc. July 31, 2000.

Health Care Financing Administration, Department of Health and Human Services.Medicare program; changes to the hospital inpatient prospective payment systems andfiscal year 1999 rates, Federal Register. July 31, 1998, Vol. 63, No. 147, p. 40954–41131.

Kane NM. The Medicare cost report and the limits of hospital accountability: improvingfinancial accounting data, Journal of Health Politics, Policy, and Law. February 2001,Vol. 26, No. 1, p. 81–105.

Levit K, Smith C, Cowan C, et al. Trends in U.S. health care spending, 2001, HealthAffairs. January/February 2003, Vol. 22, No. 1, p. 154–164.

Medicare Payment Advisory Commission. Report to the Congress: context for achanging Medicare program. Washington (DC), MedPAC. June 1998.

Medicare Payment Advisory Commission. Report to the Congress: Medicare in ruralAmerica. Washington (DC), MedPAC. June 2001a.

Medicare Payment Advisory Commission. Report to the Congress: Medicare paymentpolicy. Washington (DC), MedPAC. March 2002.

Medicare Payment Advisory Commission. Report to the Congress: Medicare paymentpolicy. Washington (DC), MedPAC. March 2001b.


Medicare Payment Advisory Commission. Report to the Congress: Medicare paymentpolicy. Washington (DC), MedPAC. March 2000a.

Medicare Payment Advisory Commission. Report to the Congress: rethinking Medicare’spayment policies for graduate medical education and teaching hospitals. Washington(DC), MedPAC. August 1999.

Medicare Payment Advisory Commission. Report to the Congress: selected Medicareissues. Washington (DC), MedPAC. June 2000b.

Merrill Lynch. Will “more of the same be enough” in the year ahead?, HospitalManagement. New York (NY), Merrill Lynch. January 9, 2003.

Moody’s Investors Service. Report: not-for-profit hospitals’ stability to continue in 2003;pressures loom beyond. New York (NY), press release. January 2, 2003.

Office of Inspector General, Department of Health and Human Services. Hospital closure2000. Washington (DC), OIG. June 2002.

Prospective Payment Assessment Commission. Report and recommendations to theCongress. Washington (DC), ProPAC. March 1, 1997a.

Prospective Payment Assessment Commission. Report to the Congress: Medicare and theAmerican health care system. Washington (DC), ProPAC. June 1997b.

Prospective Payment Assessment Commission. Report to the Congress: Medicare and theAmerican health care system. Washington (DC), ProPAC. June 1996.

Strunk BC, Devers KJ, Hurley RE. Health plan-provider showdowns on the rise, IssueBrief 40. Washington (DC), Center for Studying Health System Change. June 2001.

U.S. House of Representatives. Balanced Budget Act of 1997: conference report toaccompany H.R. 2015, House report No. 105-217. Washington (DC), U.S. House ofRepresentatives. July 30, 1997.


2BAssessing payment adequacy

and updating paymentsfor physician services

S E C T I O N

R E C O M M E N D A T I O N

The Congress should update payments for physician services by the projected change in inputprices, less an adjustment for productivity growth of 0.9 percent, for 2004.



Section 2B: Assessing paymentadequacy and updating payments forphysician services

Medicare payment rates for physician services are based on a fee schedule and

are updated annually with the so-called sustainable growth rate system, which

ties updates to growth in the national economy. Under this system, the update for

2003 is a reduction of 4.4 percent. If the Congress changes current law and in-

creases payment rates modestly for 2003, current rates would be adequate.

MedPAC would then recommend an update for 2004 that equals the estimated

change in input prices less an adjustment for productivity growth. If the Congress

does not increase rates for 2003, a higher update would be necessary in 2004 to

offset the rate reduction in 2003.

2BIn this section


• Accounting for cost changes inthe coming year

• Update recommendation


S E C T I O N

In this section, we assess the adequacy ofMedicare’s current payments forphysician services. We then recommend apayment update for 2004 that considersthe adequacy of current payments andchanges in cost for the coming year.

Recommending a payment update for2004 is complicated by the uncertainty ofthe update for 2003. Under current law,the update for 2003 is a reduction of 4.4percent. This would follow a 5.4 percentreduction in payment rates that occurredin 2002.1 A bill passed by the House lastsummer would have reversed thisreduction and required a positive updateof 2.0 percent. More recently, the Senatepassed an omnibus spending bill for fiscalyear 2003 that included a freeze ofphysician payment rates throughSeptember 30 of this year. MedPAC stillbelieves a modest positive update for2003 is appropriate, as recommended inour March 2002 Report to the Congress:Medicare Payment Policy.

In 2001, total payments (programspending and beneficiary cost sharing) forphysician services totaled $55.9 billion.These payments have been increasing atan average annual rate of 4.9 percent,since 1991, due to changes in the numberof beneficiaries, use of services perbeneficiary, and payment rates. Programspending for physician services isprojected to grow at an average annualrate of 2 to 4 percent from 2001 to 2006.2

This growth is projected to occur despite aseries of negative updates during thisperiod (Figure 2B-1). That is, despitenegative updates in payment rates, thevolume of services is projected to increase

providers furnishing services to Medicarebeneficiaries may indicate that theprogram’s payment rates are too high.Conversely, widespread providerwithdrawals could suggest that the ratesare too low.

For physician services, there are twoindicators of provider entry and exit. Oneindicator is the number of physiciansbilling Medicare. The other morecommonly used indicator is theparticipation rate. The participation rate isthe percentage of physicians who havesigned a participation agreement thatcommits them to “accept assignment” onall their Medicare billings for one year.3

Both indicators can provide evidence thatpayments were adequate. Data on thenumber of physicians billing Medicare areavailable through 2001, and theparticipation rate is available for 2002.The participation rate, as an indicator ofpayment adequacy, requires somequalification, however, for reasonsdiscussed below.

Physicians billing MedicareCounts of physicians billing Medicareshow that the number of physiciansfurnishing services to beneficiaries hasmore than kept pace with growth in thenumber of beneficiaries (Table 2B-1).4

From 1995 to 2001, the number ofphysicians billing traditional Medicaregrew by 8.1 percent, but Medicare Part Benrollment grew by only 5.7 percent. Thisdifference in growth rates led to anincrease in the number of physicians per1,000 beneficiaries, from 12.9 to 13.2.The difference also suggests that paymentrates were not too low in 2001.

72 As s e s s i ng paymen t adequacy and upda t i ng paymen t s f o r p h y s i c i a n s e r v i c e s

1 Both reductions—the 4.4 percent reduction in 2003 and the 5.4 percent reduction in 2002—apply to the fee schedule’s conversion factor, which translates the feeschedule’s relative weights into dollar payment amounts. The reductions include payment updates under the sustainable growth rate system, legislative adjustments, andbudget neutrality adjustments (CMS 2001 and CMS 2002).

2 The 2 percent growth rate is based on projections in the 2002 annual report of the Boards of Trustees of the Medicare trust funds. The 4 percent growth rate is based onprojections in the March 2002 baseline from the Congressional Budget Office.

3 Accepting assignment means that physicians accept the payment rates in the physician fee schedule as payment in full with no further billing of beneficiaries for amountsabove those rates. Under assignment, the physician receives the program payment, which is 80 percent of the total payment amount, directly from Medicare. Thebeneficiary is responsible for the other 20 percent. Without assignment, the beneficiary receives the program payment, and the physician bills the beneficiary for thetotal.

4 The counts of physicians billing Medicare are affected by multiple physicians (e.g., those in the same practice) using the same billing number. The extent of this problemis unknown. To the extent it occurs, however, it means that the counts reported here are an understatement of the number of physicians billing Medicare. In addition,there are indications that the problem of multiple physicians using the same billing number is increasing over time. This means that the growth rate reported for thenumber of physicians billing Medicare may be understated also.

at a rate sufficient to result in positiverates of growth in spending.


Some indicators of payment adequacy,such as entry and exit of providers,suggest that Medicare’s payments forphysician services were at least adequatethrough 2002. Other information presentsmore of a mixed picture of paymentadequacy. In 2002, physicians weresomewhat less willing to accept newMedicare patients than they were in 1999.In addition, Medicare’s payment rates fellfarther below private sector rates whenMedicare rates were reduced in 2002.Whether the difference between Medicareand private sector payment rates hasgrown enough to become a problem is notclear because the difference in 2002 wasabout the same as it was in 1999.

Taken together, these indicators suggestthat payments were adequate in 2002. For2003, payments should remain adequateas long as the Congress changes currentlaw to prevent the 4.4 percent paymentreduction from taking effect. If theCongress does not change current law,however, then payments may not beadequate in 2003 and a compensatingadjustment in payments would benecessary in 2004.

Entry and exit of providersProvider entry and exit is one indicator ofthe adequacy of the current level ofpayments. Rapid growth in the number of

Physicians signing participationagreementsThe other indicator of entry and exit—theparticipation rate—is a leading, oranticipatory, indicator. At the beginningof the calendar year, physicians establish anew agreement, if one is not already ineffect, or they cancel existing agreements.This occurs after CMS determinesMedicare’s payment rates for physicianservices for the coming year. Thus,physicians decide in advance whether toparticipate, based on the level of the ratesand other factors they deem relevant.

Participation rates have been risingsteadily (Figure 2B-2, p. 74). The rate was80.2 percent in 1997, and it rose to 89.7percent in 2002.5 This trend may end,however, if there is another payment


Physician services program spending and payment updates, 2001–2006FIGURE2B-1

Source: 2002 annual report of the Boards of Trustees of the Medicare trust funds.

2001 2002 2003 2004 2005 2006

40

35

30

25

45

50

1.00

0.90

0.80

0.70

0.60

1.10

1.20

Pro

gra

m s

pen

din

g (

dolla

rs in

bill

ions)

Cum

ula

tive

update

s (r

atio)

Cumulative updates

Program spending

Physicians billing traditional Medicare, 1995–2001

Number ofPart B physicians

Number of enrollment per 1,000Year physicians (millions) beneficiaries

1995 460,700 35.641 12.91996 469,915 36.104 13.01997 476,164 36.445 13.11998 478,123 36.756 13.01999 484,576 37.022 13.12000 491,547 37.315 13.22001 498,232 37.657 13.2

Note: The numerator of the ratio of physicians per 1,000 beneficiaries includes allopathic and osteopathicphysicians and excludes nurse practitioners, physician assistants, psychologists, and other nonphysician healthprofessionals. The denominator is the number of beneficiaries enrolled in Medicare Part B, including traditionalMedicare and Medicare�Choice, on the assumption that physicians are providing services to both types ofbeneficiaries.

Source: Unpublished CMS data.

T A B L E2B-1

5 There has been a delay in the availability of information on the participation rate for 2003. Preliminary information is usually available early in the calendar year,following CMS’s distribution of enrollment materials in November. For 2003, distribution of these materials did not occur until early January because of a delay indetermining this year’s payment rates.

reduction. According to an online surveyconducted by the American MedicalAssociation (AMA), 42 percent ofphysicians said they would not sign orcontinue a participation agreement withMedicare for 2003 if there is an additionalpayment cut (AMA 2002).6

Regardless of what happens in 2003, fortwo reasons the participation rate, as anindicator of payment adequacy, requiresqualification.

First, physicians have strong incentives tosign a participation agreement. Theseincentives make the participation rate lesssensitive than some other indicators ofpayment adequacy. This is particularlytrue for physician specialties that areheavily dependent on Medicare forrevenue, such as ophthalmology andcardiology.

One incentive for physicians to sign aparticipation agreement is that their names

appear in a directory that is available tobeneficiaries. The other—stronger—incentive is that, for those who sign anagreement, the allowed charge for aservice is 100 percent of the fee schedulepayment rate. For physicians who do notsign an agreement, the allowed charge fora service is only 95 percent of the feeschedule rate. Nonparticipating physicianscan charge the beneficiary an additionalamount, above the standard 20 percentcopayment, but only if they choose not toaccept assignment and forego directpayment from Medicare. Also, the amountof this so-called balance billing is limitedby statute. The total charge for a servicecannot exceed 115 percent of the allowedcharge, or 109.25 percent (115 percent of95 percent) of the fee schedule paymentrate.

The second reason the participation raterequires qualification is that it includesphysicians who are no longer billing

Medicare. This introduces a subtle bias inthe rate (see text box). It also reduces thevalue of the rate as an indicator ofbeneficiary financial liability.

To better understand the relationshipbetween participation and beneficiaryfinancial liability, it is necessary toanalyze claims data and calculate thepercentage of allowed charges that areattributable to participating physicians.When such analysis is done, it shows thatalmost all charges are submitted byphysicians who have signed aparticipation agreement. For instance,based on claims data from the first 6months of 2002, about 96 percent ofallowed charges for physician serviceswere for services furnished byparticipating physicians.7

Beneficiaries’ access to carePayment adequacy can also be evaluatedby assessing beneficiaries’ access to care.Widespread access problems forbeneficiaries may indicate that Medicare’spayment rates are too low. However,access measures may be difficult tointerpret because they are influenced bymany factors. Access to care for specificservices, for example, may be affected bybeneficiaries’ incomes, supplementalinsurance coverage, preferences, ortransportation barriers, all of which areunrelated to Medicare’s payment policies.

Physician willingness and abilityto serve beneficiaries Findings from a 2002 survey ofphysicians, sponsored by MedPAC andconducted by Project HOPE and TheGallup Organization (Schoenman andFeldman 2002), present a mixed picture.8

• Of physicians accepting some newpatients, 96 percent reported that theywere accepting at least some newMedicare patients. This percentagewas higher than for physicians


Physician participation rates, 1997–2002FIGURE2B-2


Per

cent

88

92

90

80

78

76

86

84

82

741997 1998 1999 2000 2001 2002

6 The results of this survey are based on responses from 520 physicians and a response rate of 26 percent.

7 Another 3 percent of allowed charges were for services furnished by nonparticipating physicians who accepted assignment. Only 1 percent of allowed charges were forservices furnished by nonparticipating physicians who did not accept assignment.

8 The survey was fielded from April through August 2002. About 800 physicians participated, representing a response rate of 54.5 percent.

accepting new Medicaid or privatehealth maintenance organization(HMO) patients.

However, there are some signs thatphysician willingness to accept Medicarepatients is declining.

• The percentage of physiciansaccepting all new Medicare fee-for-service patients dropped from 76percent in 1999 to 70 percent in 2002.The percentage of physiciansaccepting only some new Medicare

fee-for-service patients rose from 20percent in 1999 to 26 percent in 2002.

• Physicians reported that it was moredifficult to find appropriate referrals fortheir Medicare fee-for-service patientsthan for their private fee-for-service orpreferred provider organization (PPO)patients. Conversely, Medicare patientswere easier to refer than private HMOor Medicaid patients.

Many doctors participating in MedPAC’ssurvey expressed concerns about paymentlevels, but physicians were also concerned

about the administrative burdens imposedby Medicare. About 77 percent said thatthey were concerned about reimbursementlevels for their Medicare fee-for-servicepatients, although only 15 percent of themsaid that this concern had led them to limitacceptance of new Medicare patients.About 75 percent of physicians reportedthat they were concerned about billingpaperwork and administration, and 16percent of them said these factors ledthem to limit their acceptance of newMedicare patients.

Finally, many physicians who respondedto MedPAC’s survey reported taking stepsto reduce their practice costs.

• Two-thirds of physicians said thattheir practices had delayed or reducedcapital expenditures.

• More than one-third of physiciansreported that their practices hadincreased the number ofnonphysician clinical staff, and morethan half had increased billing andadministrative staff.

• Three-quarters of physicians said thatthey had increased the number ofpatients seen in an effort to increaserevenues.9

The relationship between changes inphysician practices and Medicare paymentpolicy is unclear. With time spentfurnishing services to Medicarebeneficiaries as a measure of theimportance of Medicare to a physician’spractice, the survey data show noconsistent relationship betweendependence on Medicare and reductionsin staff costs or capital expenditures. Moreimportantly, such practice changes maynot indicate that payments were too low.Instead, physicians could have beenmaking their practices more efficient inresponse to forces in the marketplace,such as lower private sector paymentrates. Research on patient outcomes isnecessary before policymakers can reachconclusions about whether access to high-quality care has diminished.


Interpreting the participation rate

Bias in the participation ratearises because the numeratorof the rate is more accurate

than its denominator. The numeratoris the number of physicians who havesigned a participation agreement, andthe denominator is a total number ofphysicians who may bill Medicareduring the coming year. Bothnumbers are based on lists ofphysicians maintained by thecontractors that process the claims forpayment that physicians submit toMedicare.

Because physicians have no reason tocontact Medicare to say whether theyare still billing the program, thedenominator changes only when thecontractors review their lists ofphysicians and drop those who are nolonger active. The numerator—thenumber of physicians who havesigned a participation agreement—canchange in two ways: Physicians canestablish an agreement or cancel anexisting one, or the contractors canreview their lists of physicians, asabove. Because the numerator is lesslikely than the denominator to beinflated by inactive physicians, a biasin the participation rate occurs.

This bias introduces uncertainty intointerpreting the participation rate as anindicator of payment adequacy. Forexample, the rate may fall because ofa drop in the number of physicianswho have a participation agreementwith Medicare, which would indicateprovider exit and, perhaps, inadequatepayments. On the other hand, theparticipation rate may fall because thecontractors’ lists of physicians havenot been reviewed recently, whichwould not indicate provider exit orinadequate payments.

The only way to avoid the problem ofbias in the participation rate is to,instead, use the percentage of allowedcharges attributable to participatingphysicians, or a measure such as theassignment rate (the percentage ofallowed charges paid on assignment).The disadvantage of these measures,however, is that they cannot becalculated until claims data becomeavailable. (Claims data for the first sixmonths of the year are usually notavailable until December.) �

9 When asked about increasing the number of patients seen, physicians were not asked to distinguish between Medicare and non-Medicare patients.

Private payer reimbursement forphysician servicesIn addition to sponsoring the survey ofphysicians, MedPAC contracted with tworesearch firms, Direct Research, LLC, andDyckman and Associates, LLC, to assessthe difference between Medicare andprivate-payer reimbursement for physicianservices. If Medicare’s payment rates fallrelative to the rates of other payers, somephysicians may have the ability to stopaccepting Medicare patients and insteadfocus their practices on other patients.

To assess the difference betweenMedicare and private rates, DirectResearch used claims data and otherinformation to estimate average privatepayment rates for physician services for1999 to 2001, and to compare those rateswith Medicare’s (Hogan 2002). Toprovide information on the actions ofprivate plans after Medicare’s rates werereduced in 2002, Dyckman and Associatesinterviewed private health plan executivesand collected survey data from the planson their physician payment methods andtheir changes in payment rates from 2001to 2002 (Dyckman and Hess 2002).

The key findings are:

• The difference between Medicare andaverage private rates is smaller nowthan it was in the mid-1990s,primarily because of shifts in privateplan enrollment from higher-payingindemnity plans to lower-payingPPOs and HMOs. Medicare’s rateswere about 66 percent of private ratesin 1994, but this percentage rose toabout 83 percent in 2001.

• During the recent period of volatilityin Medicare’s payment updates, thedifference between Medicare andprivate rates narrowed. In 2000 and2001, Medicare’s updates forphysician services exceeded inflation.Since 2001, the difference has startedto widen again because the shift in

private sector enrollment to HMOshas stopped and because privatepayers generally did not reduce ratesin 2002. Still, in 2002, Medicare rateswere about 77 to 79 percent ofprivate rates, which appears to be nolower than in 1999 and above thepercentage in 1994.

• Private plans report that Medicarereductions in payment rates haveincreased pressure on them to raisetheir rates. None of the plans,however, say that the reductions havehad a strong or direct impact on theirdecisions about payment rates for2002 or 2003. Some plans indicatethat the reductions have had amoderate impact on their decisions.

Additional access measures National indicators of access areimportant because they allow a generalassessment of access and inform decisionsabout payment updates that change theoverall level of payments. A limitation ofthese indicators, however, is that they donot reveal access problems that may existlocally or with regard to specific services.Such problems, if they exist, are importantbecause they may signal a need to alter thedistribution of payments amonggeographic areas, services, or providers.10

We can obtain some insight on the localpicture through the work of the Center forStudying Health System Change (HSC).For example, in a survey of physiciansconducted in 2001, HSC found that 71percent of all physicians were willing totake all new Medicare patients, but only55 percent of physicians in Seattle werewilling to take all new beneficiaries.

HSC’s latest published results are basedon survey data collected before thereduction in Medicare payment rates tookeffect. HSC is currently fielding a newround of surveys, and MedPAC willcontinue to monitor the results.

Changes in the volume of servicesChanges in the volume of services can beconsidered an indirect measure ofpayment adequacy. Medicare spending forphysician services is determined by twofactors: the rates physicians are paid forspecific services and the number ofservices performed by physicians forMedicare beneficiaries. The volume ofphysician services per beneficiary can beexpected to rise based on factors such asthe demographic profile of beneficiaries,their health status, and changes intreatment patterns for specific conditions.According to MedPAC’s payment updateframework, if the overall volume ofservices provided to beneficiaries falls, itmay indicate that physicians are providingfewer services to Medicare beneficiariesbecause Medicare payments to physiciansare inadequate. Conversely, largeincreases in volume growth may indicatethat Medicare is overpaying for services.In addition, changes in volume growth forspecific services may provide evidence ofunderpayment or overpayment byMedicare for those services. Becausevolume growth can be driven by a numberof factors, these data must be interpretedcautiously.

In the case of physician services, the needfor caution is particularly importantbecause of ambiguities in interpreting dataon changes in the volume of services.

• There is some evidence to suggestthat volume goes up when paymentrates go down, the so-called “volumeoffset.” For instance, actuaries at theCenters for Medicare & MedicaidServices have estimated thatpayments for increased use ofservices have offset projected savingsfrom past Medicare payment ratereductions by between 30 percent and50 percent (Codespote et al. 1998).


10 MedPAC has discussed the distinction between the overall level of payments and the distribution of payments (MedPAC 1999, p.15).


11 This is the same measure we used in MedPAC’s June 2001 Report to the Congress on Medicare in rural America (MedPAC 2001).

12 The analysis is based on data for the first six months of each year. Growth rates calculated with these data may differ from growth rates based on full-year databecause of seasonal variation in use of services.

• It is possible that a volume offset, if itoccurs, results in increased volumefor services other than those affectedby the payment reduction. Forexample, some services, such asoffice visits and noninvasivediagnostic procedures, are morediscretionary than others and may bemore likely to grow in volume thanother services if payment rates arereduced.

• In addition, the volume of servicesper beneficiary varies amonggeographic areas in ways that appearunrelated to patient outcomes(Wennberg et al. 2002; Welch et al.1993). These findings raise questionsabout whether some of the current, orbaseline, volume of physicianservices is necessary and whether achange in volume means that accessto needed services has changed.

With these qualifications in mind, weanalyzed the growth in the volume ofphysician services, by type of service,using claims data for 1999 to 2002 (Table2B-2, p. 78). Volume was measured as percapita use of physician services bybeneficiaries in traditional Medicare.11

The analysis shows that, across allservices, the volume growth rate was 4.3percent for 2001 to 2002.12 This growthrate is one percentage point higher thanthe average annual growth for 1999 to2001, which raises the possibility thatphysicians offset some of the negativeupdate in 2002 by increasing the volumeof services. For two reasons, however, wecannot conclude that such a volume offsetoccurred:

• Volume growth has been highpreviously, even in years whenpayment rates have increased. Forinstance, the volume growth rate wasalso 4.3 percent for 1999 to 2000,when the payment update forphysician services was a positive 5.4percent.

• Volume could have grown because oftechnological advances or otherfactors unrelated to the paymentreduction. To conclude that thepayment reduction, and not otherfactors, was the cause of some of the2001 to 2002 volume growth, itwould be necessary to contrast thebehavior of physicians whoexperienced the payment reductionwith others who did not. This is notpossible, however, because thepayment reduction applied to allservices and, therefore, to allphysicians.

When we group services into four majorcategories—evaluation and management,imaging, procedures, and tests—and lookat 2001 to 2002 growth rates for each, wesee that evaluation and management hadthe lowest rate, which was 2.9 percent.Still, this was more than double thegrowth rate for this category in 1999 to2001. Among the other services, thegrowth rate for procedures was nearest theaverage for all services, at 3.5 percent.The growth rates for imaging and testswere much higher at 9.4 percent and 9.0percent, respectively.

Relatively high growth rates for imagingservices were concentrated in severalspecific categories, all of which involvetechnology of one kind or another. Forinstance, nuclear medicine grew by 13.0percent, computerized automatedtomography (CAT) of parts of the bodyother than the head grew by 15.3 percent,magnetic resonance imaging (MRI) ofparts of the body other than the brain grewby 15.9 percent, and MRI of the braingrew by 14.6 percent. It is noteworthy,however, that none of these technologiesare new. Instead, it appears that use ofwell-established technologies isincreasing. CAT, for example, wasintroduced in the 1970s. MRI began todiffuse as a new technology in the 1980s.Thus, the indications for use of thesetechnologies may be changing.

Volume growth was most pronounced forservices related to the most commonhealth problems of the elderly. Forexample, some coronary care servicesshowed relatively high volume growth asfollows: echography of the heart (10.8percent); pacemaker insertion (8.9percent); and cardiovascular stress tests(8.7 percent).

Some of the highest growth rates wefound were for a minor-procedurescategory that includes primarily outpatientrehabilitation. Those rates included 17.6percent for 1999 to 2001 and 14.3 percentfor 2001 to 2002. This rapid growthoccurred when spending caps foroutpatient rehabilitation were temporarilylifted. Under the Balanced BudgetRefinement Act of 1999, a moratorium onthe spending caps was implemented in2000. The moratorium was later extendedthrough 2002, and CMS recentlyannounced a delay until July 2003 forending the moratorium.

Volume decreased for some services. Forexample, the volume of two types ofcardiology services—coronaryangioplasty and heart imaging, includingcardiac catheterization—went downslightly, by 1.2 percent and 0.4 percent,respectively. Given the rapid growth inuse of these services that had occurredpreviously, such small decreases may notsignal a change in access to care forMedicare beneficiaries. Reasons for someof the other volume decreases—officevisits by new patients, coronary arterybypass grafts, cystoscopy, hip fracturerepair, and colectomy—are unclear. Insome cases, volume decreases may be theresult of the substitution of one service foranother. The decrease in the volume ofcoronary artery bypass grafts, forexample, may be due to greater use ofcoronary angioplasty, which is a newerprocedure for treatment of coronary arterydisease.


Change in per capita use of physician services by beneficiaries in traditional Medicare, by selected type of service, 1999–2002

Per capita service use

PercentAverage annualof totalpercent changeservice

Type of service 1999 2000 2001 2002 1999–2001 2001–2002 use

All services 663.4 691.8 707.9 738.5 3.3% 4.3% 100.0%

Evaluation and management 353.6 359.4 361.9 372.5 1.2 2.9 50.4Office visits—established patient 127.6 131.2 130.3 133.3 1.1 2.3 18.1Hospital visit—subsequent 65.0 64.6 64.7 66.7 –0.2 3.1 9.0Consultations 39.8 41.5 42.6 44.5 3.5 4.4 6.0Emergency room visit 18.1 19.0 20.1 21.4 5.3 6.5 2.9Specialist—psychiatry 18.5 18.3 18.2 18.5 –1.0 2.1 2.5Specialist—ophthalmology 15.9 16.8 17.5 18.1 4.9 3.5 2.4Hospital visit—initial 17.6 17.4 17.2 17.2 –1.2 0.3 2.3Office visits—new patient 15.4 15.5 14.9 14.9 –1.4 –0.2 2.0

Imaging 81.1 88.2 96.1 105.1 8.9 9.4 14.2Echography—heart 12.6 13.8 14.9 16.5 8.8 10.8 2.2Standard—nuclear medicine 10.0 11.7 13.6 15.4 16.5 13.0 2.1Advanced—CAT: other 9.3 10.7 12.3 14.1 14.8 15.3 1.9Advanced—MRI: other 6.4 7.9 9.4 10.9 21.3 15.9 1.5Standard—musculoskeletal 8.5 8.8 9.2 9.5 3.9 2.9 1.3Advanced—MRI: brain 5.1 5.8 6.5 7.4 12.6 14.6 1.0Standard—chest 6.7 6.5 6.3 6.3 –3.3 0.4 0.9Advanced—CAT: head 2.7 2.8 2.9 3.0 3.2 4.5 0.4Imaging/procedure—heart, including cardiac catheterization 1.9 2.1 2.4 2.4 10.4 –0.4 0.3

Procedures 200.3 214.5 218.5 226.1 4.5 3.5 30.6Minor—other, including outpatient rehabilitation 14.7 18.9 20.4 23.3 17.6 14.3 3.2Eye—cataract removal/lens insertion 16.0 16.1 15.6 15.8 –1.3 1.3 2.1Endoscopy—colonoscopy 7.6 8.3 9.0 9.9 8.4 10.1 1.3Major, cardiovascular—coronary artery bypass graft 6.7 6.7 6.1 5.6 –4.7 –7.0 0.8Endoscopy—upper gastrointestinal 4.7 4.9 4.9 5.0 1.8 2.7 0.7Major, orthopedic—knee replacement 3.9 4.0 4.2 4.6 4.2 9.9 0.6Major, cardiovascular—coronary angioplasty 3.9 4.2 4.6 4.6 9.0 –1.2 0.6Endoscopy—cystoscopy 3.8 3.8 3.8 3.7 0.0 –3.1 0.5Eye—treatment of retinal lesions 3.5 3.7 3.5 3.6 0.9 2.6 0.5Major, orthopedic—hip fracture repair 3.6 3.6 3.5 3.2 –1.9 –9.5 0.4Major, orthopedic—hip replacement 2.7 2.8 2.8 2.9 1.8 3.1 0.4Major, cardiovascular—pacemaker insertion 1.7 1.9 2.1 2.2 8.6 8.9 0.3

Tests 22.0 22.6 23.7 25.9 3.9 9.0 3.5Other—electrocardiograms 6.5 6.5 6.4 6.6 –0.6 1.9 0.9Other—cardiovascular stress tests 3.3 3.6 3.9 4.2 8.8 8.7 0.6Lab tests—other (physician fee schedule) 1.9 2.1 2.4 2.9 12.9 22.1 0.4Other—electrocardiogram monitoring 1.9 1.9 1.9 2.0 0.2 6.0 0.3

Note: CAT (computerized automated tomography), MRI (magnetic resonance imaging). Service use is measured as the relative weights (relative value units) for services receivedmultiplied by the physician fee schedule conversion factor. To put service use in each year on a common scale, we used the relative weights and conversion factor for2002. For billing codes not used in 2002, we imputed relative weights based on the average change in weights for each type of service.

Source: MedPAC analysis of claims for a 5 percent random sample of Medicare beneficiaries from the first six months of each year.

T A B L E2B-2

It should be emphasized that furtheranalysis is required to understand thefactors underlying volume growth.MedPAC is currently conducting researchon this issue.

Accounting for costchanges in the comingyear

In order to determine the appropriatepayment update for 2004, we mustestimate how much costs will change inthe coming year. Two factors are expectedto affect the cost of physician servicesduring the coming year: input priceinflation and productivity growth.Productivity growth is expected to reducecosts through capital investment, changesin work processes, and other factors.

It is possible that other factors, includingsome scientific and technologicaladvances, may increase costs. Features ofthe physician fee schedule should accountfor those cost increases, at least partially,however. Every year, new billing codesare created and existing codes are revised.Also, by law, the fee schedule’s relativeweights are reviewed and recalibratedevery five years.

Measuring input priceinflationThe Medicare Economic Index (MEI) isused as the generally accepted measure ofinput price inflation for physicianservices. It is calculated by CMS as aweighted average of price changes forinputs used to provide physician services(Table 2B-3). Those inputs includephysician time and effort, or work,practice expense, and professional liabilityinsurance (PLI). Practice expense includesnonphysician employee compensation,office expense, medical materials andsupplies, medical equipment, and otherprofessional expenses, such as privatetransportation. In general, the weights

used to construct the MEI represent theshares of physicians’ practice revenuesattributable to each input, based primarilyon a survey conducted by the AMA for1996. Physician work has a weight of 54.5percent, practice expense has a weight of42.3 percent, and PLI has a weight of 3.2percent. CMS revises these weights andthe other components of the MEIperiodically (see text box, p. 80).

CMS currently projects that input pricesfor physician work will increase 3.4percent in 2004, based on increases of 3.4percent in wages and salaries and 3.5percent in nonwage compensation.Practice expenses are projected to increaseby 3.1 percent. This projection includes a3.7 percent increase in nonphysicianemployee compensation and a 3.0 percentincrease in office expenses.

The largest change expected in inputprices is for PLI, which is projected to

increase by 5.6 percent. Historically, thiscomponent of the MEI has followed acyclical pattern, illustrated by the changesin PLI premiums from 1990 to 2002(Figure 2B-3, p. 80).13 The recent increasein PLI premiums in 2002, estimated at11.3 percent, was the highest in over adecade.

In sum, the index shows that input pricesfor physician services are expected toincrease by 3.4 percent in 2004.

Productivity growthProductivity growth is the ratio of growthin outputs to growth in inputs. Measuringproductivity growth requires detailedinformation on the personnel, facilities,and other inputs used and on the quantity,quality, and mix of services (outputs)produced. Because such data are generallynot available, MedPAC has adopted apolicy standard or goal for achievable


13 Despite the changes in PLI premiums, the premiums have not varied much as a percentage of physician revenues. From 1990 through 1998, PLI premiums remained ina narrow range, from 3 to 5 percent of revenues (Gonzalez and Zhang 1998, Zhang and Thran 1999, and Wassenaar and Thran 2001).

Medicare Economic Index weights and forecast of input price changes for 2004

PriceWeight (percent) changes

for 2004Input Category Total (percent)

Total 100.0% 3.4%Physician work 54.5 3.4

Wages and salaries 44.2% 3.4Nonwage compensation 10.3 3.5

Practice expense 42.3 3.1Nonphysician employee compensation 16.8 3.7

Wages and salaries 12.4 3.7Nonwage compensation 4.4 3.6

Office expense 11.6 3.0Medical materials and supplies 4.5 2.3Medical equipment 1.9 1.7Other professional expense 7.6 2.7

Professional car 1.3 1.8Other 6.3 2.9

Professional liability insurance 3.2 5.6

Note: Numbers may not total exactly because of rounding.

Source: MedPAC analysis of unpublished data from CMS.

T A B L E2B-3

productivity growth that is based ongrowth in multifactor productivity in thenational economy.14

Using the current estimate of growth inmultifactor productivity from the Bureauof Labor Statistics, the productivityadjustment would be 0.9 percent.

Update recommendation

Under MedPAC’s payment updateframework, updates can include threecomponents: an adjustment for paymentadequacy, if appropriate; an estimate ofinflation in input prices; and a downwardadjustment in the update for productivitygrowth.


Revising the Medicare Economic Index

CMS revises the MedicareEconomic Index (MEI)periodically so that the index’s

weights and other components reflectcurrent conditions. A revisionoccurred most recently in 1998 basedon data primarily for 1996. Previousto that revision, the agency revised theMEI in 1992 with data for 1989.

So far, the primary data source for theweights in the MEI has been theAmerican Medical Association’sSocioeconomic Monitoring System(SMS) survey. The weights for themajor categories of inputs consideredin the MEI—physician work, practiceexpense, and professional liabilityinsurance—have all come from theSMS survey. The SMS survey hasalso been the source of the weights for

subcategories of practice expense:nonphysician employeecompensation, office expense,medical materials and supplies,medical equipment, and otherprofessional expense. Within thesesubcategories, CMS has assignedweights to inputs with data from othersources, including the EmploymentCost Index of the Bureau of LaborStatistics, the Asset and ExpenditureSurvey of the Bureau of the Census,and the Current Population Survey ofthe Bureau of the Census.

For the next revision of the MEI, itwill be necessary for CMS tosubstitute another data source for theSMS survey because the AMAdiscontinued the SMS survey after itwas conducted last, in 1999. �

0

Quarterly changes in professional liability insurance premiums, 1990–2002FIGURE2B-3


15

10

5

�5

�10

0

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

Per

cent

change

14 Multifactor productivity is based on all relevant inputs used to provide goods and services. These inputs include labor, capital, and other inputs, such as energy andmaterials.

R E C O M M E N D A T I O N 2 B

The Congress should updatepayments for physician services bythe projected change in input prices,less an adjustment for productivitygrowth of 0.9 percent, for 2004.

Payments are adequate if there is a modestpositive update in 2003. If the Congressdoes not change current law and prevent apayment reduction in 2003, however,payments may not be adequate, and acompensating adjustment in paymentsmay be necessary in 2004. The othercomponents of the update are the

projected change in input prices, which is3.4 percent, and an adjustment forproductivity growth, which is 0.9 percent.The net of these two components is anupdate of 2.5 percent.

I M P L I C A T I O N S 2 B :

Spending• This recommendation would update

physician payments more than undercurrent law. It is expected to increasecosts by more than $1.5 billion in2004.

Beneficiary and provider• Increasing payments for physician

services would help preservebeneficiary access to care.

• Increasing payments to physicianswould help to maintain the adequacyof those payments and allowphysicians to furnish high-qualityservices. �


References

American Medical Association. Member Connect Medicare physician payment cutsurvey. Chicago (IL), research brief. July 2002.

Centers for Medicare & Medicaid Services, Department of Health and Human Services.Medicare program; revisions to payment policies and five-year review of and adjustmentsto the relative value units under the physician fee schedule for calendar year 2002,Federal Register. November 1, 2001, Vol. 66, No. 212, p. 55246–55503.

Centers for Medicare & Medicaid Services, Department of Health and Human Services.Medicare program; revisions to payment policies under the physician fee schedule forcalendar year 2003 and inclusion of registered nurses in the personnel provision of thecritical access hospital emergency services requirement for frontier areas and remotelocations, Federal Register. December 31, 2002, Vol. 67, No. 251, p. 79966–80184.

Codespote SM, London WJ, Shatto JD, Office of the Actuary, Health Care FinancingAdministration. Estimated volume and intensity response to a price change forphysicians’ services. August 13, 1998. Available athttp://www.cms.hhs.gov/statistics/actuary/physicianresponse/.

Dyckman Z, Hess P, Dyckman and Associates, LLC. Survey of health plans concerningphysician fees and payment methodology. Draft report to the Medicare PaymentAdvisory Commission. Written December 2002.

Gonzalez ML, Zhang P, Center for Health Policy Research. Socioeconomiccharacteristics of medical practice (1997–1998 edition). Chicago (IL), AMA. 1998.

Hogan C, Direct Research, LLC. Medicare physician payment rates compared to ratespaid by the average private insurer, 1999–2001. Draft report to the Medicare PaymentAdvisory Commission. Written November 2002.

Medicare Payment Advisory Commission. Report to the Congress: Medicare in ruralAmerica. Washington (DC), MedPAC. June 2001.


Schoenman JA, Feldman JJ. Results of the Medicare Payment Advisory Commission’s2002 survey of physicians. Bethesda (MD), The Project HOPE Center for Health Affairs.December 2002.

Wassenaar JD, Thran SL, Center for Health Policy Research. Physician socioeconomicstatistics (2000–2002 edition). Chicago (IL), AMA Press. 2001.

Welch WP, Miller ME, Welch HG, et al. Geographic variation in expenditures forphysicians’ services in the United States, New England Journal of Medicine. March 4,1993, Vol. 328, No. 9, p. 621–627.

Wennberg JE, Fisher ES, Skinner JS. Geography and the debate over Medicare reform,Health Affairs Web Exclusives. February 13, 2002. Available athttp://www.healthaffairs.org.

Zhang P, Thran SL, Center for Health Policy Research. Physician socioeconomicstatistics (1999–2000 edition). Chicago (IL), AMA. 1999.


2CAssessing payment adequacy

and updating paymentsfor skilled nursing facility services

S E C T I O N


2C-1 The Secretary should continue a series of nationally representative studies on access toskilled nursing facility services (similar to studies previously conducted by theDepartment of Health and Human Services’ Office of Inspector General).


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2C-2 The Congress should eliminate the update to payment rates for skilled nursing facilityservices for fiscal year 2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2C-3A Consistent with previous MedPAC recommendations, the Secretary should develop anew classification system for care in skilled nursing facilities.

Because it may take time to develop this system, the Secretary should draw on new andexisting research to reallocate payments to achieve a better balance of available resourcesbetween the rehabilitation and nonrehabilitation groups.

To allow for immediate reallocation of resources, the Congress should give the Secretarythe authority to:� remove some or all of the 6.7 percent payment add-on currently applied to the

rehabilitation RUG–III groups.� reallocate money to the nonrehabilitation RUG–III groups to achieve a better balance

of resources among all of the RUG–III groups.

2C-3B If necessary action does not occur within a timely manner, the Congress shouldprovide for a market basket update, less an adjustment for productivity growth of 0.9percent, for hospital-based skilled nursing facilities to be effective October 1, 2003.



Section 2C: Assessing paymentadequacy and updating payments forskilled nursing facility services

Based on the available evidence, we conclude that aggregate Medicare payments

for skilled nursing facilities (SNFs) are adequate as of fiscal year 2003, but that

payments are not distributed appropriately to account for the expected resource

needs of different types of Medicare beneficiaries. Our estimate of the overall

Medicare margin for SNF services across all providers in fiscal year 2003 is about

5 percent, with the Medicare margin for freestanding SNFs (90 percent of all

facilities) about 11 percent and the Medicare margin for hospital-based facilities

about –36 percent. After high cost growth prior to the implementation of the

prospective payment system for SNFs, we have seen a decline in costs for free-

standing facilities in recent years in response to incentives in the SNF prospective

payment system. We expect this trend to continue into fiscal year 2004. This

decline in costs does not appear to have resulted in a lower quality of care.

Continued entry of for-profit freestanding providers, increases in the volume of

services provided, continued access to services for most Medicare beneficiaries,

and lack of systematic problems with SNFs’ access to capital that would pose

problems for beneficiaries’ access to services suggest that Medicare payments are

at least adequate to cover the costs of caring for Medicare beneficiaries. We

believe it is important to continue monitoring beneficiaries’ access to SNF

services.

2CIn this section



• Update recommendations


S E C T I O N

Background

Medicare beneficiaries who need short-term skilled care (nursing or rehabilitationservices) on an inpatient basis following ahospital stay of at least three days areeligible to receive covered services inskilled nursing facilities (SNFs).1 Theseservices may be provided either infreestanding or hospital-based facilities,with freestanding facilities representingabout 90 percent of all SNFs. Afreestanding SNF is typically part of anursing facility that also providesresidential long-term care, which is notcovered by Medicare.

Skilled nursing facilitypayment systemIn July 1998, Medicare adopted aprospective payment system for SNFservices. This system pays SNFs a casemix adjusted amount for each day of care.2

The per diem payment rates under thissystem are intended to provide fullpayment for all facility services, exceptfor the costs of approved medicaleducation programs. The rates cover allroutine, ancillary, and capital costs, aswell as those for most ancillary items andservices for which payment previouslywas made under Medicare Part B.3

Patients are assigned to 1 of 44 groups,called resource utilization groups, versionIII (RUG–III). Each RUG–III groupincludes patients with similar serviceneeds who are expected to require similaramounts of resources. Patients’ expectedresource needs are determined by periodicassessments of their condition, including

their need for intensive physical,occupational, or speech therapy; specialtreatments (such as tube feeding); andtheir functional status (their ability tomanage unassisted ordinary dailyactivities, such as eating and using thetoilet). The daily rate for each RUG–IIIgroup is the sum of three components:

• a fixed amount for routine services(such as room and board, linens, andadministrative services),

• a variable amount reflecting theintensity of nursing care and ancillaryservices patients are expected torequire, and

• a variable amount for the expectedintensity of therapy services(physical, occupational, and speechtherapies).

Payment rates for SNF services arecomputed separately for urban and ruralareas, and the labor portion of the totalrate is adjusted to reflect the wage marketconditions within the SNF’s geographiclocation. Furthermore, rates are updatedannually on the basis of the projectedincrease in the SNF market basket index,a measure of the national average pricelevel for the goods and services SNFspurchase to provide care (see Appendix Afor more information on the SNF paymentsystem).

Shortly after the SNF prospectivepayment system was implemented, theCongress responded to providers’concerns about payment rates and thedistribution of payments by granting aseries of temporary payment rateincreases:

• The Balanced Budget RefinementAct of 1999 (BBRA) increased ratesfor all 44 RUG–III groups by 4percent for care furnished betweenApril 2000 and September 2002.

• The Medicare, Medicaid, and SCHIPBenefits Improvement and ProtectionAct of 2000 (BIPA) increased thebase rate for the nursing componentby 16.66 percent for care furnishedfrom April 2001 through September2002.

• BBRA and BIPA increased paymentrates for 14 rehabilitation groups by6.7 percent and rates for 12 complexcare groups by 20 percent. Theseincreases were intended to give CMStime to refine the RUG–IIIclassification system and arescheduled to expire when CMSadopts that refinement.

Trends in Medicarepayments for skilled nursingfacility servicesTotal spending for SNF services on behalfof Medicare beneficiaries was $15.3billion in fiscal year 2001. This amountincludes benefit payments by theMedicare program and beneficiaries’payments for cost-sharing obligations.Medicare spending on SNF services grewan average of 13 percent from fiscal years1992 through 2002, with a noticeable dipin spending occurring in fiscal years 1999and 2000 (Figure 2C-1). TheCongressional Budget Office (CBO)projects that expenditures for this sectorwill grow by about 8 percent per yearfrom fiscal years 2002 to 2007.4 Medicarespending for SNF services represents

86 As s e s s i ng paymen t adequacy and upda t i ng paymen t s f o r s k i l l e d n u r s i ng f a c i l i t y s e r v i c e s

1 Medicare covers 100 SNF days in a spell of illness. Medicare pays 100 percent of the rate for the first 20 days of a SNF stay. From the 21st to the 100th day,beneficiaries are responsible for a copayment equal to one-eighth of the hospital deductible, or $105 per day in fiscal year 2003.

2 The prospective payment system differs substantially from the payment system in effect throughout most of the 1980s and 1990s, when SNFs were paid on the basis oftheir costs subject to limits on their per diem routine costs (room, board, and routine nursing care). No limits were applied for ancillary services (such as drugs andtherapy).

3 The per diem rates exclude amounts for services furnished by physicians and certain other practitioners, such as qualified psychologists, and for dialysis services andsupplies. These services continue to be paid for under Part B. Certain high cost, low probability ancillary services have also been excluded from the SNF per diem rate tolimit SNFs’ liability for services typically outside the scope of SNF care. These services include emergency room care, outpatient hospital CAT scans, MRIs and surgeries,and certain high cost chemotherapy agents and prosthetic devices. Costs for physical, occupational, and speech therapy services are included in the per diem rate evenif they are furnished by or under the supervision of a physician.

4 CBO plans to revise its projections of Medicare spending for SNF services downward after conducting an updated analysis of the relationship between the use of SNFservices and the incidence of disabilities and hospitalizations among Medicare beneficiaries. CBO’s updated projections for SNF services were not available before ourreport went to press.

about 6.5 percent of total Medicarespending for all services. Although about1.4 million beneficiaries (about 3.5percent of all beneficiaries) use SNFservices each year, Medicare’s paymentsfor these services account for only about10 to 12 percent of freestanding nursingfacilities’ revenues and less than 2 percentof total revenues for hospitals. Otherpayments for these services come fromMedicaid and private sources.


Each year, MedPAC makes paymentupdate recommendations for the comingfiscal year for SNF services. To informour recommendations, we considermultiple factors, including the relationshipof payments to costs and theappropriateness of current costs,providers’ entry and exit from theprogram, changes in the volume of

services, beneficiaries’ access to care, andSNFs’ access to capital.

After assessing all of these factors, weconclude that fiscal year 2003 paymentsto SNFs overall are adequate to cover thecosts of caring for the beneficiaries thatuse these services. We estimate theMedicare margin—a measure of therelationship between Medicare paymentsand costs—for all SNFs to be about 5percent, with the Medicare margin forfreestanding SNFs (90 percent of allSNFs) about 11 percent (Table 2C-1). The costs of providing SNF servicesappear to be decreasing, while we find noevidence of declines in the quality of care.In addition, the most recent available datasuggest no declines in the overall numberof SNFs participating in the Medicareprogram between 1998 and 2002—withincreases in freestanding providersbalancing decreases in hospital-basedproviders—or in the volume of servicesprovided. We find no evidence ofsubstantial declines in beneficiaries’

ability to access SNF services or in SNFs’access to capital.

However, while Medicare payments toSNFs appear adequate overall, the SNFclassification system appears to do a poorjob of tracking the expected resourceneeds of different types of beneficiarieswho use SNF services. This causes sometypes of beneficiaries to be moreprofitable for SNFs than others. Studies


Medicare spending for skilled nursing facility services, 1992–2002FIGURE2C-1

Note: Spending is for Part A services only.

A

nnual e

xpen

diture

s (d

olla

rs in

bill

ions)

Fiscal year

14

12

10

18

16

8

6

4

2

01992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002


Medicare marginsfor skilled nursing

facilities, 2000 and estimated 2003

Reported Estimated2000 2003

Freestanding 17 11Hospital-based –57 –36All facility types 7 5

Source: MedPAC analysis of Medicare cost reportdata from CMS.

T A B L E2C-1

have repeatedly shown that hospital-basedSNFs tend to treat a much largerproportion of the less-profitable types ofpatients—those with multiple complexneeds that do not include rehabilitationtherapy—than freestanding facilities(Dalton 2002, Liu and Black 2002,MedPAC 2001). The Medicare margin forhospital-based facilities also tends to belower (–36 percent in fiscal year 2003)than the Medicare margin for freestandingfacilities.5 This may be one of manyreasons why some hospital-basedfacilities—about 26 percent between 1998and 2002—have exited the Medicareprogram. The decline in hospital-basedfacilities does not appear to have led to adecline in beneficiaries’ access to care,though, because beneficiaries whootherwise would have been treated inthese facilities either remain in the acutecare hospital setting longer or receive carein a freestanding facility. However, thesubstantial declines in the number ofhospital-based facilities participating inMedicare may indicate an imbalance inthe distribution of payments acrossdifferent types of patients in the SNFpayment system.

Current payments and costs One of the many factors we use to informour update recommendation for fiscal year2004 is the estimated relationship betweenSNF payments and costs (margin) forfiscal year 2003.6 To produce thisestimate, we modeled fiscal year 2003SNF payments and costs using methods

similar to those we use for all settings forwhich we make update recommendations:

• We used the latest cost report dataavailable (fiscal year 2000) as thecost and payment base.

• We increased costs by the actual SNFmarket basket index for fiscal years2001 and 2002 and used CMS’sforecast of the SNF market basketindex for fiscal year 2003.

• We increased payments by the updatefactor that applied for each yearstarting after fiscal year 2000.

In modeling fiscal year 2003 paymentsand costs, we incorporated any policychanges scheduled in current law for fiscalyear 2004. We excluded the 16.66 percentincrease in the nursing component of thebase rate from our estimate because it wasimplemented after fiscal year 2000 (ourbase year) and expired before fiscal year2003 (the year we modeled). We deductedthe 4 percent increase to all payment ratesfrom the fiscal year 2000 payments beforeestimating fiscal year 2003 paymentsbecause this add-on is not scheduled to bein effect for fiscal year 2004. We includedthe 20 percent add-on for certain RUG–IIIgroups in our projections because weanticipate that this increase will still be ineffect in fiscal year 2004.7

We estimate that the overall Medicaremargin for all SNFs will be about 5percent in fiscal year 2003. This is aboutthe same as the overall Medicare margin

for all SNFs we estimated for fiscal year2002. However, the lack of a difference isdue largely to our approach to estimatingthe Medicare margin for hospital-basedSNFs—it is more conservative than lastyear’s approach.8 If we had used the samemethod as last year, we would likely haveseen an increase in the Medicare marginfor all SNFs in fiscal year 2003 bothbecause the Medicare margin forfreestanding SNFs increased substantiallyand the proportion of all SNFs that arefreestanding increased between 1998 and2002.

On average, we estimate that the Medicaremargin for the 90 percent of all SNFs thatare freestanding will be 11 percent infiscal year 2003, an increase of less than 2percentage points over the 9.4 percent weestimated for fiscal year 2002.9 Theincrease is due largely to substantialincreases in freestanding SNFs’ reportedmargins between fiscal years 1999 and2000. The reported margin forfreestanding facilities was about 9 percentin fiscal year 1999 and just under 17percent in fiscal year 2000 (see text boxnext page).

In contrast to the increase in the marginseen for freestanding facilities, theMedicare margin for hospital-basedfacilities does not appear to have changedmuch between fiscal year 1999 and fiscalyear 2000 (–56 to –57 percent). However,the fiscal year 2003 Medicare margin for hospital-based facilities is differentfrom that which we estimated for fiscal


5 Hospital-based SNFs’ higher case mix is only one factor that may explain the lower Medicare margin for these facilities. Recent research indicates that much of thedifference between freestanding and hospital-based SNF margins is due to hospital-based SNFs having higher fixed costs (Pizer et al. 2002). To some extent, thesehigher fixed costs result from hospital cost allocation methods. Hospital-based SNFs may also offer a different product than freestanding SNFs, with more licensed staffand a much shorter average length of stay (MedPAC 2001).

6 A margin is calculated as payments less costs, divided by payments.

7 The 20 percent add-on for certain RUG–III groups that became effective in April 2000 was intended to give CMS time to refine the SNF classification system. BIPAchanged this add-on, effective April 2001, by applying the 20 percent add-on only to nonrehabilitation RUG–III groups and applying a 6.7 percent add-on to all of therehabilitation RUG–III groups. However, the 20 percent add-on as originally mandated in BBRA only applies in fiscal year 2000.

8 To estimate the fiscal year 2002 Medicare margin for hospital-based SNFs last year, we used the costs for freestanding SNFs and inflated them by 30 percent (our bestestimate of the difference in costs attributable to a different case mix and product between the two types of facilities). In computing the fiscal year 2003 Medicare marginfor hospital-based SNFs, we took a more conservative approach; we used the costs for hospital-based SNFs and deducted 17.5 percent (our best estimate of the amountattributable to hospital cost accounting based on a study sponsored by the Health Care Financing Administration [HCFA, now CMS] that estimated the range to bebetween 15 and 20 percent) (CHPS Consulting 1994).

9 The 9.4 percent estimate for fiscal year 2002 was modeled assuming that the 20 percent add-on to payments for 12 complex care groups and the 6.7 percent add-on topayments for 14 rehabilitation groups would remain in current law in fiscal year 2003. Because CMS has yet to announce a refinement of the SNF classification system,this add-on remains in effect.

year 2002, primarily because we changedthe method to estimate the margin (seefootnote 6). We estimate the hospital-based facilities’ Medicare margin to beabout –36 percent in fiscal year 2003.Differences in measured margins betweenhospital-based and freestanding facilitiesare difficult to interpret, because theyresult from both the artifact of hospitals’allocation of costs to their SNFs anddifferences in case mix and productbetween the two types of facilities.

Appropriateness of currentcosts Under the cost-based Medicare paymentsystem in effect for SNFs throughout mostof the 1980s and 1990s, SNFs were paidbased on their reported costs. Both theGeneral Accounting Office (GAO) andthe Office of Inspector General (OIG)found that these costs were excessivelyhigh (GAO 1998, OIG 1999b). Accordingto that system, SNFs had limits for routineoperating costs (for example, room andboard) but no limits on costs for ancillary

services, such as physical therapy.Separate limits applied based on location(urban or rural) and whether facilitieswere hospital-based or freestanding, withhospital-based facilities having higherlimits than freestanding facilities. Inaddition, new SNFs were exempt from theroutine cost limits for up to their first fouryears of operation.

Because Medicare’s payments were basedon SNFs’ costs and SNFs had littleincentive to contain costs, Medicarespending grew rapidly during this period.


Adjustment to freestanding skilled nursing facility costs (margin)

Prior to implementation ofMedicare’s prospective paymentsystem for skilled nursing

facilities (SNFs) in 1998, many nursingfacilities designated separate anddistinct units as Medicare SNF units.Nursing facilities would generally usethe beds in these SNF units exclusivelyto care for patients during theirMedicare coverage for SNF services;the rest of the nursing facility wouldgenerally care for other types ofpatients, such as long-term care patientspaid for under Medicaid or with privateresources. Nursing facilities thatmaintained separate units for Medicareand non-Medicare patients wererequired to report the costs of caring fortheir Medicare patients to the Medicareprogram each year, and Medicarepayments were based on these reportedcosts for Medicare patients.

Under the SNF prospective paymentsystem, Medicare no longer paysnursing facilities based on theirreported costs; instead, facilities receivea fixed, case-mix adjusted per diemamount for each Medicare SNF patient.Consequently, many nursing facilitieshave abandoned their practice ofmaintaining a separate unit for

Medicare SNF patients, nowinterspersing them with non-Medicarepatients throughout their facilities. Thenursing facilities that made this changenow report the average costs of caringfor all patients in the facility toMedicare each year, instead ofreporting separate costs for MedicareSNF patients only. Facilities may havechosen to make this change for anumber of reasons: It allows them tokeep patients in the same beds whenthe Medicare SNF coverage ends andpatients must transition to other sourcesof coverage, and it allows facilitiesflexibility to accept more MedicareSNF patients.

Averaging Medicare and non-Medicarecosts results in understated costs forMedicare patients. Medicare SNFpatients generally require a higher levelof nursing care than other patients. So,Medicare payment-to-cost ratios appearhigher than they would if the SNFs’reported costs were only based on theirMedicare patients. Independentanalysis by the General AccountingOffice using a different method reachesa similar conclusion—that the use offreestanding SNFs’ unadjusted averagecosts in computing the Medicare

margin overstates SNFs’ actualMedicare margin (GAO 2002b).

To account for this understatement ofthe actual costs of caring for MedicareSNF patients, we adjusted fiscal year2000 costs. We estimated the costdifferential between Medicare and non-Medicare patients in the 54 percent ofSNF facilities that reported separatecosts for each patient group in fiscalyear 2000 and applied this adjustmentto the Medicare costs for facilities thatreported average costs across allpatients. It should be noted that thisadjustment relies on the accuracy offacilities’ reported costs of caring forMedicare patients in the distinct partunits, which are determined using cost-allocation methods. To the extent thatthese costs are overallocated, ouradjustment would underestimate thetrue margin.

Prior to the adjustment, we estimate afiscal year 2000 Medicare margin forfreestanding SNFs of almost 20percent. The adjustment brings theMedicare margin for fiscal year 2000down to just under 17 percent. �

Between 1990 and 1996, for example,SNF spending grew at an average of about23 percent per year (MedPAC 2002).Much of this growth in spending was dueto increased provision of ancillaryservices.10

Under the prospective payment system,SNFs have financial incentives todecrease their costs, and evidenceindicates that freestanding SNFs haveresponded accordingly.11 FreestandingSNFs have lowered costs in a number ofways, including negotiating lower pricesfor contract therapy (physical,occupational, and speech therapists) andpharmaceuticals, substituting lower-costlabor for higher-cost labor (Liu et al.2000), and decreasing the number oftherapy staff (White 2001).

In addition, preliminary research suggeststhat the average number of minutes oftherapy provided in freestanding SNFsmay have declined (Gifford and Angelelli2002) and that rehabilitation charges perpatient per SNF day declined substantiallyin freestanding SNFs—in some cases, byas much as 47 percent—from 1997 to2000 (White 2002c).

Freestanding SNFs appear to haveresponded to incentives in the prospectivepayment system by reducing the averagenumber of minutes of physical,occupational, and speech therapy theyprovide per week to patients in each of therehabilitation RUG–III groups. In contrastto prospective payment systems for mostother providers, payment rates under theSNF prospective payment system forpatients requiring rehabilitation therapyare determined based on the number ofminutes per week SNFs actuallyprovide—or estimate they will provide—rather than on the patients’ characteristics.So, to a certain extent, SNFs candetermine the amount they are paid bycontrolling the number of therapy minutes

they provide per week. Prospectivepayments to SNFs increase at certainthreshold amounts of therapy provided,meaning that SNFs are paid one rate forproviding between 45 and 149 minutes oftherapy to a given patient and a higher ratefor providing between 150 and 324minutes for that same patient, all elsebeing equal. Thus, under the SNFprospective payment system, facilitieshave strong incentives to provide levels oftherapy that correspond to the lower endof each range, unless they can provideenough therapy to move the patient intothe next highest RUG–III group (Figure 2C–2).

The way the RUG–III payment rates arestructured provides greater incentives forSNFs to treat patients needing moderate tohigh levels of therapy than patients inother groups because these types ofpatients tend to be more profitable forSNFs than patients in other groups.Studies have generally found that, sincethe SNF prospective payment system wasimplemented, SNFs increased theproportion of patients they care for inRUG–III groups requiring moderate tohigh levels of therapy and reduced theproportion of patients in the groupsrequiring either extremely high levels oftherapy or no therapy (GAO 2002c, White2002c).

Despite substantial evidence that the costsof caring for Medicare patients infreestanding SNFs have decreased, we canfind no evidence of decreases in thequality of care delivered to beneficiaries inSNFs. This may be because SNFs’ costswere so high before the SNF prospectivepayment system that they had room toreduce their costs without reducingquality. Preliminary research examiningnational data from 1997 to 2000 has foundno change in crude measures of quality—such as activity of daily living (ADL)scores, walking scores, rates ofrehospitalizations, or incidence of

mortality (White 2002b). In addition,preliminary evidence from a study of84,000 Medicare SNF patients in Ohiobetween 1997 and 2000 indicates thatquality of care either remained the same orimproved slightly over the period (Giffordand Angelelli 2002). Researchers foundthat rehospitalization rates improvedamong beneficiaries in certain Ohio SNFs,walking scores improved slightly, andother measures of quality remainedrelatively constant. They concluded thatthese findings were not attributable tochanges in the case mix of patients.

Furthermore, studies point to a positiverelationship between increased nursingstaff times and nursing home quality ofcare (Abt 2001, HCFA 2000), and we findno evidence of declines in the overallamount of nursing staff time provided tobeneficiaries since the SNF prospectivepayment system began. Studies by GAOand by the American Health CareAssociation (AHCA) both indicate thatslight increases in nursing staff time mayhave occurred in SNFs between 2000 and2002 (AHCA 2002, GAO 2002a). GAOreported that nursing staff time per patientper day increased by about 1.9 minutes;AHCA reported an increase of 4.8 minutesper patient day in freestanding facilities,with no change in hospital-based facilities.However, both studies indicate a shift inthe mix of nursing staff time provided,with the proportion of time delivered byregistered nurses declining and theproportion of time delivered by licensednurse practitioners and nurse aides bothincreasing. Because we do not know whatimplications these changes might have forquality, it will be important to continue tomonitor the quality of care in SNFs overtime to ensure that changes in staff mix donot lead to decreases in quality in thefuture.

SNFs may have additional incentives toimprove quality regardless of cost


10 In addition, during the 1990s, the OIG found that some SNFs were billing Medicare for therapy that was not medically necessary, that was provided by staff withoutthe proper skill level to perform the therapy, and that may not have been provided at all. They also found that, in some cases, Medicare may have been paying SNFsas much as 86 percent more than the SNFs actually paid their contractors to provide the therapy. These improper billing practices likely contributed to Medicare’sspending increases for SNFs over the period (OIG 1999b).

11 Although freestanding SNFs appear to have lowered their costs significantly since the implementation of the SNF prospective payment system, evidence indicates thatcosts for hospital-based SNFs actually increased from 1997 to 1999. In fact, GAO reported that hospital-based SNF costs increased by $29 per day from 1997 to1999, while freestanding facilities’ costs decreased by $49 per day over the same period (GAO 2002a).

pressures because CMS has recentlybegun to publish nationwide reports withquality measures at the individual nursingfacility level. CMS is also devotingresources to help nursing facilities thatwish to improve their scores on thesenationally reported measures. Nursingfacilities generally indicate that they areaware of the public reporting of thequality measures and that they would liketo improve their scores on future reports.Thus, this public reporting may serve as acountervailing force to maintain quality innursing facilities even as the incentives ofthe prospective payment systemencourage facilities to reduce costs.

We therefore conclude that SNFs havelowered the costs of inputs to providingcare to Medicare beneficiaries. At thesame time, we find no reductions in the

quality of care. Together, this points to animprovement in productivity in the SNFsector because SNFs appear to have beenable to reduce the resources needed toproduce SNF services while maintainingservice quality.

Relationship of payments tocosts Although our estimate of the Medicaremargin for SNFs provides one importantpiece of information regarding theadequacy of Medicare’s payments forSNF services, we look to other availableevidence from market factors to ensurethat Medicare payments are generallyadequate to meet the needs of providersand beneficiaries. From this analysis, wecan find no indications of overallproblems with Medicare payments toSNFs.

Entry and exit of providersThe total number of SNFs participating inMedicare remained relatively stablebetween 1998 and 2002, declining by lessthan 1 percent in each of the first threeyears of the period and increasing by lessthan 1 percent between 2001 and 2002(Table 2C-2). The patterns of entry andexit vary among different types of SNFs,however. From 1998 to 2002, the numberof freestanding SNFs participating inMedicare increased by about 3 percent,while 26 percent of hospital-based SNFshave exited the program over the sameperiod.

Recent research examining the entry andexit of SNF providers using average dailycensus measures and inflows and outflowsof providers finds similar patterns in entryand exit (White 2002a). This research also


200

Skilled nursing facility payments and minutes of therapy for rehabilitationpatients, 2003

FIGURE2C-2

Note: Payment rates are based on 2003 urban payment rates for a Medicare skilled nursing facility patient with Activities of Daily Living score of 9. The 6.7 percent payment add-on is included.

Source: CMS. Medicare program: prospective payment system and consolidated billing for skilled nursing facilities—update. Federal Register. July 31, 2002, Vol. 67, No. 147, p. 49802.

450

400

300

350

250

150

100

50

00

Med

icare

paym

ent

rate

s, 2

00

3 (

dolla

rs p

er d

ay)

705105

$212

$283 $296

$339

$409

205 305 405 505 605

Minutes of therapy per week

indicates that post-prospective paymentsystem changes in SNF payment ratesmay be one of many factors determiningwhether facilities remained in the programor exited, though perhaps not the mostimportant factor. Freestanding SNFs weremore likely to close if they were new tothe market, nonprofit, and smaller, with asmaller fraction of Medicare beds. Theywere less likely to exit if they had morepatients needing high levels ofrehabilitation therapy (more profitableunder the prospective payment system) orfewer patients requiring expensivepharmaceutical services (which are notreimbursed outside of the per diempayment). Similarly, hospital-based SNFswere more likely to exit if they were newto the market, if they were for profit(especially members of chains), or if agreater proportion of their patients hadhigh inhalation therapy costs. For the mostpart, only facilities experiencing morethan a 40 percent decline in payments

after the implementation of the SNFprospective payment system had a higherthan average exit rate from the programbetween 1998 and 2000.

The continuing entry of freestanding SNFproviders—particularly for-profitfreestanding providers—to the Medicareprogram may indicate that these providersfind the flow of revenues from Medicareto be at least adequate. However, hospital-based SNF providers continue to leave theprogram. Analysis by MedPAC and othersshows that hospital-based SNFs have asubstantially higher case mix of patientsthan freestanding SNFs and may treat adisproportionate number of patients withexpensive, nonrehabilitation therapyneeds (Dalton and Howard 2002, Liu andBlack 2002, MedPAC 2001). Because theSNF classification system appears to do apoor job of allocating resources accordingto the expected resource needs ofrehabilitation and nonrehabilitation

patients, hospital-based SNFs may havemore difficulty making a profit under theSNF prospective payment system thanfreestanding SNFs.

In addition, hospital administrators maybe responding to increased demand foracute care services. Acute care hospitaloccupancy rates have increased in recentyears at the same time that the nation isexperiencing a shortage of nurses. Somehospital administrators report an increasein demand for hospital beds and nurses onthe acute care side and may have shiftedbeds and nurses from the SNF to the acutecare units. In some cases, this may havemeant closing the SNF unit altogether.

Other possible reasons for hospital-basedSNFs to exit may include large changes inMedicare reimbursements before and afterthe SNF prospective payment system andstate and federal regulatory burden issues.

Changes in the volume ofservicesChanges in the volume of servicesdelivered by a particular set of Medicareproviders may indicate whether paymentsto those providers are too high or too lowrelative to providers’ costs. If we seeincreases in the volume of services, thislikely indicates that payments are at leastadequate. Large increases may signal thatpayments are too high relative to costs.

The most recent available data from 2000suggest that the volume of SNF serviceshas increased overall (Table 2C-3).


Change in the number of certified skilled nursingfacilities, by type, 1998–2002

Percent change Percent change1998 2001 2002 1998–2002 2001–2002

Hospital-based 2,173 1,762 1,611 –26% –9%Freestanding 12,862 12,993 13,204 3 2All facility types 15,035 14,755 14,815 –1 �1

Source: MedPAC analysis of Online Survey, Certification, and Reporting (OSCAR) system data from CMS.

T A B L E2C-2

Payment and use of skilled nursing facilities, 1996–2000

Percentchange,

1996 1997 1998 1999 2000 1999–2000

Payment (billions) $639.3 $211.0 $211.3 $029.5 $210.4 10%Average payment/day 208 233 250 223 236 6

Discharges (1,000s) 1,318 1,582 1,588 1,450 1,439 –1Covered days (1,000s) 44,639 47,295 45,240 42,535 44,103 4Average days/discharge 33.9 29.9 28.5 29.3 30.7 5

Note: Data include facilities in Puerto Rico, Virgin Islands, and “unknown.” Data do not include swing bed units.

Source: CMS.

T A B L E2C-3

The total number of discharges fromSNFs remained essentially stable between1999 and 2000, decreasing by less than 1percent. Over this same period, theaverage length of stay in SNFs increasedby more than one day. Hence, totalMedicare covered days increased by about4 percent from 1999 to 2000. TotalMedicare payments to SNFs and averagepayments per day also increased, by 10and 6 percent respectively, from 1999 to2000. These payment increases reflect, atleast in part, the payment add-ons thattook effect in April 2000 (a 4 percentincrease in payments for all RUG groupsand a 20 percent increase in payments for12 complex care groups).

Beneficiaries’ access to care In 2001, the OIG reported thatbeneficiaries generally did not haveproblems obtaining SNF care. However,the findings suggested that patientsrequiring costly services might haveexperienced delays in accessing SNF care(OIG 2001). These findings wereconsistent with those from a similar studyin 2000 (OIG 2000).

In October 2002, MedPAC convened afocus group of 15 hospital dischargeplanners to continue to monitor patients’access to SNF care. These dischargeplanners told us that beneficiaries needingrehabilitation therapy services generallyhad no problem accessing SNF services.Certain beneficiaries needing expensive,nonrehabilitation services might remain inthe acute care hospital longer than beforethe SNF prospective payment system.Hospitals are concerned about thisbecause they do not receive additionalMedicare reimbursement for theadditional time these patients spend in thehospital (decreasing the profit hospitalscan make on these patients). However, itis not clear that the additional time in thehospital is an inappropriate outcome forthese patients. Overall, we did not findevidence of widespread access problems.

Because beneficiaries’ access to care issuch an important indicator of theadequacy of Medicare payments, it isimperative that we continue to monitorthis market factor using the most currentand reliable information possible. From1999 to 2001, the Department of Healthand Human Services’ Office of InspectorGeneral conducted an annual series ofstudies assessing beneficiaries’ access toSNF services (OIG 1999a, OIG 2000,OIG 2001). The OIG did not issue a reporton SNF access in 2002, and has indicatedthat it does not plan to continue thesereports in the future. We believe thesestudies are an important and relevantaddition to the policy process.

R E C O M M E N D A T I O N 2 C - 1

The Secretary should continue aseries of nationally representativestudies on access to skilled nursingfacility services (similar to studiespreviously conducted by theDepartment of Health and HumanServices’ Office of Inspector General).

I M P L I C A T I O N S 2 C - 1

Spending• This recommendation should not

affect Medicare benefit spending.

Beneficiary and provider• To the extent that future OIG studies

allow us to monitor beneficiaries’access to SNF services closely and toreact quickly if problems develop,they contribute to preservingbeneficiaries’ access to care. Webelieve this recommendationrepresents a minimal burden toproviders.

In MedPAC’s report to the Congress inMarch 2000, we recommended that theSecretary conduct annual studies toidentify potential problems in

beneficiaries’ access to care that may arisein the evolving Medicare program,particularly from the implementation ofnew payment systems in the varioussectors (MedPAC 2000). The SNFpayment system continues to evolve,indicating a need for continuedmonitoring of beneficiaries’ access toSNF services.

Future reports do not need to be done onan annual basis; they may be necessaryonly every few years as long as noadverse trends in access are observed andas long as Medicare payments to SNFsremain relatively stable over the course ofa few years. Primarily, it is important thata consistent knowledge base be built upover time in this area. We expect that thelength of time between studies wouldgenerally be left to the discretion of theSecretary, as would the best operatingdivision to conduct these studies (forexample, OIG or CMS).

SNFs’ access to capital Overall, SNFs’ access to capital may havebeen affected by recent bankruptcies,payment uncertainties, and the costs ofliability insurance and lawsuits. However,the evidence does not suggest systematicproblems with SNFs’ access to capital thatwould pose problems for beneficiaries’overall access to SNF services.

Whereas Medicare payments for inpatienthospitalizations, for example, represent arelatively large share of hospitals’revenues, Medicare payments for SNFcare represent a small share of bothhospitals’ and nursing facilities’revenues.12 Thus, Medicare payments toSNFs have a less important role indetermining whether SNFs are able toaccess capital than other factors, such aswhether SNFs are associated with acutecare hospitals or nursing facilities and theamount of funding SNFs receive from


12 Medicare payments, on average, comprise about 10 to 12 percent of revenue for nursing facilities and about 2 percent for hospitals. Large for-profit nursing facilitycompanies derive the largest share of revenues from Medicare, about 25 percent.

other sources. Given that Medicaidpayments generally comprise the largestshare of nursing home revenues,investors’ views of the nursing homeindustry may be driven largely byperceptions of the adequacy of Medicaidpayments. Current fiscal pressures andstate budget cuts may be leading todecreases in Medicaid payments, whichwould tend to make investors more waryof investing in this sector than they havebeen in the past. However, to the extentthat this may be happening, all indicationsare that this is more a reflection of theadequacy of Medicaid payments thanMedicare payments to nursing homes (seetext box below).

As mentioned earlier, hospital-basedSNFs represent about 10 percent of allSNFs. They generally have access tocapital through their parent hospitalorganizations; the extent to which they areable to access capital depends on thefinancial condition of the hospital as a

whole. (Hospitals’ access to capital isdiscussed in more detail in Chapter 2A.)

About 90 percent of all SNFs are locatedwithin nursing facilities. The nursingfacility industry consists of many smallcompanies, with the top 10 nursingfacility companies (as measured by thenumber of beds) controlling only about 18percent of the market. Nursing facilities’access to capital may have been affectedby recent bankruptcies, uncertaintiesabout government revenues, and the costof liability lawsuits and insurance.

Five of the 10 biggest for-profit publicly-held companies are either restructuringunder Chapter 11 or have recentlyemerged from bankruptcy. Both GAO andCMS found that these bankruptciesresulted from extensive investment inancillary service lines of business andhigh capital-related costs (such asdepreciation, interest, and rent) (CMS2002, GAO 2000). Some of thesecompanies appear to be regaining

competitive ground as they emerge frombankruptcy (CMS 2002).

Most smaller- and mid-sized for-profitcompanies appear to have been able torespond to lower Medicare revenues underthe prospective payment system bylowering their costs (Fitch 2001). If thisfact is recognized by lenders, thosefacilities may have reasonable access tocapital.

Uncertainty about government revenuesand liability insurance rates and lawsuitscontinues to be a concern for the nursingfacility sector. Uncertainty about whetheror not the Congress intends to reinstatetwo temporary payment add-ons thatexpired on October 1, 2002 has causedinvestors to be generally cautious. Inaddition, nursing facilities have had anumber of recently publicized problemswith liability and lawsuits. One for-profitnursing facility chain sold 49 Floridanursing facilities last January in partbecause of liability concerns; anotherlarge nursing facility chain’s stock pricefell by over 16 points, to $11.36, when itannounced it was recording about $55million in additional costs for professionalliability claims (Charles Schwab 2002,Standard & Poor’s 2002).

Nonprofit SNFs had difficulty gettinginvestment grade ratings both before andafter the SNF prospective paymentsystem.

The evidence regarding demand forcapital in this sector is mixed. Someevidence indicates that the demand forcapital to finance new construction maybe low because of large capitalinvestments in the late 1990s and nursingfacility occupancy rates that average about81 percent (National Investment Center2001). On the other hand, we should bemindful of the need to replace oldbuildings and equipment, which mayrequire additional capital to financerenovations and improvements to theexisting capital stock.


Medicaid payments to nursing facilities

Many are concerned about potentialinadequacies in Medicaid payments tonursing facilities. For this reason,many representatives of the nursingfacility industry and others havesuggested that Medicare shouldmaintain higher payments—that farexceed the costs of caring forMedicare beneficiaries—in order tocompensate for the lower Medicaidpayments and maintain the financialstability of the industry.

However, MedPAC believes thatusing high Medicare payments tocompensate for any inadequacies thatmay exist in Medicaid payments is aninefficient way of improving thefinancial situation of this industry forthree reasons. First, Medicarepayments represent about 10 to 12

percent of total revenues for theaverage nursing facility; with such asmall base of revenues, Medicarecannot be expected to assumeresponsibility for the financial welfareof the whole industry. Second, ifMedicare does assume thisresponsibility, states may beencouraged to reduce their Medicaidfunding even further. Finally, usinghigh Medicare payments tocompensate for low Medicaidpayments targets the money to thewrong facilities (i.e., more moneywould go to facilities with fewerMedicaid patients, instead of thosefacilities that presumably would needadditional funding the most—thosewith a high proportion of Medicaidpatients). �


MedPAC’s update recommendationsdepend on two assessments: the adequacyof current payments for care in SNFs andexpected changes in the costs of providingcare in the coming year. As in othersettings, when considering changes incosts in the coming year, we start with aforecast of the market basket index. TheSNF market basket index, currentlyprojected to be 2.9 percent for fiscal year2004, provides a measure of how priceschange for a fixed set of inputs to provideSNF care.

In predicting expected changes in costs forthe coming year, we look for evidence ofthe adoption of quality-enhancing newtechnologies that put substantial upwardpressure on the costs of care. We do notfind evidence of quality-enhancingtechnological advances in the SNF sectorthat would significantly increase costs.The largest component of SNF costs islabor. The SNF market basket index isdesigned to account for any cost increasesfor labor and other inputs to the provisionof services in SNFs.13

Similarly, we look for evidence ofproductivity growth that typically lowersthe cost of providing care. Evidenceshows that SNFs have lowered the costsof caring for Medicare beneficiaries inSNFs substantially since the SNFprospective payment system wasimplemented. This appears to haveoccurred without a reduction in the qualityof care provided, indicating an overallimprovement in productivity in the SNFsector. We expect this trend to continueinto fiscal year 2004, with SNFproductivity at least matching theeconomy-wide growth in multifactorproductivity—about 0.9 percent peryear—over the coming year.

Because we do not anticipate quality-enhancing advances in new technologythat will significantly increase costs in thissector, our update recommendation isbased primarily on our assessment of theadequacy of current payments to SNFsand our assumption that growth inproductivity will continue over the nextyear.

Update recommendations

We estimate that overall Medicarepayments to SNFs are adequate to coverthe costs of caring for Medicare SNFpatients, but the evidence indicates thatthe distribution of payments in the systemmay make it more difficult for facilities toprofit from treating a higher proportion ofpatients with expensive, nonrehabilitationtherapy needs. MedPAC thereforerecommends two changes: one affectingthe base payment amount and the otherone affecting payments for SNF patientswith expensive, nonrehabilitation therapyneeds.

In our March 2002 recommendations, werecommended differential updates tofreestanding and hospital-based SNFsbecause we believed that the developmentand implementation of a new SNF patientclassification system would take too muchtime. We recommended differentialupdates as an interim measure. This year,we recommend more immediate measuresto balance the distribution of payments inthe system so they better track theexpected resource needs of SNF patients.Differential updates are no longernecessary, unless the recommendedchanges do not occur rapidly enough.

R E C O M M E N D A T I O N 2 C - 2

The Congress should eliminate theupdate to payment rates for skillednursing facility services for fiscal year2004.


Spending• Because this recommendation

provides no update to payments forskilled nursing facility services,whereas current law updatespayments for these services by theSNF market basket index, thisprovision is expected to reduceMedicare spending relative to currentlaw by between $200 million and$600 million for fiscal year 2004 andbetween $1 billion and $5 billionover 5 years.

Beneficiary and provider• Because we estimate current

Medicare payments to besubstantially above the costs ofcaring for Medicare beneficiaries inSNFs, we expect little if any effect ofthis provision on beneficiaries’ accessto care. Similarly, we do notanticipate major problems forproviders of SNF services,particularly in combination withRecommendation 2C-3.

Given that the overall Medicare marginfor all SNFs is about 5 percent and thatmarket factor evidence indicates no majorproblems in this sector, the base rate forall SNFs appears to be adequate and noupdate to payment rates is necessary atthis time.

However, while we find overall Medicarepayments to SNFs to be adequate, weremain concerned about the distribution ofexpenditures resulting from a SNF patientclassification system that makes certaintypes of Medicare beneficiaries moreprofitable for SNFs to treat than others.For this reason, we combine thisrecommendation with a recommendationdesigned to improve the allocation ofresources in the SNF payment system sothat it will recognize and better balancethe resource needs of SNF patients withrespect to rehabilitation therapy andnonrehabilitation therapy needs.


13 In the years since the SNF prospective payment system was implemented, the projected SNF market basket index used to determine SNF payment rate updates hasunderstated the actual SNF market basket index. Had CMS been able to go back and correct for this error in forecasting the market basket index, fiscal year 2003Medicare payments to SNFs would exceed the costs of caring for Medicare SNF patients by more than the 5 percent we estimate.

RECOMMENDAT ION 2C-3A

Consistent with previous MedPACrecommendations, the Secretaryshould develop a new classificationsystem for care in skilled nursingfacilities.

Because it may take time to developthis system, the Secretary shoulddraw on new and existing researchto reallocate payments to achieve abetter balance of available resourcesbetween the rehabilitation andnonrehabilitation groups.

To allow for immediate reallocationof resources, the Congress shouldgive the Secretary the authority to:

• remove some or all of the 6.7percent payment add-on currentlyapplied to the rehabilitationRUG–III groups.

• reallocate money to thenonrehabilitation RUG–III groups toachieve a better balance ofresources among all of the RUG–IIIgroups.

RECOMMENDAT ION 2C-3B

If necessary action does not occurwithin a timely manner, the Congressshould provide for a market basketupdate, less an adjustment forproductivity growth of 0.9 percent,for hospital-based skilled nursingfacilities to be effective October 1,2003.


Spending• Because part A of this

recommendation suggests aredistribution of resources already inthe system, this provision is expectedto be spending neutral.

• Part B of this recommendation wouldincrease spending relative to thecombination of Recommendation2C-2 and Recommendation 2C-3A.However, it would not change theexpectation of a reduction inspending for Recommendation 2C-2of between $200 million and $600

million over 1 year and between $1billion and $5 billion over 5 years,relative to current law.

Beneficiary and provider• To the extent that payments track

more closely the expected resourceneeds of different types of SNFpatients and increase the incentivesfor providers to accept patients withhigh nontherapy ancillary serviceneeds, beneficiaries’ access to care isexpanded.

• To the extent that this provisionredistributes payments to providersthat care for a disproportionatenumber of SNF patients with highnontherapy ancillary service needs,Medicare payments may be moreequitably distributed among SNFproviders according to the costs ofthe patients they treat. To the extentthat hospital-based SNFs treat more

of these types of patients, thisredistribution should provide themwith more resources.

The Commission remains concerned thatthe current SNF patient classificationsystem does not appropriately distributeresources among patients with differenttypes of resource needs. SNFs who carefor more patients with expensive,nonrehabilitation therapy needs may notbe able to operate as profitably under theSNF prospective payment system as SNFsthat care for a higher proportion ofpatients with short-term rehabilitationneeds. In addition, patients withexpensive, nonrehabilitation therapyneeds may experience longer delays inaccessing SNF services than other typesof patients. The Commission recommendsa series of long-, intermediate-, and short-term steps to address these problems andbetter balance the available resourcesamong patients with different types ofresource needs.


History of skilled nursing facility payment add-on

In the Balanced Budget RefinementAct of 1999 (BBRA), the Congressmandated a 20 percent increase to thepayment rates for 12 RUG–III groupscovering medically complex cases inthe extensive services, special care,and clinically complex groups, as wellas 3 select rehabilitation RUGs. Thispayment increase began on April 1,2000, and was designed to remain ineffect until CMS announced a revisedskilled nursing facility (SNF)classification system. MedPACindicated at the time that this add-onwas not a perfect solution, although itmight help offset some providerexpenses for these patients. MedPACcontinued to be concerned, however,that Medicare’s reimbursement ratesfor the rehabilitation RUGs were toohigh.

The following year in the Medicare,Medicaid, and SCHIP BenefitsImprovement and Protection Act of2000 (BIPA), the Congress altered thepayment increase mandated by BBRA

such that all 14 rehabilitationRUG–III groups, including thosepreviously receiving the 20 percentadd-on, would receive a 6.7 percentpayment add-on until CMSannounced a refinement to theclassification system. BIPA left inplace the 20 percent add-on topayment rates for thenonrehabilitation RUG–III groups.The revised payment add-ons becameeffective April 1, 2001.

The 6.7 percent payment rate add-onto the rehabilitation RUGs and the 20percent payment add-on to thenonrehabilitation RUGs both remainin effect until CMS announces arefinement to the SNF classificationsystem. At the moment, it is unclearwhen CMS might announce such arefinement, although they are requiredby law to report to the Congress onalternatives to the existing RUG–IIIpayment system by January 1,2005. �

In the long term, the problems describedhere cannot be fully addressed with thecurrent SNF patient classification system.The best solution, therefore, is to developa new system that distributes resourcesmore appropriately among patients withdifferent expected service needs.

However, a new payment system willalmost surely take time to develop andimplement. Therefore, as an intermediatestep, we feel it is important to look to allcurrently available sources of informationfor ways to improve payments until a newclassification system can be adopted.

Two key conditions must be met beforethe Secretary significantly restructures thecurrent SNF patient classification systemor implements any new SNF patientclassification system:

1. Any changes must be demonstrated tobe effective on a nationallyrepresentative sample of Medicarebeneficiaries.

2. The new system must be more effectivethan the current system, in that it mustexplain more of the variation in SNFpatients’ expected resource needs.

Substantial improvements may still beyears away, so a more immediateredistribution of resources is neededwithin the current payment system. Suchredistribution involves adjusting thepayment add-on that the Congressimplemented in BBRA (and revised inBIPA) to give CMS time to refine theRUG–III payment system (see text boxprevious page).

Accordingly, the Congress should give theSecretary the authority to redistributesome or all of the 6.7 percent paymentincrease from the rehabilitation RUG–IIIgroups to the nonrehabilitation groups.This has no effect on the 20 percent add-on to payment rates for thenonrehabilitation RUGs in current law.Payments to SNFs for rehabilitationpatients appeared more than adequateeven before the Congress implemented the6.7 percent payment add-on to therehabilitation RUGs, and payments toSNFs for nonrehabilitation patients appearnot to be adequate even with the 20percent add-on currently in effect. �


References

Abt Associates Inc. Appropriateness of minimum nurse staffing ratios in nursing homes,phase II final report. Cambridge (MA), Abt Associates, Inc. Winter 2001.

American Health Care Association. Letter to GAO in response to skilled nursingfacilities: available data show average nursing staff time changed little after Medicarepayment increase, GAO–03–176. December 3, 2002.

Centers for Medicare & Medicaid Services. CMS health care industry market update:nursing facilities. Baltimore (MD), CMS. February 6, 2002.

Charles Schwab. Charles Schwab company research report: Kindred Healthcare, Inc.Charles Schwab. October 17, 2002.

CHPS Consulting. Final report: outpatient resource costing study. Columbia (MD),CHPS. December 1994.

Dalton K, Howard HA, University of North Carolina at Chapel Hill. Market entry andexit in long term care: trends before and after Medicare’s prospective payment system forskilled nursing facilities. Unpublished paper. Written 2002.

Fitch, Inc. and the National Investment Center for the Seniors Housing & Care Industries.Fitch/NIC investment analysis of nursing home properties. Annapolis (MD), NationalInvestment Center for the Seniors Housing & Care Industries. September 2001.

General Accounting Office. Balanced budget act: implementation of key Medicaremandates must evolve to fulfill congressional objectives, No. T–HEHS–98–214.Washington (DC), GAO. July 1998.

General Accounting Office. Nursing homes: aggregate Medicare payments are adequatedespite bankruptcies, No. T–HEHS–00–192. Washington (DC), GAO. September 2000.

General Accounting Office. Skilled nursing facilities: available data show averagenursing staff time changed little after Medicare payment increase, No. GAO–03–176.Washington (DC), GAO. November 2002a.

General Accounting Office. Skilled nursing facilities: Medicare payments exceed costsfor most but not all facilities, No. GAO–03–183. Washington (DC), GAO. December2002b.

General Accounting Office. Skilled nursing facilities: providers have responded toMedicare payment system by changing practices, No. GAO–02–841. Washington (DC),GAO. August 2002c.

Gifford D, Angelelli J, Center for Gerontology and Health Care Research, BrownUniversity. Presentation at the Academy for Health Service Research conference andMedPAC staff discussions with authors. Washington (DC). June 25, 2002.

Health Care Financing Administration. Report to Congress: appropriateness of minimumnurse staffing ratios in nursing homes, phase I. Baltimore (MD), HCFA. Summer 2000.

Liu K, Black K, The Urban Institute. Hospital-based and freestanding skilled nursingfacilities: any cause for differential Medicare payments? Washington (DC). Unpublishedpaper. Written 2002.


Liu K, Harvell J, Gage B. Post-acute care issues for Medicare: interviews with providerand consumer groups, and researchers and policy analysts, Contract No. 100–97–0010.Washington (DC), Assistant Secretary for Planning and Evaluation. March 2000.




National Investment Center for the Seniors Housing & Care Industries. The case forinvesting in seniors housing and long term care properties. Annapolis (MD), NationalInvestment Center for the Seniors Housing & Care Industries. 2001.

Office of Inspector General, Department of Health and Human Services. Early effects ofthe prospective payment system on access to skilled nursing facilties: nursing homeadministrator’s perspective, No. OEI–02–99–00401. Washington (DC), OIG. October1999a.

Office of Inspector General, Department of Health and Human Services. Early effects ofthe prospective payment system on access to skilled nursing facilities, No.OEI–02–99–00400. Washington (DC), OIG. August 1999b.

Office of Inspector General, Department of Health and Human Services. Medicarebeneficiary access to skilled nursing facilities, 2001, No. OEI–02–01–00160. Washington(DC), OIG. July 2001.

Office of Inspector General, Department of Health and Human Services. Medicarebeneficiary access to skilled nursing facilities, 2000, No. OEI–02–00–00330. Washington(DC), OIG. September 2000.

Pizer SD, White AJ, White C. Why are hospital-based nursing homes so costly? Relativeimportance of acuity and treatment setting, Medical Care. May 2002, Vol. 40, No. 5, p.405–415.

Standard & Poor’s. Stock reports: Manor Care. Standard & Poor’s. October 12, 2002.

White C, Harvard University. Do skilled nursing facilities receive adequatereimbursement from Medicare? Unpublished paper. Written 2002a.

White C, Harvard University. MedPAC’s preliminary discussions with Sally Kaplan.Washington (DC), December 2002b.

White C, Harvard University. Rehabilitation therapy in skilled nursing facilities: effectsof Medicare’s new prospective payment system. Unpublished paper. Written 2002c.

White C, Harvard University. Rehabilitation therapy in skilled nursing facilities:evaluating the effects of recent changes in Medicare’s payment system. Presentationmade at Harvard University. Cambridge (MA). November 6, 2001.


2DAssessing payment adequacy

and updating paymentsfor home health services

S E C T I O N


2D-1 The Secretary should continue a series of nationally representative studies on access tohome health services (similar to studies previously conducted by the Department of Healthand Human Services’ Office of Inspector General).


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2D-2 The Congress should extend for one year add-on payments at 5 percent for home healthservices provided to Medicare beneficiaries who live in rural areas.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2D-3 The Congress should eliminate the update to payment rates for home health services forfiscal year 2004.



Section 2D: Assessing paymentadequacy and updating payments forhome health services

Our review of the evidence finds that aggregate Medicare payments for home

health services are more than adequate relative to costs, even after accounting for

the reduction in the base payment for fiscal year 2003. Our estimate of the

Medicare margin for home health services in fiscal year 2003 is 23.3 percent.

Changes in the home health product over the past five years have reduced the

costs of producing an episode of home health services. Our evidence suggests that

the costs of producing an episode of home health services will continue to de-

crease, at a slower pace, over the coming year. Medicare spending for home

health is projected to increase due to growth in both the number of users and pay-

ments per user. Other broad indicators also suggest that payments are adequate:

access to care is generally good, the rate of decline in the number of users has de-

creased, and the entry and exit of agencies has remained stable for the third year

in a row.

2DIn this section


• Accounting for providers’ costchanges in the coming year



S E C T I O N

Background

Assessing payment adequacy and makingan appropriate update for home healthservices requires consideration of thebenefit itself, how providers are paid forthe services, and the context of recenttrends in spending.

Home health care is skilled nursing, aideservice, medical social work, or therapyprovided to beneficiaries in their places ofresidence. To qualify for Medicare’shome health benefit, beneficiaries mustmeet the program’s eligibility criteria:they must need part-time or intermittentskilled care to treat their illness or injury,and they must be homebound. Medicare’scoverage does not include unskilled careto maintain a person’s health unless it isrequired in conjunction with medicaltreatment by a skilled medicalprofessional. In some instances, skilledcare over a long period of time would becovered. Also in some instances, skilledcare for patients whose medical conditionis stable would be covered. However,patients who need more or less full timeskilled nursing care over an extendedperiod of time generally would not qualifyfor Medicare home health benefits (CMS2001). To qualify for coverage,beneficiaries must also be unable to leavetheir homes without considerable effort.

Throughout the early 1990s the use of thehome health benefit changed. A growingproportion of the home health benefit wasdirected toward beneficiaries’ long termcare needs, and less to the medicalservices necessary for the diagnosis andtreatment of illness or injury that arecovered under other Medicare post-acutecare benefits. By 1996, one-third of allvisits were provided to beneficiaries whoreceived over 300 visits a year (MedPAC1998). Legislative changes to Medicare inthe Balanced Budget Act of 1997 (BBA)included refinements to the eligibilitystandards and two new payment systemsthat made home health care more similarto Medicare’s other post-acute care

services. The continuing impact of thechanges made in 1997 is evident in 2001in substantially slower but continuingdeclines in the number of home healthusers, the duration of their care, and thenumber of visits they use. This chapterexamines the change in the home healthproduct, and the implications for ourassessment of payment adequacy.

Home health servicespayment systemThe current structure of the paymentsystem continues to have a profoundeffect on home health services (see textbox). Under Medicare’s prospectivepayment system (PPS) for home healthcare implemented in October 2000, homehealth agencies receive payment for 60-day episodes of care. Neither copaymentsnor deductibles apply to home health. Thebase payment amount for an episode ofcare is $2,160 for fiscal year 2003. Thebase payment is adjusted to account fordifferences in patients’ expected resourceneeds, as reflected by their clinical andfunctional severity, recent use of otherhealth services, and therapy use. Paymentalso is adjusted for differences in localmarket conditions by the hospital wageindex. Adjustments for several otherspecial circumstances, such as outliers orepisodes with four or fewer visits, can alsomodify the payment (see Appendix A formore information on the home healthpayment system).

The structure of the home health PPSprovides financial incentives for homehealth agencies to reduce the number ofvisits delivered in an episode of care. Solong as high quality of care persists, wecan infer that such declines increase theefficiency of the provider, rather thanadversely affect patients’ outcomes.Concern about the incentives that the PPSwould introduce once it was implementedled CMS to develop the Outcome andAssessment Information Set (OASIS) tomonitor the quality of home health care.We have used OASIS measures as part ofour assessment of payment adequacy to

indicate whether high quality of care haspersisted.

Trends in Medicarepayments for home healthservices Over the past 10 years, Medicare spendingfor home health has changed fromunprecedented growth to rapid decline,only to return to projections of rapidgrowth for the next 5 years. Between 1990and 1996, spending grew nearly 400percent, with some year-to-year growth ashigh as 50 percent (Figure 2D-1, p. 6).

Previous research (Komisar and Feder1998) disaggregated the components ofgrowth in spending from 1990 to 1996and attributed it to increases in the:

• number of Medicare beneficiaries, 7percent

• proportion of home health usersamong Medicare beneficiaries, 36percent

• visits per home health user, 49percent

• average payment per visit, 9 percent.

This research suggests that the level ofpayment per unit of service is only oneinfluence among several that affect thespending and use of the home healthbenefit. At its high point in 1997,Medicare spent $18 billion on homehealth services for beneficiaries.

Changes to the home health benefit—especially changes to the system of payingfor home health—led to a rapid decline inuse, and hence spending, after 1997 (seetext box). In 2001, Medicare spentbetween $9 and $11 billion1 on homehealth services; as a sector, home healthrepresented about 4 percent of totalMedicare fee for service spending (Figure2D-1, p. 106). Spending for home healthservices is composed entirely of programspending; beneficiaries have no cost-sharing obligations for home healthservices.

104 As s e s s i ng paymen t adequacy and upda t i ng paymen t s f o r home hea l t h s e r v i c e s

1 Estimates from the Congressional Budget Office and Office of the Actuary vary.


Changes in use of the Medicare home health benefit

Use of Medicare’s home healthbenefit has changedconsiderably over the past 10

years. In 1990, fewer than 2 millionbeneficiaries used the home healthbenefit. Between 1990 and 1996, thenumber of users grew 85 percent,adding over one million beneficiaries tothe number of users of the benefit. Thetrend was reversed in 1997; by 2001the number of users had fallen toaround 2.2 million.

Three influences—changes in thecriteria for beneficiaries’ eligibility toreceive home health services,enforcement of the rules of the programfor providers, and the structure of thepayment system and incentivesassociated with it—have shaped thetrends in use and spending forMedicare’s home health benefit overthe past 10 years as much or more thanthe level of payment for a unit of homehealth service.

Leading to growth• Eligibility. In 1989, a legal decision

(Duggan v. Bowen) made the HealthCare Financing Administration(HCFA, now CMS) change itsinterpretation of eligibility for thebenefit so that persons who neededdaily, long-term-care—oftenbeneficiaries with chronicconditions—could qualify.

• Enforcement. That legal decisionalso constrained HCFA’s ability todeny coverage and payment in manyinstances. Pursuant to the decision,HCFA could no longer denypayments for some marginal visits

for a given beneficiary based upongeneral inferences about patientswith similar diagnoses, but insteadhad to review the entire case of eachbeneficiary individually.

• Incentives. Prior to the PPS, homehealth agencies were paid for eachvisit according to visit types—generally therapy, nursing, or homehealth aide. Per-visit paymentsencouraged agencies to provide asmany visits as possible as long astheir costs were less than the per-visit payment limits for that type ofvisit.

Following these changes, use of thebenefit grew. In 1996, over 3.5 millionbeneficiaries used the home healthbenefit. Concern over the rapid rate ofgrowth and the changing nature of theservices led to legislation and otheractions intended to reverse the trends.

Changing direction• Eligibility. In 1997, the BBA

clarified the acceptable frequency ofvisits and removed the drawing ofblood as a qualifying service. Bydefining the term “part-time orintermittent,” the BBA narrowedcoverage of very frequent or nearlyfull-time care from 56 hours perweek of nursing and home healthaide service to 35 hours per week(Komisar and Feder 1998).Agencies reported that excluding thedrawing of blood decreased thenumber of users “significantly” in atleast six high-use states (GAO1999).

• Enforcement. The Secretaryinitiated Operation Restore Trust,1

which scrutinized Medicare homehealth, prompted the involuntaryclosure of hundreds of agencies thatwere not in compliance with theprogram’s integrity standards, andestablished civil liabilities forphysicians who knowingly falselycertified the eligibility of abeneficiary.

• Incentives. The structure of theinterim payment system (IPS)implemented in 1997 gaveincentives for agencies to maintain amix of patients who needed fewvisits and inexpensive visits to staybelow the cost limits. Under IPS,agencies were paid the lesser ofactual costs, aggregate costs perbeneficiary subject to an agency-specific limit, or aggregate costs pervisit subject to an agency-specificlimit. There were no outlierpayments for high cost patients. InMedPAC’s survey of changes inprovider behavior, providers statedthat many tried to avoid costlypatients under the IPS (Stoner et al.1999).

In the wake of these changes, thenumber of Medicare beneficiaries usinghome health care decreased by aboutone million. The decrease in use wascaused by decreases in the number ofeligible beneficiaries, a decline in thenumber of beneficiaries who neededcontinuous care using the benefit, adecline in fraudulent or questionableuse of the benefit, and the structure andincentives of the IPS. Fifteen percent of

(continued next page)

1 Operation Restore Trust began as a demonstration project in 1995 in California, Florida, Illinois, New York, and Texas and was expanded to additional statesin 1997. It included skilled nursing facilities and other sectors of Medicare in addition to home health.


Changes in use of the Medicare home health benefit (continued)

users in 1996 had more than 150 visitsin the year; the decline in the averagevisits per user from 1997 to 2001 (see“Changes in volume,” p. 110) suggeststhat such heavy use is no longercommon.

Though there were fewer Medicarehome health users in 2001 than in1999, the rate of decline has slowed.Use of home health is projected toreturn to its pattern of growth as theeffects of the PPS are more fully felt(CBO 2002). The PPS creates anenvironment that allows providers tocare for costlier, more complex patientswith less financial risk than under theIPS.

Anticipating growth• Incentives. The PPS removes some

of the features of the IPS thatcontributed to the decline in homehealth users. Under PPS, agencies

can maximize margins by keeping costsper episode below the payment and bymaximizing the number of episodesthey provide. The PPS reflects theclinical and functional severity of thepatient in the episode payment; thus anepisode that is likely to be costlyreceives a higher reimbursement thanone for a beneficiary with lowerexpected resource needs. Reflecting theanticipated needs of the patient in thepayment removes the disincentive tocare for patients with costly care needs.

The PPS pays more for patients whoneed therapy (as long as at least 10therapy visits are provided) and formultiple episodes of home health careuse. It also has an outlier policy to payfor the costliest patients. While onecould expect more dramatic changes inuse than have been observed thus far,the new system may require somerefinements and it may take some timefor providers to adapt.

• Eligibility. The “homebound”criteria was loosened by BIPA.Some beneficiaries who would havebeen ineligible due to theirparticipation in religious services oradult day care will now be eligibleto receive the benefit. This couldincrease the number of beneficiariesusing the home health benefit,though the General AccountingOffice estimates that the impact willbe negligible (GAO 2002a).

• Enforcement. The Office ofInspector General continues tomonitor this sector for fraudulentor abusive behavior. Physiciansremain cautious due to what theyperceive to be harsh penalties forimproper home health referrals.Due to the continued diligence, itseems unlikely that inappropriateuse of the benefit will increase. �

In its March 2002 detailed baselineestimate, the Congressional Budget Office(CBO) projected an average annualgrowth for Medicare home health of 17percent from 2002 to 2007. In August, theCBO indicated that they will revise theirMarch estimate downward for homehealth spending because of a new, moremoderate projection of the growth in useof the benefit. CBO’s updated projectionsfor home health services have not yet beenreleased.


Our analysis of current payments andcosts for Medicare home health servicesconcludes that payments are more thanadequate. This conclusion is based on

Estimated spending for home health, 1992–2002FIGURE2D-1

Source: Office of the Actuary, CMS, 2002.

Spen

din

g (

dolla

rs in

bill

ions)

10

20

15

5

019941992 1993 1995 1996 1997 1998 1999 2000 2001 2002

estimates of a substantial, positiveaggregate margin; a high ratio ofpayments to charges for claims for bothurban and rural beneficiaries’ services;evidence of product change; decliningvisit volume; generally good access tocare; and a stable number of providerswith little entry or exit.

Current payments and costsWe used three different approaches toestimate the current relationship ofpayments to efficient provider’s costs.First, we estimated the aggregateMedicare margin using reported costs andpayments from a sample of agencies’ costreports from fiscal year 2001. Next, wecombined reported costs from 1999 withclaims from 2001 and 2002 to estimate theimpact of changes in visit volume oncosts. Combining the first and secondestimates allowed us to project marginsfor the current year. Finally, we usedclaims from 2001 and 2002 to calculatethe ratio of payments to charges fordifferent types of episodes as well as forurban and rural beneficiaries. We alsoreviewed the General Accounting Office’sestimate of payments and costs perepisode.

Together, these estimates show thatcurrent payments are more than adequatewhen compared to costs.

Medicare marginOne method the Commission uses toevaluate the adequacy of current paymentsis calculating the relationship betweenpayments and costs (Table 2D-1). Currentcosts and payments are estimated byupdating the most recent available data.For the home health sector, the mostrecent available cost reports cover fiscalyear 2001 (October 1, 2000 to September31, 2001), the period immediatelyfollowing the implementation of the PPS.

Seven hundred freestanding agencies’ costreports were available; as a sample theyrepresent about 10 percent of all Medicarecertified agencies. These margins do notinclude hospital-based home healthagencies because their cost reports for

fiscal 2001 were not yet available. About30 percent of all agencies were hospital-based in 2000. The sample was notrandom, though it did contain aproportional number of urban and ruralproviders and a proportional number ofproviders by type of control (voluntary,private, and government).

In modeling 2003 payments and costs, weincorporate both policy changes that wentinto effect in 2003 and those scheduled tobe in effect in 2004. For the home healthsector, the 2003 estimate includes theeffect of the so-called “15 percent cut”implemented on October 1, 2002 and theexpiration of the 10 percent rural add-onfor services provided to beneficiariesliving outside metropolitan areas. Thoughthe add-on is not scheduled to expire untilApril 2003, in our estimate we removed itfor all of 2003 to better inform ourdecision regarding the 2004 update.

We estimate that the aggregate financialMedicare margin for all home healthagencies is 23 percent in fiscal year 2003.The estimate of margins in 2003incorporates the increase in the base rateof payment in fiscal year 2002, thedecrease in the base rate due to the “15percent cut” in fiscal year 2003, theeffects of the expiration of the rural add-on, and continuing small declines in thecost of producing an episode of care.

The current estimated Medicare financialmargin of 23 percent suggests thataggregate payments are more thanadequate when compared to costs. Wewere able to measure some variations inmargins two ways: by the total volume ofvisits for each agency and by the urban orrural location of the agency. We calculatethe total number of episodes provided byan agency and divide all the agencies intoone of five equal-sized groups. The 20percent of agencies with the lowestvolume are in the “lowest 20th percentile”group; the 20 percent with the highestvolume are in the “highest 100th

percentile” group, and so on. Allestimated margins are positive; and thehighest percentile group’s margin is fivetimes that of the lowest percentile group.

Our analysis cannot exclude factors otherthan visit volume that could explaindifferences among the margins for theagencies in these percentiles. However, itdoes suggest that visit volume may havean impact on margin. The small size of thecurrent sample—10 percent of allagencies reporting—suggests caution ininterpreting the results we do have andtends to preclude further disaggregation.

Though margins are more than adequatein aggregate, there may be variations inthe experience under PPS among sometypes of agencies. For example, lowermargins for rural agencies suggest thatsome variation in their costs is notaccounted for by the current paymentsystem. Similarly, voluntary agencies thatare likely to be the provider of last resortmay have lower margins. Moreover, theremay be other groups of agencies whosemargins are significantly higher or lowerthan the aggregate margin that we havenot yet been able to identify. Finally, weknow that there is variation in how thebenefit is provided across the country. Ifdistributional issues are present and


Freestanding homehealth Medicare

margin, by agency group,2001 and estimated 2003

Agency group 2001 2003

All agencies 21.9% 23.3%

Urban 22.0 23.9Rural 21.6 19.1

Volume of episodesLowest 20th percentile 5.2 7.540th percentile 7.9 10.260th percentile 14.3 16.580th percentile 16.4 18.5Highest 100th percentile 26.3 28.1

Note: Data for 2001 are preliminary, based on 10percent of all agencies covered by prospectivepayment. Data for 2003 are estimated.

Source: MedPAC analysis of Medicare cost reportdata from CMS.

T A B L E2D-1

persist, it will be difficult for financiallystressed agencies to meet the needs of anaging population (see Chapter 3).

Impact of changes in volume onper unit costsIn our estimate of the current aggregatemargin, we applied an estimate of costchanges rather than assuming that costswould rise at the same rate as input pricesas measured by the market basket. Oursecond analysis of the relationshipbetween payments and costs—designed tomeasure the cost changes associated withdeclining visit volume—determined thatcosts per episode fell by 16 percent from1999 to 2001. Much of the 16 percentdecline occurred before the PPS; thedecline over the course of 2001 was 5percent. Taking into account the steepdecline that preceded the PPS as well asevidence that the decline continued at aslower pace after the PPS, we assumedthat costs fell 2.5 percent each yearbetween 2001 and 2003. We used thisestimate of the changes in costs in ourestimate of margins for 2003 instead ofassuming that costs rose at the same rateas the market basket.

To estimate the change in costs, we beganwith costs2 as reported on home healthagencies’ cost reports for 1999. First, wedivided total costs into fixed and variablecosts. Next, we inflated both by themarket basket for 2000 and 2001. Thenwe applied the inflated variable costs tothe number of visits by type in thebeginning of 2001 and the end of 2001.This allowed us to account for bothchanges in the number of visits as well asthe more costly, higher intensity mix oftherapy and nontherapy visits in 2001compared to 1999. Finally, we addedfixed and variable costs to estimate totalcosts.

We made two assumptions that lead ourmodel to err on the side of producing highcosts per episode and underestimating thedecreases in costs. First we assumed thatfixed costs did not decline as volume

declined but instead rose by the full rate ofincrease in input prices. Second, weassumed that variable costs per visit roseby the full rate of increase in input prices;that is, productivity had no impact oncosts per visit while such influences asrising wages would increase costs. Acaveat is warranted: this estimate can notaccount for changes in the visit itself—such as activities performed during a visit,supplies used, or the length of the visit—that may have had an impact on costs pervisit beyond changes in input prices.

Ratio of payments to chargesIn addition to our estimate of theaggregate margins, we used claims fromall of calendar year 2001 and the first sixmonths of 2002 to calculate the ratio ofaggregate payments to charges. This ratiois not the typical ratio of payments tocosts that MedPAC uses in other sectors.However, we believe it is illuminatingbecause it allows us to use very recentdata, to look at different episode types,and to compare urban and ruralbeneficiaries.

From this analysis we concluded that theratio of payments to charges was greater

than 1.0 in the beginning of 2001 and wasstill rising by the middle of 2002 (Table2D-2).

This ratio of payments to charges impliesthat the program currently pays more inthe aggregate for services than it wouldhave been charged under the previoussystem of charges per visit by visit type.

The ratio reinforces the conclusion thatpayments are more than adequatecompared to costs. To arrive at thisconclusion, we made two assumptions.First, we assumed that charges are as highor higher than costs. Basic economicswould suggest that this is usually true.Second, we assumed that current chargesare accurate. Under the cost-based system,Medicare paid agencies the lesser of theirreasonable costs or customary charges.Thus, there was a strong incentive to setcharges higher than costs. At that time, theratio of payments to charges was about0.73.3 The current payment to chargeratios for low utilization paymentadjustment (LUPA) episodes—whereinservices are paid per visit by visit type—isalmost the same as the ratios under thecost-based payment system when


2 Costs included visits, supplies, and outpatient therapy provided to home health users.

3 Under the cost-based system, the ratios were 0.74 in 1994 (Leon et al. 1997) and 0.73 in 1997 (HCFA 2000).

Ratio of payments to charges, by type of homehealth episode, 2001 and 2002

January– July– January–June 2001 December 2001 June 2002

All episodes 1.03 1.09 1.12

Urban 1.02 1.08 1.11Rural 1.04 1.12 1.16

Episodes with four orfewer visits 0.76 0.77 0.76

Outlier episodes 0.47 0.46 0.48

Note: Urban episodes include services delivered to beneficiaries who reside within a metropolitan statistical area(MSA). Rural episodes include services provided to beneficiaries who reside outside an MSA. Episodes withfour or fewer visits are paid per visit by visit type, rather than by the episode; this is the low-utilization paymentadjustment.

Source: MedPAC analysis of the 5 percent Standard Analytic File of home health claims from CMS.

T A B L E2D-2

incentives to set charges higher than costswere in place (Table 2D-2).

These aggregate charges included chargesfor visits, medical supplies, and drugsused during the episode of care. Wecalculated the ratio for full episodes aswell as high-cost outlier episodes,episodes that include a beneficiary’ssignificant change in condition andreclassification, and episodes with four orfewer visits that are paid by the visit.

We compared claims for servicesprovided to urban and rural beneficiaries.In each of the three periods, the rural ratiowas higher than the urban one. Forexample, in the first six months of 2002,agencies were paid $1.11 for each dollarin charges for services provided to urbanbeneficiaries, while agencies were paid$1.16 for each dollar in charges forservices provided to rural beneficiaries. Inthe latter two periods, the impact of the 10percent add-on for services provided tobeneficiaries living in rural areas wasevident. If the add-on were not in effect,the rural ratio would still have beengreater than one, and greater than theurban ratio. The relationship betweenurban and rural ratios was the same evenwhen we distinguish rural beneficiaries bytypes of rural areas.

General Accounting Office’sanalysis This past summer, the GeneralAccounting Office (GAO) also examined

CMS claims data. They estimated that theaverage episode payment of $2,700 was$700 above the average episode cost in2001, an overpayment of about 35percent. To create the estimate, GAObegan with CMS’s estimated visit costs byvisit type for 1999 (based upon an auditedsample of 1997 cost reports). GAO usedthe home health market basket to inflatecosts to 2001. To estimate episode costs,they used half a year of home healthclaims (January to June 2001) to calculatethe average number and type of visits ineach type of episode and multiplied theestimated visit costs by those averages.GAO concluded that the magnitude of thedisparity between payments and estimatedcosts demonstrated that a reduction inpayment rates—such as theimplementation of the “15 percent cut”—would not harm the industry.

Appropriateness of current costsMedicare home health services havechanged consistently with theimplementation of the PPS. Theprevailing mode of Medicare home healthcare post-PPS is changing from themaintenance of consistently ill or disabledpeople over time at low intensity torecovery from an acute illness or injuryover a short period of time with aconcentration on therapy. The changebegan in 1997 and continued with theimplementation of the PPS in 2000. Dueto this change, payments may no longer

be in line with costs because currentpayments are based on previouslymeasured costs of production.

There are two caveats to using the averagenumber of visits per episode as anindicator of product change. First, thedecline in the number of visits per episodehas not been similar from state to state.State by state average visits per episodevary widely. Although all states’ averageshave declined since 1997, the averagenumber of visits per episode in somestates remains high. In the first six monthsof 2001, home health users in WashingtonState received 13 visits per episode whilethose in Utah received 28 (GAO 2002b).Heavy use in some states pulls thenational average well above the mediannumber of visits per episode (Table 2D-3).

Second, counting the number of visitsdoes not give us complete informationabout the amount of time that nurses,therapists, and others are spending in theirpatients’ homes during a visit. If the timespent per visit is changing along with thenumber of visits per episode, thenmeasuring the number of visits may fail tocapture real changes in the amount ofservice beneficiaries receive.

Declines in the average number of visitsper episode are one indicator that theproduct may be changing. In 1997, homehealth users, on average, received 36 visitsin 60 days. In 1999 that number droppedto 29 visits. Over the course of the mostrecent year and a half, the average numberof visits per 60-day episode has continuedto decline at a slower rate than before thePPS, from 22 to 20 (Table 2D-3).

Another indication of the changingproduct is the dramatic decline in theaverage length of stay (LOS) of homehealth patients. The LOS measures thenumber of days between the daybeneficiaries receive their first homehealth visit and the day upon which theyare discharged from treatment.4 Unlikepatients in other settings (e.g., acute carehospitals or skilled nursing facilities),home health patients rarely receive visits


Use of home health after the PPS

January– July– January–June 2001 December 2001 June 2002

Average visits per episode 22 21 20Median visits per episode 16 15 15Average length of stay (days) 46 47 44

Note: PPS (prospective payment system). Excludes episodes subject to the low utilization payment adjustment (LUPA)that contain four or fewer visits and are reimbursed differently from regular episodes. Beneficiaries’ length ofstay may span several episodes.

Source: MedPAC analysis of the 5 percent Standard Analytic File of home health claims from CMS.

T A B L E2D-3

4 Under the PPS, a beneficiary may receive multiple 60-day episodes of home health services, as long as they remain eligible for the benefit. Thus, a single stay is theamount of time between the start of care and discharge; it may be one 60-day payment episode or several payment episodes.

on every day during their stay; and onsome days patients may receive more thanone visit. However, the home health LOSmeasures the duration of the observation,evaluation, and treatment of the patient’scondition, even though the visits areintermittent. In 1997, the LOS was 106days; by 1999, that number had fallen to69 (McCall et al. 2001). In the first sixmonths of 2002, the average length of stayfor a Medicare beneficiary was 44 days(Table 2D-3). When episodes that contain4 or fewer visits are included in the LOScalculation, the latest LOS falls further to41 days, less than half the duration of careonly 2 years earlier.

The mix of visit types has also beenchanging. As Table 2D-4 indicates, homehealth care under the PPS after October2000 has a greater concentration oftherapy compared with the paymentsystems that preceded the PPS. In 1997,the prevailing pattern was more typical ofmaintaining consistently ill or disabledpersons in their homes over a long periodof time, with much of the service providedby home health aides.

One aspect of home health services thatsurprisingly has not changed under thePPS is the provision of very short durationcare. Because of strong incentives in thepayment system, it was predicted thatepisodes of care consisting of four orfewer visits (LUPAs or low utilization

payment adjustments) would dwindleunder prospective payment. HHAs thatmake at least five visits qualify for anepisode payment and avoid the LUPA;even the highest LUPA payments aremuch lower than the lowest episodepayment. In 1997, these very shortepisodes comprised about 15 percent ofall episodes. In its construction of the newpayment system, CMS predicted that theproportion of very short episodes wouldfall to 5 percent (CMS 2000). However,our analysis of claims in 2001 indicatesthat 14 percent of all episodes for that yearhad four or fewer visits.

This section has discussed three homehealth indicators that suggest that thehome health product is changing in thewake of the implementation of the PPSand one indicator that (surprisingly) hasnot changed. The average number of visitsper episode and the LOS have declined.The mix of visits by type has shiftedtoward therapy and away from homehealth aide services. However, theincidence of LUPA episodes, despite theincentives in the payment system to avoidthem, has remained about the same. Thepersistence of LUPA episodes suggeststhat one widely anticipated behavioralresponse to the PPS has not yet occurred.Otherwise, HHAs have responded to theincentives of the new payment system.

Relationship of payments to costsOur analysis indicates that home healthagencies are paid more than adequatelyunder the PPS, even after accounting forthe impact of the 7 percent paymentreduction (the “15 percent cut”). Indeed,aggregate margins under the home healthPPS are higher than those we estimatedfor any other sector in Medicare. Also wedo not observe measurable reductions inthe quality of care—although data on thispoint are limited. Other market factorsalso indicate that payments are at leastadequate compared to costs.

Changes in volume The volume of home services in terms ofthe total number of visits provided hascontinued its post-1997 decline because adrop in the number of users hascompounded the decrease in the averagenumber of visits per user.5 In 1997, 3.3million beneficiaries used home healthservices during the year. By 1999, thatnumber had fallen to 2.5 million (McCallet al. 2001). Following theimplementation of the PPS, the number ofusers has continued to decline. Ouranalysis of CMS’s claims databaseidentified 2.2 million beneficiaries usinghome health care in 2001.

Many factors explain both the increaseand the decrease. Examples include thelevel of fraud and abuse oversight; thestringency of eligibility and medicalnecessity criteria; and the incentives of theprevailing payment systems. To the extentthat users left the system as a result offraud and abuse oversight; tighterapplications of eligibility and medicalnecessity requirements; and theelimination of payment incentives thatrewarded the inefficient use of services,reductions in the numbers of users may bewarranted. However, to the extent thatusers who qualify for the benefit cannotaccess home health services, declines inthe number of users are cause for concern.


Share of visits per home health episode, by type of visit

Pre-PPSPost-

Type of visit 1997 1998 1999 PPS

Therapy 9% 11% 15% 23%Home health aide 49 42 35 27Skilled nurse 41 45 48 49

Note: The prospective payment system (PPS) began in October 2000. “Post-PPS” refers to October 2000 throughSeptember 2001. Columns do not sum to 100 percent because data were not available for all visit types.

Source: CMS analysis of the national claims history file.

T A B L E2D-4

5 Estimates of use are based on fee-for-service claims and do not include Medicare�Choice enrollees.

Under the PPS, there are mechanisms thatshould encourage agencies to take highcomplexity patients—a case mix measurethat adjusts payments based oncomplexity, multiple episodes if patientsneed extended care, and an outlierpayment mechanism for high-costpatients. Nonetheless, the number of userscontinues to decline. These declines haveoccurred even though demographic andclinical indicators would lead us to expectan increase in home health use; in fact,estimators have repeatedly predictedannual increases in utilization. MedPACplans to extend its current analysis of costand use data to explore the variation inagencies’ experiences and the impact ofthe payment system. Additionally, wenote that CMS has plans to refine the PPSand to that end is:

• developing a database of claimsassociated with the start-of-care anddischarge OASIS assessments so thatoutcomes and utilization can belinked,

• developing a tool for medical reviewof claims to detect evidence ofstinting,

• providing case mix and adverse eventreports to agencies so that they canmonitor their processes and outcomesat the individual patient level, and

• planning to report quality informationto home health care consumers.

With respect to elements of the paymentsystem, CMS is looking into:

• the therapy threshold,

• the structure of the outlier paymentmechanism, and

• refinements to the case mix system.

The Commission strongly supports thisresearch and looks forward to its timelycompletion so that it can be considered indeveloping refinements to the PPS. Thepayment system should be amended to

accurately capture the costs of an efficientprovider.

Quality of careThe OASIS provides some evidence thatthe product changes in home healthfollowing the PPS have not had adetrimental effect on the quality of care.OASIS measures patients’ clinicalseverity and functional limitations at thebeginning and end of an episode of homehealth care. It allows HHAs to track theirpatients’ outcomes and to change their useof resources, care planning, or otherprocesses to improve their services. CMSalso uses OASIS to produce reports foragencies’ own quality improvementefforts and plans to publish OASIS-basedquality information to guide consumers tochoose high quality providers.

The decline in volume of visits perepisode has prompted many to questionthe impact of low volume on the qualityof care. Many studies have found that therelationship between volume and qualityis weak (Bishop et al. 1999, Fortinsky andMadigan 1997, Penrod et al. 1998, Welchet al. 1996). However, one study of a ruralpopulation before the implementation ofthe PPS found a correlation between verylow visit volume and quality (Schlenker etal. 2002). After adjusting for case mix andagency differences, the study indicatedthat rural home health users met the goalsof their care less frequently thancomparable urban home health users.CMS is testing a system of standards torelate outcomes for common diagnosesand functional limitations to visit volume(HCFA 2001).

Relating visit volume to quality presentstwo challenges: the home health visitremains something of a “black box,” andit is difficult to measure other sources ofcare, especially informal care, that areavailable to patients at home. First, unlikethe coding system for physician services,for example, home health claims data donot differentiate visits by purpose, e.g.,

evaluation or follow-up, teaching, ormedical procedure. Without informationon the content of the visit, it is verydifficult to relate available measures of thenumber of visits to the quality ofoutcomes. Second, unlike institutionalsettings, patients at home may have othersources of care that can have a significantimpact on the outcomes of care. Onestudy that failed to find a correlationbetween Medicare home health use andoutcomes (Penrod 1998) did find acorrelation between greater use ofinformal care and better outcomes.

An index based upon patients’ scores onthe home health outcomes assessment toolsuggests that quality has not declined overthe first year of the PPS (OutcomeConcept Systems 2002). The indexcaptures improvement, decline, orstabilization in the patients’ ability toperform activities of daily living and theseverity of their clinical condition,measured by scores on the OASIS at thestart of care and again at the end of care.Between the final three months of 2000and the final three months of 2001, themedian score had not moved significantlyup or down.6

The stability of this quality index providessome evidence that quality has notdeclined under the PPS despite the declinein the volume of visits and thecorresponding decrease in costs perepisode. This reinforces our conclusionthat home health agencies have improvedtheir productivity and current costs areappropriate. However, our analysis cannotdismiss the possibility that the patientpopulation has changed; consistent qualityat lower visit volume could also beachieved by serving a less-complex mixof patients.

Entry and exit of providersAs of October 1, 2002, about 7,000Medicare certified home health agencieswere serving beneficiaries. Following adecline of about 3,000 agencies between


6 This index was developed by researchers at Outcome Concept Systems, Inc., a private firm that collects data from 700 Medicare-certified HHAs to benchmark theirperformance. The index was developed by a team of statisticians, researchers, and clinicians. The index was based upon 350,000 patient episodes of home healthcare. Participating agencies include a cross-section of sector, geographic area, and type of control (voluntary, proprietary, and others).

1998 and 2000, this number has beensteady over the past several years (Figure2D-2).

The limited exit of home health agenciesover the past three years may suggest thatmost agencies’ payments are equal to orgreater than their costs. In 1996, under thecost-based payment system, about threenew agencies entered for each exitingagency. During 1999 under the IPS,exiting agencies outnumbered enteringones 8 to 1. Between October 2001 andOctober 2002, a little over 300 agenciesentered the program while nearly 200exited; the near-equilibrium of entry andexit led to almost no change in the totalnumber of agencies.

Entry and exit may be sensitive to less-than-adequate payments while notproviding information about overadequate payments. Exits from theprogram seem to correspond to theimplementation of the IPS, though someof those exits were involuntary. Agenciesthat involuntarily exited the program wereunable to meet one or some of the

program’s integrity standards and mayhave left the program due to OperationRestore Trust’s activities rather than theIPS. Some entries to the program mayhave been prevented or delayed by stateregulations that limit the number ofparticipating agencies. Comparing entrypre- and post-PPS may be misleadingbecause the structure of the PPS mayfavor larger agencies with the ability toaverage profit and loss over a large andvaried patient population. Also, thoughhome health is not a capital-intensivesector, starting a home health agency maybe more expensive than it was in the pastdue to tighter financial standards andgreater need for computerization tomanage the patient data collectionrequirements implemented in 1999.

A reduction in the number of Medicare-certified agencies does not necessarilyindicate a reduction in home health carecapacity. Some observers have suggestedthat having only a small number ofagencies per Medicare beneficiary in anarea may impair access, but no evidenceexists to suggest that the number of

agencies is a meaningful measure ofaccess. GAO found that neither closuresnor changes in practice patterns wereindicative of access problems (GAO1999). In fact, “In those counties that losttheir only HHA, hospital dischargeplanner supervisors as well as managersof nearby HHAs [reported] that access isnot a problem because services areavailable from HHAs in neighboringcounties or from branch offices located inthe county” (GAO 1999, p. 20).Furthermore, because the home healthindustry has been experiencing acquisitionand consolidation, the agencies stillparticipating in Medicare may be largerthan their predecessors.

Beneficiaries’ access to careThis year, our analysis of access has theadvantage of using very recentinformation, but also has twodisadvantages. First, the nationallyrepresentative, focused work of the Officeof Inspector General on access to homehealth care for Medicare beneficiaries thatwe have used in the past is not availablethis year. Also, neither we nor they


12,000

10,000

8,000

6,000

4,000

2,000

01996 1997 1998 1999 2000 2001 2002

Num

ber

of

agen

cies

Certified home health agencies, 1996–2002FIGURE2D-2

Source: MedPAC analysis of Online Survey, Certification, and Reporting (OSCAR) system data from CMS.

currently have adequate means to assessbeneficiaries’ access to home health carewithout a preceding hospital stay.

MedPAC is developing resources toprovide more information on access tocare. Our episode database will be able totrack patterns and changes in home healthuse by beneficiaries referred from thecommunity or from a skilled nursingfacility. The OIG’s work, or a regularstudy with a similar methodology andsample, would continue to be an importantparallel effort to MedPAC’s accessmonitoring because a consistent series ofstudies spanning the start of the PPSprovides a crucial baseline andcomparisons over time.

R E C O M M E N D A T I O N 2 D - 1

The Secretary should continue aseries of nationally representativestudies on access to home healthservices (similar to studies previouslyconducted by the Department ofHealth and Human Services’ Office ofInspector General).

I M P L I C A T I O N S 2 D - 1

Spending• This recommendation should not

affect Medicare benefit spending.

Beneficiary and provider• To the extent that future OIG studies

allow us to monitor beneficiaries’access to home health care, theCommission may makerecommendations to preserve orimprove their access to care.

One year ago, the OIG found thatbeneficiaries continue to maintain goodaccess to care (OIG 2001a, OIG 2001b),suggesting that payments are at leastadequate to induce agencies to serveMedicare beneficiaries. The OIG surveyedhospital and nursing home dischargeplanners in early 2001, after the PPS hadbeen in place for about six months. Mostdischarge planners reported placingbeneficiaries in home care withoutdifficulty. Of the few planners whoreported difficulties, most were unable to

place only a small fraction of dischargedbeneficiaries.

MedPAC convened a panel of hospitaldischarge planners in October to continueto monitor patients’ access to home healthcare. Generally, they offered no evidenceof increased difficulties with placing mostpatients in home health care since theimplementation of the PPS in October2000.

The discharge planners did experiencesome difficulty—ranging from a one-daydelay in placement to no servicesavailable—with a few patients in certainsubgroups. They told us that services aremore difficult to access in rural areas,especially if therapy is needed, and thatsince the implementation of PPS homehealth agencies are substituting physicaltherapy visits for occupational therapy,limiting social work visits, and providingfewer services for training diabetics inself-care. Patients requiring wound care,daily care, or expensive medication orsupplies were among those more difficultto place, as were patients with mentalillness or cognitive impairment. Membersof the panel did not indicate which, if any,of the hard-to-place subgroups werenewly difficult to place or more difficultto place in home health care following theimplementation of the PPS. They also didnot conclude that the lack of prompt homehealth placement necessarily led toclinically inappropriate care for patients.

Home health in rural areasFor most rural agencies, payments willmore than adequately cover costs in 2003.The Medicare margin for all ruralagencies in 2003 was 19.1, nearly thesame as the margin for urban agencies,even accounting for the sunset of the ruraladd-on in April 2003. However,examining agencies in more or lessdensely populated rural areas reveals awide variation in the experience of ruralagencies under the PPS; some ruralagencies have low margins.

At this point in time, our analysis cannotexplain the variation among ruralproviders—low margins are not explained

by what we know about volume orownership of the agencies in the group.The very low margin group had aproportionate share of voluntary, private,and other types of control agencies. Thesample had somewhat more low volumeproviders and fewer high volumeproviders than the entire sample generally;but the group also contained several veryhigh volume providers. The sample of lowmargin rural providers was notgeographically representative due tolimitations of the sample. Costs perpatient could be higher in rural areas thanin urban because of the small scale ofoperations, the distances to travel amongrural clients, and differences in the use oftherapy.

The difference between the ratio ofpayments to charges for urban and ruralbeneficiaries suggests that specialtreatment of beneficiaries in rural areas isnot necessary. As discussed earlier, claimsfor services provided to all ruralbeneficiaries, as well as claims groupedby the rural characteristics of thebeneficiaries’ county of residence, showthat payments are higher than charges by agreater ratio than they are for urbanbeneficiaries’ services.

Two access indicators provide mixedevidence for the special treatment of ruralareas. In 2001, the OIG found thatdischarge planners at urban and ruralhospitals were able to place Medicarebeneficiaries in home health at similarrates (OIG 2001a). However, in our panelof discharge planners, five of the fifteenpanelists had observed hospitals takingspecial measures to provide ruralbeneficiaries with home care. They wereaware of hospitals that rented hotel roomsand owned apartments in metropolitanareas to temporarily house ruralbeneficiaries who could not accessservices at their homes. The panel’sperceptions may have differed somewhatfrom the OIG’s because the panel’s much-smaller sample of discharge planners maybe less representative of dischargeplanners generally and rural hospitalswere overrepresented on our panel.


In summary, our analysis cannot dispelconcerns about some rural providers. Ouranalysis of payment-to-charge ratios (witha large sample of recent data) tends tosuggest that payments for the care of ruralbeneficiaries are adequate. However,variations among margins for some ruralagencies and the observations of somemembers of the discharge planners’ panelcontradict this conclusion and suggest thatadditional payments for care provided torural beneficiaries are appropriate.


The Congress should extend for oneyear add-on payments at 5 percentfor home health services provided toMedicare beneficiaries who live inrural areas.


Spending• This would increase spending

compared to current law between $50million and $200 million for fiscalyear 2004 and less than $1 billionover five years. The current add-on of10 percent is scheduled to expire onApril 1, 2003.

Beneficiary and provider• There is concern that payments under

the PPS may not be appropriatelydistributed for some rural providers.Temporarily extending the add-onwill provide some time for additionaldata and analysis to explore thevariation. The lower amount of theadd-on acknowledges, however, thatthe margins of rural providers are notvery different from the aggregatemargins of home health agencies as awhole.

Adjustments to currentpaymentsThree adjustments are relevant topayments for fiscal 2003: a 7 percentreduction in the base episode rate forfiscal year 2003 (“15 percent cut”), anupdate, and a rural payment provision.

The Balanced Budget Act of 1997 set inmotion many changes for the home healthsector, including the replacement of thecost-based payment system with the IPS,and a contingency for a 15 percentreduction in the payment limits under theIPS system if CMS did not replace the IPSwith a PPS. When the PPS did replace theIPS in October 2000, the reduction in theIPS limits was postponed rather thaneliminated. When this cut wasimplemented on October 2002 under thePPS, CMS had to model the effect that a15 percent reduction in IPS limits wouldhave had, build in assumed behavioralchanges by HHAs, and project the effectonto current spending. Due largely to thebehavioral assumptions in the model,CMS estimated that a 7 percent reductionin PPS rates would be needed to achievethe reduction anticipated in the originallegislation.

In addition to this reduction, rates forFY2003 were also adjusted by a marketbasket update. The legislated update wasthe percent change in the market basketminus 1.1 percent; the change in themarket basket was 3.2 percent, so the baserate was increased by 2.1 percent. Thus,the net effect of the 7 percent reductionand the update was a 5 percent reductionin the base rate for an episode, to $2,160for FY2003.

After the decreases in the number of homehealth users and providers in the late1990s, concerns about access to homehealth services in rural areas led theCongress to provide an additional 10percent payment for home health servicesprovided to beneficiaries living in ruralareas.7 This addition is scheduled toexpire in April 2003. Our model ofcurrent payments and costs (fiscal year2003) incorporates the expiration of theadd-on. To be conservative, the modelincorporates the effects as if the add-onwere unavailable for the entire fiscal yearrather than only half of the fiscal year.

Accounting for providers’cost changes in thecoming year

In addition to accounting for the adequacyof current payments, a payment updateshould account for changes in costs in thecoming year. Because the home healthproduct has changed, we have notadjusted for changes in productivity or theimpact of scientific and technologicaladvances in projecting next year’s costchanges. Our estimate of the impact ofvisit volume on costs per episode (seediscussion p. 108) suggests that costs willcontinue to decline over the coming year.

Home health, perhaps more so than othersectors, may feel the impact of a shortageof nurses or therapists because a largeportion of its total costs are for labor. Themarket basket weights reflect this laborshare; labor is 80 percent of home healthinput costs, compared to 60 percent inhospitals or 70 percent for physicianservices. The market basket for homehealth uses the same proxies for theimpact of changing wages, salaries, andbenefits used by the hospital sector.Within the update framework, we assumethat the market basket captures changes ininput prices, such as those created by anursing shortage. At this time, we have noevidence to suggest that home health laborcosts increased faster than the input pricesin the market basket.

Although home health agencies are likelyto face increasing input prices during thecoming year, we expect a decline in thecosts per episode because continuingdeclines in the number of visits perepisode will offset the effects of risingprices. We conclude that neither a positivenor a negative adjustment should be madeto the update to account for cost changesover the coming year.


7 Under the legislation, rural beneficiaries are those who reside outside a metropolitan statistical area.


To summarize, MedPAC has consideredthe update framework in the currentcontext for home health paymentdecisions. We considered the currentrelationship of payments and costs.Aggregate Medicare margins and the ratioof payments to charges suggest thatcurrent payments are more than adequatecompared to costs. Market factors suggestthat current payments are at least adequatein relation to costs: access to care isgenerally good, the rate of decline in thenumber of users has decreased, and theentry and exit of agencies has remainedstable for the third year in a row.

When we considered likely changes incost over the coming year we found thatthe chief influences over costs will be theprice of labor and the volume of visitswithin an episode. These influences willwork in opposite directions: prices willprovide upward pressures on costs whiledeclining visit volume will depress costs.These factors provide evidence that

payments will continue to be more thanadequate over the coming year.

In our March 2002 recommendations, wehandled the home health payment updatedifferently. This was because, at this timelast year, no cost report data from the PPSwere available. We did not have sufficientclaims data to estimate whether decreasesin visit volume would continue under thePPS or information on changes in qualityto assess the impact of lower volume oncare. Though market factors weregenerally positive, the Commission erredon the side of caution. Sensitive to thedramatic changes that had preceded thePPS, we recommended a year of stability.Over the course of the past year, nounforseen changes have been made toMedicare’s home health benefit and timehas allowed data to become available.


The Congress should eliminate theupdate to payment rates for homehealth services for fiscal year 2004.


Spending• Since current law provides a full

market basket update for the basepayment for home health services,this recommendation would decreasespending relative to current lawbetween $200 million and $600million for fiscal year 2004 andbetween $1 billion and $5 billionover 5 years.

Beneficiary and provider• Because we estimate that current

Medicare payments are well over thecosts of caring for Medicare homehealth users, and evidence suggeststhat the level of payment is only oneof several influences on the use of thehome health benefit, we wouldexpect little if any effect of thisprovision on beneficiaries’ access tocare. �


References

Bishop C, Kerwin J, Wallack S. The Medicare home health benefit: implications ofrecent payment changes, Care Management Journals. Summer 1999, Vol. 1, No. 3.

Centers for Medicare & Medicaid Services, Department of Health and Human Services.Home health services prospective payment system, final rule, Federal Register. July 3,2000, Vol. 65, No. 128, p. 41128–41214.

Centers for Medicare & Medicaid Services. Health insurance for the aged: home healthagency manual, Section 205.1. December 2001.

Congressional Budget Office. The budget and economic outlook: an update. Washington(DC), CBO. August 2002.

Fortinsky RH, Madigan EA. Home care resource consumption and patient outcomes:what are the relationships?, Home Health Care Services Quarterly. Cleveland (OH), CaseWestern Reserve University School of Medicine. 1997, Vol. 16, No. 3, p. 55–73.

General Accounting Office. Medicare home health agencies closures continue, with littleevidence beneficiary access is impaired, No. HEHS–99–120. Washington (DC), GAO.May 1999.

General Accounting Office. Medicare home health: clarifying the homebound definitionis likely to have little effect on costs and access. Washington (DC). April 26, 2002a.

General Accounting Office. Medicare home health care: payments to home healthagencies are considerably higher than costs, No. HEHS–02–663. Washington (DC),GAO. May 2002b.

Health Care Financing Administration. Active projects report research anddemonstrations in health care financing. Baltimore (MD), HCFA. 2001.

Health Care Financing Administration. Healthcare chartbook. Baltimore (MD), HCFA.2000.

Komisar H, Feder J. The Balanced Budget Act of 1997: effect on Medicare’s home healthbenefit and beneficiaries who need long-term care. Washington (DC), Institute for HealthCare Research and Policy, Georgetown University. February 1998.

Leon J, Neuman P, Parente S. Understanding the growth in Medicare’s home healthexpenditures. Bethesda (MD), The Project HOPE Center for Health Affairs. June 1997.

Medicare Payment Advisory Commission. Report to Congress: context for a changingMedicare program. Washington (DC), MedPAC. June 1998.

McCall N, Komisar H, Petersons A, Moore S. Medicare home health before and after theBBA, Health Affairs. May/June 2001, Vol. 20, No. 3, p. 189–198.

Office of Inspector General, Department of Health and Human Servies. Access to homehealth care after hospital discharge 2001, No. OEI 02–01–00180. Washington (DC), OIG.July 2001a.


Office of Inspector General, Department of Health and Human Services. Medicare homehealth care community beneficiaries 2001, No. OEI 02–01–00070. Washington (DC),OIG. October 2001b.

Outcome Concept Systems, Inc. PPS & patient outcomes: a year in review. Seattle (WA),Outcome Concept Systems, Inc. 2002.

Penrod JD, Kane RL, Finch MD, Kane RA. Effects of post-hospital Medicare homehealth and informal care on patient functional status, Health Services Research. August1998, Vol. 33, p. 513–529.

Schlenker RE, Powell MC, Goodrich GK. Rural-urban home health care differencesbefore the Balanced Budget Act of 1997, Journal of Rural Health. Spring 2002, Vol. 18,No. 2, p. 359–72.

Stoner D, Goldberg HB, McCallum-Keeler G, Robinson C. Medicare Payment AdvisoryCommission (MedPAC) home health agency survey, 1999. Cambridge (MA), AbtAssociates. 1999.

Welch H, Wennberg DE, Welch WP. The use of Medicare home health care services,New England Journal of Medicine. 1996, Vol. 335, p. 324–329.


2EAssessing payment adequacy

and updating paymentsfor outpatient dialysis services

S E C T I O N


The Congress should update the composite rate payment by the projected change in input prices,less 0.9 percent, for calendar year 2004.



Section 2E: Assessing paymentadequacy and updating payments foroutpatient dialysis services

Current aggregate Medicare payments for outpatient dialysis services appear to

be adequate. Together, payments for composite rate services and injectable drugs

exceeded providers’ costs by about four percentage points in 2001. We conser-

vatively estimate that the aggregate payment-to-cost ratio will be no lower than

1.01 in 2003. However, aggregate payments relative to costs will probably de-

cline by less than three percentage points between 2001 and 2003 because pay-

ments for injectable drugs and their profitability relative to composite rate ser-

vices will continue to increase during this period. Market conditions—such as

continued entry of for-profit freestanding providers, increases in the volume of

services provided, lack of evidence of beneficiaries facing systematic problems

in accessing care, continued improvements in the quality of dialysis care, and ad-

equate access in providers’ access to capital—strongly suggest that Medicare’s

outpatient dialysis payments are adequate, relative to efficient providers’ costs.

Based on this evidence, we see no need to adjust the base rate for composite rate

services. To account for changes in providers’ costs in the coming year, the

Congress should update the composite rate for outpatient dialysis services by the

change in input prices, currently estimated at 2.5 percent, less an 0.9 percent ad-

justment for growth in multifactor productivity, for calendar year 2004.

2EIn this section





S E C T I O N

End-stage renal disease (ESRD) is achronic illness characterized bypermanent kidney failure. Occurring at thelast stage of progressive impairment ofkidney function, the illness is caused by anumber of conditions including diabetes,hypertension, glomerulonephritis, andcystic kidney disease. Persons with ESRDrequire either chronic dialysis or a kidneytransplant to maintain life. Because of thelimited number of organs available fortransplantation, the majority of ESRDpatients receive chronic dialysis. The 1972amendments to the Social Security Actextended Medicare benefits to people withESRD, and more than 350,000 patientswere enrolled in 2001.1

Medicare pays dialysis providers aprospective payment—the composite

rate—for each dialysis treatment theyprovide in dialysis facilities (in-center) orin patients’ homes.2 The averagecomposite rate in 2002 was about $130 forfreestanding facilities. Providers receive aseparate payment for furnishing certaininjectable drugs during dialysis. TheCongress has set the payment forerythropoietin, the costliest of these drugsin terms of spending by Medicare andbeneficiaries, at $10 per 1,000 unitswhether it is administered in dialysisfacilities or in patients’ homes. Providersreceive 95 percent of the averagewholesale price (AWP) for separatelybillable injectable medications other thanerythropoietin administered during in-center dialysis. Medicare’s payments forinjectable drugs averaged about $80 perdialysis treatment in 2001.

Medicare spending for outpatient dialysisservices furnished by freestandingfacilities increased by about 10 percentper year between 1991 and 2001 (Figure2E-1).3 Two factors that contribute to thegrowth in Medicare spending are theincreasing size of the ESRD populationand the diffusion of new technologies.

• Incident rates per million populationhave been increasing steadily since1980 (United States Renal DataSystem [USRDS] 2002). Forexample, the number of new ESRDpatients increased by about 7 percentannually between 1992 and 2000.Increasing incident rates have beenlinked to improvements in survival,as well as increases in the number of

122 As s e s s i ng paymen t adequacy and upda t i ng paymen t s f o r o u t pa t i e n t d i a l y s i s s e r v i c e s

1 To qualify for the ESRD program, individuals must be fully or currently insured under the Social Security or Railroad Retirement programs, entitled to monthly benefitsunder the Social Security or Railroad Retirement programs, or the spouse or dependent child of an eligible beneficiary.

2 The composite rate was designed in 1983 to include all nursing services, supplies, equipment, and drugs associated with a single dialysis session.

3 Medicare spending includes program outlays and beneficiary cost-sharing.

Medicare spending for outpatient dialysis services furnished by freestandingdialysis facilities, 1991–2001

FIGURE2E-1


Spen

din

g (

dolla

rs in

bill

ions)

3

7

4

2

1

01991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

5

6

people with diabetes, which is a riskfactor for ESRD.

• New technologies—particularlyinjectable drugs, such aserythropoietin, iron supplements, andvitamin D analogues that were notavailable when the outpatient dialysispayment system was implemented in1983—have also increasedMedicare’s spending for dialysisservices. MedPAC estimates thatspending for injectable drugsincreased from $1.3 billion in 1998 to$2.3 billion in 2001.

The growth in spending for all Medicare-covered services for ESRD patients hasincreased from about $10 billion in 1994to more than $15 billion in 2001. BecauseMedicare has kept the nominal price forcomposite rate services essentially fixedsince the inception of the payment rate in1983, spending for other services—particularly inpatient hospital services andcare for vascular access complications andother chronic conditions (e.g., diabetes)—has significantly contributed to the growthin total spending.4 Thus, it is importantalso to consider these services whenthinking about ways to improve thequality of care and to control totalspending for ESRD patients.

In addition, the growth in spending hasbeen fueled by the increase in the numberof people in the two most costly ESRDcohorts: (1) older beneficiaries, and (2)beneficiaries with multiple chroniccomorbidities such as diabetes,hypertension, and congestive heart failure.The proportion of new ESRD patientswho are 75 years and older grew from 18percent in 1991 to about 25 percent in2001; the proportion of new ESRDpatients with diabetes grew from 36percent of all new patients to 46 percent inthe same period. Both of these cohorts areheavy users of the health care system. TheUSRDS found that total payments were23 percent higher for older ESRDbeneficiaries (75 years and older) than foryounger beneficiaries (0 to 19 years ofage). They also found that total Medicare

payments were 18 percent higher fordialysis beneficiaries with renal failurecaused by diabetes than for beneficiarieswithout diabetes (USRDS 2002).


The first question in applying MedPAC’sapproach to updating payments is whetherthe current level of Medicare’s paymentsfor outpatient dialysis services is at leastadequate. The Commission answers thisquestion by assessing aggregate Medicarepayments and costs for both dialysisservices and injectable medicationsadministered during dialysis treatment forwhich providers receive separatepayments from Medicare. Our assessmentincludes the payments and costs forinjectable medications because their usehas increased significantly throughout the1990s and their effect on the financialperformance of dialysis providers issignificant. Including payments and costsfor separately billable medications gives amore accurate picture of the financialperformance of dialysis providers.

MedPAC concludes that total paymentsfor outpatient dialysis services will beadequate in 2003 and that no adjustmentfor payment adequacy is needed as part ofthe 2004 update for outpatient dialysisservices. To estimate current Medicarepayments and costs, we assessedaggregate 2001 payments and costs foroutpatient dialysis services and thenprojected both to 2003. We adjusted theunaudited 2001 cost data based on ourfindings that the allowable cost pertreatment was about 96 percent of thereported costs in 1996, the most recentyear for which audited cost data areavailable. Current payments for compositerate services and separately billable drugscombined exceeded costs of freestandingfacilities by about 4 percentage points in2001, and our estimate of the payment-to-cost ratio for 2003 is that it will be nomore than 3 percentage points lower than

the 2001 level (reflecting 2002 to 2004payment rules).

To further study the question of paymentadequacy, we looked at several marketindicators, including the growth in thecapacity of providers to furnish dialysisand changes in the financial health ofdialysis providers. Because Medicare isthe largest purchaser of outpatient dialysisservices, Medicare payment adequacyshould be reflected in these broadindicators. The findings from this analysisstrongly suggest that aggregate Medicarepayments appear to be sufficient relativeto efficient providers’ costs. Between1994 and 2001, the number of facilitiesand in-center hemodialysis stationsincreased by about 7 percent annually.There was a net increase of 156 facilitiesbetween 2000 and 2001. The number offor-profit freestanding facilities continuesto increase, suggesting that furnishingdialysis services to ESRD patients isfinancially attractive to for-profitproviders. Data from the Centers forMedicare & Medicaid Services (CMS)show that providers continued to improvethe quality of care furnished tobeneficiaries, as assessed by measures ofdialysis adequacy and anemiamanagement. Furthermore, the large for-profit, multicenter dialysis companies(chains) that account for 65 percent of allfacilities appear to have adequate accessto capital, as shown by the continuedgrowth in the number of facilities.

Current payments and costsThe Commission assesses currentpayments and costs for dialysis servicesby comparing Medicare’s payments forcomposite rate services and injectablemedications with providers’ Medicare-allowable costs. Cost reports submitted byproviders provide data on the costs theyincur to furnish dialysis services andinjectable drugs. We use data from costreports to estimate Medicare’s paymentsfor dialysis services and erythropoietinand claims data to estimate Medicare’spayments for separately billable injectable


4 In constant dollars, the composite rate has decreased by more than half of its original 1983 base rate of $127 for hospital facilities and $123 for freestanding facilities.

drugs other than erythropoietin. TheCommission has traditionally expressedthe relationship of aggregate payments tocosts as a payment-to-cost ratio.

As described in the opening of thischapter, MedPAC’s analysis of currentcosts uses only Medicare-allowable costs.Each year, CMS’s contractors—fiscalintermediaries (FIs)—regularly audit costreports submitted by certain institutionalproviders to ensure that the costs reportedby providers are Medicare allowable. TheBalanced Budget Act of 1997 (BBA)required the Secretary to audit the costreports of each dialysis provider at leastonce every three years beginning in 1996.CMS’s recent review of the 1996 dataresulted in 62 percent of submitted costsreported being reopened and audited. Theauditing of more recent cost reports iscurrently underway but not complete.5

MedPAC compared the audited costreport data for 1996 to unaudited 1996data. Our analysis showed that theallowable cost per treatment for compositerate services and injectable drugs forfreestanding facilities was about 96percent of the reported cost of treatment.As shown in Table 2E-1, all types offacilities were affected by the audit. Forexample, allowable costs as a percentageof reported costs were 96 percent formedium-sized facilities and 97 percent forsmall and large facilities. Our finding thatallowable costs are less than reportedcosts is consistent with an audit performedby CMS in 1988 that determined that theallowable cost per treatment forfreestanding facilities was 88 percent ofthe reported cost per treatment(Prospective Payment AssessmentCommission 1993).

If history is any guide, a portion of thereported costs for services furnishedbetween 1997 and 2001 will most likelybe found nonallowable when these reportsare audited by CMS. MedPAC believes itis important to consider the effect of thedifference between reported and allowable

costs when assessing the relationshipbetween current payments and costs.Consequently, we assessed providers’costs for services furnished between 1997and 2001 in two ways. First, we used theactual costs reported by providers thathave not yet been audited by CMS.Second, we adjusted the actual costsreported by providers by the ratio ofallowable costs to reported costs derivedfrom the analysis of the 1996 cost reports,the most recent year for which audited

data are available. We calculated the ratioof allowable costs to reported costs in1996 by each type of facility and appliedthis adjustment to the 1997 to 2001 costsof the corresponding facility type. Ourapproach assumes that the ratio ofallowable costs to reported costs for 1997to 2001 will be the same as 1996; thisrelationship may or may not be the caseonce the cost reports for this period areaudited. However, based on the results ofthe earlier audits of providers’ cost


Payment-to-cost ratios for composite rate services andseparately billable drugs for freestanding dialysis

facilities, 1996 and 2001

1996 2001

Not Not Adjusted forFacility type audited Audited audited audit effect

Composite rate services only

All 1.04 1.09 0.93 0.97Small 0.94 0.97 0.85 0.89Medium 1.02 1.08 0.91 0.95Large 1.08 1.12 0.97 1.01Nonprofit 1.02 1.04 0.86 0.89For profit 1.05 1.09 0.94 0.98Urban, in an MSA 1.04 1.09 0.93 0.97Rural 1.03 1.07 0.92 0.96

Composite rate services and injectable drugs

All 1.10 1.14 1.01 1.04Small 1.02 1.05 0.96 0.99Medium 1.09 1.13 1.00 1.03Large 1.12 1.16 1.03 1.06Nonprofit 1.07 1.09 0.95 0.98For profit 1.10 1.14 1.02 1.05Urban, in an MSA 1.10 1.14 1.01 1.04Rural 1.10 1.13 1.02 1.05

Note: MSA (metropolitan statistical area). These mean payment-to-cost ratios are weighted by the number of in-centerand home dialysis sessions furnished by each facility. The size of the facility is defined in each year based onthe 25th and 75th percentiles of dialysis sessions. Small facilities are those reporting dialysis sessions � the25th percentile of all dialysis sessions; medium facilities are those reporting dialysis sessions between the 25th

and 75th percentiles of all dialysis sessions; and large facilities are those reporting dialysis sessions � the 75th

percentile of all dialysis sessions.

Source: Data compiled by MedPAC from 1996 and 2001 cost reports and the outpatient institutional file from CMS.

T A B L E2E-1

5 For example, the proportion of 1997 to 2001 cost reports that have been reopened or audited range from 0.1 percent in 2001 to 11 percent in 1998. During fiscalyear 2003, the FIs will audit one third of facilities with cost report years ending between January 1, 2001 and December 31, 2001. In fiscal years 2004 and 2005, theFIs will audit the remaining ESRD cost reports for this time period (CMS 2002).

reports, we believe that once the costreports for 1997 to 2001 have beenaudited, the ratio of allowable costs toreported costs will be less than 1.0.

For 2001, we estimate that Medicare’spayments for composite rate services andinjectable medications exceededproviders’ costs by about 4 percentagepoints when the effect of the audit isconsidered (Table 2E-1). There is littlevariation in the aggregate payment-to-costratios for urban and rural facilities. Ourfinding that the payment-to-cost ratiosvary considerably based on a facility’ssize and profit status stems fromdifferences in the cost per dialysistreatment.

As shown in Figure 2E-2, aggregatepayments for composite rate services andinjectable drugs relative to providers’

costs have steadily declined during themost recent five-year period available,1997 to 2001. This decline is occurringbecause the composite rate was updatedtwice during this time period, 1.2 percentin 2000 and 2.4 percent in 2001. Duringthis time period, providers’ costs forcomposite rate services have increased byabout 3.0 percent annually. In addition,the manufacturer of erythropoietin raisedthe price in 2000 and 2001, while the perunit payment of this injectable drug hasremained unchanged by the Congress.

A different picture of financialperformance emerges when we isolatecomposite rate services. In 2001,Medicare’s payments for composite rateservice costs did not cover the costs ofproviding dialysis services. This finding,when taken together with the earlier one

about the aggregate payment-to-cost ratio,demonstrates that payments for separatelybillable drugs significantly exceedproviders’ costs.6 Additionally, thisfinding strongly suggests that theprofitability of erythropoietin and otherseparately billable drugs is subsidizing thelower margins under the composite rate.

To estimate the aggregate payment-to-costratio for 2003, we assumed that providers’costs will grow at the same rate predictedby MedPAC’s dialysis market basketindex in 2002 and 2003, less anadjustment for productivityimprovements. This assumption seemsreasonable given our analysis showingthat providers’ average per unit costsincreased at a rate lower than the increasein the dialysis market basket indexbetween 1997 and 2000. Our payment


Aggregate payment-to-cost ratios for dialysis services, adjusted and unadjusted, 1997–2001

FIGURE2E-2

Source: MedPAC analysis of 1997–2001 cost reports and outpatient institutional claims of freestanding dialysis facilities from CMS.

Ratio

1.04

1.16

1.08

1.12

1.00

0.96

0.921997 1998 1999 2000 2001

Unaudited data

Adjusted for audit effect

Note: The aggregate payment-to-cost ratio includes payments and costs for composite rate services and injectable drugs.

6 Two studies by the Office of Inspector General (OIG) concluded that Medicare’s payment rates for these drugs were high relative to providers’ costs and the rates paidby the Department of Veterans Affairs and state Medicaid programs (OIG 2000, OIG 1997).

estimate for 2003 reflects current law,which does not provide any update fordialysis services between 2002 and 2004.Based on these assumptions, payments forcomposite rate services and injectablemedications relative to providers’ costs in2003 are likely to be no more than 3percentage points lower than the 2001level. This estimate is conservativebecause we also assumed that revenuefrom injectable medications relative tothat from composite rate services wouldnot change between 2001 and 2003.However, based on historical trends, thepercentage of revenue from injectabledrugs relative to composite rate serviceswill most likely increase between 2001and 2003. Assuming the increasing use ofinjectable drugs and their continuedprofitability between 2001 and 2003, theaverage aggregate payment-to-cost ratiowill probably decline by less than threepercentage points in 2003.

Although the payment-to-cost ratio forcomposite rate services and injectablemedications is the most comprehensivemeasure we have to assess the financialperformance of dialysis facilities, it doesnot account for the potential profitabilityof other services associated withoutpatient dialysis. For example, severalnational dialysis chains own laboratoriesand receive Medicare payments forlaboratory tests outside the composite ratepayment bundle. In addition, providershave begun to provide diabetes outpatientself-management training services,payment for which was implemented bythe BBA. In the future, MedPAC willregularly monitor the extent to whichthese training services are furnished bydialysis providers.

Appropriateness of currentcostsAt issue is whether aggregate dialysiscosts provide a reasonable representationof the costs that efficient providers wouldincur in furnishing high-quality care.Because the composite rate ispredetermined, providers have anincentive to restrain their costs forcomposite rate services. In contrast,because injectable medications are paid

per unit, providers have little incentive toimprove efficiency.

To address this issue, MedPAC assessedthe factors explaining the growth inproviders’ costs for furnishing compositerate services and injectable medications. Itis too soon to tell whether the spike inaverage costs for composite rate servicesin 2001, which exceeded the increase inproviders’ costs predicted by the dialysismarket basket, will continue in futureyears. Our analysis of selectedproductivity measures showed littlechange in the composite rate servicesfurnished to beneficiaries between 1997and 2000–2001. MedPAC generallyexpects average cost growth toapproximate the rate of increase in themarket basket index given little change inthe services furnished to beneficiaries.

Costs for composite rate servicesProviders’ costs for composite rateservices increased by 5.7 percent between2000 and 2001. This rate of increaseexceeded the 3.8 percent increasepredicted by the dialysis market basketindex for this same time period.MedPAC’s analysis shows that twocategories of costs spiked in 2001:

• Labor costs increased by about 7percent, compared with a 2 percentincrease between 1997 and 2000.

• General and administrative costsincreased by about 9 percent,compared with a 2 percent increasebetween 1997 and 2000.

Historically, dialysis providers have beenable to adopt efficiencies in servicedelivery, enabling them to keep their costsat or below the dialysis market basketindex. It is too soon to tell whether thegrowth in providers’ labor andadministrative costs between 2000 and2001 is an anomaly. Like other health careproviders, dialysis providers contend thattheir labor costs have increased becausethey face increased competition forrecruiting registered nurses andtechnicians (driven by the possibleemergence of labor shortages). Inaddition, providers claim that recent

changes in licensure and scope of practicelaws in certain states means that certainservices previously furnished by dialysistechnicians must be provided by eitherregistered nurses or licensed practicalnurses. Finally, providers contend thatsince 2000 they have faced significantincreases in the cost of utilities and ofliability and property insurance.Unfortunately, the cost report data do notallow for an analysis of the specificcomponents comprising the costs reportedas general and administrative.

Thus, it is too soon to draw conclusionsabout the appropriateness of thecomposite rate cost base. To conclude thatproviders’ costs are not appropriate, theCommission would need to see that thelong-term growth in cost per casecontinues to significantly exceed thegrowth predicted by the market basket.MedPAC will monitor future trends inproviders’ costs and also changes in thedialysis product, which we discuss in thefollowing section.

Changes in composite rateservicesOne way to assess whether the cost basefor composite rate services is appropriateis to examine changes in the servicesfurnished by providers. MedPACexamined possible changes in the productby looking at changes over time in thestaff furnishing in-center hemodialysiscare and the productivity of the staff in1997 to 2000–2001.

From 1997 to 2001, few changes weremade in the composition of the stafffurnishing in-center dialysis care (Table2E-2). The proportion of technicians topatient care staff has not significantlychanged between 1997 and 2001, and theratio of patients to registered nurses andtechnicians has remained relativelyconstant between these two years. Also,the productivity of patient care staff wasfairly stable during this period. Forinstance, the average duration ofhemodialysis sessions slightly increasedfrom 210 minutes in 1997 to 215 minutesin 2000. The productivity of patient carestaff, as measured by the number of


in-center hemodialysis treatments perstation and the total number ofhemodialysis treatments per staff, alsoremained relatively constant between1997 and 2001.

The cost of incremental changes in thetechnologies used during dialysis areprobably not significantly contributing tothe growth in providers’ costs. Data fromproviders’ cost reports show that the twocategories that probably include the costsof new technologies, capital and otherdirect costs, increased by only 2 percentbetween 2000 and 2001. In comparison,labor costs increased by 7 percent, andgeneral and administrative costs increasedby 9 percent during this time period.

Costs for separately billablemedicationsBased on MedPAC’s previous findings,we expect that the costs of separatelybillable drugs have grown more rapidlythan the costs of composite rate services.Costs for separately billable drugsincreased by about 12 percent between2000 and 2001. This change is consistent

with the trends between 1998 and 2000.The payment method for separatelybillable drugs gives providers noincentives to improve efficiency. Incontrast, prospective payment methodsprovide incentives to control costsbecause payment is based on apredetermined rate unaffected by incurredcosts or posted charges. Substituting new,more costly drugs for older, lessexpensive medications may be anotherreason why providers’ costs for injectablemedications per dialysis treatmentincreased during the 1997 to 2001 period.For example, the price of a vitamin Danalogue (paricalcitol) newly approved in1998 is twice that of the older agent it hasdisplaced (calcitriol). Between 2000 and2001, spending for paricalcitol increasedfrom $172 million to $386 million; incontrast, spending for calcitriol decreasedfrom $127 million to $67 million duringthis same time. Finally, a 3.9 percentincrease in the price charged by themanufacturer of erythropoietin in 2000and 2001 also increased providers’ costsper treatment.

Relationship of payments tocostsNext we assess the relationship ofpayments to appropriate costs foroutpatient dialysis services and find thataggregate Medicare payments appear tobe sufficient. We base this conclusion, inpart, on the following evidence aboutmarket conditions throughout the 1990s:(1) the average annual growth in thenumber of hemodialysis treatments haskept pace with the average annual growthin the number of hemodialysis patients;(2) the number of for-profit freestandingdialysis facilities is increasing; (3) therehas been no widespread access problemfor beneficiaries; (4) the quality of dialysiscare has improved; and (5) there has beenno change in providers’ access to capital,as evidenced by continued growth in thenumber of providers and their capacity tofurnish dialysis.

Changes in volumeBetween 1993 and 2001, the growth in thenumber of in-center hemodialysistreatments generally kept pace with thegrowth in the number of dialysis patients.The number of dialysis treatmentsincreased, on average, by 8 percentannually; in comparison, the number ofdialysis patients increased, on average, by7 percent during this time period.

The growth in payments for injectabledrugs increased more rapidly than thegrowth in payments for dialysis treatmentsin the 1990s.7 Between 1998 and 2001,total payments for erythropoietinfurnished by freestanding dialysisfacilities increased by about 15 percentper year, and total payments for otherinjectable drugs increased by about 30percent per year. In contrast, payments forcomposite rate services increased by 9percent per year during this same period.The Commission anticipates that thegrowth in the use of injectable drugs paidfor outside the composite rate willcontinue to increase. For example, CMSrecently made a national coverage


Indicators to assess changes in services furnishedduring in-center hemodialysis treatments,

1997 and 2001

Indicator 1997 2001

Ratio of:Patients to technicians 19.2 18.0Patients to registered nurses 17.6 15.7Technicians to patient care staff 0.54 0.54

Length of hemodialysis treatment (minutes) 210 215*

Number of:Treatments per in-center hemodialysis station 654 658In-center hemodialysis treatments per patient care staff member 695 742In-center hemodialysis shifts per week 12.3 12.0

*The average length of an in-center hemodialysis session in 2000.

Source: MedPAC analysis of 1997, 2000, and 2001 cost reports and data on clinical performance measures fromCMS.

T A B L E2E-2

7 We express volume in terms of total Medicare payments because each injectable drug has its own unit of measurement.

decision to cover injections oflevocarnitine for patients with ESRDbeginning in January 1, 2003.8

Use of injectable medications has grownfor several reasons. First, many of theagents—including erythropoietin and ironsupplements—were only approved by theFood and Drug Administration in theearly 1990s. Since their approval, their usehas been advocated in clinical guidelinesset forth by the National KidneyFoundation (NKF). The use of many ofthese medications has enhanced thequality of care furnished to dialysisbeneficiaries. For example, the increaseduse of erythropoietin has reduced theproportion of dialysis patients sufferingfrom anemia, which contributes tomorbidity if not treated effectively.However, the profitability of certaininjectable medications has providedincentives in how they are used. Forexample, Medicare pays $10 per 1,000units for erythropoietin administeredeither intravenously or subcutaneously(under the skin). Paying on a per unitbasis promotes the use of the intravenousform of this medication, which requireshigher average doses (more units) toachieve target hematocrit levels.9 Thepredominant use of intravenouserythropoietin persists despite thepublication of the NKF’s DialysisOutcome Quality Initiative ClinicalPractice Guideline for the treatment ofanemia, which advocated subcutaneousadministration (NKF 1997).

Revenue from injectable medications hasbecome more important relative torevenue from composite rate servicesduring the past five years. Forfreestanding dialysis providers, revenuefrom injectable medications relative tothat from composite rate services hasincreased from about 33 percent of totalpayments in 1997 to 40 percent of totalpayments in 2001. As noted earlier, the

positive payment margins for injectabledrugs are subsidizing the lower paymentmargins under the composite rate.

Broadening the payment bundle to includefrequently used injectable drugs that arenow paid for separately would provide astrong incentive for providers to furnishthese services more efficiently. In ourMarch 2001 report, MedPACrecommended that the Congress requirethe Secretary to: (1) include in theprospective payment bundle services thatare frequently used for dialysis butcurrently excluded from this bundle, and(2) revise the payment system to accountfor factors that affect providers’ costs,including dialysis method, dose,frequency, and patient acuity.

Entry and exit of providers Reports of facility closings tend to belinked to local issues, such as rising realestate prices in certain areas, shortages oftechnicians and nurses, and states’certificate of need regulations. MedPACexamined the characteristics of dialysisfacilities that closed during 2001 usingdata from CMS’s facility survey. Between2000 and 2001, there was a net increase of156 facilities. Facilities that closed weremore likely to be smaller, in terms of boththe number of patients they treated and thenumber of in-center hemodialysis stationsthey maintained, than facilities thatremained in business in 2001. In addition,facilities that closed were more likely tobe nonprofit and hospital-based. Someproviders contend that they are limitingtheir exposure to Medicare patients.However, our data show little correlationbetween proportions of facility patientloads attributable to Medicare and facilityclosings between 2000 and 2001.

Our finding—that facilities that closedwere more likely to be small, nonprofit,and hospital-based than facilities thatremained open—is consistent with the

changes in the characteristics of dialysisproviders in the 1990s. As shown in Table2E-3, freestanding and for-profit facilitiesgrew at the expense of hospital-based andnonprofit facilities. Between 1993 and2001, freestanding facilities increasedfrom 70 percent to 83 percent of allfacilities, while for-profit facilitiesincreased from 61 percent to 79 percent ofall facilities. In addition, dialysis chainscontinue to acquire independentlyoperated facilities. MedPAC estimatesthat about 65 percent of all facilities wereoperated by the four national for-profitchains in 2001. Our finding thatfreestanding facilities have steadilyincreased as a share of the totalthroughout the 1990s suggests thatdialysis facilities are sufficientlyprofitable to stand on their own. Ourfinding that for-profit facilities continue togrow at the expense of nonprofit facilitiessuggests that furnishing dialysis servicesto ESRD patients is financially attractiveto for-profit providers.

Beneficiaries’ access to careA review of the published literature showsno evidence of beneficiaries facingsystematic problems in obtaining neededdialysis care in 2001 and 2002.MedPAC’s analysis of data from CMS’sfacility survey shows that the capacity ofproviders to furnish care has increasedsteadily between 1993 and 2001. The totalnumber of dialysis facilities grew by about7 percent during this time, as did thenumber of in-center hemodialysis patients(Table 2E-3). With about 25 percent of allfacilities located in rural areas between1993 and 2001, the capacity to furnishdialysis in rural areas appears to havestayed relatively constant during this timeperiod.

The Commission finds that providers havekept up with the demand for dialysis byincreasing the number of facilities ratherthan increasing capacity within facilities.


8 Levocarnitine supplements the loss of carnitine, a naturally occurring body substance that helps transport long-chain fatty acids for energy production by the body.Patients on hemodialysis can suffer carnitine deficiencies from dialytic loss, reduced renal synthesis, and reduced dietary intake. Patients must show improvement fromthe levocarnitine treatment within six months of initiation of treatment for Medicare to continue to pay for the treatment.

9 Some providers contend that erythropoietin is predominately furnished intravenously because patients experience less discomfort than when it is furnishedsubcutaneously.

We based this finding on our analysis oftrends in the following:

• average hemodialysis stations perfacility

• average in-center hemodialysistreatments per facility

• average in-center hemodialysistreatments per dialysis station10

The total number of in-centerhemodialysis treatments provided bydialysis facilities has increased by about 8percent per year between 1997 through2001, but the average number ofhemodialysis stations per facility hasremained relatively constant at about 21per facility. Average total dialysistreatments also have remained relatively

constant, ranging from 15,500 to 16,000during this time period. Finally, averagehemodialysis treatments per station haveremained relatively constant during thistime period, ranging from 648 to 658.

Opening new facilities may improveaccess to care by reducing the time thatbeneficiaries have to travel to obtain carethree times per week. Researchers havenoted that transportation to and from thedialysis facility can affect patients’compliance with their prescribedtreatment, with some patients shorteningtheir dialysis treatments or skippingtreatments (Rocco and Burkart 1993,Sehgal et al. 1998, USRDS 1997).However, the sustained growth in thenumber of dialysis facilities raisesquestions about the optimal efficiencies of

scale and the tradeoff between openingnew facilities versus increasing thecapacity of existing facilities.

Quality of careClinical performance indicators collectedby CMS show continued improvements inthe quality of dialysis care, as measuredby the percentage of hemodialysis patientsreceiving adequate dialysis and sufferingfrom anemia (Table 2E-4, p. 130). Forexample, the proportion of in-centerhemodialysis patients receivinginadequate dialysis declined from 26percent in 1996 to 14 percent in 2000.However, no clinically important changesor improvements were found in thepercentage of hemodialysis patients withadequate or optimal serum albumin levelsin 2000 compared to previous years.11


Characteristics of dialysis facilities, 1993–2001

1993 1994 1995 1996 1997 1998 1999 2000 2001

Total number of dialysis facilities 2,343 2,502 2,732 2,940 3,172 3,394 3,619 3,805 3,961

Percent of all facilities

For profit 60.8% 62.2% 64.6% 67.4% 71.1% 75.0% 77.3% 78.3% 79.4%Nonprofit 33.4 32.2 30.3 28.1 25.2 21.9 19.8 19.1 18.1Government 5.8 5.6 5.0 4.4 3.8 3.2 2.9 2.7 2.5

Freestanding 70.0 71.6 73.7 75.1 77.0 78.8 80.7 81.6 82.6Hospital-based 30.0 28.4 26.3 24.9 23.0 21.2 19.3 18.4 17.4

Urban, in an MSA 77.3 76.8 76.8 76.2 75.6 75.1 75.1 75.1 74.8Rural, total 22.7 23.2 23.2 23.8 24.4 24.9 24.9 24.9 25.2

Adjacent to an MSAIncludes a town with at least

10,000 people 6.7 6.8 6.5 6.8 6.7 6.6 6.6 6.5 6.4Does not include a town with

at least 10,000 people 5.0 5.4 5.5 5.8 6.1 6.5 6.6 6.8 7.0Not adjacent to an MSA

Includes a town with at least10,000 people 6.6 6.4 6.4 6.3 6.1 6.1 5.9 5.7 5.7

Does not include a town withat least 10,000 people 4.4 4.5 4.8 5.0 5.5 5.8 5.9 5.9 6.1

Source: MSA (metropolitan statistical area). Data compiled by MedPAC from the 1993–2001 facility survey file from CMS. Numbers may not total exactly because of rounding.

T A B L E2E-3

10 Average hemodialysis stations per facility, treatments per facility, and treatments per dialysis station are weighted by the number of dialysis sessions at each facility.

11 Mean serum albumin levels have been shown to be a marker for diminished patient survival.

Some providers and researchers contendthat increased use of certain types ofmedical interventions, particularlyparenteral nutrition, would improve theoutcomes of certain patients. Medicare’scoverage policies limit the number ofdialysis patients who qualify for theseinterventions.12

A recent study raised important issuesabout the quality of dialysis care in theUnited States (Devereaux et al. 2002).The authors reported a death rate 8percent higher among kidney failurepatients receiving dialysis at for-profitcenters than among those treated atnonprofit facilities, for an estimated 2,500additional deaths each year. Thisconclusion was based on a meta analysisof 8 retrospective studies that examinedthe risk of mortality for more than500,000 patients. Seven of these studiesused data from 1990 through 1997; onestudy was based on data from 1973 to1982.

Past research by CMS, USRDS, andothers has shown that many factors,including patients’ clinical characteristicsand providers’ characteristics, affectoutcomes of dialysis patients. Studies

underway using more recent data areevaluating whether patient outcomes varyby facility profit status and other providercharacteristics. Two abstracts recentlypublished using post-1997 data show nosignificant difference in mortality at for-profit versus nonprofit facilities (Held etal. 2002, Wolfe et al. 2002).

Two MedPAC studies currently underwaywill partly address the issue of the qualityof care furnished to dialysis patients. Thefirst study will explore the use ofincentives—both financial andnonfinancial—for Medicare to encourageproviders to improve care. Strategies forencouraging more-focused providerattention to improving quality are beingdiscussed in national forums such as theInstitute of Medicine and the NationalQuality Forum and in numerous purchasercoalitions across the country. The secondstudy will examine the relationshipbetween quality of care and providers’costs per treatment. No publishedinformation is available regarding theinfluence of dialysis facility costs onpatient outcomes. Previous MedPACanalysis has shown significant variation inthe cost per dialysis treatment amongfreestanding dialysis facilities.

The findings by Devereaux et al. onquality demonstrate the importance ofMedicare’s continuing efforts to monitorthe quality of care furnished by dialysisproviders. Beginning in 1993, CMS hasannually published information about thequality of care furnished to dialysispatients, including adequacy of dialysisand anemia management. The USRDSalso collects, analyzes, and distributesinformation on different aspects of thecare of patients with ESRD, includingtrends in disease incidence andprevalence, patient survival and causes ofdeath, modality of treatment, and use ofhospital services.

Providers’ access to capitalDialysis facilities need access to capital toimprove their equipment and to open newfacilities to accommodate growth in thenumber of patients requiring dialysis.About 80 percent of all dialysis facilitiesare for-profit, and the four largest for-profit chains account for about 65 percentof all facilities. These for-profit chainsappear to have adequate access to capital,as demonstrated by growth in the numberof clinics, the number of patients theytreat, and their earnings. Data fromindustry sources show that the growth inrevenues between 1996 and 2000 for thesefour chains ranged from 36 to 62 percent.A bond analyst described the sector ashaving no problems with access to capitaland ratings for the bonds of two of thelargest chains, although below investmentgrade, are neutral going forward. Inaddition, industry reports have stated thatrevenues for dialysis service are fairlypredictable, given the recurringrequirement for treatment. However, theyalso have noted that dialysis providersface potential pressures from privatepayers, and are highly susceptible to anyfuture changes in Medicare’s paymentpolicies. Finally, the stocks of these for-profit chains have in large part enjoyedpositive ratings by financial analysts overthe last year.


Clinical performance indicators, 1994–2000

Performance indicator 1994 1995 1996 1997 1998 1999 2000

Percent of hemodialysis patients receiving inadequate dialysis N/A N/A 26% 22% 20% 16% 14%

Percent of hemodialysis patients with low hematocrit levels N/A N/A N/A 57% 41% 32% 26%

Percent of hemodialysis patients who are malnourished 20% 16% 19% 16% 18% 20% 20%

Note: N/A (not available), Kt/V (urea clearance multiplied by the time normalized by total body water divided bythe volume of distribution of urea), gm/dL (grams per deciliter). Patients receiving inadequate dialysis arethose with Kt/V � 1.2. Patients with low hematocrit levels are those with hemoglobin levels � 11 gm/dL.Patients malnourished are those with serum albumin levels � 3.5 gm/dL.

Source: MedPAC analysis of 1994–2000 data on clinical performance measures from CMS.

T A B L E2E-4

12 Daily parenteral nutrition is limited to patients “with severe pathology of the alimentary tract which does not allow absorption of sufficient nutrients to maintain weightand strength commensurate with the patient’s general condition” (CMS 2003).


As noted earlier, the Commissionaccounts for expected cost changes in thecoming year primarily through theforecast of input price inflation. CMS hasnot developed a market basket index foroutpatient dialysis services.13

Consequently, MedPAC uses an index fordialysis services comprising price indexesfor hospitals, skilled nursing facilities, andhome health agencies. MedPAC’s indexindicates that the prices dialysis facilitiespay for their inputs included in thecomposite rate will rise an estimated 2.5percent between 2003 and 2004.

Another factor considered by MedPAC’supdate framework that may affectproviders’ costs in the next payment yearis scientific and technological advances.This factor is designed to reflect onlythose new technologies that are qualityenhancing and costly, and haveprogressed beyond the initial stage of usebut have not yet fully diffused intomedical practice. Based on our review ofthe literature, we believe that the costs ofmost medical advances will be accountedfor primarily through payments forseparately billable drugs. Therefore, thereis no need for an addition to the update formedical advances.

Finally, MedPAC’s update frameworkreflects the expectation that, in theaggregate, providers should be able toreduce the quantity of inputs required toproduce a unit of service whilemaintaining service quality. Prospectivepayment is designed to promoteefficiency, and productivity increasesshould be expected from providers. Toestimate productivity increases, MedPACuses the 10-year moving average ofmultifactor productivity in the economy asa whole, which is 0.9 percent.


Based on our review of the adequacy ofpayments for outpatient dialysis servicesand expected cost changes in the comingyear, the Commission recommends thefollowing:

R E C O M M E N D A T I O N 2 E

The Congress should update thecomposite rate payment by theprojected change in input prices, less0.9 percent, for calendar year 2004.

As noted earlier, MedPAC’s dialysismarket basket projects that input priceswill rise by 2.5 percent between 2003 and2004. The Congress should consider using

CMS’s dialysis market basket index toupdate the composite rate payment once itbecomes available because it may be amore current projection than theCommission’s market basket index.

I M P L I C A T I O N S 2 E


spending between $50 and $200million in one year. Over 5 years,spending would increase between$250 million and $1 billion.

Beneficiary and provider• This recommendation would result in

a payment increase sufficient to coverexpected increases in efficientproviders’ costs for dialysis servicesin 2004. Dialysis providers should beable to realize productivity gains topartially offset the increases in inputprices reflected in the dialysis marketbasket index.

• To the extent that adequate paymentallows providers to meetbeneficiaries’ health care needs,beneficiaries will continue to haveaccess to medically necessary care ofhigh quality. �


13 In our March 2000 report, MedPAC recommended that the Congress instruct CMS to consider a periodic update for outpatient dialysis services. The Medicare,Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 instructed the Secretary to submit a report on methods to update the outpatient dialysis paymentsystem, including a market basket for dialysis services, by July 2002. This study is currently being reviewed within the agency.

References

Centers for Medicare & Medicaid Services. System tracking for audit and reimbursementinstructions: end stage renal disease audits and hospice cost reports. Programmemorandum transmittal A–02–083. Baltimore (MD), CMS. 2002.

Centers for Medicare & Medicaid Services. Coverage issues manual. January 23, 2003.Available at http://cms.hhs.gov/manuals/06_cim/ci60.asp#_65_101.

Devereaux PJ, Schunemann HJ, Ravindran N, et al. Comparison of mortality betweenprivate for-profit and private not-for-profit hemodialysis centers, Journal of the AmericanMedical Association. November 20, 2002, Vol. 288, No. 19, p. 2449–2457.

Held PJ, McCullough KP, Gillespie BW, et al. Survival among dialysis patients in for-profit and not-for-profit hemodialysis facilities. American Society for Artificial andInternal Organs abstracts March–April 2002. Available at www.asaiojournal.com.

National Kidney Foundation. Clinical practice guidelines for the treatment of anemia ofchronic renal failure. New York (NY), NKF. 1997.

Office of Inspector General. Medicare reimbursement of end-stage renal disease drugs,OEI–03–00–00020. Washington (DC), OIG. June 2000.

Office of Inspector General. Review of EPOGEN reimbursement, A–01–97–00503.Washington (DC), OIG. November 1997.

Prospective Payment Assessment Commission. Report and recommendations to theCongress. Washington (DC), ProPAC. March 1993.

Rocco MJ, Burkart JM. Prevalence of missed treatments and early sign-offs inhemodialysis patients, Journal of the American Society of Nephrology. November 1993,Vol. 4, No. 5, p. 1178–1183.

Sehgal AR, Snow RJ, Singer ME, et al. Barriers to adequate delivery of hemodialysis,American Journal of Kidney Diseases. April 1998, Vol. 31, No. 4, p. 593–601.

United States Renal Data System. USRDS 2000 annual data report. Bethesda (MD),National Institute of Diabetes and Digestive and Kidney Diseases. 2002.

United States Renal Data System. USRDS 1997 annual data report. Bethesda (MD),National Institute of Diabetes and Digestive and Kidney Diseases. 1997.

Wolfe R, Ashby V, Hulbert-Shearon T, et al. Mortality and transplantation rates, andpractice patterns at for-profit vs non-profit and at chain vs non-chain facilities. Presentedat the 2002 annual meeting of the American Society of Nephrology. Available atwww.abstracts-online.com/abstracts/asn.


2FAssessing payment adequacy

and updating paymentsfor ambulatory surgical

center services

S E C T I O N


2F-1 The Secretary should expedite collection of recent ASC charge and cost data for thepurpose of analyzing and revising the ASC payment system.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2F-2 The Congress should eliminate the update to payment rates for ASC services for fiscal year2004.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2F-3 Until the Secretary implements a revised ASC payment system, the Congress should ensurethat payment rates for ASC procedures do not exceed hospital outpatient PPS rates for thoseprocedures, after accounting for differences in the bundle of services covered.



Section 2F: Assessing paymentadequacy and updating payments forambulatory surgical center services

An ambulatory surgical center (ASC) is a distinct entity that exclusively furnishes

outpatient surgical services. The most recent data on the cost of providing ASC

services to Medicare beneficiaries are from a 1994 survey by CMS of ASCs’

costs and charges. Because we lack recent data on ASCs’ costs, our analysis of

the adequacy of current Medicare payments for ASC services is based only on

market factors, such as entry and exit of providers, changes in the volume of ser-

vices, and providers’ access to capital. Through our analysis of these factors, we

find that current payments for ASC services are more than adequate. There has

been rapid growth in the number of ASCs; between 1991 and 2001, the number

of Medicare-certified ASCs more than doubled. The volume of procedures pro-

vided by ASCs to beneficiaries increased by over 60 percent between 1997 and

2001. In addition, ASCs have sufficient access to capital. We estimate that ASCs’

per-service costs will increase during the coming year at the rate of inflation in

input prices, less an adjustment for expected productivity growth. Current

Medicare payments for ASC services are at least adequate to cover this estimated

increase in unit cost. The Commission is concerned that the existence of ASC

payment rates that exceed hospital outpatient department rates for the same pro-

cedures could create financial incentives to shift services between settings.

2FIn this section

• Collecting recent ASC costdata




• Variations in payment forambulatory surgicalprocedures by setting


S E C T I O N

Background

Since 1982, Medicare has covered thefacility costs of certain surgicalprocedures provided in freestanding orhospital owned and operated ambulatorysurgical centers (ASCs). An ASC is adistinct entity that exclusively furnishesoutpatient surgical services. Theprocedures that are eligible for Medicarepayment when provided in an ASC arealso furnished to Medicare beneficiaries ininpatient and outpatient hospital settings,and sometimes in physician offices. In2001, ASCs provided almost 3 millionsurgical procedures to Medicarebeneficiaries and received about $1.6billion in related payments. Medicareaccounts for 20 to 30 percent of revenuesreceived by the largest for-profit ASCchains.

To receive payments from Medicare,ASCs must meet Medicare’s conditions of

coverage for ASCs, which requirecompliance with state licensure law andspecify minimum standards for: administration of anesthesia, qualityevaluation, operating and recovery rooms,the medical staff, nursing services, andother areas. ASCs are deemed to be incompliance with the conditions ofcoverage if they are licensed by a stateagency or accredited by a privateaccreditation body.1 Most Medicare-certified ASCs are for-profit, freestanding(as opposed to hospital owned andoperated) facilities located in urban areas(Table 2F-1). Almost 40 percent ofMedicare-certified ASCs are concentratedin four states that account for 25 percentof beneficiaries: California, Florida,Maryland, and Texas (Figure 2F-1).

ASC procedures eligible forMedicare paymentThe Centers for Medicare & MedicaidServices maintains a list of surgical

procedures eligible for Medicare facilitypayment when performed in an ASC.CMS is required by law to update the listevery two years in consultation withappropriate medical organizations. Since1995, however, with the exception ofupdates resulting from coding changes,the list has not been modified. The mostcommon categories of proceduresfurnished to Medicare beneficiaries inASCs in 2001 were cataract removal/lensinsertion, colonoscopy, and other eyeprocedures (Table 2F-2, p. 138).2

Surgical procedures must meet severalcriteria to be added to the list ofprocedures eligible for Medicare paymentwhen performed in an ASC:

• Site-of-service volume. Proceduresmust meet two site-of-service volumestandards to be added to the list: (1)The procedure must be performed inhospital inpatient settings at least 20percent of the time but can also be

136 As s e s s i ng paymen t adequacy and upda t i ng paymen t s f o r ambu l a t o r y s u r g i ca l c e n t e r s e r v i c e s

1 If an ASC is privately accredited, it must still comply with state licensure requirements. The Centers for Medicare & Medicaid Services has approved four privateaccreditors: the American Association for Accreditation of Ambulatory Surgical Facilities, the Accreditation Association for Ambulatory Health Care, the AmericanOsteopathic Association, and the Joint Commission on Accreditation of Healthcare Organizations.

2 These procedure categories are based on CMS’s Berenson-Eggers Type of Service classification scheme, which groups several related procedures in each category. Thecategory of other eye procedures includes after cataract laser surgery (Healthcare Common Procedure Coding System (HCPCS) code 66821).

Characteristics of Medicare-certified ambulatory surgical centers, 1991–2001

1991 1996 1997 1998 1999 2000 2001

Number of facilities 1,460 2,265 2,462 2,644 2,786 3,028 3,371New facilities 237 228 162 295 446Exiting and merged facilities 40 46 20 53 103

Net percent growth from previous year 8.7% 7.4% 5.4% 8.7% 11.3%

Percent of all centers

For profit 94% 93% 93% 94% 94% 94% 94%Nonprofit 6 6 6 6 6 6 5

Freestanding 99 99 99 99 99 99 99Hospital owned and operated 1 1 1 1 1 1 1

Urban, in MSA 88 90 90 89 89 88 88Rural 12 10 10 11 11 12 12

Note: MSA (metropolitan statistical area, as defined by the Office of Management and Budget).

Source: MedPAC analysis of provider of services file from CMS.

T A B L E2F-1

safely performed in outpatientfacilities. (2) The procedure can notbe performed more than 50 percent ofthe time in physician offices(procedures usually provided inphysician offices are generallyassumed not to require the moreelaborate facilities of an ASC).3

• Time needed to performprocedure. To be payable byMedicare in an ASC, a proceduremust not exceed 90 minutes ofsurgery or 4 hours of recovery time;anesthesia for the procedure cannotlast longer than 90 minutes.

• Clinical criteria. A procedure isexcluded from Medicare payment inan ASC if it (1) generally results inextensive blood loss, (2) requires

major or prolonged invasion of bodycavities, (3) directly involves majorblood vessels, or (4) is generallyemergent or life-threatening innature.

In 1998, CMS proposed revising itscriteria for determining which proceduresare eligible for Medicare facility paymentwhen provided in an ASC and expandingthe list of procedures approved forpayment (Health Care FinancingAdministration 1998). CMS proposedeliminating the surgery, anesthesia, andrecovery time limits but continuing to usespecific clinical standards for determiningwhether a procedure could safely beperformed in an ASC. CMS also proposedeliminating site-of-service volume as aprincipal criterion of approval for the ASC

list but proposed continuing to consider itas one of the factors in the approvalprocess. This change would have allowedprocedures that are frequently performedin physician offices to be considered foraddition to the ASC list. Thus, it couldhave led to the shift of some procedures toASCs from the physician office setting,where the practice expense fee isgenerally less than the ASC facility fee.CMS has been planning to release apartial final rule that would update theASC list (but not modify the criteria fordetermining eligibility for the list) in early2003 (Scully 2002). Expanding the list ofprocedures payable by Medicare in ASCswould likely increase the volume ofprocedures provided to beneficiaries inASCs.


Per

centa

ge

10

16

12

8

4

2

6

0California TexasFlorida Maryland

14

Percent of total Medicare-certified ASCs (2001) Percent of total Medicare enrollment (1999)

States with the most Medicare-certified ASCs, 2001FIGURE2F-1

Source: MedPAC analysis of provider of services file from CMS and Health Care Financing Administration, Department of Health and Human Services. Health Care Financing Review, Medicare and Medicaid Statistical Supplement, 2000. Baltimore (MD), HCFA. June 2001.

Note: ASC (ambulatory surgical center).

3 There are different site-of-service criteria applied to procedures that are already on the list of services eligible for Medicare payment. To remain on the list, proceduresmust have combined inpatient, hospital outpatient, and ASC volume greater than 46 percent, physician office volume of less than 50 percent, and inpatient hospitalvolume of greater than 10 percent (Health Care Financing Administration 1998).

ASC payment systemMedicare uses a fee schedule to pay forfacility services provided in an ASC, suchas nursing, recovery care, anesthetics, andsupplies (see Appendix A for moreinformation on the ASC payment system).The ASC fee schedule divides proceduresinto nine payment groups based on similarcosts.4 For fiscal year 2003, the paymentrates for these groups range from $333 to$1,399. Medicare pays for related

physician services separately under thephysician fee schedule.

CMS is statutorily required to conduct asurvey of costs and charges for individualprocedures from a sample of ASCs everyfive years. These data are used to reviseASC payment rates. Although the mostrecent cost survey was conducted in 1994,the payment rates based on this surveywere never implemented because of

legislative action (see discussion below).Thus, current payment rates are based ona 1986 cost survey and are probably nolonger consistent with ASC costs.

Between revisions to the payment system,the payment rates generally are required tobe updated annually using the consumerprice index for all urban consumers(CPI–U). From fiscal year 1998 throughfiscal year 2002, however, the BalancedBudget Act of 1997 (BBA) limited annualupdates to the CPI–U minus 2 percentagepoints (but not less than zero).5 ASC rateswere updated by 3 percent for fiscal year2003.

In 1998, CMS proposed restructuring theASC payment system to make it moreconsistent with the outpatient hospitalprospective payment system (PPS), whichwas then under development. The agencyproposed replacing the 8 ASC paymentgroups with 105 ambulatory paymentcategories (APCs) that classifiedprocedures based on cost and clinicalcharacteristics.6 The payment rates for theAPCs would have been based on datafrom the 1994 cost survey.

In response to CMS’s proposed rule, theCongress included a provision in theMedicare, Medicaid, and State Children’sHealth Insurance Program BenefitsImprovement and Protection Act of 2000that required CMS to do the following:

• delay implementing the new paymentsystem until 2002;

• phase in the payment system overfour years; and

• base payment rates on cost surveydata from 1999 or later.7


Most common categories of procedures provided toMedicare beneficiaries in ASCs, 2001

MedicareVolume payments Medicare

(as percent (as percent paymentsProcedure category of total) of total) (millions)

Cataract removal/lens insertion 29.1% 49.5% $799Colonoscopy 18.0 13.4 217Other eye procedures 12.0 9.7 156Upper gastrointestinal endoscopy 10.1 6.6 106Minor procedures—musculoskeletal 10.1 5.2 84Other ambulatory procedures 4.5 3.0 48Ambulatory procedures—musculoskeletal 3.5 2.8 42Cystoscopy 3.1 2.0 32Arthroscopy 1.9 1.7 27Ambulatory procedures—skin 1.8 1.3 21

Total 94.1 95.2 $1,532

Note: ASC (ambulatory surgical center). Each category includes several procedure codes. Table does not include allprocedures provided to beneficiaries in ASCs.Other eye procedures include after cataract laser surgery.Minor procedures—musculoskeletal include interventional pain management procedures (such as epiduralinjection and facet joint block), soft tissue biopsy, tumor excision, and closed treatment of certain fractures.Other ambulatory procedures include services such as breast biopsy, nasal polyp excision, abscess drainage,dilation of esophagus, and septoplasty.Ambulatory procedures—musculoskeletal include services such as hammertoe operation, tendon sheathincision for finger, arthrotomy, tenotomy, and tendon repair.Ambulatory procedures—skin include services such as skin debridement, excision of lesion, wound repair,and skin graft.

Source: MedPAC analysis of the 5 percent Standard Analytical File of ASC facility claims, 2001, and the Berenson-Eggers Type of Service classification scheme from CMS.

T A B L E2F-2

4 The highest payment group ($1,399) currently has only one code (HCPCS code 50590, Extracorporeal Shock Wave Lithotripsy). Payments have not yet been made forthis procedure due to a court order (American Lithotripsy Society v. Sullivan) that required CMS to reconsider the payment rate. CMS is planning to add severalprocedures to the ASC list that will be placed in this payment group (CMS 2002).

5 The Omnibus Budget Reconciliation Act of 1993 had eliminated the annual CPI–U update for 1994 and 1995.

6 The APCs proposed for the ASC payment system were those included in the outpatient payment system proposed in 1998. Subsequently, CMS modified the APCdefinitions for the outpatient PPS and expanded the number of APCs.

7 In the first year of the new payment system’s implementation, 25 percent of the payment would be based on the new system and 75 percent on the current system. Theproportion of the payment from the new payment system would increase to 50 percent in the 2nd year of implementation, 75 percent in the 3rd year, and 100 percentin the 4th year.

As of early 2003, CMS has not conductedthe new cost survey that is needed torevise the ASC payment system.

Trends in Medicarepayments for ASC servicesBetween 1996 and 2001, Medicarepayments (program spending andbeneficiary cost sharing) for ASC facilityservices doubled while payments tophysicians increased by 25 percent andpayments to outpatient departments grewby 17 percent. Medicare payments toASCs more than quadrupled between1991 and 2001, increasing from $375million to $1.6 billion (Figure 2F-2).Payments to ASCs are projected toincrease at an average annual rate of 11 to

12 percent between 2002 and 2007.8

Payments to ASCs were less than 1percent of total Medicare spending in2001.

Factors affecting growth ofASC servicesIn addition to Medicare payment policy(discussed in the next section), severalother factors have influenced the rapidgrowth in Medicare payments for ASCservices:

Shift of services from inpatientsettings to ambulatory caresettingsTo some extent, the growth in ASCservices is part of the general shift of

services from inpatient hospital toambulatory care settings. Between 1994and 1998, several high-volume proceduresthat can be provided in multiple settings—such as upper gastrointestinal (GI)endoscopy, colorectal endoscopy, andarthroscopy—migrated from the inpatientsetting to one or more ambulatory caresettings (MedPAC 2000).

Growth in ASCs’ share ofambulatory servicesASCs’ share of certain ambulatorysurgical procedures has been increasing incomparison to that of hospital outpatientdepartments and physician offices. Forexample, our analysis of Medicare claimsdata found that between 1997 and 2000,


Growth in total Medicare payments for ASC services, 1991–2001FIGURE2F-2

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

800

600

200

400

0

1,400

1,200

1,000

1,600

Paym

ents

(dolla

rs in

mill

ions)

Medicare payments (nominal dollars) Medicare payments (1991 dollars)

Note: ASC (ambulatory surgical center). Medicare payments include program spending and beneficiary cost sharing for ASC facility services. Average annualgrowth of nominal payments (1991–2001) � 15.6 percent. Average annual growth of payments in 1991 dollars (1991–2001) � 12.5 percent.

Source: CMS, Office of the Actuary.

8 This estimate is based on projections from the Congressional Budget Office’s March 2002 baseline and the 2002 annual report of the Boards of Trustees of theMedicare trust funds.

ASCs’ share of cataract removal/lensinsertion procedures increased from 37percent to 42 percent. ASCs’ share ofcolonoscopies, upper GI endoscopies, andother eye procedures (such as aftercataract laser surgery) also grew.

Changes in practice patterns andmedical technologyChanges in clinical practice and healthcare technology have expanded the use ofambulatory procedures. For example,colonoscopy and upper GI endoscopicprocedures, which together account for 20percent of Medicare payments to ASCs(Table 2F-2, p. 138), have increasedbecause of the development of flexiblefiberoptic scopes and expanded Medicarecoverage of colon cancer screening. Thegrowth in cataract lens replacement,which accounts for about half of Medicarepayments to ASCs, has been spurred byadvances in microsurgery and ultrasoundtechniques and the aging of the population(MedPAC 2000).

Benefits to patientsAn ASC may offer patients moreconvenient locations, shorter wait times,and lower coinsurance than a hospitaloutpatient department (20 percent in anASC compared with up to 55 percent inan outpatient department).

Benefits to physiciansBecause ASCs are specialized settings forambulatory surgery, physicians may beable to perform procedures moreefficiently than in a hospital outpatientdepartment. For example, the surgicalenvironment in an ASC is oftencustomized for a specific procedure, suchas cataract lens replacement. In addition, itmay be easier for physicians to reservesurgical time in an ASC than an outpatientdepartment that may be subject tounpredictable demands.

Physicians also may be able to increasetheir revenues by investing in ASCs. Thereare fewer legal restrictions on physicianownership of ASCs than on other types ofhealth care facilities, such as clinicallaboratories. The laws prohibitingphysicians’ referral to health care entitieswith which they have financialrelationships (Section 1877 of the SocialSecurity Act) do not apply to surgicalservices provided in an ASC (Health CareFinancing Administration 2001). Inaddition, the Department of Health andHuman Services Office of InspectorGeneral has published safe harborregulations that protect physicians whoinvest in ASCs from prosecution under theanti-kickback statute, if certain conditionsare met.9 Among other conditions, the safeharbor regulations generally protectphysician investors for whom the ASC isan extension of their office practice (Officeof Inspector General 1999). Physicianswho invest in an ASC can receive a shareof the ASC’s profits that is related to theirportion of the investment. The CEO of alarge ASC chain has claimed that aphysician’s ASC revenues can “replace . . .the decline in his or her professional feethat has occurred in the last three to fiveyears because of pressure from managedcare, insurance companies, and Medicare”(Physician Compensation Report 2002).10

However, data on the relative profitabilityof ASCs and the extent of physicianownership of ASCs are difficult to obtain.

Collecting recent ASC costdata

As discussed earlier, CMS is statutorilyrequired to conduct a survey of ASCs’costs and charges every five years. Thesedata are used to revise the ASC paymentrates. However, CMS has not conducted anew cost survey since 1994. Thecollection of recent ASC cost data would

allow the Congress and CMS to evaluatecurrent ASC payment rates and to revisethe ASC payment system. Once they arecollected, MedPAC would use recent costdata to assess the adequacy of ASCpayment rates.

R E C O M M E N D A T I O N 2 F - 1

The Secretary should expeditecollection of recent ASC charge andcost data for the purpose ofanalyzing and revising the ASCpayment system.

I M P L I C A T I O N S 2 F - 1

Spending• The collection of ASC charge and

cost data would not affect Medicarebenefits spending. However, therevision of ASC payment rates basedon recent data would probably affectMedicare spending. Until new ratesare developed, however, we areunable to project whether they wouldincrease or decrease spending.

Beneficiary and provider• The collection of recent charge and

cost data should not affectbeneficiaries. There could be smalladministrative costs for ASCs toprovide the data to CMS.


The first question in applying MedPAC’sapproach to evaluating payment adequacyis whether the current level of Medicare’spayments for ASC services is adequaterelative to providers’ costs. However,there is no recent information on the costof ASC services that would allow us tocompare Medicare’s payments to ASCs’costs. The revised ASC payment rates


9 The anti-kickback statute prohibits health care providers from receiving or paying anything of value to influence the referral of services covered by Federal healthprograms.

10 The Medicare payment changes to which this statement refers may include the phase-in of the resource-based practice expense relative value units, which ended in2002 and reduced payment rates for surgical services, on average, and the 5.4 percent cut in physician payment rates in 2002.


proposed by CMS in 1998 (which werebased on data from the 1994 ASC costsurvey) would have reduced 1998payment rates for high-volume servicessuch as cataract-related procedures andgastrointestinal endoscopies, whichsuggests that 1998 payments exceededcosts for these procedures.11 Although welack recent data on ASC costs,information on market factors allows us tojudge the adequacy of Medicare paymentsfor ASC services. Rapid growth in thenumber of ASCs and the volume ofprocedures they provide to beneficiaries,together with ASCs’ sufficient access tocapital, lead us to conclude that currentMedicare payments to ASCs are morethan adequate.

Entry and exit of providersRapid growth in the number of providersfurnishing services to beneficiaries mayindicate that Medicare’s payment rates areat least adequate and potentially too high.Conversely, rapid provider withdrawalsfrom Medicare could suggest that rates aretoo low.

The number of Medicare-certified ASCsmore than doubled between 1991 and2001, from 1,460 to 3,371 (Table 2F-1, p. 136). After slowing down in 1998and 1999, growth in the number offacilities accelerated in 2000 and 2001.Each year from 1997 through 2001, anaverage of over 270 new facilities enteredthe market, while an average of only 52closed or merged with other facilities.Most of the new and existing ASCs arefor-profit entities.

Changes in the volume of servicesLarge increases in the volume of servicesprovided could indicate that payment ratesare at least adequate and potentially toohigh, and small increases could signalunfavorable rates. The volume ofprocedures provided by ASCs toMedicare beneficiaries increased by over

60 percent between 1997 and 2001. Thisgrowth occurred despite annual updates toASC payment rates of less than 1 percentbetween 1998 and 2002, as mandated bythe BBA.

The growth in the volume of ASCprocedures has paralleled increases in thenumber of ASCs (Figure 2F-3, p. 142).The growth in the number of facilities,volume of procedures, and Medicarepayments to ASCs appears to beaccelerating.

Beneficiaries’ access to careAlthough ASCs are growing in number,they are not available in all areas.Beneficiaries who are unable to access anASC may receive ambulatory surgicalservices in a hospital outpatientdepartment, and, in some cases, aphysician’s office. Thus, even thoughsome beneficiaries do not have access tosurgical services in an ASC, they canreceive the same services in other settings.

Providers’ access to capital Rapid growth in the number of bothindependently-owned ASCs and ASCs thatare part of investor-owned chains impliesthat they have sufficient access to capital.The relatively small start-up costs of ASCsand their quick returns on investment havemade them attractive to physicians andother investors (Versel 2002).

Several ASCs acquire capital, as well asmanagement expertise, by partnering withfor-profit ASC chains. Companies thatinvest in or manage ASCs have increasedtheir acquisition of new facilities andexperienced strong revenue and earningsgrowth in the last few years. The fourlargest investor-owned ASC chains had afinancial stake in about 13 percent of allASC facilities in 2001. New ASC chainshave recently entered the market andothers are poised to follow. The stockvalue of at least two large chains has beengrowing faster than that of the overallhealth care industry (Borden 2002).

Although the stock value of the largestowner of ASCs has recently fallenbecause of factors unrelated to its ASCline of business, other ASC firms havereceived positive investment ratings byfinancial analysts over the past year.


Given the information about the adequacyof the current level of Medicare payments,the next step in determining paymentupdates is to ask how much providers’unit costs will change in the coming year.Several factors will affect the change inthe unit cost of ASC services.

The most important factor that will affectthe cost of ASC services is inflation ininput prices. Medicare’s payment systemfor ASCs uses the CPI–U to approximatechanges in input prices per unit of servicefaced by ASCs. Currently, CMS projectsthat the CPI–U will increase by 2.7percent in fiscal year 2004.12

ASC costs also may increase because ofscientific and technological advances thatenhance the quality of care but also raisecosts. The ASC payment system, unlikethe hospital outpatient PPS, has no pass-through payment mechanism to accountfor the cost of new technologies.However, among procedures eligible forMedicare payment in an ASC, we lackevidence that the ASC payment systemhas created barriers to the use of newtechnologies. For example, proceduresthat use new technologies have notexperienced reductions in the volume ofservices provided to beneficiaries. Thus,we do not make an allotment for costincreases due to scientific andtechnological advances when estimatingASC cost changes in the coming year. Weplan to continue monitoring changes inthe volume of ASC procedures associatedwith new technologies to ensure that

11 The revised 1998 payment rates proposed by CMS would have increased payments for several lower-volume procedures, such as arthroscopic surgery and herniarepair, which suggests that actual 1998 payment rates were less than the costs of these services.

12 This estimate is subject to revision by CMS as more recent CPI–U data become available.

payments are adequate to cover the cost ofnew technologies that enhance quality.

Productivity growth (the ratio of growth inoutputs to growth in inputs) should reducethe cost of ASC services. Measuringproductivity growth requires detailedinformation on the personnel, facilities,and other inputs used and on the quantity,quality, and mix of services (outputs)produced. Because such data are generallynot available, MedPAC has adopted apolicy standard for expected productivitygrowth that is based on growth inmultifactor productivity in the nationaleconomy. The current estimate of growthin multifactor productivity from theBureau of Labor Statistics is 0.9 percent.

By subtracting productivity growth frominput price inflation (2.7 percent), itappears that the unit cost of ASC serviceswill increase by about 1.8 percent duringthe coming year. We believe that currentpayments for ASC services are at leastadequate to cover this cost increase.13



The Congress should eliminate theupdate to payment rates for ASCservices for fiscal year 2004.

Under current law, CMS will update ASCpayment rates for fiscal year 2004 by the

projected increase in the CPI–U. Ouranalysis of ASC market factors suggeststhat current Medicare payments for ASCservices are more than adequate andshould be at least adequate to cover theexpected increase in ASC costs in fiscalyear 2004. Thus, we conclude that noupdate to ASC payment rates is necessaryfor next year.


Spending• Because this recommendation would

eliminate the current law update toASC payment rates for fiscal year2004, we estimate that it wouldreduce payments by less than $50


13 Even if we were to assume that ASCs’ input prices per unit of service will increase by the hospital market basket (projected to increase by 3.5 percent in fiscal year2004), we believe that current payments for ASC services are at least adequate to cover this cost increase.

Growth in the number of ASCs and volume of procedures provided to Medicare beneficiaries in ASCs, 1996–2001

FIGURE2F-3

Note: ASC (ambulatory surgical center).

Source: MedPAC analysis of provider of services file and 5 percent Standard Analytical File of ASC facility claims from CMS.

1996 1997 1998 1999 2000 2001

2,500

2,000

1,500

1,000

500

0

3,000

3,500

Num

ber

of

pro

cedure

s (in t

housa

nds)

/ASC

s

Procedure volume Number of ASCs

million in the first year and by lessthan $250 million over 5 years.

Beneficiary and provider• Because current Medicare payments

for ASC services are more thanadequate, we do not expect that thisrecommendation would reduceASCs’ ability to provide ambulatorysurgical services to beneficiaries.

Variations in payment forambulatory surgicalprocedures by setting

Procedures payable by Medicare whenprovided in ASCs are also performed inhospital outpatient departments and, insome cases, physician offices. Asdiscussed in the accompanying text box,many other ambulatory services can beprovided in multiple settings (see text box,p. 144). Generally, Medicare facilitypayment rates for the same surgicalprocedure vary depending on the site ofcare. For example, ASCs and hospital

outpatient departments receive differentpayment rates for the same surgicalprocedures. The 2003 ASC payment rateexceeds the 2003 outpatient departmentrate for 13 percent of the procedure codesfor which ASCs received Medicarepayments in 2001.14 These codesaccounted for 35 percent of Medicarepayments to ASCs in 2001. ASC rates arehigher than outpatient department rates for8 of the 10 procedure codes with thehighest share of Medicare payments toASCs (Table 2F-3). However, the ASCrate is lower than the hospital outpatientrate for cataract removal/lens insertion,the procedure that accounted for thelargest share (half) of Medicare paymentto ASCs in 2001.

Payment differences may reflectunderlying cost differences amongsettings, such as levels of staffing or themix of patients, or they may be due to thehistorical development of each paymentsystem. If payment variations are due tofactors other than differences inunderlying costs, there could be financialincentives to shift services between

settings, which might increase costs to theprogram and beneficiaries.

Although ASCs receive higher paymentrates than outpatient departments forcertain procedures, it does not appear thatASCs incur higher costs, on average, thanoutpatient departments for theseprocedures. In fact, outpatient departmentsare probably more costly than ASCs forsimilar procedures because they mustmeet additional regulatory requirementsand treat patients who are more medicallycomplex. Unlike ASCs, hospitals aresubject to the Emergency MedicalTreatment and Active Labor Act, whichrequires outpatient departments tostabilize and transfer patients who believethey are experiencing a medicalemergency, regardless of their ability topay. In addition, Medicare’s conditions ofparticipation for hospitals require them tocomply with patients’ rights requirements,such as establishing a patient complaintprocess, and to implement qualityimprovement programs (CMS 2003).Medicare’s conditions of coverage forASCs, which have not been updated since


Hospital outpatient department and ASC payment rates for ambulatory surgery services, 2003

Share of MedicareHospital Percent payments to ASCs,

Procedure code Description outpatient rate ASC rate difference 2001

66984 Cataract removal/lens insertion $1,160 $973 –19% 49%66821 After cataract laser surgery 246 446 81 745378 Colonoscopy, diagnostic 413 446 8 543239 Upper gastrointestinal endoscopy, biopsy 387 446 15 545385 Colonoscopy with removal of lesion by snare 413 446 8 362311 Epidural injection, lumbar or sacral 250 333 33 345380 Colonoscopy with biopsy 413 446 8 245384 Colonoscopy with removal of lesion by forceps 413 446 8 243235 Upper gastrointestinal endoscopy, diagnostic 387 333 –14 152000 Cystoscopy 329 333 1 1

Note: ASC (ambulatory surgical center). Procedures are arranged by share of Medicare payments to ASCs in 2001, from highest to lowest.

Source: CMS, program memo on update of rates and wage index for ambulatory surgical center payments effective October 1, 2002 (AB–02–124); CMS, Final rule: Medicareprogram; changes to the hospital outpatient prospective payment system and calendar year 2003 payment rates (CMS–1206–FC).

T A B L E2F-3

14 These figures are based on MedPAC’s analysis of 2003 ASC and hospital outpatient payment rates and the 5 percent Standard Analytical File of ASC facility claims,2001, from CMS.


Medicare payments for services provided in multiple settings: a larger issue

In addition to ambulatory surgicalservices, many other ambulatoryservices—including clinic visits,

many diagnostic tests, and sometherapies—can also be provided inmultiple settings. The proliferation ofsettings that provide similar servicescan improve access to care forbeneficiaries. Medicare should strive toensure that clinical considerations,rather than financial incentives, drivedecisions about the setting in whichcare is delivered.

What are some paymentdifferences by setting?Medicare payment differences bysetting do not consistently favor onesetting over another. For example, in2002, the practice expense payment toa physician for a magnetic resonanceimaging (MRI) of the brain was $403,whereas a hospital outpatientdepartment was paid $365 for the sameservice.1 In 2002, the practice expensepayment for a low-level clinic visit in aphysician’s office was $25, while thehospital outpatient department facilityfee for the same visit was $54.2

Hospital outpatient departmentpayment rates for chemotherapy drugs,which are based on hospitals’ reportedcosts, are lower than payment rates forchemotherapy drugs delivered inphysician offices and clinics (which arebased on 95 percent of the drug’saverage wholesale price).

Do payment policies influence thesetting and organization of care?Payment differences may affectproviders’ decisions regarding whichorganizational structures to adopt andwhich services to provide in a givensetting. Differences in payment that aredriven by differences in the cost ofproviding a service should notinfluence these decisions. However,differences in payment that affect theprofitability of providing a specificservice in one setting versus anothermay do so. Fully assessing the impactof payment differences on how care isorganized and where it is deliveredrequires a better understanding of thecosts of providing care in each setting,the types of patients who receive carein each setting, and how physicians andbeneficiaries decide where care isreceived.

How is the provision of serviceschanging?In recent years, settings that specializein certain services have grown. Forexample:

• The number of ambulatory surgicalcenters, which often specialize inparticular surgical procedures,doubled between 1991 and 2001(MedPAC analysis of provider ofservices file from CMS).

• Single-speciality hospitals, whichprovide both inpatient andoutpatient care, are emerging for

cardiac care, orthopedics, andcancer care (Hospitals and HealthNetworks 2002).

• Providers are also developingspecialized ambulatory facilities foroncology and cardiac care (Deverset al. 2001).

The growth of specialized settingscould be driven by the higherprofitability of certain services in onesetting versus another or by providers’desire to specialize in higher profitservices within a setting (such ascardiac care in an inpatient hospital).Particular services have shifted fromone setting to another. For example, arecent MedPAC analysis shows that anumber of more sophisticated services,including MRI, radiation therapy, andmany cardiac services, are increasinglyprovided in physicians’ offices orclinics rather than hospitals’ outpatientdepartments.3

What are the implications forpatient care?A better understanding of the quality ofcare provided in alternative settings—including safety, regulatory oversight,and clinical considerations—is needed.Existing clinical guidelines typically donot address the site of care. Originalresearch is required to develop the toolsnecessary to determine what impact thesetting of care may have on quality andoutcomes. �

1 The practice expense payment accounts for the cost of office-based resources used in providing the service.

2 These are payments for Healthcare Common Procedure Coding System (HCPCS) code 70551, MRI of brain without contrast, and HCPCS code 99213,office/outpatient visit, established patient.

3 MedPAC analysis of the 5 percent Standard Analytical File, 1999 and 2000, from CMS.

1982, do not contain these requirements(Office of Inspector General 2002).15

By comparing the characteristics ofpatients who received similar proceduresin ASCs or hospital outpatientdepartments, we found that outpatientdepartments serve patients who are moremedically complex than ASCs. It isprobably more costly to provide surgicalprocedures to patients with more healthproblems. For example, patients in worsehealth may require additional monitoringduring the surgery and recovery period.We first compared the average risk scoresof patients who received similarprocedures in an ASC or outpatientdepartment in 1999.16 The risk scores

represent beneficiaries’ expected serviceuse given their health status, relative tothat of the national average beneficiary.Expected use is based on the beneficiary’srisk category, which reflects age, sex, anddiagnoses from hospital inpatient, hospitaloutpatient, and physician visits during theprevious year (1998), and on the nationalaverage historical spending perbeneficiary in each risk category.

Because outpatient departments are morelikely than ASCs to perform services suchas cardiovascular procedures that areassociated with higher-risk patients, it isimportant to control for the type ofsurgical procedure provided whencomparing risk scores between settings.

Thus, we calculated average risk scoresfor patients who received similar types ofprocedures, such as cataract removal orcolonoscopy. For the 10 categories ofprocedures with the highest share ofMedicare payments to ASCs, patients whowere treated in outpatient departments hadsomewhat higher average risk scores thanASC patients (Table 2F-4).

We also compared average total Medicarepayments for all services for beneficiarieswho received similar procedures in ASCsand hospital outpatient departments in1999. Total payments represent healthcare use and could reflect beneficiaries’health status: Use of services shouldincrease as health status declines.


15 Medicare’s conditions of coverage for ASCs require them to assess and maintain the quality of care they provide, which is less stringent than the requirement forhospitals to conduct specific performance improvement projects.

16 The risk scores were derived from the hierarchical condition category risk adjustment model.

Average risk scores for Medicare beneficiaries receiving surgical procedures in ASCs and outpatient departments, 1999

Average risk score for beneficiaries inShare of

Outpatient Percent Medicare paymentsProcedure category ASCs departments difference to ASCs, 1999

Cataract removal/lens insertion 1.25 1.28 2% 54%Other eye procedures 1.31 1.37 5 11Colonoscopy 1.15 1.22 6 11Other ambulatory procedures 1.33 1.38 4 7Upper gastrointestinal endoscopy 1.32 1.44 9 6Ambulatory procedures—musculoskeletal 1.09 1.22 12 3Cystoscopy 1.43 1.50 5 2Ambulatory procedures—skin 1.45 2.26 56 1Arthroscopy 0.90 0.99 10 1Minor procedures—other 1.58 1.73 9 1

Note: ASCs (ambulatory surgical centers). Procedure categories are based on CMS’s Berenson-Eggers Type of Service classification scheme.Each category includes several procedure codes. This table includes the 10 procedure categories with the highest share of Medicare payments to ASCs in 1999.These categories accounted for 97 percent of payments to ASCs in 1999. This analysis includes only procedures that were payable by Medicare in ASCs in 1999.Risk scores are based on the hierarchical condition category risk adjustment model, which predicts beneficiaries’ expected service use in 1999, given their health status,relative to that of the average beneficiary. Expected use is based on each beneficiary’s age, sex, and diagnoses from inpatient, outpatient, and physician visits in 1998.The risk score differences between settings are statistically significant (1 percent level). The average risk score across all Medicare beneficiaries is 1.0.Other eye procedures include after cataract laser surgery.Other ambulatory procedures include interventional pain management procedures (such as epidural injection and facet joint block), dilation of esophagus, and septoplasty.Ambulatory procedures—musculoskeletal include services such as hammertoe operation, tendon sheath incision for finger, arthrotomy, tenotomy, and tendon repair.Ambulatory procedures—skin include services such as skin debridement, excision of lesion, wound repair, and skin graft.Minor procedures—other include certain nasal, oral, urological, and nerve procedures.

Source: MedPAC analysis of the 5 percent Standard Analytic File of Medicare claims, 1998 and 1999, from CMS, and CMS’s Berenson-Eggers Type of Service classificationscheme.

T A B L E2F-4

However, health care use also could beaffected by other factors, such asvariations in supplemental coverage,access to providers, and regional practicepatterns. Total Medicare payments includeboth Medicare spending and beneficiarycost sharing for all services used by thebeneficiary, including inpatient,ambulatory, and post-acute care. Wecontrolled for geographic adjustments topayment rates by using nationallystandardized rates. As with our analysis ofrisk scores, we controlled for the differentmix of services in each setting byseparately calculating average totalpayments for beneficiaries who receivedservices in each category of procedures.

For each of the 10 categories ofprocedures with the highest share ofMedicare payments to ASCs, beneficiarieswho received care in outpatientdepartments had substantially higher totalservice use than patients who were treatedin ASCs (Table 2F-5). These results are

consistent with the results of our analysisof beneficiaries’ average risk scores ineach setting. Together, these studiesindicate that, compared to ASCs,outpatient departments serve patients whoare more medically complex.

Our comparison of regulatoryrequirements and patient characteristics inASCs and outpatient departmentsindicates that outpatient departments areprobably the more costly setting. Thus, theexistence of ASC rates that are higherthan hospital outpatient rates is probablynot due to higher costs in the ASC settingbut instead related to the separatedevelopment of the payment systems foreach setting. The ASC payment systemcurrently sets rates for 9 payment groupsbased on 1986 cost data, while the 2003hospital outpatient PPS sets rates for 570APC groups based on 2001 cost data.Because the higher payment rates forcertain procedures performed in ASCs donot appear to be related to higher costs in

the ASC setting, these payment variationscould create financial incentives toinappropriately shift services fromoutpatient departments to ASCs.


Until the Secretary implements arevised ASC payment system, theCongress should ensure that paymentrates for ASC procedures do notexceed hospital outpatient PPS ratesfor those procedures, after accountingfor differences in the bundle ofservices covered.


Spending • Because this recommendation would

lower ASC payment rates forprocedures in which the ASC ratecurrently exceeds the hospitaloutpatient PPS rate, after adjustingfor differences in the bundle ofservices covered, we estimate that it


Average total Medicare payments for all services for beneficiaries receiving surgical procedures in ASCs and outpatient departments, 1999

Average total payments for beneficiaries inShare of

Outpatient Percent Medicare paymentsProcedure category ASCs departments difference to ASCs, 1999

Cataract removal/lens insertion $6,948 $8,044 16% 54%Other eye procedures 6,584 7,796 18 11Colonoscopy 6,254 7,088 13 11Other ambulatory procedures 8,494 11,033 30 7Upper gastrointestinal endoscopy 8,672 10,784 24 6Ambulatory procedures—musculoskeletal 6,236 9,410 51 3Cystoscopy 9,508 11,194 18 2Ambulatory procedures—skin 9,759 24,990 156 1Arthroscopy 5,539 8,109 46 1Minor procedures—other 10,035 12,600 26 1

Note: ASCs (ambulatory surgical centers). Procedure categories are based on CMS’s Berenson-Eggers Type of Service classification scheme.Each category includes several procedure codes. This table includes the 10 procedure categories with the highest share of Medicare payments to ASCs in 1999.These categories accounted for 97 percent of payments to ASCs in 1999. This analysis includes only procedures that were payable by Medicare in ASCs in 1999.Total payments include both Medicare spending and beneficiary cost sharing for all services used by beneficiaries, including inpatient, physician, ambulatory, and post-acute care.Medicare payments are based on nationally standardized payment rates.The differences in average total payments between settings are statistically significant (1 percent level).Other eye procedures include after cataract laser surgery.Other ambulatory procedures include interventional pain management procedures (such as epidural injection and facet joint block), dilation of esophagus, and septoplasty.Ambulatory procedures—musculoskeletal include services such as hammertoe operation, tendon sheath incision for finger, arthrotomy, tenotomy, and tendon repair.Ambulatory procedures—skin include services such as skin debridement, excision of lesion, wound repair, and skin graft.Minor procedures—other include certain nasal, oral, urological, and nerve procedures.

Source: MedPAC analysis of the 5 percent Standard Analytic File of Medicare claims, 1999, from CMS, and CMS’s Berenson-Eggers Type of Service classification scheme.

T A B L E2F-5

would reduce Medicare payments bybetween $50 million and $200million in the first year and between$250 million and $1 billion over 5years. These estimates are based oncomparisons of the ASC andoutpatient base rates for the sameprocedures. They do not includeadjustments to account fordifferences in the bundle of servicescovered in each setting or changes inthe provision of ASC services thatmight result from payment ratechanges.17

Beneficiary and provider• We estimate that this

recommendation would lower ratesfor about half of ASC services

(weighted by the volume of servicesprovided to beneficiaries in 2001).These procedures, which account forabout 35 percent of Medicarepayments to ASCs, would experienceaverage payment reductions of 20percent. Overall, ASC paymentswould be reduced by about 7 percent.

• The impact of this recommendationon individual ASCs would vary bythe services offered by the facility.Table 2F-6 shows the paymentimpact of implementing thisrecommendation by procedurecategory. Each category includesseveral procedure codes. Althoughpayments for cataract removal/lensinsertion (the highest-volume

category of ASC services) would notbe affected, payments for other eyeprocedures (primarily after cataractlaser surgery) would be reduced byalmost 30 percent. Almost half ofASCs provide ophthalmologyprocedures (Table 2F-7, p. 148).Payments for gastrointestinalprocedures would be reduced byabout 8 to 11 percent. About 40percent of ASCs furnish theseprocedures. Single-specialty ASCsproviding a limited range of servicesfor which payments are reducedwould be disproportionately affectedcompared to multispecialty ASCs,which could spread paymentreductions across a broader service


17 The estimates are from a model of ASC payments that is based on 2003 ASC and hospital outpatient PPS payment rates and 2001 ASC service volume.

Estimated impact of limiting ASC payment rates to hospital outpatient rates, by procedure category

Estimated percent Share of Average 2003 Average 2003reduction in 2003 Medicare payments ASC rate ASC rate

Procedure category ASC payments to ASCs, 2001 (current law) (if rates limited)

Cataract removal/lens insertion 0% 50% $971 $971Colonoscopy 8 13 446 411Other eye procedures 29 10 493 351Upper gastrointestinal endoscopy 11 7 425 377Minor procedures—musculoskeletal 19 5 335 273Other ambulatory procedures 2 3 435 425Ambulatory procedures—musculoskeletal 1 3 505 501Cystoscopy 2 2 390 382Arthroscopy 0 2 604 603Ambulatory procedures—skin 15 1 480 410

Average across all procedures 7

Note: ASC (ambulatory surgical center). Procedure categories are based on CMS’s Berenson-Eggers Type of Service classification scheme. Each category includes severalprocedure codes. This table includes the 10 procedure categories with the highest share of Medicare payments to ASCs in 2001. These categories accounted for 95percent of payments to ASCs in 2001. Average ASC rates are the average of the rates for the procedure codes in each category, weighted by each code’s servicevolume. The estimated reductions in 2003 ASC payments assume that ASC payment rates would not exceed hospital outpatient base rates for the same procedure. Theestimates do not include adjustments to account for differences in the bundle of services covered in each setting or changes in the provision of ASC services that might resultfrom payment rate changes. Other eye procedures include after cataract laser surgery.Minor procedures—musculoskeletal include interventional pain management procedures (such as epidural injection and facet joint block), soft tissue biopsy, tumor excision,and closed treatment of certain fractures.Other ambulatory procedures include services such as breast biopsy, nasal polyp excision, abscess drainage, dilation of esophagus, and septoplasty.Ambulatory procedures—musculoskeletal include services such as hammertoe operation, tendon sheath incision for finger, arthrotomy, tenotomy, and tendon repair.Ambulatory procedures—skin include services such as skin debridement, excision of lesion, wound repair, and skin graft.

Source: MedPAC model based on 2003 ASC and hospital outpatient payment rates and 2001 volume of ASC services from 5 percent Standard Analytic File of ASC facilityclaims, 2001, from CMS.

T A B L E2F-6


Distribution ofambulatory surgicalcenters by specialty

type, 2001

Specialty type Percent of ASCs

Ophthalmology 48%Plastic surgery 45Gastroenterology 40Orthopedic 38General surgery 35Gynecology 35Otolaryngology 35Podiatry 35Urology 28Pain management 22

Note: ASC (ambulatory surgical center). ASCs mayoffer services in more than one specialty.Data include both Medicare-certified and non-Medicare-certified ASCs.

Source: SMG Marketing Group, Inc., 2002.

T A B L E2F-7

line. Single-specialty andmultispecialty facilities each accountfor roughly half of all ASCs(Federated Ambulatory SurgeryAssociation 2002).

• Based on our analysis of paymentadequacy, we do not expect ASCs toprovide fewer procedures as a resultof this recommendation. However,even if ASCs provide fewerambulatory surgical services, we donot expect beneficiaries’ access tothese services to be reduced becausethey can be received in alternativesettings. Reductions to ASC paymentrates also would lower beneficiarycost sharing.

This recommendation refers to the totalMedicare payment received by ASCs andhospital outpatient departments (theprogram’s portion of the payment plus thebeneficiary’s cost sharing). Becausedifferent payment systems apply to ASCsand outpatient departments, the servicebundle for the same procedure may not beequivalent in each setting. Differences inthe bundle of services should be taken intoaccount when comparing ASC andoutpatient hospital payments for the sameprocedure. For example, if a surgicalprocedure does not normally require animaging or radiology service, theprocedure’s payment rate in each settingwill not reflect the cost of this additionalservice. In some cases, however, thephysician performing the procedure maydecide that it is clinically important to usean imaging service (such as usingfluoroscopy to enhance the surgeon’s fieldof vision). Although an outpatientdepartment could bill Medicare for both

the surgical procedure and the imagingservice, an ASC is not permitted to billseparately for ancillary services, such asimaging or radiology services. Thus, anASC that provided an imaging service inconjunction with a surgical procedurewould not be separately reimbursed for itscost. Payments for services that aresometimes provided in connection with asurgical procedure but are not part of theprocedure payment rate should beaccounted for when comparing paymentrates in ASCs and outpatient departments.

Another issue that affects thecomparability of payment rates betweensettings is whether the cost of drugs ordevices used in a procedure is part of thepayment bundle. Outpatient departmentsmay receive pass-through payments forcertain new technology items, such asdrugs and devices, that are used in thedelivery of services (see Appendix A).18

Pass-through payments are provided inaddition to the service’s base payment.ASCs do not receive pass-throughpayments. To the extent that newtechnology items are used for proceduresprovided in ASCs, their costs are includedin the procedure payment rate and notreimbursed separately. On the other hand,ASCs can receive separate payments forprosthetic devices used in conjunctionwith surgical procedures, whereasoutpatient departments cannot. The cost ofprosthetic devices is included in theoutpatient PPS base payment rate.Separate payments for items used inconnection with a surgical procedureshould be considered when comparingASC and outpatient rates.

18 Most of the payments for pass-through items have been incorporated into the outpatient PPS base rates for 2003.

19 In 2002, price information from manufacturers was used to incorporate some pass-through costs into base APC rates. In 2003, hospital cost data was used to calculateall payment rates. This change in methodology generally led to lower payment rates for services using pass-through items in 2003 than in 2002. CMS took steps tolimit the change in payment from 2002 to 2003.

Because the outpatient PPS is relativelynew and its payment rates have fluctuatedin the last few years, there could be aconcern with using these rates to set aceiling for ASC payment rates. However,outpatient rates have recently fluctuateddue to technical reasons and we expectrates to stabilize in future years.19 The useof cost data from hospitals operatingunder the outpatient PPS to set outpatientrates, which was done for the first time for2003 rates, also should enhance thestability of the system. Previously,outpatient PPS rates were based on costdata from hospitals operating under theprior, cost-based, payment system. �

References

Borden W. Outpatient surgery centers stir investor interest, Reuters market news.September 30, 2002. Available athttp://biz.yahoo.com/rf/020930/health_surgerycenters_1.html.

Centers for Medicare & Medicaid Services. Medicare and Medicaid programs; hospitalconditions of participation: quality assessment and performance improvement; final rule,Federal Register. January 24, 2003, Vol. 68, No. 16, p. 3435–3455.

Centers for Medicare & Medicaid Services. Program memo on update of rates and wageindex for ambulatory surgical center payments effective October 1, 2002, No.AB–02–124. Baltimore (MD), CMS. August 28, 2002.

Devers KJ, Brewster, LR, Casalino, LP, Center for Studying Health System Change.Changes in hospital competitive strategy: a new medical arms race? Washington (DC),Emerging Health Care Market Trends: Insights from Communities. December 10, 2001.

Federated Ambulatory Surgery Association. Email to Ariel Winter. Alexandria (VA).December 31, 2002.

Health Care Financing Administration, Department of Health and Human Services.Medicare and Medicaid programs; physicians’ referrals to health care entities with whichthey have financial relationships; final rule with comment period, Federal Register.January 4, 2001, Vol. 66, No. 3, p. 855–965.

Health Care Financing Administration, Department of Health and Human Services.Medicare program; update of ratesetting methodology, payment rates, payment policies,and the list of covered surgical procedures for ambulatory surgical centers effectiveOctober 1, 1999; proposed rule, Federal Register. June 12, 1998, Vol. 63, No. 113,p. 32290–32521.

Hospitals and Health Networks. Navigating the niche. Chicago (IL), American HospitalAssociation. August 2002, Vol. 76, No. 8, p. 30–35.


Office of Inspector General. Quality oversight of ambulatory surgical centers,OEI–01–00–00450. Washington (DC), OIG. February 2002.

Office of Inspector General, Department of Health and Human Services. Medicare andstate health programs: fraud and abuse; clarification of the initial OIG safe harborprovisions and establishment of additional safe harbor provisions under the anti-kickbackstatute; final rule, Federal Register. November 19, 1999, Vol. 64, No. 223,p. 63517–63557.

Physician Compensation Report. ASCs’ main impact on income comes through timesavings. Washington (DC), Atlantic Information Services. June 1, 2002, Vol. 3, No. 6, p. 7.

Scully T. Letter to the Honorable Pete Stark. April 16, 2002.

Versel N. Banks see ASCs as good loan risks, Modern Physician. Chicago (IL), CrainCommunications, Inc. May 1, 2002, Vol. 6, No. 5, p. 19.


Access to care in theMedicare program

C H A P T E R 3

basic goal of Medicare is to ensure that elderly and disabled

Americans have access to appropriate, high-quality health

care. In this chapter, MedPAC evaluates beneficiary access

along three dimensions: (1) the health system’s capacity; (2)

beneficiaries’ ability to obtain care; and (3) access to appropriate care.

As noted in Chapter 2, our analysis finds no widespread problems with benefi-

ciaries’ access to care. Although more selective about their patients than in the

past, most physicians are accepting at least some Medicare beneficiaries. Post-

acute services are generally available, although it has become more difficult to

place the most complex patients in skilled nursing facilities. Shortages of regis-

tered nurses could affect the availability or timeliness of certain services, how-

ever, and demographic trends raise concerns about the future capacity of the

health system.

General measures of access show that elderly beneficiaries are more satisfied

with access to care than other age groups. However, as is the case for other pop-

ulations, certain beneficiaries—those in poor health, with low incomes, and

without supplemental insurance—are more likely to report difficulty than others.

In addition, some beneficiaries are not receiving appropriate preventive or pri-

mary care services.

A

C H A P T E R

Access to care in the Medicareprogram

3In this chapter

• Evaluating access to care: anoverview

• The capacity of the healthsystem to meet beneficiaries’needs

• Beneficiaries’ ability to obtaincare

• Beneficiaries’ ability to obtainappropriate care

• Conclusion


A basic goal of the Medicare program isto ensure that elderly and disabledAmericans have access to appropriate,high-quality health care. As part of itscongressional mandate, MedPACmonitors Medicare beneficiaries’ accessto care generally and the impact ofMedicare payment policies on access toMedicare covered services.

Evaluating access is a complex anddifficult task, in part because there is noagreed upon measure of what constitutesappropriate access. The Institute ofMedicine (IOM) has defined access tocare as “the timely use of personal healthservices to achieve the best possibleoutcome” (IOM 1993). In this chapter,MedPAC evaluates access using aframework that relies on three interrelateddimensions: (1) the capacity of the healthsystem to provide health care forMedicare beneficiaries, (2) Medicarebeneficiaries’ ability to obtain healthservices, and (3) Medicare beneficiaries’experiences obtaining clinicallyappropriate health care.

Evaluating access to care:an overview

Measuring access requires analysts andpolicymakers to piece together manytypes of information to create a balancedpicture. There is no simple definition ofaccess because the concept involvesquestions about both the availability andthe actual use of services. A sufficientsupply of providers does not guaranteethat beneficiaries will be able to obtaincare. Further, knowing that beneficiariesare obtaining care does not tell us whetherthey are receiving the right mix ofservices.

In addition to access beingmultidimensional, it is difficult to findvalid and precise measures of access.National data may mask problems inspecific regions or for certain types ofbeneficiaries, while data focusing on

targeted areas may not reflect the situationin other areas. Conclusions about accessdepend greatly on the types of questionsasked. And different people may answerthe same questions differently. Thelimitations of data require policymakers togather and evaluate information on accessfrom a variety of viewpoints.

Dimensions of access Taking these factors into account,MedPAC evaluates Medicarebeneficiaries’ access to care from as manyperspectives as possible along threeinterrelated dimensions.

• Capacity of the health system tomeet Medicare beneficiaries’needs. There is no generally acceptedstandard for the health systemcapacity needed to provide care forMedicare beneficiaries (e.g., a ratioof providers or specialists tobeneficiaries). One alternative is torely on indirect indicators of capacity(e.g., the supply of providers, rates ofentry and exit of providers). Theefficiency and productivity ofindividual providers may also affectthe capacity of the health system. Inaddition, it is important to evaluatethe geographic distribution ofproviders and to considerbeneficiaries’ anticipated health careneeds in both the short term and thelong term.

• Medicare beneficiaries’ ability toobtain health care. Large numbersof hospitals or physicians nationallyor in a specific region may indicatethe presence of enough healthprofessionals to provide access to theMedicare population. Such numbersdo not, however, answer the questionof whether beneficiaries are actuallyobtaining care. Even if capacity issufficient, a variety of factors, such asfinancial barriers or the presence ofcomplex medical needs, may posebarriers to beneficiaries’ obtainingcare.

• Appropriateness of the careMedicare beneficiaries receive.1

The most complex dimension ofaccess is appropriateness of care—that is, whether Medicarebeneficiaries are receiving the rightcare in the right setting at the righttime. Defining appropriate care isdifficult, but evidence-basedguidelines have been developed foran increasing number of clinicalconditions. Such guidelines, whichcall for specific procedures ortreatment regimens, can be used tomeasure appropriateness of care insome settings. They can also be usedto determine if beneficiaries arereceiving beneficial preventiveservices. In addition, certainconditions termed “ambulatory caresensitive conditions,” if treatedappropriately in the ambulatorysetting, need not result inhospitalizations. Hospital admissionsor emergency department (ED) usefor these conditions may indicateinadequate access to ambulatorysettings or services, or inadequatecare management.

Measures of access to care Conclusions about access to care dependheavily on which data are used and whichquestions are asked. Some measures focuson whether beneficiaries can find any typeof care, whereas others focus more on thewillingness of physicians to acceptMedicare patients. Other measures look atthe care experience through waiting timesor through delays in obtaining care.

Many current data are designed toproduce national estimates—providing ageneral impression of access to care—butmay mask local variation. For example,although a Center for Studying HealthSystem Change (HSC) survey ofphysicians in 2001 found that 71.1 percentof physicians overall were willing to takeall new Medicare patients—meaning thatthey accepted all new Medicare patientswho wished to make appointments—only

154 Acce s s t o ca r e i n t h e Med i ca r e p r og ram

1 Measures in this dimension of access overlap significantly with measures of quality. However, the reason behind identified problems could either be an access concern—the beneficiary did not take the time to obtain the necessary care or tried, but was not able to obtain it from a provider; or a quality concern—the beneficiary did obtaincare, but was not given the right type of care.

55 percent of physicians in Seattle werewilling to do so. Because access to care isoften driven by local market conditions, itis also important to distinguish betweenisolated problems and those that couldsignal emerging systemic problems.

Different questions may lead to differentconclusions. In the physician survey notedabove, although only 55 percent ofphysicians in Seattle said they werewilling to take all new beneficiaries, only8 percent of Medicare beneficiaries saidthey delayed or put off obtaining care. So,it is unclear from these apparentlyconflicting findings whether beneficiarieshave a problem obtaining care in Seattle.

Assessments of access to care are alsosubjective, to some degree. For example,in the MedPAC analysis discussed in alater section of this chapter, highlyeducated persons reported more problemsaccessing care than less educated persons.It seems unlikely, at least intuitively, thatthese self-reports capture a true differencein beneficiaries’ ability to obtain care.Rather, they are more likely the result ofdifferent expectations.

Measures of use are also limited in whatthey can tell us about access. Data onthese measures tell us more about howoften beneficiaries use certain servicesthan about the appropriateness of thoseservices.2

Data on access to care The ability to measure access alsodepends on the availability and utility ofdata. The three main sources of data onaccess for Medicare beneficiaries areadministrative data on the use of services,data generated directly from providers,and data generated from beneficiaries.Each of these sources providesinformation useful for evaluating severaldimensions of access.

A common concern across all these typesof data is timeliness—it is often difficultto find data that are recent enough toallow unambiguous conclusions about

current beneficiaries’ experienceaccessing care. While focus groups andsmaller surveys provide more timelyinformation, results from these sources are often not as generalizable as large,multiyear surveys or administrative data.

That said, administrative utilization dataoffer several advantages. First, such dataare routinely collected, thus minimizingthe costs of obtaining them. Second, thedata are usually extensive and provideinformation on all beneficiaries usingservices, so they can often answer manyquestions in a statistically valid manner.On the other hand, because administrativedata are collected for billing purposes, andfor tracking beneficiary eligibility andenrollment information, such data are notalways organized in a manner thataddresses policy or research questions.

Administrative data on claims paid canprovide information on the capacity of thehealth system and on the needs ofbeneficiaries over time. These data tellpolicymakers how often a certain serviceis being used, whether use has increasedor decreased, and which type ofbeneficiaries used certain types ofservices. Finally, they can provide someinformation on appropriateness of care byrevealing whether beneficiaries are usingthe right types of services. For example,they can tell policymakers how manybeneficiaries received appropriatepreventive services, such asimmunizations, and whether diabetics inthe program received a test to measuretheir glucose levels. When combined withmedical record review and clinicaljudgement, administrative data canprovide even richer information onappropriateness of care.

Data collected directly from beneficiariesor providers may be obtained throughbroad surveys, targeted surveys, structuredfocus groups, or focused interviews withindividuals. These type of data provideinformation on beneficiaries’ andproviders’ unique perceptions of access.Different types of these data have distinct

advantages and disadvantages. Large,carefully designed surveys may providebroad, valid information and—dependingon size and sample design—make itpossible to identify variations amonggroups within the surveyed population.However, large surveys can be veryexpensive and take time to administer andanalyze.

Smaller surveys and focus groups orinterviews can provide rapid response totargeted questions, but the results may beless reliable and, because samples aresmall, are not generalizable to the wholepopulation. But, because smaller surveysand focus groups or interviews make itpossible to gather more in-depthinformation, they are useful in learningmore about the reasons behind accessbarriers. They can also be used to providemore detailed in-depth targeted analysis ofsubpopulations.

The capacity of the healthsystem to meetbeneficiaries’ needs

The sector-by-sector analysis presented inChapter 2 for purposes of determining theadequacy of payment generally finds thatthere are sufficient hospitals, physicians,skilled nursing facilities, home healthagencies, outpatient dialysis facilities, andambulatory surgical centers at the nationallevel to provide Medicare beneficiarieswith access to Medicare-covered services.In the discussion that follows, we expandon analyses presented in Chapter 2. Inparticular, we focus on three areas ofparticular concern to policymakersbecause of recent payment systemchanges or other reasons—the availabilityof:

• physicians,

• post-acute services, and

• registered nurses.


2 For example, because many Medicare beneficiaries with supplemental coverage often have first dollar coverage, it is possible that these beneficiaries could be usingsome services of marginal value in addition to necessary care (MedPAC 2002).

We conclude, based on informationcurrently available, that physicians andpost-acute providers are available to mostMedicare beneficiaries. Both providertypes need to be monitored closely,however. Physicians appear to be growingmore selective about the types of newpatients they take from all insurancesources, including Medicare. Medicarebeneficiaries’ level of need for certaincomplex services seems to be a factor inskilled nursing facilities’ decisions aboutaccepting new patients. Our analysis alsofinds evidence of shortages in theavailability of nurses, which may lead toaccess problems in the future.

In assessing the capacity of the healthsystem to provide access in the comingyears, it is important to consider the futureneeds of Medicare beneficiaries. Ouranalysis suggests that the Medicareprogram will face increasing pressures onresources as the numbers of beneficiariesincrease and the needs of futurebeneficiary populations differ from thoseof current beneficiaries.

Availability of physicians Physicians treat patients in all settings andare the major directors of health care. Forthat reason, their willingness to treatMedicare beneficiaries is a criticalcomponent of access. Recent reductions inMedicare physician payment rates haveraised new concerns about beneficiaryaccess to physicians.3 However, ouranalysis does not find widespreadproblems with Medicare beneficiaries’access to physician services. According toour findings:

• most physicians are still acceptingMedicare beneficiaries in theirpractices;

• some physicians are being moreselective, but they are also beingselective about patients insured byother payers;

• physicians are as concerned about theadministrative burden of Medicare asthey are about reimbursement levels;and

• physician availability varies byregions.

Over the past several years, Medicarebeneficiaries’ access to physicians at thenational level has been good. PooledMedicare Current Beneficiary Survey(MCBS) data from 1996 to 1999 showthat only 2.4 percent of beneficiaries saidthey had trouble getting care,4 and 91percent said they had a usual doctor.While these beneficiary survey data arenot yet available for more recent timeperiods, administrative and physiciansurvey data through 2001 and 2002 do notsuggest a decline in overall access toservices.

As noted in Chapter 2B, the number ofphysicians furnishing services tobeneficiaries has kept pace with thegrowth in the beneficiary population inrecent years. From 1995 to 2001, thenumber of physicians per 1,000beneficiaries grew slightly from 12.9 to13.2. In addition, the volume of physicianservices beneficiaries use has also grown.Between 2001 and 2002 volume percapita grew by 4.3 percent. Almost all ofthese services are delivered byparticipating physicians.5 Based on claimsdata from the first six months of 2002,about 96 percent of allowed charges forphysician services were for servicesfurnished by participating physicians.

MedPAC has sponsored surveys ofphysicians in 1999 and 2002. MedPAC’s2002 survey looked at the impact of recentpayment rate reductions on physicians’willingness to accept Medicarebeneficiaries and their overall impressionof the Medicare program in comparison toother payers on a variety of aspects. Wealso compared these findings to findingsof other surveys. HSC surveyed bothbeneficiaries and physicians, but HSC’sphysician survey was conducted beforethe reduction in payment rates. TheAmerican Medical Association (AMA)sponsored an internet-based survey ofphysicians that was fielded after thepayment rate reductions were in place.

The 2002 MedPAC survey reveals that alarge majority of physicians are stilltaking some or all new Medicarebeneficiaries. In the 2002 survey, 95.9percent of physicians accepting any newpatients from any insurer were acceptingsome or all new Medicare patients.

However, each of the three surveys didshow that physicians are increasinglylimiting the proportion of their patientcare load insured by Medicare (Table3-1). Between MedPAC’s 1999 and 2002survey, the percentage of physiciansaccepting all new fee-for-service (FFS)Medicare patients fell 6.3 percentagepoints from 76.4 percent to 70.1 percent.HSC’s results were similar. Between 1997and 2001, the percentage of physicianssurveyed by HSC who said they acceptedall new Medicare patients fell from 74.6percent to 71.1 percent.6

The AMA survey, fielded after thepayment rate reductions between Februaryand April 2002, found a higher percentageof physicians—83 percent—willing to take


3 Fee updates for physicians will be –4.4 percent in 2003 and –5.1 percent in 2004 under current law. However, Medicare expenditures for physician services are stillincreasing. They rose from $42 billion in 1996 to $56 billion in 2001 and are expected to grow at an annual rate of 2 to 4 percent from 2001 to 2006, assumingthese negative updates under current law.

4 It is important to note that “trouble accessing care” applies to more than physician services. As such, it may only be an indirect indicator of beneficiaries’ ability to obtaincare from a physician.

5 The number of participating physicians is often used as an indicator of whether physicians are available to beneficiaries. However, in this chapter we use the percentageof allowed charges because it is a more direct measure of beneficiary use of participating physicians. For a more detailed discussion of the relevance of this indicatorsee Chapter 2B.

6 These numbers may not represent an appropriate comparison because the HSC survey was fielded before the physician payment rate reductions in 2002.

all new Medicare beneficiaries.7 But,similar to the other surveys, the AMAsurvey also found that physicians werebecoming increasingly selective aboutMedicare patients. When physicians wereasked whether they had decreased orrestricted the number or type of Medicarepatients they treat in the last six months, 16percent responded that they had. Another 8percent said they planned to implementsuch restrictions in the next 12 months.

The phenomenon of physician selectivityin accepting new patients is not unique toMedicare, however. The MedPAC surveyfound that physicians’ limiting theirpatient load was even more pronouncedfor patients with Medicaid or privatehealth maintenance organization (HMO)coverage. Results show that, in general,physicians view patients insured byprivate sector FFS or preferred providerorganization (PPO) options morefavorably than those of any other payer(Table 3-2).

While the HSC survey did not distinguishamong types of private insurance, it foundthat physicians were limiting theiracceptance of both Medicare-covered andprivately insured patients. The decline inthe percentage of physicians willing totake all new privately insured patients wassimilar to the decline in their willingnessto take all new Medicare patients, fallingfrom 76.2 percent in 1997 to 70.8 percentin 2001.

Analysis of the MedPAC survey resultsreveals that the level of reimbursementwas more often the reason physiciansreported for limiting acceptance of newMedicaid or HMO patients than it was forMedicare patients. However, thepercentage limiting their acceptance ofnew patients due to concern overreimbursement was slightly higher forMedicare—15.6 percent—than for privateFFS/PPO patients—15.0 percent.8

Furthermore, the share of physiciansconcerned about reimbursement who saidthey limited new Medicare patientsbecause of reimbursement levels—15.6percent—was slightly less than the sharewho said they did so based on concernsabout the administrative burden ofMedicare—16.0 percent.

Medicare beneficiaries’ access tophysicians also varied by region. TheHSC survey found that, while Bostonranked among the highest of the 12markets in physician willingness to acceptall new Medicare patients (at about 70percent), Seattle ranked near the bottom(at about 55 percent). This measure would


Share of physicians accepting all new Medicarepatients, 1997–2002

1997 1999 2001 2002

MedPAC* N/A 76.4% N/A 70.1%Center for Studying Health System Change 74.6% 72.5% 71.1% N/AAmerican Medical Association N/A N/A N/A 83.0%

Note: N/A (not applicable). *MedPAC’s survey results reflect physician acceptance of new fee-for-service beneficiaries.

Source: MedPAC Survey of Physicians 1999 and 2002; Center for Studying Health System Change 1997 and2001; and American Medical Association Survey of Physicians 2002.

T A B L E3-1

7 It is interesting that this statistic is higher than the MedPAC or HSC findings. Some have suggested that because the physicians responding to the AMA survey agreeahead of time to participate in AMA surveys, they might be more likely to voice concern over Medicare policies. Clearly, the physicians in the AMA survey haveconcerns, but they do not appear any more negative towards Medicare than respondents to the other surveys who were chosen to be more nationally representative. TheAMA survey was based on 520 respondents with a response rate of 26 percent, compared with about 800 respondents in the MedPAC survey with a 54.5 percentresponse rate and 12,500 respondents in the HSC survey with a response rate of 61 to 65 percent.

8 Seventy-seven percent of physicians voiced concern about reimbursement. Of these physicians, 15.0 percent said they limited their acceptance of new private FFS/PPOpatients, 15.6 percent did so for Medicare patients, 38.0 percent did so for Medicaid patients, and 32.4 percent did so for all other HMO patients.

Acceptance of all or some new patients, by type of patient: MedPAC physician survey

results, 1999 and 2002

PercentType of patient 1999 2002 change

Private FFS and PPO patients 97.9% 99.3% 1.4*FFS Medicare patients 96.8 95.9 –0.9Uninsured patients1 90.5 92.8 2.3HMO and other capitated plan patients2 87.6 86.3 –1.3Medicaid patients3 73.7 69.5 –4.2*

Note: FFS (fee-for-service), HMO (health maintenance organization), PPO (preferred provider organization). Analysislimited to physicians who were accepting new patients (regardless of type) in the year. The response showsthe percentage of doctors with patients of a type who are accepting new patients of that type into theirpractice. 1999 percentages were weighted to account for oversampling of selected surgical specialties.Missing values excluded from all calculations.1In 2002, uninsured included charity and self-pay patients; in 1999, it did not.2In 2002, the Medicaid category included both HMO and fee-for-service patients; in 1999, this categoryincluded only fee-for-service patients.3In 2002, the HMO category did not include Medicaid patients; in 1999, it did.*Change since 1999 is significantly different from zero at the 95 percent confidence level.

Source: MedPAC Survey of Physicians, analysis of responses to Question 27B (2002) and Question 19 (1999).

T A B L E3-2

lead one to believe that Medicarebeneficiaries have better access tophysician care in Boston than in Seattle.However, taken together with results fromother HSC questions, the difficulty ofdrawing conclusions about access fromany single measure becomes apparent. Ona different measure—delay for a check-upexceeding three weeks—Seattlebeneficiaries appeared to have betteraccess to physician services. In Boston,54.6 percent of beneficiaries reported thatthey had to wait more than three weeksfor an appointment for a check-up. InSeattle, only 24.2 percent of Medicarebeneficiaries reported such a delay.

A survey of State Health InsuranceAssistance Programs (SHIPs) and otherMedicare advocacy organizations by theMedicare Rights Center, a consumeradvocacy organization, identified eightstates where access to physiciansworsened after the payment ratereductions went into effect in 2002. TheSHIPs and others reported an increase inthe volume of calls from beneficiarieshaving difficulty finding doctors whowould accept new Medicare patients inTennessee, Missouri, Arizona, Virginia,New Hampshire, Texas, Rhode Island,and New Mexico. The Medicare RightsCenter cautions, however, that they didnot evaluate the level of increased calls—the characterization of an increase couldbe based on a handful of beneficiaries or alarge volume of calls.

Availability of post-acuteservicesPost-acute services covered by Medicareinclude skilled nursing facility servicesprovided after a hospital stay, as well ashome health care. MedPAC’s review ofavailable evidence, as discussed below,generally supports the conclusion thatMedicare beneficiaries’ access to skillednursing care and home health servicesremained stable after the implementationof prospective payment for skilled nursingand home health services. Nevertheless,Department of Health and HumanServices (HHS), Office of Inspector

General (OIG) surveys and a small focusgroup sponsored by MedPAC did find thathospital discharge planners reportedincreasing problems in placing patientswith particularly complex health problemsin skilled nursing facilities (SNFs), and toa lesser extent home health agencies(HHAs), since the implementation ofprospective payment systems (PPSs) inthese sectors.

SNF services

Available evidence suggests that thecapacity of SNFs to meet Medicarebeneficiaries’ post-acute care needs hasremained relatively stable over the lastseveral years from 1998 to 2002 (seeChapter 2C). Although 26 percent ofhospital-based SNFs closed, the increasein the number of freestanding SNFsappears to have offset these closures. Infact, the number of covered daysincreased 4 percent from 1999 to 2000.

Opinion data from discharge planners alsosuggest that most Medicare beneficiarieshave access to SNF services, althoughcertain types of beneficiaries mayexperience more problems than others. Ina series of studies by the OIG from 1999to 2001, hospital discharge plannersreported that beneficiaries generally hadaccess to SNF care.9 About 5 percent ofhospital patients who needed SNF carewere described as being difficult to place,as defined by whether the surveyeddischarge planner reported a delay inplacement. Patients for whom SNFplacements were difficult werecharacterized as patients for whom carewas costly. Discharge planners said thatpatients needing rehabilitation services—for whom Medicare pays moregenerously—were not difficult to place.

In October 2002, MedPAC convened afocus group of 15 hospital dischargeplanners from a variety of regions andtypes of hospitals to discuss the impact ofMedicare’s prospective payment systemsfor skilled nursing facilities and homehealth care on Medicare beneficiaries’access to post-acute care (see text box).The findings from this focus group were

consistent with the OIG findings. Sincethe implementation of Medicare’s SNFprospective payment system, hospitaldischarge planners reported they have hadno problems getting SNFs to acceptpatients requiring rehabilitation services.However, they reported increaseddifficulty in getting SNFs to acceptpatients with particularly complex andcostly health problems, even when bedswere available.

What happens to beneficiaries who stay inthe hospital longer because they cannot beplaced in a SNF? Focus group participantstold us that some patients are eventuallyplaced in a SNF, but some are neverplaced and stay in the hospital until theycan be discharged home. However, it isnot clear whether longer hospital staysshould be characterized as an accessproblem. Even though discharge plannersmay believe that the patient is ready toleave the hospital and be admitted to askilled nursing facility, these patients maybe able to obtain appropriate care in thehospital.

Home health care

There has been a sizeable drop in homehealth agencies and use of home healthservices, but this drop followed a periodof dramatic increases in each. Twenty-four percent of home health agenciesclosed between 1997 and 1999. Since thattime the number of home health agencieshas remained stable, with the numbersentering the Medicare program roughlyequivalent to those leaving. In addition,fewer beneficiaries have been usingMedicare’s home health benefit since1997.

The declines in the number of agenciesand the use of services occur in ahistorical context that includes severalyears prior to the implementation ofMedicare’s interim payment system andPPS for home health. During this timeperiod the number of agencies,beneficiaries who used home health, andvisits per beneficiary were increasingdramatically. The percentage of Medicarefee-for-service beneficiaries using home


9 The OIG did not repeat their survey in 2002 and has no plans to do so in 2003. MedPAC is recommending in this report that this series of surveys be continued.

health services in 2001 was 5.5 percent,similar to the percentage in 1991 (6.5percent). But, at the high point of usage in1996, 9.0 percent of Medicarebeneficiaries used home health services.The increase in use of services in the earlyand middle 1990s was the primary reasonthat Congress implemented the PPS,clarified eligibility rules, and strengthenedfraud and abuse enforcement in theprogram.

Despite the decrease in use due to policychanges, the OIG surveys and MedPACdischarge planner focus group (see textbox) did not identify widespread problemsfinding home health care for beneficiaries

discharged from hospitals. The OIGsurvey did reveal that it was difficult toplace a small subset of hospital patientswho needed home health care. Thesepatients tended to be those with morecomplex care needs.

We have less information on beneficiariesreferred to home health care from thecommunity. While the MedPAC andearlier OIG surveys did not address thistopic, another 2001 OIG survey did. TheOIG generally found the reportedexperiences of “community beneficiaries”to be similar to those of beneficiariesdischarged from the hospital intoMedicare home health services.10

Availability of nursesThe supply and retention of registerednurses is an important concern for theentire health system. The Bureau ofHealth Professions within HHS hasreported a growing shortage of nurses,which is expected to worsen by 2010 andthereafter. In a recent survey, hospitaladministrators report historically highvacancy rates for nurses, as well as othertypes of personnel (First ConsultingGroup 2001).11 Nursing homes, homehealth agencies, health systems, and otherorganizations have also reporteddifficulties filling nursing positions.


Impact of Medicare’s SNF and home health PPS on access to post-acute care:findings from MedPAC’s discharge planner focus group

In October 2002, MedPACconvened a focus group of 15hospital discharge planners from a

variety of regions and types of hospitalsto discuss the impact of the skillednursing facility (SNF) and home healthprospective payment systems onbeneficiaries’ access to post-acute care.Focus group participants told us thatSNF placements are delayed for certaintypes of patients at least one daybetween 5 percent and 25 percent of thetime, with some beneficiaries withdelayed placements remaining in thehospital for significant periods of time.It was unclear from the discussionwhether this delay resulted in theinability of beneficiaries to obtainappropriate care. According to thedischarge planners, under the PPS,patients requiring rehabilitationservices—for whom Medicare paysmore generously—have no problembeing placed in SNFs. Hard-to-placepatients included those:

• needing dialysis,

• needing expensive medications,

• needing ventilator services,

• requiring total parenteral nutrition,

• having infectious diseases, and

• having mental illness or cognitiveimpairment.

Patients with infectious diseases, end-stage renal disease, and mentalimpairments were difficult to placebefore the PPS was implemented, butthe discharge planners said placementwas even more difficult afterimplementation. They suggested thatSNFs were eager to take rehabilitationpatients because payments for theseservices were more generous.

The delays in placing patients post-PPSdo not necessarily relate to the lack ofavailable beds in freestanding SNFs,according to the discharge planners.

Even when beds are available,freestanding SNFs often will not takecomplex patients.

The focus group was not as concernedabout placing beneficiaries needinghome health care as they were aboutplacing those needing SNF care.However, a few planners said that itwas harder to place beneficiaries forhome health services if they:

• lived in rural areas, especially iftherapy, such as physical or speech,was needed;

• required extensive supplies, such aswound treatments; or

• were unable to remain safely athome.

The group did not indicate that thesebeneficiary groups were either newlyhard-to-place or more hard-to-placeafter the implementation of the PPS. �

10 These findings were based on surveys of 21 physicians, 30 home health agencies, 60 aging network representatives, and beneficiaries already receiving services in10 states. The OIG has no plans to repeat the study. MedPAC is recommending in this report that the Secretary continue this type of study on beneficiaries’ access tohome health services.

11 These administrators also noted difficulty filling positions for other personnel such as clinical pharmacists and imaging technicians. MedPAC will continue to monitorthese capacity issues, but limits the discussion in this chapter to nurses.

In 2000, the Bureau of Health Professionsat HHS calculated a shortage of 110,000nurses, which represents a 6 percent gapbetween the supply of full-time equivalentregistered nurses and the demand forthose nurses. As illustrated in Figure 3-1,this gap is expected to grow to 12 percentby 2010 and then worsen dramatically toaround 20 percent by 2015, whenproviders will face unprecedented demandfrom Medicare beneficiaries seekingservices.

Available data do not provide firmevidence that this shortage hascompromised access for Medicarebeneficiaries. However, in a recent survey,hospital administrators cited emergencydepartment diversions, emergencydepartment overcrowding, and lesserability to staff beds as the top threeproblems caused by nursing and otherpersonnel shortages (First ConsultingGroup 2001). These problems could leadto delays in receiving inpatient and urgent

care, as well as increased pressure forearly discharges.12 In addition, as thenumber of beneficiaries increases in thefuture, ensuring an adequate supply ofnurses and other health professionals tomeet the growing needs of Medicarebeneficiaries will be critical.

The reasons behind these shortages havebeen documented in several studies andreports. Factors include an aging existingworkforce, fewer young persons choosingand graduating from nursing school,dissatisfaction with the work environment,an increasing number of choices of placesto work both in and out of nursing, anddeclining relative earnings. The averageage of working registered nurses hasincreased, resulting in a higher proportionof nurses who are approaching retirementage. Nurses also retire at an earlier agethan other workers, often in their mid- andlate-50s (Berliner and Ginzberg 2002).And not all of these older nurses are beingreplaced by new graduates entering the

workforce. Twenty-six percent fewerpeople graduated from nursing schools in2000 than in 1995 (Bureau of HealthProfessions 2002). One reason youngerpersons may not be choosing nursing isthat real earnings, the amount availableafter adjusting for inflation, have beenrelatively flat since 1991 (Bureau ofHealth Professions 2002).

Some observers suggest that the nursingshortage may be cyclical and thereforeaddressed over time by market forces,such as increases in wages. Most expertson the shortage of nurses, however,suggest that the gap between individualsentering the nursing workforce and theaging of the current nursing workforce istoo large to be addressed by higher wages.They also point to data suggesting thatdissatisfaction with working conditions,rather than low wages, is one of theprimary reasons nurses are retiring earlyand fewer persons are entering theprofession. Aiken and colleagues report in


Supply and demand projections for full-timeequivalent registered nurses, 2000–2020

FIGURE3-1

Source: Bureau of Health Professions, registered nurse supply and demand projections.

2000 2004 2008 2012 2016

1,500

1,000

500

0

2,000

2,500

3,000

Num

ber

of

nurs

es (

in t

housa

nds)

SupplyDemand

12 While not directly discussed in this chapter, much research has also focused on the impact that nursing shortages have on the quality of care, both in nursing homesand hospitals.

a recent article that more than 40 percentof nurses working in hospitals aredissatisfied with their jobs. Recent datafrom New York City indicate that 50percent of new nurses in hospitals leavebefore their second year of employment(Berliner and Ginzberg 2002).

Beneficiaries’ needs forhealth services: healthsystem capacity to provideaccess in the coming yearsIn assessing the capacity of the health caresystem to provide access to Medicarebeneficiaries, it is important to understandbeneficiaries’ needs for services and howtheir needs can be expected to change inthe coming decades. In this section weconsider beneficiaries’ future needs bylooking at anticipated changes in theprevalence of certain conditions, race andethnicity, age, and gender, and how thesechanges will challenge the health caresystem.

MedPAC’s assessment of currentcapacity, as described in Chapter 2, is thatfor the most part the health system isadequate to meet beneficiaries’ needs.However, the rate of increase in thenumber of Medicare beneficiaries isexpected to be higher than the rate ofincrease in the overall population,doubling over the period 2000 to 2030.The dramatic rise in the number ofbeneficiaries and changing demographicsof Medicare beneficiaries may alter thetypes of services needed in the future.

More beneficiaries will mean that, basedon current patterns of use, demand foralmost all services will rise.13 Althoughnew technology breakthroughs andtreatment modalities could change theway care is delivered, current utilizationpatterns provide a reasonable baseline forpredicting which types of services will bein greater demand. Combined withinformation on the types of beneficiaries

who have greater difficulty obtaining care,this analysis may also identify populationsthat could need more careful monitoringin the future.

In addition to specific types of services forbeneficiaries, a healthcare workforce withthe skills necessary to treat an older anddisabled population will be needed.Although it may be difficult to providegeriatric training to enough newphysicians to appropriately treat anincreasingly elderly patient population, itmay be possible to train currentprofessionals to better manage thesepatients.14

Beneficiaries’ care needs by age

Analysis of census data shows that thefastest growing segment of Medicarebeneficiaries has been individuals over

age 85, typically referred to as the “old,old.” Even though the rate of growth forthis subpopulation is expected to fall fromits recent high levels, this population willbe significantly larger in the future.

Beneficiaries over age 85 use more of allservices than younger beneficiaries, someof which are Medicare-covered and otherswhich are not. They use disproportionatelymore home health and SNF services thanother populations (Figure 3-2).Beneficiaries this age who use the ED doso more frequently than youngerbeneficiaries. The greatest driver of totalhealth care costs for the over-85population is nursing home expenditures.However, the vast majority of these arepaid for by Medicaid and out-of-pocketpayments (Spillman and Lubitz 2000).


13 While there is some evidence that disability and functional capacity have improved on some measures among the elderly over the last decade, research has not showna clear trend with regard to the most severe forms of age-related disability, such as cognitive impairment and the ability to perform activities of daily living; thereforethese improvements may not affect usage patterns in the future (Freedman, Martin, Schoeni 2002).

14 A report issued jointly by the Merck Institute of Aging and Health and the Gerontological Society of America speculates that even if geriatric training was mandatory inevery medical school, it would take more than 40 years for physicians with geriatric training to replace those without such training. They suggest this would be too longto meet beneficiaries’ needs and that training the existing health work force would be more effective.

Per capita Medicare spending for beneficiariesages 65–69 and 85 and over,

selected services, 1998

FIGURE3-2

Source: MedPAC analysis of National Center for Health Statistics data.

7,000

6,000

5,000

8,000

4,000

3,000

2,000

1,000

065–69 85+

Spen

din

g (

dolla

rs)

Skilled nursing Hospital inpatient

PhysicianHome health

Beneficiaries’ care needs by race and ethnicity

The racial and ethnic composition of theMedicare population is expected tocontinue to change in coming years.Census data show that, from 2000 to2050, the percentage of the Medicarepopulation that is white, non-Hispanic isexpected to decrease from 84 to 64percent. The growth in the number ofHispanics is even greater than the increasein the number of African Americans, sothat by 2050 there will be more Hispanicsthan African Americans in Medicare.

Minorities in Medicare are less likely toreport having a usual doctor or a usualsource of care other than either an urgentcare center or ED. CMS data from the2000 MCBS show that approximately 10percent of both Hispanics and AfricanAmericans report they use an urgent carecenter or the ED as their usual source ofcare (compared with 2 percent of non-Hispanic whites). As described later inthis chapter, the ED provides importanturgent care, but some of this urgent care isfor acute manifestations of chronicconditions that could be more efficientlymanaged through an ongoing relationshipwith a physician. If minorities’ reliance onthe ED as a usual source of carecontinues, access problems may growwith increases in the numbers of minoritybeneficiaries.

National Center for Health Statistics(NCHS) data also show that AfricanAmericans over 65 use more home healthservices than other populations so thisservice use will also need to be monitoredin the future.

Beneficiaries’ care needs byhealth conditions

Data from the Federal Interagency Forumon Aging-Related Statistics show that theleading causes of death for those over 65are heart disease, stroke, cancer, diabetes,

chronic obstructive pulmonary disease,pneumonia, and influenza. Five of theseare chronic conditions. While death rateshave fallen for heart disease and stroke,their prevalence and that of otherimportant chronic conditions has not.Between 1984 and 1995, prevalence ratesof heart disease, stroke, diabetes, andcancer all increased.

The prevalence of chronic conditions alsovaries by race and ethnicity. AfricanAmericans in 1995 were more likely tohave diabetes, stroke, or hypertension thanwhites or Hispanics. However, whiteswere more likely to report cancer thaneither African Americans or Hispanics(Federal Interagency Forum on Aging-Related Statistics 2000).

The increasing prevalence of chronicconditions and the ability to manage themcould have several effects on the servicesneeded by Medicare beneficiaries in thefuture. Because these conditions are long-term diseases, they require ongoing caremanagement to prevent acute episodesfrom occurring and may affect the type ofcare provided for acute episodes for otherconditions. Caregivers will need to be ableto coordinate multiple needs and treatmentregimens across settings and over time. Inaddition, services that prevent acuteepisodes and/or increased disease severity,such as blood pressure screening andmanagement, diabetic checks, andmammograms, will be increasinglyimportant as the prevalence of chronicconditions increases.

Care needs of disabledbeneficiaries

The number of persons eligible forMedicare on the basis of disability isincreasing. The population under age 65who qualify for Medicare on the basis of adisability has grown from 2.2 millionpeople in 1975 to 5.6 million in 2000; thisnumber is projected to reach 8.8 million in2017 (MedPAC 2002).

While nearly two-thirds of the currentMedicare disabled have physicaldisabilities, the remainder qualify on thebasis of a mental disorder. Thesebeneficiaries account for adisproportionate amount of Medicarespending (Foote and Hogan 2001).Because people with mental conditionsusually qualify for Medicare at a muchyounger age, and therefore are eligible forMedicare for a longer period of time, theywill continue to become a largerproportion of the disabled population.15

This could mean an increase in the needfor psychiatric services and forappropriate management ofpharmaceuticals specific to mentalconditions.

Beneficiaries’ care needs by sex

Women make up a disproportionate shareof Medicare beneficiaries and willincrease as a percentage of the over-65population in the future (FederalInteragency Forum on Aging-RelatedStatistics 2000). Females make up 56percent of the overall Medicare populationand 71 percent of the over-85 population.Women live alone more often than men.According to 2000 MCBS data, 72percent of the nearly 30 percent ofMedicare beneficiaries who lived alonewere women. Women are also more likelyto have lower incomes and less likely tohave employer-sponsored supplementalinsurance (Schoen et al. 1998).

MedPAC analysis presented in the nextsection finds that lower incomes areassociated with difficulty obtaining care,so the increase in the number of womenwith lower incomes could heighten accessconcerns in the future. In addition,women’s lack of a caregiver at home maymean that the need for home healthservices will increase and that morebeneficiaries may need to be admitted tonursing homes.


15 The reason the age of eligibility affects the proportion is that every year fewer persons with mental conditions leave the program compared with the elderly and thephysically disabled. The physically disabled tend to sign up for Medicare closer to the age of 65 and thus are usually eligible for Medicare benefits for a shorter periodof time.

Beneficiaries’ ability toobtain care

Medicare beneficiaries in general reportgood access to health care services.Compared with younger populations,Medicare enrollees appear to have betteraccess to care and, over time, they reportimproved access. Certain subpopulations,however, report higher levels of accessproblems than others. All else being equal,those in poor health, those who live inpoverty, and those without supplementalhealth coverage report higher levels ofaccess problems. Finally, the disabledunder-65 population reports substantiallyhigher rates of access problems than theaged Medicare population.

In this section, we first review current dataon access to care among the Medicarepopulation. Second, we present the resultsof a MedPAC study on the influence ofbeneficiary characteristics on access tocare for the Medicare population.

What is currently knownabout Medicarebeneficiaries’ ability toobtain care? Overall, Medicare beneficiaries say theyhave good access to services. Thisperception has become more prevalentover time and is stronger for the elderlythan for any other age group. Beinginsured by Medicare has made it possiblefor an otherwise difficult-to-insure andfrailer population to have access to care.According to 1999 MCBS data, 94.3percent of beneficiaries reported that theyhad a usual source of care and only 4percent reported that they had troublegetting care. MCBS data also indicate thatthe percentage of Medicare beneficiariesdelaying care because of cost declinedover the time period 1991 to 2000.

Another access measure, the percentagenot seeing a doctor during the past year,also declined during this time period. Datafrom the National Health InterviewSurvey (NHIS) from 2001 also show thatthe percentage of people over the age of65 who report that they failed to receivecare because of financial barriers is verysmall, at 2.1 percent.

Overall, national surveys show thatMedicare beneficiaries report fewerproblems than other adults with access tocare. Medical Expenditure Panel Survey(MEPS) data show that, of those requiringurgent care, older persons were morelikely than adults ages 18 to 64 to reportthat they always received the care as soonas they wanted (66 percent vs. 51percent). Persons over 65 also report thatthey delay care less often than those closeto the Medicare age. On the 2001 NHIS,5.6 percent of those age 55 to 64 reporteddelaying care because of cost versus 2.1percent of those over age 65 (Cohen2003).

One recent survey of beneficiaries foundan increasing rate of access problems, butthe increases were not limited to Medicarebeneficiaries. The 2001 HSC surveyfound that 11 percent of Medicare seniorsreported that they “did not get or put offcare.” This was an increase from 1997,when 9.1 percent of seniors reported suchoccurrences. On the same populationsurvey, 40.3 percent of Medicare seniorsreported waiting a week or more for anappointment for a specific illness, anincrease over 34.3 percent in 1997.Privately insured near-elderly alsoreported increasing access problems,although not on the same measure. Theproportion of privately insured personsbetween the ages of 50 and 64 reportingaccess problems increased from 15.2percent in 1997 to 18.4 percent in 2001.

Multivariate analysis ofbeneficiary characteristicsthat influence beneficiaries’ability to obtain careAlthough Medicare appears to have beenlargely successful in ensuring access tocare for many beneficiaries, certainsubgroups seem to have more accessproblems than others. A large body ofpublished research suggests that personsof low income, persons with no orinadequate insurance, and individualsfrom racial and ethnic minority groupsmay have lower access to care regardlessof their insurer (Mayberry, Mili, and Ofili2002; Aday, Fleming, and Anderson1984; Gornick, Eggers, and Reilly 1996;Gornick 2000).16

To determine empirically whichbeneficiary characteristics have the mostinfluence on access to care, MedPACconducted a multivariate analysis usingfive different outcome measuresrepresenting different dimensions ofaccess to care (see text box, p. 164).

Across most measures of access,MedPAC found that, all other factorsbeing equal, beneficiaries who were inpoor health, those who were living inpoverty, and those without anysupplemental insurance most consistentlyreported access problems. Specificfindings related to each of thesebeneficiary characteristics are discussedfurther below (Table 3-4, p. 166).

• Health status. Beneficiaries inexcellent health were only 20 percentas likely to report trouble getting careand only 30 percent as likely to reportdelaying care because of costs as wellas not seeing a doctor when needingto, compared with those in poorhealth. Interestingly, beneficiaries inexcellent health were more likely toreport not having a usual source ofcare or usual doctor than those in


16 There has been considerable interest in variations across subgroups in access to care. The Institute of Medicine (IOM) was commissioned to develop a set of indicatorsfor monitoring access to personal health services over time at the national level. In its publication, Access to health care in America, the IOM noted that because mostelderly people are entitled to Medicare benefits, they are frequently neglected in discussions of access (IOM 1993). Given that Medicare benefits are notcomprehensive, the IOM noted that disparities in access among this population may exist and should be explored further. In a subsequent study on race and ethnicity inthe U.S., Unequal treatment: confronting racial and ethnic disparities in health care, the IOM concluded that racial and ethnic minorities tend to receive a lower qualityof healthcare, compared with nonminorities, even after controlling for access-related factors (IOM 2002).

worse health. However, this findingmay reflect the fact that suchbeneficiaries do not require manyhealth services.

• Supplemental insurance status. Abeneficiary’s supplemental insurancestatus is also an important variableinfluencing self-reported access tocare.17 All other factors being equal,beneficiaries with supplementalcoverage were 13 to 75 percent aslikely (depending on the type ofadditional coverage and the specificmeasure examined) to report accessproblems as beneficiaries withMedicare FFS coverage only. Thisresult is not unexpected given thelimits of the Medicare core benefitpackage and the cost-sharing

requirements. Although representingonly about 6 percent of beneficiariesin our sample, those with coveragelimited to Medicare fee-for-servicereported higher levels of accessproblems than any other agedbeneficiary subgroup: 11.9 percentreported not seeing a doctor whennecessary, 16.1 percent reporteddelaying care because of costs, 22.3percent reported no usual source ofcare, and 27.9 percent reported nousual doctor.18

Although all forms of supplementalcoverage improved Medicarebeneficiaries’ self-reported access tocare, there was little difference inaccess to care based on the four typesof supplemental insurance reported.

Medicare HMO beneficiariesreported better access than those withMedicare fee-for-service coverageonly.

• Income. All other factors beingequal, beneficiaries with the highestincomes were less likely to reportaccess problems than those withlower incomes across most measuresof self-reported access. Beneficiarieswith the middle and highest incomeswere about 75 percent as likely asbeneficiaries at or below the povertylevel to report not seeing a doctorwhen they needed to, not having ausual source of care, and not having ausual doctor. Beneficiaries with themiddle and highest incomes were 25to 50 percent as likely as beneficiariesat or below the poverty level to reportdelaying care because of costs.Finally, the poorest beneficiaries—those below 50 percent of the povertylevel—were a third more likely thanthose between 50 and 100 percent ofthe poverty level to report needing tosee a doctor but not doing so. Otherresearchers have suggested that therole of income in influencing accessto care, as well as receipt ofappropriate care examined in otherstudies, may be related to factorssuch as better transportation, betterenvironment, and additionalresources that may be available towealthier beneficiaries (Gornick2000).

• Race/ethnicity and socioeconomicstatus. Studies have found importantdifferences in access by race/ethnicityand socioeconomic status (Schoen1998; Fiscella 2000; MedPAC 2002;Gornick, Eggers, and Reilly 1996;Schulman et al. 1995; IOM 1993;IOM 2002). Race/ethnicity andsocioeconomic status are closelyintertwined, however, and it is oftendifficult to isolate their respective


Beneficiary characteristics and outcome measures in the MedPAC analysis

To assess Medicare beneficiaries’ability to obtain care, MedPACused beneficiaries’ self-reports

of the following measures: (1) troublegetting care; (2) delaying care becauseof costs; (3) having a health conditionand needing to see a doctor but notseeing one; (4) having no usual sourceof care; and (5) having no usual doctor.Four years of Medicare CurrentBeneficiary Survey data (1996–1999)were pooled to yield a sample largeenough to examine differences acrosssubgroups of beneficiaries. Thesociodemographic and othercharacteristics of Medicarebeneficiaries examined in MedPAC’sstudy include gender, race, age, livingarrangements, income (income-to-poverty ratio), insurance status, andprescription drug coverage (Table 3-3).

On each measure of access, MedPACcalculated two types of statistics foreach subgroup of beneficiaries:

• the proportion who reported anaccess to care problem (theunadjusted percentage); and

• the likelihood of reporting anaccess problem after controllingfor the remaining beneficiarycharacteristics listed in Table 3-4(the adjusted odds ratio). Forexample, an adjusted odds ratio of0.70 for delaying care amongHispanics can be interpreted asfollows: all factors other thanethnicity held constant, Hispanicswere 70 percent as likely as whites(the reference group) to reportdelaying care because of costs. �

17 We define someone as having supplemental insurance if they are enrolled in Medicare FFS and have either Medicaid, Medigap, or employer-sponsored insurance, orif they are in a Medicare HMO instead of Medicare FFS.

18 The 6 percent reflects 1996–1999 MCBS data, in contrast to 2000 data cited elsewhere in this report that suggest a higher proportion of Medicare beneficiarieshaving no supplemental insurance coverage.

roles in affecting access to care.19

Notably, in MedPAC’s multivariateanalysis, race and ethnicity receded inimportance when other factors, suchas income and health status, weretaken into account. Specifically, raceand ethnicity were not as consistentpredictors of access problems asincome, health status, orsupplemental insurance status in fourof the five access measures, but raceand ethnicity were highly significantin influencing whether a beneficiaryreported having a usual doctor.African Americans were one-and-a-half times more likely than whites toreport not having a usual doctor.Similarly, when all other factors wereheld constant, Hispanics were almosttwice as likely as whites to report nothaving a usual doctor and almostone-and-a-half times as likely aswhites to report not having a usualsource of care. However, Hispanicswere less likely than whites, when allother factors were statisticallycontrolled, to report other accessproblems.

• Education. A beneficiary’seducation level influenced self-reported access measures inunanticipated ways. Specifically,beneficiaries with the highesteducation levels were most likely toreport concerns with accessing care.Beneficiaries with a collegeeducation were 20 to 60 percent morelikely to report having trouble gettingcare, delaying care because of costs,or not seeing a doctor whennecessary. However, they were aslikely as those with only a highschool diploma to report both a usualsource of care and a usual doctor.Perhaps this finding reflects thehigher expectations of individualswho have higher education levels.


Characteristics of the noninstitutionalized agedMedicare population, 1996–1999

Percent of the Characteristics Medicare population

Total (N � 34,561) 100%Sex

Male 42Female 58

Race/ethnicityWhite, non-Hispanic 84African American, non-Hispanic 7Hispanic 6Other 3

Age65–74 5075–84 3985� 11

Health statusExcellent/very good 46Good/fair 48Poor 6

UrbanicityUrban 75Rural 25

Living arrangementAlone 32With spouse 53With others 14

EducationNo high school diploma 36High school diploma only 31Some college or more 33

Income to poverty ratio1

� .5 extreme poverty 5.5–1 poverty 151–2 low income 332–4 middle income 324� high income 15

Supplemental insurance statusMedicare only 6Medicare and Medicaid 10Medicare and Medigap 27Medicare and employer-sponsored 38Health maintenance organization 16Other2 3

Prescription drug coverageYes 72No 28

Note: All end-stage renal disease beneficiaries and institutionalized beneficiaries are excluded from the analysis.1Calculated by dividing self-reported family income by the poverty threshold.2Other includes Medicare and Department of Defense, and Medicare and Department of Veterans Affairs.

Source: Medicare Current Beneficiary Survey, Access to Care, and Cost and Use files, 1996–1999 combined.

T A B L E3-3

19 Since socioeconomic data are unavailable in the Medicare administrative databases, race and ethnicity are often used as proxies, although they have been shown torepresent different issues. When socioeconomic data have been used in Medicare studies, they are often reported as ecologic variables using ZIP code levelinformation which may not necessarily correlate to the income of the specific individual included in the study. A major advantage in using MCBS data, therefore, is theavailability of both race/ethnicity and individual income data.


Aged

Med

icare

popula

tion r

eport

ing a

cces

s to

care

pro

ble

ms,

by b

enef

icia

ry c

hara

cter

istics

, 1996–1

999

Del

ayed

care

Nee

ded

to

No u

sual

Had t

rouble

bec

ause

of

see

doct

or

sourc

e of

No u

sual

Chara

cter

istic

get

ting c

are

cost

but

did

not

care

doct

or

Unadju

sted

Unadju

sted

Unadju

sted

Unadju

sted

Unadju

sted

per

cent

Odds

ratio

1per

cent

Odds

ratio

1per

cent

Odds

ratio

1per

cent

Odds

ratio

1per

cent

Odds

ratio

1

All

2.4

N/

A4.

5N

/A

6.7

N/

A6.

1N

/A

9.1

N/

A

Sex M

ale

(R)

2.2

1.0

3.9

1.0

6.1

1.0

7.0

1.0

10.8

1.0

Fem

ale

2.5

1.09

5.0

1.20

*7.

21.

16*

5.5

0.68

*7.

90.

61*

Race W

hite

(R)

2.2

1.0

4.2

1.0

6.5

1.0

5.7

1.0

7.9

1.0

Afri

can

Am

eric

an3.

61.

077.

10.

858.

90.

997.

91.

0715

.81.

54*

Hisp

anic

3.4

0.96

5.6

0.70

*7.

30.

79*

9.0

1.42

*17

.01.

84*

Age 65

–74

(R)

2.6

1.0

5.3

1.0

7.4

1.0

6.6

1.0

9.8

1.0

75–8

42.

20.

82*

4.1

0.66

*6.

10.

75*

5.6

0.85

*8.

40.

85*

85�

2.2

0.71

*2.

80.

34*

6.0

0.67

*5.

70.

71*

8.5

0.72

*

Hea

lth s

tatu

sEx

celle

nt/

very

goo

d1.

60.

19*

2.8

0.30

*1.

60.

32*

7.8

1.94

*10

.51.

56*

Goo

d/fa

ir2.

50.

32*

5.3

0.51

*4.

80.

52*

4.5

0.97

7.6

0.92

Poor

(R)

7.8

1.0

11.3

1.0

7.7

1.0

5.1

1.0

9.5

1.0

Urb

anic

ityU

rban

(R)

2.5

1.0

4.1

1.0

6.4

1.0

5.8

1.0

8.9

1.0

Rura

l2.

00.

865.

81.

097.

61.

127.

00.

999.

60.

94

Livin

g ar

rang

emen

tA

lone

(R)

2.6

1.0

5.3

1.0

7.0

1.0

6.9

1.0

10.2

1.0

With

spo

use

1.9

0.79

*3.

60.

79*

6.3

0.95

5.2

0.69

*7.

70.

68*

With

oth

ers

3.7

1.19

6.5

0.89

8.0

0.94

7.7

0.95

11.9

0.90

TA

BL

E3

-4

cont

inue

d on

nex

t pag

e


Aged

Med

icare

popula

tion r

eport

ing a

cces

s to

care

pro

ble

ms,

by b

enef

icia

ry c

hara

cter

istics

, 1996–1

999

Del

ayed

care

Nee

ded

to

No u

sual

Had t

rouble

bec

ause

of

see

doct

or

sourc

e of

No u

sual

Chara

cter

istic

get

ting c

are

cost

but

did

not

care

doct

or

Unadju

sted

Unadju

sted

Unadju

sted

Unadju

sted

Unadju

sted

per

cent

Odds

ratio

1per

cent

Odds

ratio

1per

cent

Odds

ratio

1per

cent

Odds

ratio

1per

cent

Odds

ratio

1

TA

BL

E3

-4

Educ

atio

nN

o hi

gh

scho

ol d

iplo

ma

2.8

1.16

6.7

1.46

*8.

11.

19*

7.3

1.10

11.6

1.09

Hig

h sc

hool

di

plom

a on

ly (R

)1.

81.

03.

41.

05.

91.

05.

71.

08.

21.

0So

me

colle

ge

or m

ore

2.5

1.60

*3.

21.

38*

6.1

1.21

*5.

10.

92

7.4

0.92

Inco

me

to p

over

ty ra

tio2

�.5

ext

rem

e po

verty

3.9

1.00

8.5

1.15

11.1

1.34

*8.

91.

1015

.61.

19.5

–1 p

over

ty (R

)3.

71.

08.

11.

08.

91.

08.

21.

013

.31.

01–

2 lo

w in

com

e2.

20.

865.

70.

897.

00.

926.

50.

899.

40.

85*

2–4

mid

dle

inco

me

1.9

0.90

2.6

0.51

*5.

60.

80*

4.9

0.74

*7.

00.

74*

4�hi

gh in

com

e2.

11.

001.

30.

25*

5.1

0.74

*4.

90.

72*

6.8

0.72

*

Supp

lem

enta

l ins

uran

ce s

tatu

sM

edic

are

only

(R)

4.4

1.0

16.1

1.0

11.9

1.0

22.3

1.0

27.9

1.0

Med

icar

e an

d M

edic

aid

4.5

0.92

7.9

0.43

*9.

90.

74*

7.1

0.29

*13

.30.

36*

Med

icar

e an

d M

edig

ap1.

30.

33*

3.8

0.30

*5.

60.

53*

5.9

0.29

*7.

70.

29*

Med

icar

e an

d em

ploy

er-sp

onso

red

1.7

0.49

*2.

90.

31*

6.1

0.66

*5.

20.

28*

7.0

0.26

*H

ealth

mai

nten

ance

or

gani

zatio

n 3.

61.

042.

70.

27*

6.1

0.66

*2.

70.

13*

5.4

0.17

*

Pres

crip

tion

drug

cov

erag

eYe

s (R

)2.

41.

03.

61.

04.

81.

04.

81.

07.

91.

0N

o2.

41.

217.

11.

42*

9.5

1.17

*9.

51.

20*

12.1

1.02

Not

e:N

/A

(not

app

licab

le),

R (re

fere

nce

grou

p). A

ll en

d-sta

ge re

nal d

iseas

e be

nefic

iarie

s an

d in

stitu

tiona

lized

ben

efic

iarie

s ar

e ex

clud

ed fr

om th

e an

alys

is.

1Ea

ch o

dds

ratio

is a

djus

ted

in th

e m

ultip

le re

gres

sion

anal

ysis

for a

ll ot

her v

aria

bles

in T

able

3-3

.2C

alcu

late

d by

div

idin

g se

lf-rep

orte

d fa

mily

inco

me

by th

e po

verty

thre

shol

d.*I

ndic

ates

that

the

odds

ratio

is s

igni

fican

t at a

n ap

prox

imat

e P�

0.05

. To

miti

gate

the

desig

n ef

fect

s of

the

Med

icar

e C

urre

nt B

enef

icia

ry S

urve

y (M

CBS

), a

signi

fican

ce le

vel o

f P�

0.02

5 w

as u

sed

to a

ppro

xim

ate

a 0.

05sig

nific

ance

leve

l. Su

bseq

uent

adj

ustm

ent f

or th

e M

CBS

des

ign

effe

cts

may

alte

r the

sta

tistic

al s

igni

fican

ce fo

r odd

s ra

tios

of b

orde

rline

sig

nific

ance

.

Sour

ce:

MC

BS A

cces

s an

d C

ost a

nd U

se fi

les,

199

6–19

99 c

ombi

ned.

• Disabled. MedPAC conducted aseparate analysis to examine accessto care among the disabledpopulation. The analysis was alsodone using 1996–1999 pooled MCBSdata. Compared with the agedMedicare population, the disabledunder-65 population reportedconsiderably higher access problems:8.9 percent of the disabled versus 2.4percent of the aged populationreported trouble getting care; 19.8percent versus 4.5 percent reporteddelaying care because of costs; 18.3percent versus 6.7 percent reportednot seeing a doctor when they neededto; and 16.3 percent versus 9.1percent reported not having a usualdoctor (data not shown). Similarproportions of each populationreported no usual source of care.

Beneficiaries’ ability toobtain appropriate care

Up until this point, we have discussedwhether the health system has the capacityto meet Medicare beneficiaries’ currentand future needs, and we have examinedthe experiences of different types ofbeneficiaries obtaining care. But justbecause beneficiaries are able to obtaincare does not necessarily mean that theyare obtaining appropriate care. Evaluatingvarious measures of appropriate use ofservices we found:

• many beneficiaries are not receivingpreventive and primary care servicesthat can help manage a condition thatmight otherwise result in an acuteepisode;

• some beneficiaries are ending up inthe hospital with conditions thatmight have been prevented if theircare had been managed moreeffectively; and

• trends in the types of ED servicesused and the types of beneficiarieswho use them may suggest a lack ofavailability of ambulatory serviceselsewhere.

Preventive and primary careservicesUse of preventive and primary careservices known to be effective is onepossible indicator of access to appropriatecare.20 Declines in use of these servicescould signal that the Medicare populationmay have access problems. In addition,variations in use rates for these servicesby population subgroup might revealdisparities that do not appear in aggregatemeasures. For example, an annualinfluenza vaccination is recommended forall persons 65 years and older. However,in 2000, 70 percent of all whitebeneficiaries, 52 percent of AfricanAmericans, and 54 percent of Hispanicsreceived flu shots (MCBS Access to carefile 2000). Thus, significant portions ofthe elderly population are not receiving aservice that could prevent one of the sixleading causes of death among the elderly,and minorities seem to have the biggestgap between the amounts ofrecommended and received preventiveservices.

The Medicare population is alsounderusing preventive services fordiabetes and other chronic conditions.One tool MedPAC has employed tomonitor use of appropriate services is the

Access to Care for the Elderly Project(ACE–PRO) indicators.21 MedPACanalysis of 1996–1999 MCBS Cost andUse files using these indicators revealedthat only 46 percent of beneficiaries withdiabetes received an eye exam at leastonce a year and only 41 percent received atest to measure glucose levels every sixmonths. Although the number ofbeneficiaries obtaining these preventivetests has increased since that analysis, asignificant number still do not receive theappropriate tests.22 Diabetic beneficiarieswith no supplemental coverage were evenless likely to receive appropriatepreventive services (Table 3-5).

Preventable hospitalizationsSome of the ACE–PRO indicators usemeasures of preventable hospitalizationsto identify those beneficiaries who maynot have received the right service. Thesemeasures are based on the premise thatpatients go to the emergency departmentor are admitted or readmitted to thehospital for some conditions, such asasthma, if they have not receivedappropriate primary care. Researchershave identified a number of conditionssensitive to ambulatory care, includingcongestive heart failure, pneumonia,asthma, diabetes, gastroenteritis, anddehydration (Rutstein et al. 1976, Billingset al. 1993, Epstein 2001). Table 3-5contains two of these examples, includingthe percentage of Medicare beneficiarieswith known angina who went to the EDthree or more times in one year forcardiovascular-related diagnoses.

We know of no national or Medicare-specific benchmark that describes theright level of hospitalizations for these


20 A low or decreasing utilization rate for these services could be due to access barriers like the availability of providers or willingness to obtain care, or could indicate aproblem with the quality of care—that is, the beneficiary did go to the physician, but did not receive the right test or vaccine. Therefore, it is important to use theseindicators cautiously when making conclusions about access to care. Whether the problem is an access or quality concern, the level of coverage or payment for theservice is an important concern.

21 The ACE–PRO indicators were developed by RAND for the Physician Payment Review Commission. The indicators use Medicare claims and enrollment data to identifywhether patients with certain conditions have received the minimally necessary services (as defined by clinicians).

22 Data from the CMS quality improvement organization program show that the rates of provision of tests to measure glucose levels in a one-year period for the medianstate in 2000–2001 was 78 percent; 70 percent of the diabetics received an eye exam within a two-year period (Jencks SF, Huff EO, Cuerdon T 2003). One of theprimary reasons for this difference in rates between the ACE–PRO and the CMS data is that the time frames differ. The CMS data reported on rates of provision acrossa one-year period for glucose testing and the ACE–PROs used a six-month period. For eye exams, CMS used a two-year period while the ACE–PROs were based on aone-year period.

conditions. However, some studies in theVeterans Health Administration systemand specific geographic regions have usedclinical protocols and medical recordreviews to establish rates of preventablehospitalizations for a variety ofconditions.

In 1997, a study authored by researchersfrom NCHS reported that in 1990 almosthalf of all potentially avoidablehospitalizations were for those aged 65and over (Pappas et al. 1997). For personsover 65, 15 percent of adjusted totaldischarges were for potentially avoidablehospitalizations. The rate of potentiallyavoidable hospitalizations for persons inthis study aged 45 to 64 was 10 percent of

adjusted total discharges. Most of thesehospitalizations for those over 65 were forcongestive heart failure (40 percent) andpneumonia (35 percent).

MedPAC analysis of unpublished nationalestimates for 1999 from the HealthcareCost and Utilization Project of the Agencyfor Healthcare Research and Quality(AHRQ) shows that of the totalpotentially avoidable hospitalizations forconditions identified by AHRQ as“prevention quality indicators,” fiveconditions accounted for 88 percent of allof the potentially avoidablehospitalizations for those over 65. Thoseconditions were: congestive heart failure(30 percent), bacterial pneumonia (25

percent), chronic obstructive pulmonarydisease (16 percent), urinary tractinfections (9 percent), and dehydration (8percent).

Emergency department useMedicare beneficiaries use EDs moreoften than people under 65 who are noteligible for Medicare, with the oldestbeneficiaries and minorities using themmore than other beneficiaries. This careappears to be appropriate on one level; theproportion of visits assessed as“nonurgent” at the time of admissionamong the elderly is quite low.23

However, this use of EDs may alsoindicate that these beneficiaries are notgetting appropriate care elsewhere thatmight have prevented the need for an EDvisit. NCHS analysis of patientcharacteristics that act as barriers toobtaining care show that high users ofEDs are more likely to report no usualdoctor and no usual source of care.Because much of the increase in the use ofEDs by older Americans in the 1990s wasto treat illness or complications of medicaltreatment, including problems withmedications, older Americans may not beusing a regular source of care tocontinually monitor and manage theirhealth conditions.

Emergency care is essential when peoplebecome critically ill and becomesincreasingly important as people age.Slightly more than 20 percent of all adultsover the age of 18 in the United States hadone or more ED visit in 2000 (NCHS2002). However, more than 25 percent ofpeople age 75 and older had at least oneED visit in 2000 and people 75 and olderwere almost twice as likely as those age55 to 64 to have two or more visits to theED. Data collected by NCHS in theNational Hospital Ambulatory MedicalCare Survey (NHAMCS)24 have explored


Share of beneficiaries using selected clinically necessary services,

by supplemental coverage

No Somesupplemental supplemental

Indicators Total coverage coverage

Use of necessary care for specific conditionsEye exam every year for patients with diabetes 46.0% 29.9% 47.1%Glycosylated hemoglobin or fructosamine test

every six months for patients with diabetes 41.3 36.3 41.7Follow-up visit within four weeks of initial

diagnosis of gastrointestinal bleeding 72.2 54.0 73.3Arthroplasty or internal fixation of hip during

hospital stay for hip fracture 88.9 80.0 89.7

Incidence of avoidable outcomesAmong patients with known diabetes:

admissions for hyperosmolar or ketotic coma 0.1 0.6 0.1Among patients with known angina:

three or more emergency room visits forcardiovascular-related diagnoses in one year 5.2 6.0 5.2

Source: MedPAC analysis of the 1996–1999 Medicare Current Beneficiary Survey Cost and Use files using theAccess to Care for the Elderly Project Indicators.

T A B L E3-5

23 Data from the 2000 NHAMCS show that, for those visits for which the immediacy of care need was known, the proportion of visits that were “nonurgent” decreaseswith age: For people age 75 and older, 5.3 percent of visits were nonurgent, compared with 5.8 percent for visits by those age 65 to 74, and 9.0 percent for peopleage 45 to 64 (McCaig and Ly 2002). Nonurgent is defined in the survey as “a visit in which the patient should be seen within 121 minutes to 24 hours.” This definitionis stricter—care needed within two hours or less—than the standard used by some health systems for their urgent care protocols.

24 This survey has been conducted annually by NCHS since 1992. This survey obtains detailed data on all ED and outpatient visits, for all patients, provided in non-Federal, short-stay hospitals providing general (medical or surgical) care (Burt and McCaig 2001). This makes it a particularly valuable source for comparing the use ofEDs across all populations, and over time. Because the survey is based on visits, rather than people, however, the rates cannot fully explain variations in the use of EDservices. Each visit is an independent observation, and visits by a particular individual cannot be linked.

the use of ED services in depth (Burt andMcCaig 2001; McCaig and Ly 2002).25

Several issues related to access to careemerge from the analysis of these data.

• Beneficiaries used the ED forurgent care. Beneficiaries tend touse ED care for care related toexisting medical conditions that havereached the stage where urgent care isnecessary. The illness-related visitrates, as compared with visits forinjuries, for persons over 65 as awhole increased by 21 percentbetween 1992 and 1999. It appearsthat these ED users were under sometype of medical treatment and weretaking an increasing number ofprescription drugs. During thisperiod, the rate of visits caused byadverse effects from medicaltreatment increased from 4.8 to 10.2visits per 1,000 persons per year andthe rate of visits in which 5 or moreprescription drugs were mentioned inthe visit record increased by 59percent.

• ED use varies with beneficiaries’characteristics. African Americansand beneficiaries with Medicaidcoverage (a poorer population) useEDs to a greater extent than otherMedicare populations. MedPACanalysis of several years of MCBSdata also show that the oldestbeneficiaries, those with end-stagerenal disease, disabled beneficiaries,and those using some type of nursingfacility care were also heavier usersof EDs.

These data cannot provide the level ofinformation required to evaluate how caremanagement could mediate beneficiaries’need for emergency services. It is very

difficult to distinguish between use ofservices that is necessary and appropriateand use of services that may be necessarybut could have been avoided withappropriate primary care or bettermanagement of complex medicalconditions. However, lack of access toappropriate care management forvulnerable populations may contribute toacute episodes that require visits to theED.26

Conclusion

Our analysis finds no widespreadproblems in beneficiaries’ access to care.On some important measures,beneficiaries enjoy better access toservices than is the case for older adultsnot yet eligible for Medicare. There are,however, some areas of concern regardingthe availability of appropriate, effectiveservices for a growing beneficiarypopulation that the Commission willmonitor closely. First, recent researchsuggests that some physicians inMedicare’s fee-for-service program arebecoming more selective about thepatients they accept into their practices.This selectivity does not appear to betargeted exclusively to Medicarebeneficiaries, but trends in physicianparticipation in Medicare and inbeneficiaries’ ability see physicians on atimely basis are important indicators totrack. Second, there is some evidence thatpatients with particularly complex careneeds may have problems gaining accessto appropriate post-acute care services. Itwill be important to monitor the effect that

delayed placement in skilled nursing caremay have on patients with more complexneeds.

Advances in medical technology andimprovements in the management ofcomplex health care problems maychange the landscape of the health careservices people use. If current trendspersist, however, the beneficiarypopulation will not only be larger, but itwill also include a growing number ofdisabled beneficiaries, people over age 85,more minorities, and more women livingalone. These beneficiary groups arecurrently among the most vulnerable, interms of prevalence of serious chronicconditions, low incomes, and adequatesupplemental insurance. MedPACanalysis also shows that they are morelikely than other beneficiaries to reportproblems across measures of access tocare. MedPAC will continue to monitorthese and other Medicare beneficiaries’access to care issues to evaluate whetherthe health care system is responding totheir health care needs.

Finally, closer examination of data on theuse of health services across populationssuggests the importance of focusing notonly on access to care, but on access to theright kind of care, in the right setting. Theevidence suggests a need to evaluatewhether better access to appropriatepreventive and primary care, as well asbetter management of complex chronicillnesses, might help prevent or delayserious complications, including the needfor emergency services and subsequentinpatient care. �


25 The 2001 NCHS report on ED trends examined data from the NHAMCS from 1992 to 1999 and also drew on NHIS and Medicare data to explore some of the trendsidentified in the analysis of ED service use (Burt and McCaig 2001). Unless noted otherwise, NCHS analysis of ED use rates discussed here is drawn from Burt andMcCaig, 2001.

26 Recent research points to ways that care managed outside the ED can prevent the need for ED visits. See, for example, Coleman EA, Eilersten TB, Kramer AM et al.Reducing emergency visits in older adults with chronic illness, Effective Clinical Practice. March–April 2001, Vol. 4, No. 2, p. 49–57.

References

Aiken LH, Clarke SP, Sloane DM, Sochalski J. Hospital nurse staffing and patientmortality, nurse burnout, and job dissatisfaction, Journal of the American MedicalAssociation. October 23/30, 2002, Vol. 288, No. 16, p. 1987–1993.

Aday LA, Fleming R, Anderson R. Access to medical care in the U.S.: who has it, whodoesn’t? Chicago (IL), Pluribus Press, Inc. and the University of Chicago. 1984.

Berliner HS, Ginzberg E. Why this hospital nursing shortage is different, Journal of theAmerican Medical Association. December 4, 2002, Vol. 288, No. 21, p. 2742–2745.

Billings J, Zeitel L, Lukomnik L, et al. Impact of socioeconomic status on hospital use inNew York City, Health Affairs. Spring 1993, Vol. 12, No. 1, p. 162–173.

Bureau of Health Professions, Health Resources and Services Administration,Department of Health and Human Services. Projected supply, demand, and shortages ofregistered nurses: 2000–2020. Washington (DC), Bureau of Health Professions. July2002.

Burt CW, McCaig LF. National Center for Health Statistics. Trends in hospitalemergency department utilization: United States, 1992–99, Vital and Health Statistics.November 2001, Series 13, No. 150.

Cohen A, National Center for Health Statistics data. E-mail communication with JillBernstein. Hyattsville (MD). January 2003.

Epstein AJ. The role of public health clinics in preventable hospitalizations amongvulnerable populations, Health Services Research. June 2001, Vol. 36, No. 2, p. 405–420.

Federal Interagency Forum on Aging-Related Statistics. Older Americans 2000: keyindicators of well-being. Available at www.agingstats.gov/chartbook2000.

First Consulting Group. The healthcare workforce shortage and its implications forAmerica’s hospitals. Unpublished slide show. Prepared Fall 2001. (Results from a surveyof hospital administrators commissioned by The American Hospital Association, theAssociation of American Medical Colleges, the Federation of American Hospitals and theNational Association of Public Hospitals and Health Systems.)

Fiscella K, Franks P, Gold MR, Clancy CM. Inequality in quality: addressingsocioeconomic, racial and ethnic disparities in health care, Journal of the AmericanMedical Association. May 17, 2000, Vol. 283, No. 19, p. 2579–2584.

Foote SM, Hogan C. Disability profile and health care costs of Medicare beneficiariesunder age sixty-five, Health Affairs. November/December 2001, Vol. 20, No. 6, p.242–253.

Freedman VA, Martin LG, Schoeni RF. Recent trends in disability and functioningamong older adults in the United States: a systematic review, Journal of the AmericanMedical Association. December 25, 2002, Vol. 288, No. 24, p. 3137–3146.

Gornick ME, Eggers PW, Reilly TW, et al. Effects of race and income on mortality anduse of services among Medicare beneficiaries, New England Journal of Medicine.September 12, 1996, Vol. 335, No. 11, p. 791–799.


Gornick ME. Disparities in Medicare services: potential causes, plausible explanationsand recommendations, Health Care Financing Review. Summer 2000, Vol. 21, No. 4, p.23–43.

Institute of Medicine. Access to health care in America. Washington (DC), NationalAcademy Press. 1993.

Institute of Medicine. Unequal treatment: confronting racial and ethnic disparities inhealth care. Washington (DC), National Academy Press. 2002.

Jencks SF, Huff ED, Cuerdon T. Change in the quality of care delivered to Medicarebeneficiaries, Journal of the American Medical Association. January 15, 2003, Vol. 289,No. 3, p. 305–311.

Mayberry RM, Mili F, Ofili E. Racial and ethnic differences in access to medical care. In:LaViest TA (ed.). Race, ethnicity and health: a public health reader. San Francisco (CA),Jossey-Bass. 2002, p. 163–197.

McCaig LF, Ly N. National Hospital Ambulatory Medical Care Survey: 2000 emergencydepartment summary, advance data from Vital and Health Statistics. April 22, 2002, No.326.

Medicare Payment Advisory Commission. Report to the Congress: assessing Medicarebenefits. Washington (DC), MedPAC. June 2002.

Merck Institute of Aging and Health, The Gerontological Society of America. The stateof aging and health in America. Fall 2002. Available at www.geron.org orwww.miahonline.org.

National Center for Health Statistics. Health, United States, 2002. Chartroom on Trendsin the Health of Americans. Hyattsville (MD), U.S. Government Printing Office. 2002.

Office of Inspector General, Department of Health and Human Services. Access to homehealth care after hospital discharge, No. OEI 02–01–00180. Washington (DC), OIG. July2001.

Office of Inspector General, Department of Health and Human Services. Early effects ofthe prospective payment system on access to skilled nursing facilities, No.OEI–02–99–00400. Washington (DC), August 1999.

Office of Inspector General, Department of Health and Human Services. Medicarebeneficiary access to skilled nursing facilities, No. OEI–02–01–00160. Washington (DC),OIG. July 2001.

Office of Inspector General, Department of Health and Human Services. Medicarebeneficiary access to skilled nursing facilities, No. OEI–02–00–00330. Washington (DC),OIG. September 2000.

Office of Inspector General, Department of Health and Human Services. Medicare homehealth care community beneficiaries, No. OEI 02–01–00070. Washington (DC), OIG.October 2001.

Office of Research, Development and Information, Centers for Medicare & MedicaidServices. Data from Medicare Current Beneficiary Survey 1991–2000 Access to Carefiles.


Pappas G, Hadden WC, Kozak LJ, Fisher GF. Potentially avoidable hospitalizations:inequalities in rates between US socioeconomic groups, American Journal of PublicHealth. May 1997, Vol. 87, No. 5, p. 811–816.

Rutstein DD, Berenberg W, Chalmers TC, et al. Measuring the quality of medical care: aclinical method, New England Journal of Medicine. March 11, 1976, Vol. 294, No.11,p. 582–588.

Schoen C, Neuman P, Kitchman M, et al. Medicare beneficiaries: a population at risk—findings from the Kaiser/Commonwealth 1997 survey of Medicare beneficiaries. NewYork (NY), The Commonwealth Fund. December 1998.

Schulman KA, Rubenstein E, Chesley FD, Eisenberg JM. The roles of race andsocioeconomic factors in health services research, Health Services Research. April 1995,Vol. 30, No. 1, p. 179–195.

Spillman BC, Lubitz J. The effect of longevity on spending for acute and long-term care,The New England Journal of Medicine. May 11, 2000, Vol. 342, No. 19, p. 1409–1415.


Payment for new technologies in Medicare’s prospective

payment systems

C H A P T E R 4


The Secretary should introduce clinical criteria for eligibility of drugs and biologicals to receivepass-through payments under the outpatient prospective payment system.



edicare has a responsibility to pay enough for beneficial

new technologies to ensure beneficiaries’ access to care,

but must also be a prudent purchaser. The hospital inpa-

tient and outpatient prospective payment systems cur-

rently incorporate the costs of new technologies through an annual review of pay-

ment rates, as well as through special payment mechanisms for specific new

technologies. The details of these new technology payments—such as the crite-

ria technologies must meet to be eligible for them—are the mechanisms through

which Medicare balances the goals of ensuring adequate payment for beneficial

new technologies and being a prudent purchaser. To increase the program’s abil-

ity to be a prudent purchaser and to ensure fair treatment across technologies and

payment systems, MedPAC recommends that the clinical criteria currently ap-

plied to all new technology applicants under the inpatient PPS, and to new med-

ical device applicants under the outpatient PPS, be extended to applications for

new outpatient drugs and biologicals. Finally, our review of how other private

and public sector payers deal with the issue of paying for new technologies sug-

gests that many of their approaches—such as negotiation and competitive bid-

ding—may not easily be adopted into Medicare’s current administered pricing

systems, but point to value-based purchasing as a concept to pursue.

M

C H A P T E R

Payment for newtechnologies in Medicare’sprospective payment systems

4In this chapter

• Payment for new technologiesin prospective paymentsystems: an overview

• Hospital inpatient services

• Hospital outpatient services

• Lessons from other health carepurchasers


New medical technologies can improveclinical outcomes and quality of care.They are also considered a major sourceof escalating health care costs. Since theimplementation of Medicare’s prospectivepayment system (PPS) for inpatienthospital services in 1983, questions havearisen about how to pay for beneficial newtechnologies (Garrison and Wilensky1986).

Medicare has a responsibility to payenough for beneficial new technologies toensure beneficiary access to care, but mustalso be a prudent purchaser. Achievingthese two goals is technically challengingand often involves tradeoffs. Payingprudently and adequately for beneficialnew technologies requires, at a minimum,the ability to determine a technology’smerit, accurate and verifiable informationon which to base a price, and a paymentsystem that can incorporate newtechnologies in a timely fashion.However, the evidence on the value of atechnology is not always clear.Furthermore, information on the marketprice of new technologies and their effecton the costs of providing services is oftennot available. Also, it takes time forMedicare’s administered pricingmechanisms to reflect the costs of newtechnologies.

For the purposes of this chapter, we definenew technologies as those that have beenon the market for a short period of time.They may be true innovations, significantincremental improvements on existingtechnologies, or expanded uses of an oldtechnology for a new indication.1 We donot include those that provide no orinsignificant incremental improvementson existing technology. Some are newdrugs or medical devices, others are newsurgical techniques or imaging devices.

Although the need to incorporate newtechnologies applies to all prospectivepayments, the topic may be more relevantto hospital payment systems because newtechnologies are often first adopted in thatsetting. In addition, in the past severalyears, Medicare has integrated newtechnology payments into the hospitalinpatient and outpatient prospectivepayment systems through specificpayment mechanisms applied when aclaim is submitted. Therefore, this chapterprovides a brief review of those systems.The chapter then presents information onhow other large purchasers of health carepay for new technologies and considersthe relevance of those approaches forMedicare. Though the chapter focusesprimarily on payment, it notes thatpayment for new technologies oftenrelates closely to coverage decisions (seeAppendix B).

Payment for newtechnologies inprospective paymentsystems: an overview

The incentives built into PPSs promotethe use of new technologies that reducecosts, but they may slow the adoption ofnew technologies that increase costs. Inresponse to those concerned about delaysin the incorporation of new technologiesinto Medicare’s payment systems, theCongress implemented special paymentsfor new technologies used in the hospitalinpatient and outpatient settings.

Prospective payment systems define afixed payment for a bundled service. Thatis, CMS establishes a set payment fortreating a case or providing a servicemeant to cover all the costs of providing

the service. Clinicians and providersdecide how the service will be delivered,including decisions regarding the use ofspecific technologies. This systemprovides an incentive for providers toadopt technologies that decrease costs. Itcan also fairly easily accommodate mostincremental technologies that increasecosts only modestly.2

In considering how payment systems dealwith technology, it is useful to distinguishbetween new technologies that are inputsto an existing service and newtechnologies that result in a new service:

• Technologies that are inputs to anexisting service. In the case of newtechnologies that are inputs to anexisting service (e.g., monoclonalantibodies for treatment of cancer ordrug-eluting coronary artery stentsused in angioplasty), Medicare paysproviders using the payment categoryfor that service. Most newtechnologies fall into this category.Technologies that decrease costs orraise them only modestly can enterthe payment system withoutadditional decision making. Fortechnologies that are very expensive,additional decisions may be required.The Centers for Medicare &Medicaid Services (CMS) mightrevisit how the service is classified orpay for the technology through aspecial payment mechanism.

• Technologies that result in a newservice. In the case of newtechnologies that result in newservices (e.g., laser angioplasty,positron emission tomography (PET)scanning, or digital mammography),the appropriate coding group mustfirst assign it a code (e.g., an

178 Paymen t f o r n ew t e c hno l og i e s i n Med i ca r e ’ s p r o spe c t i v e paymen t s y s t em s

1 This definition applies primarily to new technologies that are involved in clinical care. As discussed later, each payment system has criteria that define the newtechnologies to be covered by a specific new technology payment provision. Additional forms of new technology, such as information systems, might affect the entireorganization of a hospital.

2 This section provides an overview of the process through which new technologies are incorporated into the payment system. A more detailed discussion can be found inChapter 3 of MedPAC’s March 2001 Report to the Congress.

ICD–9–CM3 code or a HCPCS4

code). The coding process takes timebecause the code sets must bereviewed by medical professionals,with public input, and updated in amanner that preserves the integrity ofthe system. Medicare uses multiplecoding systems and is only one ofmany players involved in maintainingand updating the code sets. After acode is assigned, CMS mustincorporate the technology into thepayment system and set a paymentrate.

Prospective payment systems classifyservices and set a unit of payment thatincorporates all of the inputs needed toprovide the service.5 Relative weightsapply to each service, reflecting therelative resource costs of providing thatservice compared with the others in theclassification system. Generally, theserelative weights are recalibrated annuallyto reflect changes in the relative costs ofone service versus another, using the mostrecent cost and claims data. Thecombination of the coding andrecalibration processes generally takes atleast two years because of the multipleparties involved and the open process thatmust be followed.

Some critics of prospective payment arguethat the pace of technology adoption, thetime required to make code assignments,and the time needed to collect and processcost data can make it difficult to reflectthe costs of new technologies immediatelyand may, therefore, slow diffusion. Newtechnologies enter the marketplacecontinuously. Between 1995 and 2001, forexample, the Food and DrugAdministration (FDA) approved about950 new drugs, biologicals, and medicaldevices (American Hospital Associationand The Lewin Group 2002). However,

the process of placing a new technology inthe payment system also gives CMS anopportunity to better understand itsclinical merits and obtain accurateinformation about its costs. In addition tothe incremental cost of the technologyitself, using a new technology may lead toefficiency gains that lower the total costsof providing the service. Conversely, useof a new technology may result inadditional requirements that increase thetotal costs of providing a service beyondthe cost of the technology itself. Finally,any dampening effect these paymentsystems might have on technologicaldiffusion is often balanced by thecompetitive and clinical forcesencouraging physicians and hospitals touse new technologies.

The general prospective paymentapproach described above worksespecially well when the unit of paymentcovers a broad bundle of services, as withthe inpatient PPS, but less so with narrowbundles, for which technology canrepresent a large share of the totalpayment. For example, a new scalpel maynot represent a large share of the paymentfor a surgical stay on the inpatient side,but the costs of a new cancer drug coulddominate payment for outpatientchemotherapy administration. In addition,this approach does not immediatelycapture the rapid decline in prices thatoften occurs shortly after the introductionof a new technology or when competitorsenter the market. Payments are setannually using hospital charge data. Pricechanges during the course of the year,therefore, are not built into the paymentsuntil the next payment review. In addition,it is unclear whether or not hospitalsdecrease their charges, which Medicareuses to approximate costs, as quickly asinput prices decline.

At least partly in response to thosearguing that Medicare’s prospectivepayment systems have not adequatelyincorporated the costs of newtechnologies, the Congress introducedspecific payment mechanisms for hospitaloutpatient services (Balanced BudgetRefinement Act of 1999 and BenefitsImprovement and Protection Act of 2000)and hospital inpatient services (BenefitsImprovement and Protection Act of 2000).

Providing separate payment promotesadoption of new technologies (as long asthe payment is sufficient), thereby helpingto ensure beneficiary access to them.However, the process involves thegovernment heavily in determining whichitems receive separate payment and whichdo not. These choices may influence boththe marketplace for technologies andclinical decision making. Separatepayment unbundles the unit of service,diminishing the efficiency incentives ofprospective payment. It may alsoaccelerate unnecessary use of expensivetechnologies, leading to increased costsfor Medicare beneficiaries and taxpayers.From a systems administrationperspective, separate payment streamsalso put a burden on both CMS andhospitals to incorporate new codes intotheir billing systems. There is a tensionbetween timely inclusion of separatepayment for new technologies thatrequires frequent system changes and thestability of the payment system. Finally, ifthe separate payments are financed in abudget-neutral manner, they will directresources away from other inputs, such asnursing, to fund new technologies.

Given the potential drawbacks ofintroducing separate payment mechanismsfor new technologies, it is important totarget them to technologies that are trulynew, costly, and beneficial. It is throughthe details of the new technology payment


3 ICD–9–CM refers to the International Classification of Diseases, Ninth Revision, Clinical Modification. This is a two-part system of coding patient information used inabstracting systems and for classifying patients into diagnosis-related groups (DRGs) for Medicare. The first part of the ICD–9–CM is a list of diseases; the second partcontains procedure codes, independent of disease codes.

4 HCPCS refers to the Healthcare Common Procedure Coding System. There are three kinds of HCPCS codes. Level I codes are based on the Current ProceduralTerminology coding system developed by the American Medical Association. Level II and III codes, which include many supplies, drugs, and devices, are developed byCMS in collaboration with the Health Insurance Association of America and the Blue Cross Blue Shield Association.

5 In some payment systems, Medicare pays separately for certain inputs. For example, blood products are paid separately under the outpatient PPS.

mechanisms that Medicare balances thegoals of ensuring adequate payment forbeneficial new technologies and being aprudent purchaser.

Hospital inpatient services

Medicare incorporates the costs of newtechnologies into the inpatient PPSthrough the standard systems used to codeservices and set payment rates, as well asthrough the recently implemented add-onpayments for new technologies.

Medicare’s standard systemfor coding and settingpayment rates for inpatienthospital servicesThe unit of payment in the hospitalinpatient PPS is the inpatient discharge, asclassified by diagnosis related group(DRG).6 The DRG system provides for abroad patient classification, encompassingall routine nursing, support service, andancillary costs incurred in patients’ stays.Most technologies are bundled into theDRG payment system. The standardsystem incorporates new technology coststhrough three processes.

First, a technical advisory panel assignscodes to new technologies using theInternational Classification of Diseases,Ninth Revision, Clinical Modification(ICD–9–CM). In response to criticismabout delays in the recognition of thecosts of new technologies, CMS recentlyshortened the coding process to speed theentry of new technologies (CMS 2001).7

Second, CMS responds to requests forrefinements in the classification of costswithin DRGs and analyzes variation in thecostliness of cases within DRGs.

Reassignment to a higher-paying DRGmay occur if certain cases are consideredto be systematically more costly, perhapsbecause of the use of a new technology.For example, for fiscal year (FY) 2003,CMS established new DRGs and higherpayment rates for angioplasty thatinvolves the use of drug-eluting coronaryartery stents, a new technology expectedto be widely adopted once approved bythe FDA.

Third, the annual recalibration of DRGcase weights corrects relative payments bylooking at the most recent year’s claims.

Medicare’s add-onpayments for newtechnologies used ininpatient settings

The Congress authorized add-onpayments for new technologies, whichCMS began to make in fiscal year 2003.The add-on payments described below aresummarized and contrasted with the newtechnology provisions of the outpatientPPS in Table 4-1.

A team of clinical experts within CMSevaluates applications for technologiesthat may raise the cost of a case so muchthat it merits additional payment beyondthe base DRG payment. The eligibilitycriteria for these payments, set forth inregulation, are considered to be fairlystringent, encompassing the newness,clinical benefit, and cost of a newtechnology.

The newness criterion states that a giventechnology will be eligible for add-onpayments until data reflecting its costs areused to recalibrate the DRG weights,generally two to three years from marketentry.

Clinical considerations require that thetechnology substantially improve—relative to technologies previouslyavailable—the diagnosis or treatment ofbeneficiaries (see text box, p. 182).

Cost considerations require the applicantto provide data showing that thetechnology is expensive relative to thecost of the entire case. The applicant mustdemonstrate that the average charge for acase using the technology is one standarddeviation above the geometric mean of thestandardized charges for all cases in therelevant DRG.8 Since the charges per casevary considerably for any given DRG, thestandard deviation is generally large, andtechnologies that meet the cost criteriawill be relatively unusual. These criteriabring together payment issues (How muchdoes it cost?) and those generallyconsidered part of the coverage process (Isit better than technologies previouslyavailable?). Only one technology (abiologic to treat severe sepsis) has metthese criteria to date (CMS 2001, CMS2002b).

Some critics contend that these criteria setthe bar too high. However, strict criteriaprovide a mechanism for ensuring thatlimited funds for new technologies aredirected to those with clinical benefit andhigh costs. They are one tool for balancingMedicare’s need to be a prudent purchaserwith quicker recognition of the costs ofnew technologies.

When a technology is eligible foradditional payment, CMS will notautomatically make an additional paymenteach time it is used. Instead, CMS basesthe payment on the costs incurred for thewhole case, as determined by the fiscalintermediary that processes the claim. Inorder for additional payment to be made,the costs of the case must be above the


6 For a more detailed description of the inpatient PPS, see Appendix A. Regarding the standard process for responding to technology costs under the inpatient PPS, seeChapter 3 of MedPAC’s March 2001 Report to the Congress.

7 The process was shortened by moving the twice-yearly meetings of the ICD–9–CM coordinating committee from May and November to April and December andallowing some new codes to be incorporated into the final rule without first appearing in the proposed rule. The revised schedule allows more new codes to take effectunder the October 1 start of the fiscal year, but restricts the time for public comment on some new codes.

8 Hospitals’ reported charges are standardized through use of the hospital wage index to remove the effects of regional differences in medical costs. To obtain thegeometric mean, the standardized charges are first transformed onto a logarithmic scale before the average is taken. This mathematical approach provides a betterestimate of the average because the distribution is highly skewed and bounded by zero.

standard DRG payment. To preserveincentives for judicious use oftechnologies, Medicare does not pay thefull extra costs of a case using the newtechnology. Rather, the additionalpayment covers only 50 percent of ahospital’s costs above the standard DRGpayment, up to a maximum of 50 percentof the estimated cost of the newtechnology. Some argue that partialpayment is not great enough to ensurebeneficiary access to new technologies.However, this method also provides an

incentive for hospitals to weigh carefullythe benefits of a technology against itscosts. Countervailing competitive andclinical forces also push technologydiffusion.

The add-on payment mechanism is budgetneutral, meaning that CMS lowers thebase payment rate prospectively by thesame percentage for all services to financethe add-on payments. This introduces theneed to balance the impact of the paymentmechanism on payments for all services

against the need for additional paymentsfor new technologies. Expenditures for theadd-on payments are capped at 1 percentof total operating payments.

Hospital outpatientservices

Medicare’s hospital outpatient PPSincorporates costs of new technologiesthrough the standard systems used to code


Design of the hospital inpatient and outpatient new technology payment mechanisms

Outpatient pass-through paymentsOutpatient new

Inpatient add-on Medical Drugs and technologyDesign element payments devices biologicals APCs

Note: APC (ambulatory payment classification).

Source: MedPAC analysis.

T A B L E4-1

What kinds of newtechnologies can receiveadditional newtechnology payments?

Which specifictechnologies receiveadditional newtechnology payments(criteria applied byCMS)?

How are paymentsfinanced?

What is the unit ofpayment?

How are paymentsdetermined?

New technologies thatare an input to an existingservice.

Eligibility based onclinical benefit, newness,and cost.

Budget neutral financing.

Payment based on thecost of the newtechnology.

Payment equal to 100percent of reported costsminus device costsalready built into the basepayment rate.

New technologies thatare an input to an existingservice.

Eligibility based onnewness and cost.


Payment based on thecost of the newtechnology.

Payment equal to 95percent of averagewholesale price.

New technologies thatrepresent a new service.

Eligibility based onnewness.

New expenditures.

Payment based on thecost of providing theservice.

Payment equal to themidpoint of the paymentrange for the newtechnology APC group.For example, payment fora service in the $1,000to $2,000 newtechnology APC is$1,500.

New technologies thatrepresent a newprocedure or are an inputto an existing diagnosis-related group.

Eligibility based onclinical benefit, newness,and cost.


Payment based on theadditional costs oftreating a case using thenew technology.

Payment equal to 50percent of the additionalcosts, capped at 50percent of the estimatedcost of the newtechnology.

services and set payment rates, as well asthrough two mechanisms that specificallytarget new technologies: new technologyambulatory payment classification (APC)groups and pass-through payments. Thetwo new technology payment mechanismsdescribed in this section are summarizedand contrasted with the inpatient add-onpayments in Table 4-1 (p. 181).

Medicare’s standard systemfor coding and settingpayment rates for outpatientservicesThe unit of payment in the hospitaloutpatient PPS is the service provided, asclassified by ambulatory paymentclassification groups.9 The APC system

mixes fairly broad bundles of inputs usedto provide a service such as ambulatorysurgery, with fairly narrow bundles ofinputs to provide an ancillary service suchas an X-ray. In cases where the bundle isnarrow, a specific technology canrepresent a fairly large share of the costsfor providing an outpatient service.Therefore, the outpatient PPS includestwo mechanisms targeted specifically tonew technologies, as well as a standardapproach for maintaining the paymentsystem, similar to that described for theinpatient sector.

Services are classified into APC groupsbased on their Healthcare CommonProcedure Coding System (HCPCS)codes. There are three kinds of HCPCScodes. Level I codes are based on theCurrent Procedural Terminology codingsystem developed by the AmericanMedical Association. Level II and LevelIII codes, which include many supplies,drugs, and devices, are developed byCMS in coordination with the HealthInsurance Association of America and theBlue Cross Blue Shield Association. Bothcoding systems accept applications fornew codes. In the case of newtechnologies, CMS has assignedtemporary codes on a expedited basis tofacilitate payments.

As with the inpatient PPS, CMS respondsto requests for refinements in theclassification of services within the APCsystem. The outpatient PPS is unique inthat it also has an external advisory body,composed of hospital representatives, thatis charged with aiding the agency indefining the APCs. The advisorycommittee also serves as a public forumfor considering requests for changes inAPC groupings.

Finally, the annual recalibration processshould reflect the costs of newtechnologies as reported by hospitals. Therecalibration process undertaken to set thecalendar year 200310 payment rates led tosignificant swings in payment, particularly


Clinical criteria for new technology add-onpayments for inpatient services and the establishment of outpatient pass-through categories for new devices

To be eligible for newtechnology add-on paymentsunder Medicare’s prospective

payment system (PPS) for inpatienthospital services, a new technologymust, among other characteristics,substantially improve the diagnosis ortreatment of beneficiaries. To beeligible for new-technology pass-through payments under the hospitaloutpatient PPS, medical devices mustmeet the same clinical criteria. CMShas established the followingexamples of how a technology canmeet these criteria:

• It offers a treatment option for apatient population unresponsive to,or ineligible for, currentlyavailable treatments.

• It offers the ability to diagnose amedical condition in a patientpopulation in which their medicalcondition is currently undetectable,or offers the ability to diagnose amedical condition earlier in apatient population than allowed bycurrently available methods. Theremust also be evidence that use ofthe technology to make a diagnosis

affects the management of thepatient.

• Use of the technology significantlyimproves clinical outcomes for apatient population as compared tocurrently available treatments. Forexample, improvements mightinclude:

– Reduced mortality rate

– Reduced rate of complications

– Decreased rate of subsequentdiagnostic or therapeuticinterventions (e.g., due toreduced rate of recurrence ofthe disease process)

– Decreased number of futurehospitalizations or physicianvisits

– More rapid beneficialresolution of the diseaseprocess

– Decreased pain, bleeding, orother quantifiable symptom

– Reduced recovery time

Extracted from CMS 2001. �

9 For a more detailed description of the outpatient PPS, see Appendix A. Regarding the standard process for responding to technology costs under the outpatient PPS, seeChapter 3 of MedPAC’s March 2001 Report to the Congress.

10 The hospital inpatient PPS runs on a fiscal year, while the hospital outpatient PPS runs on a calendar year.

for services incorporating newtechnologies. These problems may betransitional, in that this was the first yearCMS used data from hospitals operatingunder the PPS to set the payment rates,and hospitals reported significantdifficulties in coding for technologies.11

However, the small bundles used in theclassification system may make paymentrates inherently less stable.

New technology APC groupsfor technologies used inoutpatient settingsTo be placed in a new technology APC, atechnology must be a complete service orprocedure that cannot be adequatelydescribed by an existing paymentcategory. In addition, it must be a coveredservice that is new and does not meet thecriteria for pass-through payments(described below). For example, PETscans—a newly covered service that doesnot fall into any existing paymentcategory—currently fall in the newtechnology APCs. The new technologyAPCs are grouped into heterogeneouscategories by cost (for example, $0–50, or$5,000–6,000), with payment at themidpoint of the range. No cap or budgetneutrality provision governs the newtechnology APCs. In addition, since thecategories are defined solely by costranges, CMS does not include these APCgroups when recalibrating payment rates.Therefore, each service results inadditional payment. CMS moves servicesout of the new technology APCs and intothe standard system during the annualreview of the APC classification andrecalibration process, when sufficient dataon hospitals’ costs for the technology haveaccumulated.

The lack of a budget-neutrality constraintfor new technology APCs means thatpayments for new technologies couldincrease dramatically. Currently, 75services (as denoted by discreet HCPCScodes) fall into new technology APCs.CMS also has five applications pendingreview. In 2001, services in newtechnology APCs accounted for about 1percent of total payments (MedPACanalysis of 2001 outpatient claims fromCMS).

Medicare’s transitional pass-through payments fortechnologies used inoutpatient settings Transitional pass-through payments covertechnologies that are an input to anexisting service. They are limited tomedical devices, drugs, and biologicals.12

For example, alemtuzumab is amonoclonal antibody used in the treatmentof breast cancer and is paid under thetransitional pass-through mechanism.There are already codes governingchemotherapy administration, but they donot reflect the cost of this new technology.In this case, the pass-through paymentsupplements the base payment forchemotherapy.

Pass-through eligibility criteria aresomewhat different for drugs and devices.Eligibility for devices has been tightenedrecently to introduce clinical criteria inaddition to newness and cost. The clinicalcriteria are essentially the same as thosefor the inpatient add-on payments andrequire that the technology underconsideration substantially improve,relative to technologies previouslyavailable, the diagnosis or treatment ofbeneficiaries (see text box on p. 182).Including clinical considerations for

establishing new categories of medicaldevices raises concerns by somemanufacturers that the bar has now beenset too high, whereas it was previouslythought to be too low by most observers.The cost considerations require CMS toassess the cost of the technology inrelation to the base APC payment rate andcompare its cost to that of similartechnologies it replaces.13

The treatment of drugs and devices isinconsistent, in that only newness and costcriteria are applied to pass-through drugs.This difference in the criteria representsunequal treatment between types oftechnology within the outpatient paymentsystem. It also leads to a discrepancybetween the treatment of drugs under theinpatient and outpatient payment systems,since the clinical criteria are applied to alltechnologies, including drugs, on theinpatient side. Furthermore, withoutconsidering clinical benefit, the criteriaapplied to pass-through drugs may over-emphasize the goal of paying adequatelyfor new technologies at the expense ofprudent purchasing.

Payment for pass-through items is tied touse of the technology itself, withoutconsidering the impact on total costs ofproviding the service, such as efficiencygains or additional incremental costsassociated with use of the technology.That is, each use of a pass-through itemresults in the hospital receiving additionalpayment, whether or not total costs for theentire service actually rose or fell. Formedical devices, the pass-throughpayment is based on 100 percent ofcharges reduced to costs through use of ahospital-specific cost-to-charge ratio. Fordrugs or biologicals, payment is based on95 percent of average wholesale price.14


11 In revising payments for 2003, CMS for the first time used claims and cost reports from hospitals operating under the outpatient PPS. Previously, CMS relied on pre-PPSclaims data and information from manufacturers on the costs of technologies to set the payments. For 2002, manufacturers’ price data were used to incorporate 75percent of medical device pass-through costs into the base APC rates. Using these data led to much higher payment rates in 2002 for services using medical devicesthan the rates calculated for 2003 using cost data reported by hospitals. In addition, it appears that some hospitals did not code accurately for pass-through items,particularly in the beginning of the outpatient PPS. CMS modified the rates for 2003 to limit changes in payment from 2002 (CMS 2002c).

12 Until recently, radiopharmaceuticals could also be considered for pass-through eligibility. However, beginning in January 2003, CMS will no longer considerradiopharmaceuticals (CMS 2002c).

13 The three cost considerations are that the device represent at least 25 percent of the related APC rate, that it be at least 25 percent more expensive than a device itreplaces, and that the increase in cost associated with using the device represents at least 10 percent of the related APC rate.

14 Payments for all pass-through items may be reduced if estimated total payments exceed a statutory cap discussed below.

These payment mechanisms provide anincentive to manufacturers and hospitalsto overstate prices and charges to increasepayments. Analysis of the 2001 claimsindicates that hospitals initially hadtrouble billing for pass-through devices;however, over time both the share ofhospitals billing for these items and theirreported costs increased (CMS 2002c).

The pass-through payment provision isbudget neutral, meaning that payments forall services are reduced by the samepercentage to finance these payments.Budget neutrality redistributes funds toservices or cases that include newtechnologies and away from those that donot. This mechanism can havedistributional effects across types ofhospitals depending on the service mix ofa given provider. The text box on p. 185discusses the distribution of pass-throughpayments among providers.

The Congress capped spending on pass-through items at 2.5 percent of totaloutpatient PPS payments (both programand beneficiary).15 However, the cap wasnot enforced until the last nine months of2002. Because of congressional andadministrative actions, the number andcosts of pass-through items far exceededoriginal expectations in the early years ofimplementation. In 2001, pass-throughitems accounted for over 8 percent of totalpayments (MedPAC analysis of 2001claims). The number of pass-throughitems has subsequently slowed and shouldcontinue at a modest pace, at least in thenear term. The 2003 final rule includesabout two dozen drugs and five devicecategories (CMS 2002c). CMS currentlyhas fewer than 10 applications pendingreview (personal contact with CMS staff).

Despite this modest growth in the numberof pass-through items in the near term,continued medical advances are likely,perhaps leading to pressures to relaxeligibility criteria or increase payments.

For example, a recent review of WallStreet analyses states that medical deviceand supply manufacturers are in goodfinancial health and have increasedspending on research and development,indicating that the pipeline of newproducts will continue (CMS 2002a). Inaddition, the recent passage of legislationauthorizing user fees for FDA review ofmedical devices will likely accelerateapproval of additional technologies.

MedPAC has documented a number ofproblems with the pass-throughmechanism, some of which relate to theeligibility criteria, whereas others involvepayment. The Commission hasrecommended that pass-through paymentsbe selectively targeted to technologies thatare truly new (MedPAC 2001, MedPAC2002). The changes to the criteria for newcategories of medical devices are a step inthat direction. Though CMS has moved totighten the criteria for pass-throughdevices, it is likely that the agency and theCongress will face pressures to relax themin the future. However, MedPAC believesthat it is appropriate to reserve additionalpayments for technologies that provideclinical benefit and do not have clinicalsubstitutes. It may even be appropriate tolimit payments to technologies thatprovide additional benefits commensuratewith their costs. At a minimum, clinicalcriteria should apply to all newtechnologies.

R E C O M M E N D A T I O N :

The Secretary should introduce clinicalcriteria for eligibility of drugs andbiologicals to receive pass-throughpayments under the outpatient PPS.

I M P L I C A T I O N S :

Spending• This recommendation would have no

impact on spending because the pass-through payments are budget neutral.

Beneficiary and provider• The clinical criteria would apply only

to eligibility for additional payment.New drugs and biologicals notmeeting the criteria may still be usedand be paid for at the base APC rate.Therefore, the recommendationshould not affect beneficiaries’ accessto care.

• The recommendation should have noimpact on providers’ paymentsbecause the pass-through paymentsare budget neutral. Limitingadditional payments to drugs andbiologicals that have clinical benefitwill marginally reduce hospitals’administrative burden.

MedPAC has previously noted thatpayments for devices are based onhospitals’ charges (reduced to costs byapplying a cost-to-charge ratio), providingincentives for manufacturers and hospitalsto raise their prices and charges,potentially resulting in overpayments.CMS calculates payments for drugs basedon average wholesale price (AWP). Anumber of studies by the GeneralAccounting Office and the Office ofInspector General have provided ampleevidence that payment based on AWPgenerally results in Medicare paying farmore than market price.16 Incorporatingdata based on inflated costs will lead todistortions in the relative weights(MedPAC 2002). The problems we havenoted previously with the paymentformulas continue and merit further study.

Lessons from other healthcare purchasers

Various private and public sector payersother than Medicare deal with the issue ofpaying for new technologies. To assistdeliberations on how best to pay for newtechnologies in Medicare, MedPACcontracted with Project HOPE to conduct


15 The cap falls to 2.0 percent in 2004 and subsequently.

16 As indicated in MedPAC’s comment letter on the proposed rule for the outpatient PPS, the Commission is concerned about all Part B payments for drugs that are basedon AWP, not just the outpatient pass-through payments. The Commission understands that the Congress and CMS are considering ways of reforming the current system.If the Congress or CMS implements a new system this year, the Commission will monitor the impact of payment changes and their implications for beneficiary access.We will recommend refinements if necessary. If the Congress does not act on the issue this session, we will focus our efforts on analyzing options for change.

a survey of large public and private sectorpurchasers to learn what strategies theyuse to get the best possible prices for newtechnologies. The interviewees includedwell-informed representatives of healthcare insurers, group purchasingorganizations (GPOs), pharmaceuticalbenefit management organizations(PBMs), large integrated deliverysystems, the Department of Defense, theDepartment of Veterans Affairs (VA), theNew York Medicaid program, the United

Kingdom (UK), and Australia (Mohr et al.2002). We also convened an expert panelwith representatives from hospitals,manufacturers, insurers and other payers,academia, and CMS (Mohr 2002). Thepanel discussed the following threequestions:

• What principles should Medicarefollow in paying for new medicaltechnologies?

• What constraints does Medicare facein paying for new technologies?

• What options might Medicareconsider for paying for new medicaltechnologies?

As described below, MedPAC’sstructured interviews found that theapproaches used by other payers includenegotiation, competitive bidding, andother strategies that incorporate value into


Distribution of new technology payments among providers

One issue raised by newtechnology payments isdistributional: Which providers

will receive them? The question is evenmore salient in a budget-neutral system,where payments for all other servicesgo down while those for newtechnologies increase. Generally,hospitals will benefit or not dependingon their service mix. Hospitals thatprovide many technology-drivenservices will benefit whereas thoseproviding more basic services will not.Given that the outpatient prospectivepayment system (PPS) pass-throughpayments represent the first newtechnology provision to beimplemented under a Medicare PPS,the breakdown of payments under thatmechanism illustrates the distributionaleffects of new technology payments.

The outpatient PPS pass-throughpayments were not evenly distributedin 2001 (Table 4-2). Some of the resultsare to be expected. For example,though teaching hospitals receivedabout 50 percent of total outpatient PPSpayments, they received 56 percent ofpass-through payments. Similarly,cancer hospitals, which received only 1percent of total outpatient PPSpayments, received 4.3 percent of thepass-through payments and 5.5 percentof payments for pass-through drugs.However, other results are more

surprising. For example, urbanhospitals did not receive adisproportionate share of pass-throughpayments. Urban hospitals received asimilar share of total payments (81percent) and pass-through payments(80 percent); however, they received ahigher share of payments for pass-through devices (91 percent). Ruralhospitals, in contrast, receivedproportionate shares of total payments(19 percent) and pass-throughpayments (20 percent) but had a

somewhat higher share of payments forpass-through drugs (23.1 percent). Thedistribution of payments was similarfor smaller rural hospitals (less than100 beds). One reason the distributionalimpact was not as marked as might beexpected could be the large number ofitems, many of which were not trulynew, that received pass-throughpayments in 2001. In addition, manycancer drugs that are often provided inrural hospitals were eligible for pass-through payments. �

Outpatient PPS payments by service type andhospital group, 2001

Share of payments by service type

Nonpass- Pass- Pass- Pass-Total through through through through

Hospital group payments payments payments drugs devices

Urban 80.7% 80.8% 79.8% 76.9% 90.6%Rural 19.3 19.2 20.2 23.1 9.4

1–100 beds 9.5 9.5 9.5 11.5 2.5101� beds 9.8 9.7 10.6 11.7 6.9

Cancer 1.0 0.7 4.3 5.5 0.0

Teaching 50.2 49.7 55.9 54.2 62.1Nonteaching 49.7 50.2 44.1 45.8 38.0

Note: PPS (prospective payment system). Numbers may not sum to 100 due to rounding.

Source: MedPAC analysis of the 100 percent Special Analytical File of 2001 outpatient PPS claims from CMS.

T A B L E4-2

decisions about covering and paying fornew technologies. The expert panelconvened by MedPAC suggested thatalthough other payers’ approaches maynot easily be adopted into Medicare’sadministered pricing systems, the programshould pursue the concept of value-basedpurchasing.

Other payers’ approaches topaying for new technologiesEvidence from the interviews and otheranalysis by MedPAC suggest that largepurchasers other than Medicare useseveral strategies to ensure prudentpurchasing of new technologies:

• Staying informed. All respondentsreported that they invest considerableresources in tracking newtechnologies and understanding themedical evidence regarding theirbenefits to bolster their position innegotiations with manufacturers.They monitor the clinical trials beingperformed to obtain FDA approval,plus technology hotlines developedby commercial technologyassessment organizations, and theymay have their own internalcapabilities as well. Price informationmay be obtained from industryanalysts, commercial databases,European experience, or informationgathered from within an integrateddelivery system or health plan, suchas purchase contracts of memberhospitals or claims data for affiliatehealth plans.

• Direct negotiation and contractingwith manufacturers. Some largeintegrated health care systems(including military health care),GPOs, and PBMs negotiate andcontract directly with themanufacturers of new technologies.They use information about atechnology’s clinical effectivenessand costs during their negotiations. Ifa product is a “blockbuster”technology that has great clinicalbenefit and no competitors, themanufacturer is at an advantage insetting its price. Purchasers then try

to limit the length of a contract andintroduce competition clauses torenegotiate prices if a competingproduct enters the market.

• Use of coverage policies and othertools to limit exposure to highprices. Early in the diffusion of anew technology, other payers andpurchasers use various tools torestrict use of new technologies to themost appropriate cases. Examplesinclude tiered copayments,dissemination of guidelines for theuse of a technology, step therapy, inwhich use of a new technology isapproved only if existingtechnologies have been tried andfailed, and prior authorization.

• Competitive bidding. Competitivebidding is used when similar, ortherapeutically equivalent, productsare available. It is especiallysuccessful in closed systems likeintegrated delivery systems or themilitary health services that can limitprocurement. Insurers are less likelyto use competitive bidding. Ifpurchasers know that a new productoffers similar clinical outcomes toexisting or other new therapies, thenthey can offer guaranteed volume to amanufacturer in exchange for a lowerprice. This process generally resultsin lower prices but also limits thechoice of products to be used.Interviewees representing closedsystems suggested that involvingend-users of technologies (usuallyphysicians) in the development ofproduct specifications and guidelinesfor a product’s use makescompetitive bidding more viable andsuccessful.

• Invoice submission. When insurersdecide to cover a new technology thatis not already built into the paymentrates they have negotiated withproviders, they may require providersto submit an invoice showing theircosts. This approach is most effectivefor technologies like medical devices.The insurer will then pay the invoice

cost plus a percentage to coveroverhead. Using this approach, theinsurer avoids the need to pay billedcharges—which often reflect aconsiderable mark-up—and canbenefit from any reductions in pricethat may occur over time. However,invoices generally do not reflect anyrebates that a purchaser has receivedand may, therefore, overstateacquisition costs.

• Cost-effectiveness analysis. Manypayers, both public and private,invest substantial resources indetermining the cost-effectiveness ofnew and existing technologies. Thiswork supports coverage decisionsand plays into payment decisions. Forexample, in Australia, manufacturerswishing to place a pharmaceutical onthe national schedule for the nationalhealth insurance system must submitan application that includes cost-effectiveness information. Pricingdata are considered, and if the costsare considered too high, thegovernment may restrict use of thedrug or negotiate with themanufacturer to reduce the price. TheAustralian health care system alsoapplies cost-effectiveness analysis toother health care interventions,including devices, procedures,diagnostics, and blood products,although the link to pricing is lessclear. In the United Kingdom, theNational Institute for ClinicalExcellence (NICE) provides guidanceto the National Health Service (NHS)about the use of individual healthtechnologies. Although NICE is notdirectly involved in establishingprices for new technologies, it doesinfluence manufacturers’ pricingdecisions indirectly by examiningcost-effectiveness analyses whenmaking their recommendations. If atechnology exceeds a threshold that isloosely set at 30,000 pounds (almost$50,000 given current exchangerates) per quality-adjusted life year,NICE is less likely to recommend theproduct.


• Return on equity. In the UnitedKingdom, prices for pharmaceuticalsare subject to a cap based on areasonable return on equity.Manufacturers may set any price theywish at product launch, subject to theconstraint that the total rate of returnon capital invested in the UK on alltheir products reimbursed by theNHS does not exceed a pre-set limit.The return on equity is limited to arange of 17–21 percent. If the rate ofreturn exceeds these targeted rates,the manufacturer must grant the NHSa rebate or reduce the price of thedrug. Manufacturers must submitaudited financial returns detailingtheir investment in the UK. Thereturn on equity approach appliesonly to companies based in the UK.

Applicability of otherpurchasers’ strategies toMedicareEvidence from the expert panel discussionand other analysis by MedPAC suggestthat other payers’ approaches may noteasily be adopted into Medicare’sadministered pricing program, but point tovalue-based purchasing as a futuredirection to pursue.

Constraints unique to Medicare

Medicare faces constraints that otherpayers do not and that may limit its abilityto use the alternative strategies outlinedabove. These constraints have to do withthe size and national scope of theMedicare program, its role as an insurer,and program issues like public disclosurerequirements and limited administrativecapacity.

Medicare covers more than 40 millionAmericans. This large market means thatdecisions made by the Medicare programcan have a large impact. In the area ofnew technologies, Medicare’s decisionscan greatly affect the financial status of amanufacturer and also have an impact onfuture innovation. Restricting Medicare’spurchasing to one or two suppliers, as is

generally done under the competitivebidding arrangements of the organizationswe interviewed, could determine whichsuppliers flourish and which do not. Ofcourse, Medicare could structurecompetitive bidding to involve moreplayers. In addition, other payers oftenfollow Medicare in setting payment rates,leading to an even greater influence on themarket. Furthermore, the Medicareprogram is national in scope. Undercurrent law, payments are set nationally.This makes it difficult for Medicare totake advantage of local market conditions,such as market share, that might allow theprogram to negotiate better prices in onearea as compared to another.

The Medicare program acts as an insurer,reimbursing hospitals and physicians fortheir services using administered pricingsystems required under law.Consequently, the program has no directrole in negotiating with manufacturers ordistributors. It also has little control overthe choices made by providers servingMedicare beneficiaries. Of the strategieslisted above, both negotiation strategiesand competitive bidding are done best byclosed delivery systems, such as the VAor an integrated delivery system, which dohave the ability to negotiate prices and toinfluence the delivery of care to enrollees.However, the Medicare competitivebidding demonstration project forpurchase of durable medical equipmentmay provide lessons that can be applied toother parts of the program.

Administrative issues also constrain theprogram. For instance, Medicare mustfollow rule-making processes that involvepublic comment unless there is a specificexception in law, such as the Medicaidprescription drug rebate program. Publicdisclosure requirements limit theprogram’s ability to obtain and useproprietary information. For example,Medicare would be less successful innegotiating the best price for an item ifthat price then becomes public becausemanufacturers would face pressure tooffer that price to all purchasers. Even if

the program had authority to negotiateprices in confidence, administration of thepayment system currently requires theprogram to publish payment rates for useby the fiscal intermediaries and hospitals.The rule-making process also adds time toany decision-making as time must beallowed for comment by interest groupsand response from CMS.

Finally, Medicare has limitedadministrative capacity to implement thealternative strategies noted above. Mostprivate payers devote considerableresources to monitoring the newtechnology pipeline and conductingtechnology assessments. Large systems,such as the VA, even conduct clinicaltrials to evaluate technologies. Givencurrent resources, CMS may not be ableto make the same level of investment inthese activities as other organizations havedone.

Other environmentalconsiderations

In addition to the system constraints notedabove, other factors prevent the Medicareprogram from engaging in the strategiesused by many other large payers andpurchasers. Since Medicare is anentitlement program, beneficiaries and thegeneral public have expectations aboutaccess and choice, making decisions aboutlimiting access to specific itemscontroversial. Similarly, Medicare as apublic sector payer is expected to ensure alevel playing field among competingmanufacturers, which limits its ability tobe selective, a major tool used by otherpayers. Selectivity also runs afoul of thelaw stating that Medicare will not interferewith the practice of medicine.17

One strategy that Medicare might considerdespite these constraints is limiting returnon equity, as done in the UK. Thisapproach does not limit the number ofsuppliers or establish a specific price, butregulates the return to the manufacturer,and is one factor that must be taken intoaccount when setting a product’s price.One advantage of this approach is that it


17 The Medicare�Choice program and the competitive bidding demonstrations are governed by separate statutory provisions that allow some level of selectivity.

operates directly on the manufacturer.Since manufacturers of new technologiesare generally at an advantage in pricenegotiations, this approach providesincentives to limit the price to the mostappropriate actor.

The use of return on equity in the UK,however, is based on a number of factorsthat may not be applicable to Medicare.The UK limits this approach to a singleindustry, pharmaceuticals, and conductsits return-on-equity calculations based onthe whole portfolio of products that amanufacturer sells to the National HealthService, including both existing and newtechnologies. Therefore, there is no needto allocate investments to a specificproduct. By contrast, Medicare serves alimited population (the NHS covers thewhole population), and would only wantto use the return on equity approach to setpayments for new technology.Consequently, Medicare would want todetermine a return on equity for a specificproduct, or, alternatively, all of themanufacturer’s products used in providingMedicare services. This would requiresubstantial review of manufacturers’finances and sophisticated accounting toseparate out expenses and revenue streamsfor a subset of products. A return onequity calculation might also need to takeinto account some share of firms’investments in unsuccessful products.Given the large number of unsuccessfulproducts manufacturers pursue in additionto the successful ones, this calculationcould prove complex.

Furthermore, since the UK establishes areturn for all products, manufacturers arefree to price new drugs well above theestablished range of return to takeadvantage of market position as amonopoly provider of a new product.Applying return on equity to a subset ofproducts would limit a manufacturer’sability to do this. Another wrinkle is thatthe return-on-equity approach is used onlyfor firms based in the UK and applies onlyto investments made in the UK. If

Medicare were to adopt this approach, theprogram would need to decide how totreat investments overseas and firms notbased in the United States. CMS has alsonoted that it does not have theadministrative capacity or legal authorityto develop a return-on-equity approachand doubts that the resources needed todevelop one are warranted given that newtechnology payments are meant to belimited to a small number of technologies(CMS 2002d). Despite these complexities,return on equity may be a reasonableapproach for setting payment rates incertain situations, such as when there is asingle producer of a technology with clearclinical benefits and no substitutes, andMedicare is the predominant purchaser ofthe technology. It would, however, signala major break with Medicare paymentpolicy, which generally avoids regulatingprofits.

Another possibility is the use of third-party purchasers. Can Medicare contractwith multiple GPOs and PBMs tonegotiate better prices for these items? Itseems clear that the limited volume ofnew technology items makes thisapproach less viable. However, the use ofthird-party purchasers by Medicare hasbeen discussed in the context of payingfor Part B drugs and outpatientpharmaceuticals under a Medicare drugbenefit. If these strategies are pursued,third-party purchasing of newtechnologies might be considered as anadditional role.

Value-based purchasing asa future directionAlthough the specific techniques used byother payers seem to have limitedapplicability in an administered pricingprogram that is national in scope, likeMedicare’s prospective payment systems,together they embody a concept that couldprove useful to the program. In paying fornew technologies, other payers strive forvalue-based purchasing. That is, they tryto limit coverage to those technologies

that provide a demonstrated clinicalbenefit, and assess the level of additionalbenefits over existing technologies againstthe additional costs for the newtechnologies. For example, cost-effectiveness information is used innegotiations with manufacturers, andestablishment of therapeutic equivalenceis key to competitive bidding. Mostparticipants in the expert panel on howMedicare should pay for new technologiesagreed that the program should pursuevalue-based purchasing, although they didnot agree on specific approaches for doingso.

Value-based purchasing involves makingjudgments about the benefit of a newtechnology compared to other availabletherapies and considering the value of theadditional costs associated with use of thenew technology. Under value-basedpurchasing, additional payments would beless likely for a new technology that hasexisting substitutes, even if the newtechnology is substantially more costly. Ifthe same clinical outcome is achieved, is itnecessary to pay more than is paid for theexisting technology? If there are modestclinical gains at a great increase in price,should the program pay?

The clinical criteria introduced for add-onpayments under the inpatient PPS and formedical devices under the outpatient PPSmove in the direction of value-basedpurchasing by having Medicare determinethe clinical benefit of a new technologybefore it receives additional payment. Thenext step, however, of assessing the valueof that clinical benefit, or the relationshipof the clinical benefit to the extra cost, hasnot been taken systematically.18

Several methodological issues surroundvalue-based purchasing. These include,among others: establishing the level ofevidence needed to assess value;specifying a measure for assessing benefit,such as quality-adjusted life-years; anddefining the scope of the costs andbenefits to be included in assessing value,


18 While value-based judgments are not made systematically within the Medicare program, they have been used in at least one case. In the November 1 final rule for theoutpatient PPS (CMS 2002c), CMS did declare a new anemia treatment, darbepoeitin alpha, “functionally equivalent” to an existing treatment, epoeitin alpha, andreduced the pass-through payment for the new biologic to $0. The agency used its authority to ensure equity of payments in taking this step, which was novel.

such as impact on future wage earnings orcost-savings because of a reduced needfor future medical interventions. Inaddition, the choice of a threshold valuethat a technology must exceed to receiveadditional payment would likely become apolitical issue, leading to extensive debateamong manufacturers, clinicians,beneficiaries, and other interested parties.In fact, previous attempts by Medicare tointroduce cost-effectiveness analysis intothe coverage process have been blocked.For example, in 1989 CMS (then theHealth Care Financing Administration)put forth a notice of proposed rule-making

that included cost-effectiveness as acoverage criterion. The rule was neverfinalized. Later, in 2000, CMS published anotice of intent of proposed rule-makingthat outlined a four-step process forconsidering the value of an item or servicewhen making national coverage decisions.The agency has yet to follow up on thisissue. In both instances, resistance byaffected interest groups was consideredone element in delaying action (Foote2002).

Despite methodological and otherchallenges to its development, value-based purchasing provides a frameworkfor deciding where to spend scarcedollars. Expanding its ability to pursuevalue-based purchasing would allowMedicare to better balance the goals ofpaying enough for beneficial newtechnologies to ensure beneficiary accessto appropriate care, and being a prudentpurchaser. �


References

American Hospital Association, The Lewin Group. Cutting edge costs: hospitals and newtechnology, AHA TrendWatch. September 2002, Vol. 4, No. 4.

Centers for Medicare & Medicaid Services. Wall Street’s view of medical device andsupply manufacturers, Health Industry Market Update. October 10, 2002a.

Centers for Medicare & Medicaid Services. Medicare program; changes to the hospitalinpatient prospective payment system and fiscal year 2003 payment rates; final rule,Federal Register. August 1, 2002b, Vol. 67, No. 148, p. 49982–50289.

Centers for Medicare & Medicaid Services. Medicare program; changes to the hospitaloutpatient prospective payment system and calendar year 2003 payment rates; andchanges to payment suspension for unfiled cost reports; final rule, Federal Register.November 1, 2002c, Vol. 67, No. 212, p. 66717–67046.

Centers for Medicare & Medicaid Services. Medicare program; changes to the hospitaloutpatient prospective payment system and calendar year 2003 payment rates; andchanges to payment suspension for unfiled cost reports; proposed rule, Federal Register.August 9, 2002d, Vol. 67, No. 154, p. 52091–52371.

Centers for Medicare & Medicaid Services. Medicare program; payments for newmedical services and new technologies under the acute care hospital inpatient prospectivepayment system; final rule, Federal Register. September 7, 2001, Vol. 66, No. 174, p.46902–46925.

Foote SB. Why Medicare cannot promulgate a national coverage rule: a case of regulamortis, Journal of Health Politics, Policy, and Law. October 2002, Vol. 27, No. 5, p.707–730.

Garrison LP, Wilensky GR. Cost containment and incentives for technology, HealthAffairs. Summer 1986, p. 46–58.



Mohr PE. Paying for new medical technologies: what options might Medicare consider?Submitted to MedPAC. October 31, 2002.

Mohr PE, Neumann PJ, Bausch S. Paying for new medical technology: lessons for theMedicare program from other large health care purchasers. Submitted to MedPAC.October 31, 2002.


Health insurance choicesfor Medicare beneficiaries

C H A P T E R5

ince the Medicare program began, beneficiaries have been able to

make limited choices about their health coverage. Policymakers

have sought to broaden these choices; some want to use choice as

a platform for a system of competition among Medicare and

private plans. Many Medicare beneficiaries now have available to them an in-

creasingly complex array of options beyond traditional Medicare fee-for-service

and varying forms of supplemental coverage. How and when beneficiaries

choose among these options depends on a number of factors, including specific

market conditions and the circumstances of individual beneficiaries.

The determinants of how supply and demand for health insurance meet in the

marketplace are both national and local. They reflect the tension between

Medicare as a national program and the reality that it is only at the local level that

medical care is organized and delivered, beneficiaries choose insurance options

and delivery systems, and decisions to enter the insurance market are made. In

this chapter we review the entire spectrum of options as a first step in MedPAC’s

larger effort to better understand beneficiaries’ choices and market conditions.

S

C H A P T E R

Health insurance choices forMedicare beneficiaries

5In this chapter

• What health insurance optionsdo Medicare beneficiarieshave?

• Medicare beneficiaries andhealth plans in the marketplace

• When supply and demandmeet in the marketplace


Medicare beneficiaries face a complexarray of health insurance options,including the traditional Medicare fee-for-service (FFS) program; various forms ofinsurance that supplement the traditionalprogram; and alternatives to the traditionalprogram such as managed care, privatefee-for-service (PFFS), and preferredprovider organization (PPO) plans. Whichof these options, other than the nationally-available FFS program, are available tobeneficiaries depends on local marketconditions. Which they choose—orwhether they decide to choose at all—depends on the circumstances andmotivations of individual beneficiariesand the information available to them.

Although supplemental insurance andoptions for receiving care in managed careplans have been available to beneficiariessince the Medicare program began, thearray of choices for receiving Medicareand supplemental coverage has becomeincreasingly important both forbeneficiaries and for Medicare programspending. Policymakers have sought toexpand Medicare beneficiaries’ healthinsurance options for a variety of reasons.Some sought to offer Medicarebeneficiaries a wider choice of plans thatmight better meet their perceived need forhealth insurance and provide access tohealth care delivery system options thatare popular among the employedpopulation. Some sought to build aplatform for a system of competitionamong plans that might provide bettermanagement of care, market-determinedrates for providers, and better quality.1

The theory is that if plans compete on thebasis of product, quality, and price, and ifmarkets work well, beneficiaries andproviders will have the incentive to takethe costs and quality of health care intoaccount, which could help controlMedicare spending in the long run.

Choice has evolved over the years fromhealth maintenance organizations (HMOs)paid on a cost basis, to HMOs paid on arisk basis, to the currentMedicare�Choice (M�C) program, to

newly developed demonstration programs.The M�C program was established by theBalanced Budget Act of 1997 (BBA).When the program became effectiveJanuary 1, 1999, it allowed private plansto offer Medicare beneficiaries optionsbeyond the traditional FFS Medicareprogram, including HMOs and othermanaged care plans, private fee-for-service plans, and Medical SavingsAccounts. However, during the last fiveyears, many plans left the M�C program,and few new non-HMO optionsmaterialized. Enrollment declined sharplyas private plans withdrew, andbeneficiaries were upset by the instabilityin plan choices and reductions in benefitsoffered by plans. There have beenconcerns that the program has failed.

In response, the Congress and the Centersfor Medicare & Medicaid Services (CMS)have been trying several approaches toencourage greater plan participation.Plans’ regulatory concerns have beenaddressed; the Congress extended the lifeof Medicare HMOs that are paid on a costbasis; and CMS undertook ademonstration program to encouragePPOs to participate in M�C.

In this chapter, MedPAC examines thestatus of the Medicare program withrespect to health insurance options forMedicare beneficiaries on a national level,including not just M�C options but allforms of supplemental insurance. Thechapter begins by describing the healthinsurance options available to someMedicare beneficiaries, as well as the waythe options have evolved over the last fewyears. The second section of the chapterdescribes constraints on Medicarebeneficiaries’ choices in the healthinsurance market and examines Medicarebeneficiaries’ actual choices andsatisfaction, as well as the perspective ofhealth insurers, highlighting changesinsurers might like to see in order tostimulate participation.

In the final section of the chapter, weanalyze how potentially conflictingpreferences might play out in the healthinsurance marketplace. The productsavailable to beneficiaries varyconsiderably across regions and states andeven within metropolitan areas. Further,competition between options is not limitedto M�C versus traditional fee-for-serviceMedicare alone. There is also competitionbetween comprehensive plans andtraditional Medicare plus supplementalpolicies that are available to manyMedicare beneficiaries. The availability ofoptions, their costs, plus variations inM�C benefits and premiums can createvery different market dynamics in localmarkets across the nation. Furtherresearch is needed to help understandmore about how local markets arestructured and how they might work forMedicare.

What health insuranceoptions do Medicarebeneficiaries have?

Although most of the concern and debateabout the availability of health insurancechoices for Medicare beneficiaries haverevolved around the participation ofprivate managed care plans—predominantly HMOs—in theMedicare�Choice program, beneficiariesalso make choices about other Medicare-related insurance products available tothem. Therefore the discussion herecovers the broad range of health insuranceoptions available to Medicarebeneficiaries. We describe two generaltypes of insurance products:

• Insurance products that replacethe traditional Medicare FFSbenefit package. Such productsinclude M�C managed care plans,called coordinated care plans (CCPs);M�C private fee-for-service plans;Medicare demonstration PPO plans;and Medicare cost plans.

194 Hea l t h i n s u r an ce c ho i c e s f o r Med i ca r e bene f i c i a r i e s

1 The Institute of Medicine (IOM) report, Crossing the quality chasm, points out that the fragmented nature of Medicare’s traditional fee-for-service program makes theimplementation of some quality improvements more difficult (IOM 2001).

• Insurance products thatsupplement the traditionalMedicare FFS benefit package.Products designed to fill in or “wraparound” the basic Medicare FFSbenefit package include Medigapplans, Medicare Select plans,employer-sponsored retiree plans,and Medicaid.

The availability and attractiveness of theseproducts varies by geographic area and bybeneficiaries’ individual circumstances.Products that replace the traditionalMedicare FFS benefit packages, forexample, are available only in some areasof the country. Further, the cost to thebeneficiary and the benefits provided varysignificantly—even among areas wherethese replacement products areavailable—depending on factors such asMedicare payment, market characteristics,and beneficiaries’ need for services.Though generally more widely available,even some products that supplement theMedicare benefit package are availableonly to certain beneficiaries. For example,retiree supplemental coverage is limited tobeneficiaries who have worked for theemployers or unions that offer thiscoverage. Medicaid coverage is availableonly to beneficiaries who meet the low-income and other standards set by thestate in which they live. Finally, evensupplemental products available to almostall beneficiaries have premiums that canvary by market and beneficiary age.

Insurance products thatreplace the traditionalMedicare FFS benefitpackageMedicare beneficiaries can enroll in someinsurance products which serve asalternatives to the traditional Medicareprogram. When beneficiaries enroll inmost of these alternatives, they must giveup their traditional benefits (though theycan disenroll at the end of any month andreturn to FFS Medicare). In addition to

providing beneficiaries with Medicarebenefits, most of these alternatives offersome supplemental benefits.

M�C coordinated care plans

Under M�C, Medicare beneficiaries havethe option of joining a private CCP, whichthen receives payment from Medicare forproviding all Medicare-covered services.Generally, members of M�C CCPs mustuse plan providers to get their care. Theseprivate plans are allowed to provideadditional benefits and to chargebeneficiaries an additional premium forthem. However, if a plan’s projected costsfor Medicare benefits are lower than itsMedicare payments, the plan is requiredby law to either return the difference toenrollees in the form of additional benefits(or lower premiums) or contribute themoney to a reserve fund for future use(few plans choose this option).Historically, beneficiaries have been ableto join these plans and receive extrabenefits at no additional premium.

M�C CCPs have been the core of theM�C program, but they are not availableeverywhere and their benefit packagesvary considerably. Currently, M�C CCPsare available to about 58 percent of theMedicare population, down from 74percent availability at the peak in 1998.However, less than 20 percent of ruralbeneficiaries have a plan available.Currently about 5 million beneficiaries areenrolled in an M�C CCP, down fromabout 6 million in 1998. In explainingM�C plan participation trends, it isimportant to note that the CCP model isdominated by HMOs, which have beenwithdrawing in the private sector as well.

Medicare payments for M�C CCPsMedicare pays M�C CCPs a monthlycapitated rate for each enrolled Medicarebeneficiary based on the beneficiary’scounty of residence and relative healthcost risk. (See Appendix A.)

As a result of this payment system,Medicare has paid more to M�C plans,on average, than it would have paid toinsure demographically similarbeneficiaries under the traditional FFSprogram for the basic benefit package.MedPAC has calculated that in 2001,Medicare’s payments were about 104percent of average FFS costs. Thiscalculation assumes there are no riskselection differences (other than thosesuch as age, sex, and Medicaid status thatare included in the rate-setting model)between the M�C plans and traditionalMedicare. For 2003, we project the ratewill also be 104 percent.

Benefits and costs to beneficiaries ofM�C CCPs The benefit packages andbeneficiary premiums for the packagesvary quite a bit. Almost 30 percent ofMedicare beneficiaries have a planavailable in their county in 2003 thatcharges no premium. In fact, about 4percent of beneficiaries have access to aplan that will, in essence, pay them tojoin.2 At the other end of the spectrum,some plans charge premiums in excess of$200 per month. Premiums reaching thatlevel result, at least partially, from theplan providing benefits in addition to thebasic Medicare benefits. The data do notallow us to calculate the average premiumpaid, but the lowest premium available tobeneficiaries averages $40 per monthacross all M�C markets.3

The additional benefits offered and co-payments required also vary considerably.Plans can and do charge deductibles, flatcopayments, and percentage coinsuranceon days, stays, or benefit periods. Becauseof the complexity of the benefit offerings,we focus on a few indicators to compareacross packages. We looked at threesupplemental benefits sometimes offeredby plans:

• some coverage for outpatientprescription drugs,


2 The actual transaction will involve the plan paying some or all of the beneficiary’s Part B premium.

3 The Centers for Medicare & Medicaid Services (CMS) reports the number of beneficiaries enrolled under a managed care organization’s (MCO) contract. An MCO mayhave several different plans (each with different benefit packages and premiums) under a single contract, but CMS has not reported the number of enrollees in eachplan. CMS has begun collecting the plan-level enrollment information.

• inpatient hospital services coveredwithout any cost-sharing, and

• physician office visits coveredwithout any cost-sharing.

Almost half of all Medicare beneficiarieshave an M�C CCP available that coverssome prescription drugs. Almost 30percent of beneficiaries have a planavailable that does not charge any cost-sharing for inpatient hospital services.About 10 percent of beneficiaries have aplan available without any cost-sharingfor physician services.

M�C private fee-for-serviceplans

M�C PFFS non-network plans operatelike traditional FFS insurance plans in thecommercial sector. They allowbeneficiaries to use any provider who willaccept the plan’s reimbursement rates.(Although allowed by law, there arecurrently no network PFFS plans.)Medicare pays these plans the same ratesas it pays other M�C plans. They aresubject to most of the same conditions ofparticipation as other M�C plans, butsome quality data reporting requirementsare less stringent.4 As is the case withother M�C plans, PFFS plans may alterthe cost-sharing arrangements forMedicare benefits, subject to approval byCMS. CMS reviews the structure in anattempt to ensure that selection bias willnot occur.5 It is unclear how a non-network plan would compete financiallywith the traditional Medicare programexcept in areas where the payment ratesare above FFS spending. The Medicareprogram currently pays approximately102 percent of what it would be expectedto pay to insure demographically similarenrollees under the traditional program.

There are currently three M�C PFFSplans: One of them is a demonstrationplan that operates in only one county;

another, established in 2000, operates inmost of 25 states and is available to aboutone-third of all Medicare beneficiaries,and the third has just been approved andwill operate in six states. Enrollment inthe established multi-state plan is low(about 20,000 enrollees) but has beengrowing steadily since its inception in2000. However, the plan has pulled out ofsome areas in each of the last two years.

Benefits and costs to beneficiaries ofthe M�C PFFS plan The establishedmultistate M�C PFFS plan sets astandard benefit package across its entireservice area. For 2003, this M�C PFFSplan charges a monthly premium of $88.The plan does not cover outpatientprescription drugs. For inpatient hospitalservices, the beneficiary has a copaymentof $100 per day, up to a maximum of$500 per stay. (There is no limit to thenumber of days in a stay under this plan.)The beneficiary must notify the planbefore a planned admission; otherwisethere is an additional copayment of $50per day, up to a maximum of another $500per stay. For physician services, thebeneficiary’s copayment is $15 perprimary care visit and $30 per specialistvisit.

The newly approved plan charges amonthly premium of $19 and providessome coverage for outpatient prescriptiondrugs.

Medicare preferred providerorganization (PPO)demonstration plans

Although the statutory language thatestablished the M�C program specificallymentioned PPOs as examples of CCPs,only a few PPOs have ever participated inthe program. CMS wants to encouragePPOs to enter the M�C program, for atleast two reasons: (1) to enhancecompetition in the Medicare marketplaceand (2) to make the most popular form of

insurance in the commercial sector morereadily available to Medicarebeneficiaries.

CMS identified several barriers to PPOs’participation in the M�C program(Centers for Medicare & MedicaidServices April 2002):

• Low M�C payment rates in someareas. M�C payment rates were toolow in some areas for PPOs to recruitproviders into networks.

• PPOs’ reluctance to participate ina fully capitated program. Anotherbarrier to PPOs’ participation inM�C has been their wariness aboutentering the fully capitated M�Cprogram. In the commercial world,PPOs often share the risk on medicalcosts with the employers who offerthe PPOs to their employees. In manycases, the PPOs carry no medical riskand offer administrative-services-only contracts to self-insuredemployers.

• The M�C limit on premiums andcost-sharing. The M�C limit on costsharing (designed to protectbeneficiaries from paying highercost-sharing in M�C than under thetraditional program) hinders benefitdesign in some geographic areas. Theactuarial value of all cost-sharing,including premiums and copaymentsrelated to basic Medicare services,cannot exceed the national averagecost-sharing amount for thetraditional fee-for-service Medicareprogram, which is about $102 permonth for 2003. Because this cap isbased on a national average, it hasbeen troublesome for HMOs inhigher-than-average cost areas, andwould be even more of a problem forPPOs, which often include substantialout-of-network cost sharing.6


4 Because non-network PFFS plans do not have a network, their control over provider behavior is limited. Therefore, the plans are not required to report some of the qualitymeasures or participate in quality improvement projects that relate to provider practices.

5 CMS does not review or approve PFFS plans’ premiums, as it must with CCP premiums.

6 Beneficiary cost sharing is correlated with Medicare payments: the more Medicare pays for services, the higher beneficiary cost sharing. For Part B services, cost sharingis generally 20 percent of Medicare-allowable charges. Thus, in areas where Medicare spending is higher than average, it can be expected that beneficiary costsharing would be higher than average.

CMS has initiated a Medicaredemonstration program for PPOs in orderto encourage plans to enter M�C. TheMedicare PPO demonstration program isscheduled to run for three years beginningin January 2003. CMS has approveddemonstration waivers for 33 plans in 23states. The plans will be available to 11million Medicare beneficiaries(Department of Health and HumanServices 2002). Under the demonstrationprogram, payment rates will be higherthan M�C rates in some areas, the limiton cost sharing will be waived, and theMedicare program will offer to sharesome of the cost risk with the plans.

While the PPO demonstration programmay provide an additional option to manybeneficiaries, it is not likely to increasethe choices available to beneficiaries whodo not already have other alternatives toMedicare FFS. Of the more than 11million beneficiaries who will have a PPOavailable, only about a half million do notalready have a CCP available. Generally,demonstration plans are going into urbanareas, but a couple of the plans aretargeted to rural areas. As a result, out ofapproximately 10 million ruralbeneficiaries, about 600,000 will haveaccess to PPOs, but 450,000 of themalready have a CCP available. It remainsto be seen whether those who enroll inPPOs will come from the coordinated careplans, or have fee-for-service coverageonly, or have FFS plus Medigap.

Medicare payments for PPOdemonstration plans Under theMedicare PPO demonstration program,plans will be paid the higher of the M�Crate in the county or 99 percent of theaverage risk-adjusted per capita spendingunder the traditional FFS Medicareprogram. Demonstration plans will alsohave the opportunity to individuallynegotiate risk-sharing arrangements withMedicare. If beneficiaries enroll in thePPOs at the same rate in each countywhere they are offered (e.g., if 1 percentof beneficiaries in each county enroll),

PPO spending will average 109 percent ofthe cost of insuring the enrollees in theFFS Medicare program.7 The reason thatthe Medicare costs would be so high isthat the PPOs are going into manycounties where M�C payment ratesexceed fee-for-service spending.

Benefits and costs to beneficiaries inPPO demonstration plans Almost allof the PPO demonstration plans willcharge premiums, ranging from $32 to$184 per month. All but one of the PPOswill offer some coverage for outpatientprescription drugs. About one-fifth ofbeneficiaries who have a demonstrationPPO available will have one that chargesno cost-sharing for inpatient hospitalservices in network hospitals. Plans thatcover physician visits without any cost-sharing will be available to only about 2percent of beneficiaries who have a PPOavailable.

Medicare cost plans

Cost HMOs have been authorized toparticipate in the Medicare program since1972 (National Academy of SocialInsurance 1998). They were designed toallow Medicare beneficiaries who were inHMOs before they became eligible forMedicare to stay in those HMOs.Medicare pays the HMOs their cost, asdetermined by a cost report, for providingMedicare benefits for their members, lessthe actuarial value of traditional Medicarecost sharing. The beneficiaries in costHMOs generally cover this cost sharingthrough monthly premiums rather thanpayments as services are delivered. Inaddition, members are free to seekMedicare-covered services outside of theHMO’s network. If a beneficiary goes to anon-network provider, Medicare pays theprovider the same as if the beneficiarywere in the traditional FFS program, andthe beneficiary is responsible for the usualMedicare FFS cost sharing. To thebeneficiary, this structure is similar tobeing in a point-of-service (POS) HMO.

Although Medicare cost plans have beenattractive to some beneficiaries, paststudies have shown that this option coststhe Medicare program significantly morethan serving beneficiaries in thetraditional fee-for-service program (Singet al. 1998). However, those studies arebased on old data and compared costsonly relative to the traditional program.Though that comparison may be the bestone to examine, it may also be relevant tocompare cost plan performance to theperformance of M�C plans, because inareas where the M�C plans are paid morethan FFS costs, the cost plans might resultin Medicare spending less than for theM�C plans. The cost plan program is setto expire at the end of 2004, but theprogram has already been extendedseveral times, and there has beencongressional interest in extending itfurther.

Currently, 30 Medicare cost plans are inoperation, with a total of 290,000members. Those numbers should risebecause two M�C CCPs are shifting theirmembership to Medicare cost plans thatthey also operate.

Benefits and costs to beneficiaries inMedicare cost plans Premiumsgenerally range from $29 per month to$326 per month (there is one zero-premium plan). Half of the cost planofferings have monthly premiumsbetween $72 and $116. While less thanhalf of the plans include coverage foroutpatient prescription drugs, some of theones that do not provide coverage offerhigh-option choices that do include drugcoverage. Most of the plans charge nocost-sharing for inpatient hospital servicesin a plan hospital, and about one-third donot charge cost-sharing for visits to planphysicians.


7 The 109 percent figure was calculated by comparing the projected per capita FFS spending in each participating county with the rate the PPO would be paid in eachcounty. Those county-level comparisons were then aggregated and weighted by the number of Medicare beneficiaries in each county. The calculation assumes that thereis the same level of health risk in the PPO and non-PPO populations and that the risk-sharing arrangements have no aggregate net effect.

National and local availability ofalternatives to Medicare’straditional FFS program

About 80 percent of Medicarebeneficiaries nationwide live in countieswhere an alternative to Medicare’straditional FFS program—an M�Ccoordinated care plan, an M�C privatefee-for-service plan, a PPO demonstrationplan, or a Medicare cost plan—isavailable to them (Table 5-1). Thesealternatives are available to 85 percent ofurban beneficiaries but only 61 percent ofrural beneficiaries. Furthermore, whileurban beneficiaries may have a range ofplans to choose from, the only option forrural beneficiaries is generally the PFFSplan. Looking at availability ofalternatives to the traditional MedicareFFS program in terms of M�C countypayment rates, we find that 86 percent ofbeneficiaries who live in counties withpayment rates above the floors8 (asdetermined in 2002) have a plan available,while 74 percent of beneficiaries in floorcounties have a plan available. In additionto these alternatives, which are open to allMedicare beneficiaries,9 there are somespecialized plans that offer benefitsattractive to the frail elderly which aresometimes available only to categories offrail beneficiaries. (See text box.)

Insurance products thatsupplement the traditionalMedicare FFS benefitpackageIn addition to choosing among insuranceproducts just discussed which are intendedas an alternative to (and sometimes addto) the traditional Medicare FFS benefitpackage, beneficiaries can also chooseamong products designed solely to wraparound, or supplement, the basic Medicarebenefit package. All aged beneficiarieshave the option of buying a Medigap planwhen they first enroll in Medicare (this is

not the case for disabled beneficiariesunder age 65; see p. 199). Manybeneficiaries can also choose to buy aMedicare Select plan. Some beneficiariesmay be fortunate enough to have theoption of participating in an employer-sponsored retiree plan. Other beneficiariesmay be eligible to receive supplementalbenefits from state Medicaid programsand other programs designed to assistlow-income individuals.

Medigap plans

Medigap insurance is private coveragedesigned specifically to wrap around theMedicare benefit package. Most Medigapinsurance is marketed directly toindividual Medicare beneficiaries,although some employers and associationshelp enroll their retirees and members inthese publicly available plans (Chollet andKirk 2001).

Private supplemental insurance, similar towhat we now call Medigap insurance, hasexisted since Medicare began, but theOmnibus Budget Reconciliation Act of1990 (OBRA 1990) imposed somestructure on the market, simplifying andclarifying offerings for beneficiaries.10

Pursuant to OBRA 1990, the NationalAssociation of Insurance Commissioners(NAIC) created 10 standard plans,commonly labeled A through J, and statesretained primary responsibility forregulating Medigap policies andinsurers.11

For the most part, all standardized plansare available to all beneficiaries as theyturn age 65, although not every plan issold in every state. When beneficiariesturn age 65, they have a one-time openenrollment period during which Medigapinsurers must allow the beneficiary to


Availability of alternatives to the traditional Medicarefee-for-service program, 2003

Percent of beneficiaries with plans available,by type of county of residence, 2003

Percent of PPO Costbeneficiaries M�C CCP PFFS demo contracts Any plan

National 100% 58% 36% 23% 23% 80%

County payment rateFloor 55 40 50 15 16 74

Large urban floor 31 61 43 24 19 82Other floor 23 12 58 3 12 63

Non-floor 45 80 20 32 30 86

Rural areas 23 13 56 4 9 61Urban areas 77 72 30 28 25 85

Note: CCP (coordinated care plan), M�C (Medicare�Choice), PFFS (private fee-for-service), PPO (preferredprovider organization). For 2003, the large urban floor is $564.10 and other floor is $510.38.

Source: MedPAC analysis of data from CMS website, August 2002 and September 2002.

T A B L E5-1

8 The floor payment rates are described in the M�C section of Appendix A.

9 Beneficiaries who have end-stage renal disease (ESRD) and are being maintained by chronic dialysis may not enroll in an M�C plan, unless they were previously in aplan before developing ESRD.

10 Many beneficiaries had been subject to questionable sales practices and had purchased multiple policies that often duplicated existing coverage (Super 2002). TheCongress found that the policy offerings needed to be standardized.

11 Insurers in three states (Massachusetts, Minnesota, and Wisconsin) are not subject to the standards for plans A–J. These states were granted waivers because they hadpreexisting standards which they continue to maintain.

enroll in any open product. During thisperiod insurers are prohibited frommedically underwriting the beneficiary—

meaning that they cannot consider thebeneficiary’s health and medical history indeciding whether to offer a policy and

how much to charge. Medigap plans areoften unavailable to disabled beneficiaries(under age 65) because these federalguaranteed-issue requirements are limitedto beneficiaries turning 65 or in an M�Cplan that no longer participates in theprogram. Except in the few states thatrequire pure community rating, Medigapplans can be prohibitively expensive forolder or sicker beneficiaries seekingcoverage. (See text box, p.200, for age-rating methodologies.) After the six-month open enrollment period, Medigapinsurers in most states can medicallyunderwrite new applicants. This practiceis common, particularly for Medigapplans that include prescription drugcoverage. Once enrolled, however,beneficiaries can not be dropped fromtheir Medigap plan, as the policies provideguaranteed-renewal protection.

Over 10 million, or about 27 percent ofMedicare beneficiaries living in thecommunity in 2000 were enrolled in aMedigap plan.

Benefits and costs to beneficiaries inMedigap plans Medigap plans generallyprovide coverage of Medicare’s cost-sharing requirements. All standardizedplans (A through J) cover cost-sharing forphysician and inpatient hospital services,except for the $100 Part B deductible andthe $840 inpatient hospital stay deductible(Table 5-2, p. 201). Plans B through Jcover the inpatient deductible, and plansC, F, and J cover the Part B deductible.Three of the standard plans (H, I, and J)offer limited coverage of outpatientprescription drugs, but all come with a$250 annual deductible, 50 percentcoinsurance, and a cap on benefits of$1,250 per year (plans H and I) or $3,000per year (plan J). Relatively fewbeneficiaries enroll in the three plans thatoffer prescription drug coverage, and mostare in either plan C or plan F. About 25percent of Medigap enrollees have stayedin their prestandardized plans that havebeen closed to new enrollment since 1992.The benefits in the nonstandardized planstend to be similar to those found in thestandardized plans.


Managed care programs for frail beneficiaries

Over the years, the Congresshas created a variety ofmanaged care programs to

meet the needs of beneficiaries withimpairments in activities of dailyliving. These programs generally havebeen available in relatively fewlocations. Three of the more long-lived examples are the Program ofAll-Inclusive Care for the Elderly(PACE), the Social HealthMaintenance Organization (S/HMO)program, and EverCare.

Program of All-Inclusive Care forthe ElderlyPACE is a permanent program underMedicare and a state option underMedicaid. Most PACE enrollees areeligible for both Medicare andMedicaid, and the program is targetedto enrollees with substantialfunctional impairments. A primaryobjective of PACE is to delay orprevent use of hospital and nursinghome care. The program provides acomprehensive range of preventive,primary, acute, and long-term care,beyond what is available throughMedicare and Medicaid. PACEservice delivery and coordination areusually organized through adult dayhealth centers. There are now 15permanent PACE sites in 8 states,enrolling around 2,000 beneficiaries.Another group of PACE sites is stilloperating under demonstrationauthority while CMS considers theirapplications to join the permanentprogram.

Social Health MaintenanceOrganizationThe S/HMO demonstration programhas had two phases, called

generations. Both generations havetaken a traditional HMO model thatenrolls a wide spectrum ofbeneficiaries, and added a limitedlong-term care benefit. The secondgeneration program was startedlargely to address perceivedshortcomings with the first. The twogenerations of S/HMO programsdiffer in the way that Medicare paysthem, the degree to which theycoordinate care across benefits andproviders, and their targetingmechanisms for long-term carebenefits. The demonstration project isslated to end on August 1, 2003.There are now 4 S/HMO sites in 4states, enrolling around 112,000beneficiaries; 37 percent of enrolleesare in the single second generationS/HMO plan.

EverCareThe EverCare demonstration programenrolls permanent nursing homeresidents into managed care. Thedemonstration builds on theexperience of the United Health CareEverCare company in subcontractingwith Medicare HMOs to providemedical care for enrollees who live innursing homes. Unlike PACE andS/HMO, EverCare does not expandthe Medicare benefit packagesignificantly; instead, it focusesprimarily on providing moreMedicare-covered outpatient servicesto reduce residents’ use of hospitaland emergency room care. Thedemonstration project is slated to endon December 31, 2003. Six EverCaredemonstration sites now operate in 6states, enrolling around 17,000beneficiaries. �

The average premium for individualMedigap insurance across all plan types—standardized and nonstandardized—was$129 per month in 2001. The averagepremium for plan F, the most commonstandardized plan option, was $122 permonth; premiums for standardized plansthat include outpatient prescription drugcoverage ranged from $119 for plan H to$196 for plan J. Medigap premiums varyconsiderably by state.12 Premiums alsovary substantially according to the age ofthe beneficiary and the ratingmethodology used (see text box at left).For example, policies for olderbeneficiaries in attained-age rated policiesmay cost considerably more than policiesthat use issue-age or pure communityrating.

Medicare Select plans

The Medicare Select program began as ademonstration in the early 1990s and wasmade permanent in 1998. Medicare Selectpolicies are Medigap policies that covermore of the cost-sharing whenbeneficiaries use network providers. Fromthe beneficiaries’ point of view, aMedicare Select policy is exactly the sameas a Medigap policy when they use anetwork provider, but coverage is not ascomplete as with a comparable Medigapplan when they use non-networkproviders. In exchange for giving up somecoverage for non-network providers, theSelect policies usually have lowerpremiums than comparable Medigappolicies.13 Insurers are able to offer theseless-expensive products because providersagree to accept rates lower thanMedicare’s in order to participate in thenetwork. Because Medicare continues topay its share of the claims from Selectmembers, the reductions really are in theform of waiving all or part of thebeneficiary cost-sharing.

Current Medicare regulations, however,have allowed these cost-sharingreductions only for hospital services. TheOffice of Inspector General (OIG) of the


Medigap age-rating

Generally, insurancecompanies use three differentmethods to determine the

prices, or rates, for their plans, basedon the age of the enrollee:

• Pure community rating: Allenrollees in the same geographicarea pay the same premium,regardless of age.

• Issue-age rating: Enrollees paypremiums based on their age whentheir policy was first issued tothem.

• Attained-age rating: Enrolleespay premiums based on theircurrent age.

State insurance rules regulate whichmethod, or methods, insurers may use.The methods determine the relativelevels of premiums beneficiaries willface as they age.

Under pure community rating,younger policyholders generally paymore than their expected costs whileolder policyholders pay less than theirexpected costs. This cross-subsidization may be desirable forolder beneficiaries who may be lessable to afford higher premiums tied totheir expected costs. Insurers mayface special challenges undercommunity rating, however. In orderto keep premiums low, insurers needto maintain an enrollee population thatis balanced between older andyounger policyholders as theiroriginal policyholders age. Thatmeans they need to attract a steadystream of younger beneficiaries,which usually requires keepingpremiums low. If the premium is toohigh, younger beneficiaries may feelthat they will not get good value from

a policy, and they may wait until theyare older to purchase a policy orpurchase a policy that is rateddifferently. Such delaying behaviorcould lead to increases in the cost ofthe policies.

Under issue-age rating, beneficiarieshave a stronger incentive to buy apolicy without delay, because thepremium is based on their age whenthey first buy the policy. For example,if a beneficiary buys a policy at age65, the premium will continue to bethe same as that offered to new 65-year-old beneficiaries. This ratingstructure also provides incentives forbeneficiaries to stick with a planbecause in many states some of theirpremiums are put into a reserve tofund their higher expected costs asthey age.

Attained-age rating reduces cross-subsidies between groups of youngerand older beneficiaries. The premiumsfor younger beneficiaries willgenerally be lower than under anyother rating structure. However, thepremiums for older policyholders willbe higher than under any otherstructure and can becomeprohibitively expensive for manybeneficiaries.

In addition to rating by age, insurersin some states can rate by otherbeneficiary variables, including sex,whether or not the beneficiarysmokes, and the geographic areawhere the beneficiary lives. Finally, ifbeneficiaries want to enroll in plansoutside of the time periods in whichthey have guaranteed-issue rights,plans in the majority of states mayunderwrite them, charging more forbeneficiaries with certain healthconditions or denying coverage. �

12 For further discussion of Medigap products and reasons behind the variation in premiums, see Appendix B in MedPAC’s Report to the Congress: Assessing MedicareBenefits, June 2002.

13 GAO found that in 1999 the average annual premium for a Select plan was more than $200 lower than the average premiums for non-Select plans (GeneralAccounting Office July 2001).

Department of Health and HumanServices (HHS) had ruled that Part Bproviders could not waive cost-sharingwithout violating anti-kickback rules.Studies of Medicare Select found that theprogram was limited because plans couldnot include physicians in their networks,which kept them from any real possibilityof saving money through managing care(Lee et al. 1997). The OIG has nowproposed regulations that would allowphysicians and suppliers to waive Part Bcost-sharing if they participate in anetwork. If physicians are willing toaccept lower total Medicare payments toparticipate, then insurers might be able topass along savings in the form of lowerpremiums. Network creation may alsoallow plans to pursue managed careobjectives within their networks. In anyevent, if this regulatory change allowsinsurers to lower premiums on Selectplans, they may become a more attractiveoption for beneficiaries.

Benefits and costs to beneficiaries inMedicare Select plans Select benefitsare the same as Medigap benefits exceptthat beneficiaries may be limited in theirchoice of providers. For the most part, thepremiums are lower because the insurersget lower rates from network providers.More than one million Medicarebeneficiaries are enrolled in Select plans.

Employer-sponsored retireeplans

The most common form of supplementalcoverage is employer-sponsoredinsurance, which covers 33 percent ofnoninstitutionalized Medicarebeneficiaries. Some of these beneficiarieshave access to employer-sponsoredcoverage in their current jobs or through aspouse’s employer, but the majorityreceive coverage as part of their retireebenefit packages. While some employersenroll their retirees in M�C or other

managed care plans, most of the planswrap around the Medicare benefitpackage.

While employer-sponsored insurance hasbeen the largest source of supplementalcoverage, it has been declining. Over thepast decade, the proportion of employersoffering retiree health coverage hasdeclined, even during the strong economyof the late 1990s.

A nationally representative survey ofpublic and private employers with 500 ormore employees found that 23 percentoffered health coverage to Medicare-eligible retirees in 2001, down from 40percent in 1994 (Mercer 2002). Thedeclines have accelerated in recent years:The percentage of firms with 200 or moreworkers offering coverage to retirees overage 65 declined by 10 percentage pointsbetween 1999 and 2001. The same surveyfound that the percentage of small firms


Benefits, enrollment, and average premiums in standardized Medigap plans, 2001

Standardized Medigap plan

Benefits, enrollment, and premiums A B C D E F G H I J

Cost sharing

Part A hospital coinsurance ● ● ● ● ● ● ● ● ● ●

365 additional hospital days ● ● ● ● ● ● ● ● ● ●

Part B coinsurance ● ● ● ● ● ● ● ● ● ●

Blood products ● ● ● ● ● ● ● ● ● ●

Part A deductible ● ● ● ● ● ● ● ● ●

Part B deductible ● ● ●

Skilled nursing facility copayments ● ● ● ● ● ● ● ●

Part B balance billing ● ● ● ●

Additional benefits

Foreign travel ● ● ● ● ● ● ● ●

Home health care ● ● ● ●

Preventive medical care ● ●

Prescription drugs ● ● ●

Enrollment 11% 9% 23% 6% 3% 37% 3% 2% 3% 4%

Average monthly premium $91 $102 $117 $114 $108 $122 $121 $119 $170 $196

Note: Percentages do not sum to 100 because of rounding.

Source: Medicare Payment Advisory Commission analysis of 2001 Medicare Supplemental Exhibits from the National Association of Insurance Commissioners.

T A B L E5-2


Federal programs that provide supplemental coverage to retirees

Three Federal programs providesupplemental coverage toretirees.

Department of Defensesupplemental health benefitsThe National Defense AuthorizationAct for Fiscal Year 2001 created theprogram TRICARE For Life (effectiveOctober 1, 2001) to wrap aroundMedicare benefits. TRICARE For Lifeprovides supplemental coverage formilitary personnel and retirees enrolledin Medicare. Approximately 1.5million people are eligible for thisbenefit. The 2001 National DefenseAuthorization Act also created a newprescription drug benefit that provideseligible Medicare beneficiaries with thesame pharmacy benefit enjoyed bymilitary personnel not eligible forMedicare. Medicare beneficiaries whomeet the eligibility criteria areautomatically enrolled in TRICAREand in the pharmacy benefit program,with no application process.

TRICARE covers virtually all ofMedicare’s cost-sharing requirements,including deductibles and coinsurancefor inpatient and outpatient services. Itprovides unlimited coverage forinpatient hospitalizations and skillednursing facility stays, with beneficiariesresponsible for 20 to 25 percentcoinsurance for stays beyond thenormal Medicare-covered allowance.The program also offers acomprehensive prescription drugbenefit that gives beneficiaries theoption of obtaining prescription drugs atno cost from military treatment facilitiesor with only nominal copays from anypharmacy. In general, for mostMedicare-covered services, Medicarewill pay first and TRICARE will paythe beneficiaries’ remaining out-of-pocket expenses. If beneficiaries have

other sources of coverage, TRICAREpays after the other sources have paid.The program includes a $3,000 annualout-of-pocket limit (Politi 2002).

To be eligible for TRICARE,beneficiaries must pay the MedicarePart B premium but are not required topay any additional premium. Eligiblebeneficiaries include uniformed serviceretirees (including retired guard andreservists) who served at least 20 yearsin the military, family members ofuniformed service retirees (includingwidows/widowers), and certain formerspouses of uniformed service retirees, ifthey were eligible for TRICARE beforeage 65.

Department of Veterans Affairshealth benefitsIn 2003, an estimated 3.3 millionbeneficiaries will be enrolled in theDepartment of Veterans Affairs (VA)health care system (CongressionalBudget Office 2002). For individualswho qualify, the VA program providesgenerous benefits at little or no chargeto the beneficiaries, including broadcoverage of most inpatient andoutpatient services; preventive care;and prescription drug coverage. TheVA program has become increasinglypopular in recent years, with more than1 million new enrollees in the past 5years. The growth has been fueledlargely by elderly veterans seekingprescription drug coverage (Simmons2002).

To receive health care from the VAsystem, veterans generally must beenrolled with the VA. (Thoughdisabled veterans do not have to enroll,the VA encourages them to enrollformally to help the agency’s planningand resource allocation process.)Veterans are enrolled subject to

available appropriated funds, based ona priority system of eligibilitycategories, with veterans with service-connected disabilities rated 50 percentor higher accepted first. Veteransdeemed unable to make copayments fortheir treatment are given higher prioritythan others who do not have service-connected disabilities and who agree topay copayments. To qualify based oninability to defray the costs of theircare, veterans must supply the VA withincome and net worth information,which is compared to a financialthreshold. Enrollment is reviewed eachyear. Those in the lowest priority grouppay the Medicare hospital deductiblefor the first 90 days of care during any365-day period, and one-half of theMedicare deductible for each additional90 days of hospital care, as well as a$10 per day charge for each hospitalday. This group is also responsible forcopayments for most outpatient care.

Outpatient pharmacy services areprovided free to eight categories ofveterans (subject to available VAfunds), based on service-connecteddisability and other special needscriteria; others pay a fixed copayment($7 per prescription in 2002). For mostpriority groups, there is also an annualcap on copayments for drugs, includingboth prescription and over-the countermedications and supplies dispensed bya VA pharmacy ($840 in 2002); thosein the lowest priority group who areresponsible for copayments for otherhealth services are not protected by thecap (Department of Veterans Affairs2002).

The Federal Employees HealthBenefits Program (FEHBP)In addition to providing employment-based group insurance to active federalworkers, FEHBP provides group

(continued next page)

(those employing 3–199 workers) offeringretiree health coverage fell from 9 percentin 2000 to 3 percent in 2001 (Henry J.Kaiser Family Foundation,Commonwealth, HRET 2002). Few, ifany, employers have added healthcoverage for Medicare-eligible retirees(Mercer 2002).

These declines generally affect future,rather than current, retirees. In 2001, 5percent of large employers had plans thatcovered only current retirees, or thosehired before a certain year (Mercer 2002).Employers also have increased thenumber of years of service required toqualify for retiree health benefits (WatsonWyatt Worldwide, 2002). Most of theimpact of this change has yet to be felt. Itis not apparent in current coverage trends,but will appear gradually over time astoday’s workers, who have less-generousemployer contributions or no retireehealth benefits at all, begin to retire(General Accounting Office May 2001).

Not only has the number of firms offeringcoverage to their retirees declined, butthose firms that offer coverage have beenscaling back on drug benefits and

increasing retirees’ premiumcontributions. Among firms that offerretiree health benefits, 32 percentincreased cost-sharing for prescriptiondrugs, and 53 percent increased retirees’share of the premium between 1999 and2001. About 36 percent of largeemployers have capped their contributionstowards retiree coverage for either currentor future retirees (Hewitt Associates, LLC2001).14

Special attention is often paid to federalretiree health programs, but they areessentially employer-sponsored plans.15

(See text box at left.)

Benefits and costs to beneficiaries inemployer-sponsored retiree plansThe average premium paid for employer-sponsored health insurance by newretirees over age 65 was $79 per month in2002, up 20 percent from 2001 (Henry J.Kaiser Family Foundation, HewittAssociates 2002). About 20 percent ofemployers providing coverage do notrequire new retirees to pay a premium.Currently, benefits provided by employer-sponsored plans tend to be

comprehensive. Almost all retiree plans(96 percent of those issued by large firms)provide some coverage for prescriptiondrugs (Henry J. Kaiser FamilyFoundation, Hewitt Associates 2002).Further, about 90 percent of the plans thatcover prescription drugs have no upperlimit on that coverage. Although we donot have specific information on requiredcost-sharing for hospital or physicianservices, the average retiree with coveragehas an out-of-pocket cap of $1,500 peryear for all covered service (Henry J.Kaiser Family Foundation, HewittAssociates 2002).

Medicaid

In 2000, about 11 percent of beneficiariesliving in the community were enrolled inthe federal/state Medicaid program whichsupplemented their Medicare coverage.Medicaid offers several levels ofsupplemental coverage to eligible low-income beneficiaries. In addition, somelow-income individuals who do not meetall of the requirements for dual eligibilityreceive Medicaid coverage for part or allof their Medicare premiums or cost-sharing requirements.16


14 These caps were put in place to limit employers’ future liability for retiree health insurance. Employers began setting caps in the early 1990s in response to theFinancial Accounting Standards Board’s approval of Financial Account Statement No. 106 in 1990. It required employers to report annually on their current and futureretiree health benefit liabilities and include them on their balance sheets, beginning with fiscal years after December 15, 1992. The Governmental AccountingStandards Board Statement No. 34 makes a similar requirement for state and local governments, which is now being phased in.

15 Health benefits for retirees receiving care through the Indian Health Service (IHS) are an exception. For Native American and Alaska Native beneficiaries, the IHS isthe primary payer. The IHS does not technically “supplement” Medicare; rather, it provides a wide range of health services, some of which are paid for by Medicare.About 60,000 Medicare beneficiaries were served by 47 IHS or tribal-operated hospitals in 2001. Since the passage of BIPA in 2000, Medicare reimburses IHS forservices in hospitals and skilled nursing facilities, and also pays for services of physicians and nonphysician practitioners furnished in hospitals and ambulatory clinics.Noncovered services are provided by the IHS. Services may be provided through provider-based or freestanding tribal federally qualified health centers, hospitals,ambulatory care centers, or individual practitioners employed by the IHS. Native Americans and Alaska Natives using IHS health care may also be eligible forMedicaid benefits (Health Care Financing Administration April 10, 2001).

16 The Balanced Budget Act of 1997 (Public Law 105–33) allowed states to pay providers the lower of Medicare’s cost-sharing requirements or the states’ Medicaid rates,although providers are not permitted to charge beneficiaries the difference. In 1999, only 16 states reimbursed providers for the full amount of Medicare’s cost-sharingrequirements (Nemore 1999).

Federal programs that provide supplemental coverage to retirees (continued)

insurance to federal retirees. About 31percent of the 8.3 million peoplecovered by FEHBP are retired, and 1.8million (21 percent) are enrolled inMedicare (Quayle, 2003). FEHBPoffers retirees a range of commercialhealth plans, including both national

and local fee-for-service plans,preferred provider organizations, point-of-service plans, and managed careplans. The benefits included in theplans, when coordinated with MedicareFFS, are generally comparable to thoseof retiree health insurance supplements

offered by other large public- andprivate-sector employers—i.e., theygenerally fill in Medicare cost-sharing,plus offer some additional coverage forpreventive care, routine physicals, andprescription drugs. �

The benefit package for Medicarebeneficiaries who are fully eligible toreceive Medicaid (dual-eligiblebeneficiaries) is one of the mostcomprehensive of all Medicaresupplemental options. The vast majorityof dual-eligible beneficiaries do not paypremiums for Medicare or Medicaid, andany cost-sharing requirements aregenerally nominal (see text box above). Inaddition, dual-eligible beneficiariesgenerally receive a comprehensiveprescription drug benefit throughMedicaid.17

Despite the generosity of benefitsavailable to dual-eligible beneficiaries,participation in Medicaid by eligibleMedicare beneficiaries is low in most

states. An estimated 24 percent of allnoninstitutionalized beneficiaries areeligible for or enrolled in one of theMedicaid programs. However, fewer thanhalf of those eligible to receive Medicaidassistance actually do (Laschober andTopoleski 1999).

Common explanations for the lowparticipation rate include lack ofknowledge of the programs, the stigmaassociated with Medicaid, and barriers toenrollment (such as a complex applicationprocess). Beneficiaries commonly believethat Medicaid is for only “poor people”and that applying could put their estates atrisk (General Accounting Office 1999).Medicare beneficiaries who are eligible

but not enrolled in Medicaid are morelikely to be 80 years or older, married, andotherwise insured (through Medicaremanaged care or private supplementalinsurance) than are enrolled beneficiaries(Laschober and Topoleski 1999). The waya state implements its Medicaid programsalso affects participation rates. In 1999,more than half of states did not use asimplified enrollment application; morethan three-quarters of states did notprovide outreach materials in languagesother than English; and about two-thirdsof states did not make eligibility screeningtools available to outside agencies, clinics,or senior centers (Nemore 1999). Otherresearch has shown that enrollment inMedicaid is higher in states that havemore generous Medicaid programs(Pezzin and Kasper 2002).18

In 1999, the proportion of Medicarebeneficiaries classified as dual eligiblevaried by state, ranging from a high ofalmost 28 percent in Mississippi andTennessee to less than 8 percent inArizona, Idaho, and Utah (Ellwood andQuinn 2002). Compared with the rest ofthe eligible Medicare population, dual-eligible beneficiaries tend to bedisproportionately female (63 percentversus 55 percent), over age 85 (18percent versus 10 percent), and membersof racial or ethnic minority groups (38percent versus 14 percent) (CMS 2002).

Medicare beneficiariesand health plans in themarketplace

In this section, we describe constraints onMedicare beneficiaries’ choices in thehealth insurance marketplace, andexamine Medicare beneficiaries’ actualchoices and satisfaction. We then look atthe health insurance marketplace from theperspective of the health plans that serveMedicare beneficiaries.


17 Some low-income beneficiaries who do not qualify for Medicaid receive assistance for the purchase of outpatient prescription drugs through Medicaid 1115 waivers.The programs can involve considerable cost-sharing.

18 The measures of state Medicaid program generosity were based on the percentage of state Medicaid long-term care expenditures allocated to home and community-based care (HCBC), and on Medicaid per capita expenditures per elderly enrollee on HCBC waiver programs designed to help beneficiaries remain in the communityand avoid being institutionalized.

Medicaid benefits available to Medicarebeneficiaries not eligible for full Medicaid benefits

Several mandatory Medicaidprograms pay beneficiaries’Medicare premiums or cost-

sharing requirements:

• Qualified Medicare Beneficiary(QMB) program. Under the QMBprogram, states pay Medicare’spremiums, deductibles, andcoinsurance for all beneficiarieswhose income is at or below 100percent of the federal poverty leveland whose assets are at or belowtwice the Supplemental SecurityIncome limit. In providingcoverage for Medicare premiumsor cost-sharing, QMB coverageresembles a Medigap plan C orplan F (covering most ofMedicare’s cost-sharingrequirements without providingadditional benefits).

• Specified Low-Income MedicareBeneficiary (SLMB) program.Under the SLMB program, statespay the Medicare Part B premiumfor beneficiaries with incomesbetween 100 percent and 120percent of poverty.

• The Qualifying Individuals-1(QI-1) program. Under the QI-1program, states pay the Part Bpremium for beneficiaries withincomes between 120 and 135percent of poverty. Because theQI-1 program’s federal funding islimited, assistance is available on afirst-come, first-served basis(General Accounting Office 1999).

Although Medicaid’s premium andcost-sharing assistance programs aredefined by federal law, states havediscretion in how they implementthese programs (Nemore 1999). �


19 Some states require guaranteed issue and/or community rating on some or all plans (requiring insurers to charge the same premium to all insured persons, regardlessof age or health status) for disabled Medicare beneficiaries.

20 As plans have revised or scaled back additional benefits, the array of benefits, cost-sharing arrangements, and exclusions can become very complicated. According toone study that compared options for actual plans in two cities, “differing plan packages make it nearly impossible to compare plans on costs” (Dallek and Edwards2001).

Medicare beneficiariesWhen viewed at the national level, thehealth insurance market for Medicarebeneficiaries appears to offer manychoices, including whether to enroll in anM�C plan, or whether, or how, tosupplement Medicare FFS. As we havediscussed, however, the availability ofoptions varies tremendously depending oneach beneficiary’s geographic location,work history, income, health care needs,and other factors. Beneficiaries may notbe able to afford some of the healthinsurance coverage options available tothem, especially the options with thebroadest scope of benefits. Beneficiaries’coverage options are constrained not onlyby availability of the M�C plansdescribed above but also by factors suchas underwriting restrictions on Medigappolicies for some beneficiaries, financialresources, and by what is available tothem in employer-sponsored supplementalinsurance programs. Beneficiaries’preferences and health care needs mayalso affect the extent to which they arewilling to change providers or healthplans, or are interested in consideringoptions at all.

As noted above, statutory provisionsallow for a 6-month period of openenrollment for all of the standardizedMedigap options for beneficiaries enteringthe Medicare program at age 65, and (fora subset of plans) for beneficiariesaffected by the withdrawal of M�C plansfrom their market area. Beneficiariesentitled to Medicare by reason ofdisability do not have this federalguaranteed access to Medigap until theyreach age 65 and may therefore be deniedcoverage.19 Beneficiaries who want toenter the Medigap market after the openenrollment period ends may be subject tounderwriting based on age or healthcondition, depending on state law.Further, many states allow insurers to ratepolicies based on beneficiaries’ ages.

Some beneficiaries, particularly those whohave existing health care problems or areolder, may have only a small number ofpolicies open to them and those policiesmay not be affordable.

For beneficiaries with employer- or union-sponsored retiree health insurance,choices among insurance alternatives mayalso be constrained. Employers may notoffer Medicare managed care options. In2002, about half of all large employersoffered a Medicare managed care option(Henry J. Kaiser Family Foundation,Hewitt Associates 2002). Employers whodo offer Medicare managed care may beable to take advantage of the supplementalbenefits offered by the plans, loweringtheir own costs. This may lead someemployers to require higher premiums forretiree benefits that supplement MedicareFFS, and lower premiums for managedcare options. In fact, while manyemployers do not offer M�C options,those who do offer them play an importantrole in the M�C market. UnpublishedCMS data from 2002 show that 18 percentof M�C enrollees were in employer orunion-sponsored groups (Zarabozo 2003).

The ability to pay for insurance tosupplement Medicare is clearly a limitingfactor for some beneficiaries. Researchhas generally shown that the main reasonpeople choose to join M�C plans is toobtain better benefits for less cost thanthey can get from Medicare plus privatesupplemental insurance (Gold 2000;Young and Mittler 2002). A survey ofbeneficiaries conducted in 2000 byMathematica Policy Research, Inc. (MPR) showed that the majority ofbeneficiaries who had no supplementalinsurance (Medicaid or private) reportedthat supplemental insurance was tooexpensive or that they could not afford it(Gold and Mittler 2001). Analysesreported in MedPAC’s June 2002 reportshow that beneficiaries with incomes

below 200 percent of poverty are morethan twice as likely as higher-incomebeneficiaries to go without any form ofsupplemental insurance (MedPAC June2002).

Beneficiaries’ decisions about health plansand supplemental insurance also reflecttheir health care needs and preferences.Choice of a doctor, access to specialists,or a desire to stay with the same doctormay be particularly important to peoplewith health care problems and long-standing relationships with particularproviders. For many, coverage forservices not covered by traditionalMedicare—notably prescription drugs—iscritically important. For some, particulardetails of plan offerings (e.g., provisionsrelated to dental services, hearing aids oreyeglasses, or particular aspects of plandrug formularies) may be important.20

Finally, some research suggests that manyMedicare beneficiaries are not highlymotivated to make choices about theirinsurance coverage. MPR’s 2000 surveyof beneficiaries found that mostbeneficiaries (in both FFS and M�Cplans) did not give serious thought tooptions for insurance coverage. Only 14percent thought seriously about options oractually changed plans, and, of those,more than one-third were either newbeneficiaries (who had to make a choice)or beneficiaries who switched from oneM�C plan to another. Of those who didnot consider options seriously, by far themost common reason offered (65 percentof respondents) was “I like what I have”(Gold et al. 2002). Other research suggeststhat retirees may be less likely thanyounger workers to make decisions abouthealth insurance options based primarilyon cost, in part because of concerns thatretirees—especially those with health careproblems—may have about changingdoctors (Buchmueller 2000; Strombom etal. 2002).

Insurance choices made byMedicare beneficiaries

Although Medicare beneficiaries’insurance choices have been shaped by avariety of constraints, the resulting systemof multiple insurance coverage has, for themost part, provided supplementalcoverage for most beneficiaries. MedPACanalysis shows that only 9.3 percent ofMedicare beneficiaries living in thecommunity had traditional FFS Medicarecoverage only for most of the year in2000.

Note, however, that figures on the types ofsupplemental insurance held by Medicarebeneficiaries are based on survey dataavailable only through the year 2000.Because some M�C plans havewithdrawn and some employers havereduced retiree benefits, these estimates ofcoverage may not accurately reflectbeneficiaries’ current insurancecoverage.21 MedPAC’s analysis of the2000 Medicare Current BeneficiarySurvey (MCBS) Cost and Use file showsthat about one-third of all beneficiariesliving in the community have employer-

sponsored supplemental insurance, andnearly 30 percent of beneficiaries havepurchased Medigap (Table 5-3).

Analysis of choices about health careoptions also suggests that Medicarebeneficiaries are particularly interested inobtaining prescription drug coveragewhen it is available. CMS data show thatwhen plans offer a choice in benefitdesign, most beneficiaries in those planschoose to pay the higher premium for thepackages that include drug coverage(Zarabozo 2002).


Sources of additional coverage by selected beneficiary characteristics, 2000

Percent distribution

Percent ofbeneficiaries Employer- Medicareliving in the sponsored Medigap managed Medicarecommunity insurance insurance Medicaid care Other only

All beneficiaries 100.0% 32.0% 27.0% 11.6% 18.3% 1.8% 9.3%Age

Under 65 13.6 27.2 5.0 34.4 9.7 3.2 20.565–69 23.9 35.6 23.0 7.4 21.0 1.9 10.970–74 22.2 34.1 30.2 7.3 19.7 1.7 7.175–79 19.0 32.3 33.0 8.4 19.4 1.3 5.680–84 12.1 30.9 35.4 8.0 19.4 1.1 5.285� 9.2 25.4 38.1 10.2 17.2 1.8 7.3

Income statusBelow poverty 15.9 9.8 13.9 46.2 12.0 2.2 15.9100 to 125% of poverty 10.3 15.3 23.6 22.6 19.8 3.1 15.0125 to 200% of poverty 22.1 27.7 30.9 6.2 21.6 2.3 11.4200 to 400% of poverty 33.0 42.5 28.4 1.1 20.5 1.5 5.9Over 400% of poverty 18.4 46.6 32.9 0.6 15.0 0.8 4.1

ResidenceUrban 76.1 33.7 23.0 10.5 22.9 1.6 8.1Rural 24.9 26.7 39.8 12.8 3.9 2.6 13.1

Note: Income status is defined in relationship to the poverty level in 2000 ($8,259 if living alone and $10,419 if living with a spouse). Urban includes beneficiaries inmetropolitan statistical areas (MSAs). Rural includes beneficiaries living outside MSAs. Beneficiaries according to the type of coverage they held for at least six months of theyear.

Source: MedPAC analysis of 2000 Medicare Current Beneficiary Survey, Cost and Use file.

T A B L E5-3

21 A large share of the beneficiaries who no longer have Medicare managed care coverage probably now have Medigap plans. Data from 2000 suggest that Medigapenrollment is increasing as managed care enrollment declines. A 1999 survey found that 75 percent of beneficiaries who were involuntarily disenrolled from M�Cplans, and did not join a different managed care plan, found a different source of supplemental coverage (Barents 1999). The benefits offered may not have been asrich as in their M�C plans, however, or the premiums may have been higher. If we assume that people disenrolled from the M�C market between 1999 and 2002obtained supplemental coverage in the same proportions as the survey respondents reported, then the fraction of beneficiaries with no additional coverage has grownfrom 9 percent in 1999 to an estimated 11 percent in 2002. These are MedPAC estimates based on the distribution in 1998, the change in Medicare managed careenrollment between 1998 and 2002, and the survey results regarding the sources of supplemental coverage obtained by those who lost their M�C plan. Note that thisestimate of uncovered beneficiaries may be conservative. One survey of beneficiaries conducted in 2000 found that 17 percent had no supplemental coverage at thetime of the survey (Gold and Mittler 2001).

Medicare beneficiaries’satisfaction with their planchoices

Most Medicare beneficiaries report thatthey are satisfied with their healthcoverage. A survey of beneficiariesconducted in 2000 found that 61 percentof all beneficiaries in fee-for-service (withor without supplemental coverage) and 69percent of beneficiaries enrolled in M�Cplans rated the value of their currentcoverage as excellent or very good, and86 percent of FFS and 90 percent of M�Cbeneficiaries would recommend theirplans to a friend (Gold et al. 2001). Datafrom a 2002 insurance industry-sponsoredsurvey indicated that 89 percent ofrespondents were satisfied or verysatisfied with their Medigap coverage, and76 percent said that, considering thepremiums they were paying, their policieswere a good or excellent value. Accordingto this same survey, over 80 percent saidthey would recommend Medigapcoverage to a friend or relative turning 65and enrolling in Medicare (Young 2002).

CMS has devoted significant resources tothe development of programs to monitorbeneficiaries’ experiences and satisfactionwith Medicare options. The ongoingConsumer Assessment of Health PlansSurvey (CAHPS) was first fielded in 1998to obtain information from beneficiaries inM�C plans. CMS is now also fielding aversion of CAHPS designed to obtaincomparable information frombeneficiaries in the traditional FFSMedicare program.22

Data from the M�C CAHPS haveconsistently shown that beneficiariesgenerally report high levels of satisfactionwith their health plans and with the healthcare they receive. In 1999, across 69MSAs for which data were analyzed, 79.7percent of M�C enrollees gave their

plans an overall rating of 8 or more out ofa possible 10 (Lake and Rosenbach 2001).These scores, however, differedsignificantly across the geographic regionsin the first three years of the survey23

(Goldstein et al. 2001; Zaslavsky et al.2000).

Comparing data from the FFS and M�CCAHPS raises conceptual andmethodological issues. The FFS sampleincludes beneficiaries with various typesof supplemental coverage, and thosewithout any supplemental insurance. TheM�C sample reflects the nature of thecurrent M�C market—the beneficiarieswho are included in the sample are thosewho have access to, and have chosen toenroll in, M�C plans. This means thatthere are some significant differences inthe populations included in either the FFSor M�C samples across geographic areas.There are few (or no) M�C options insome areas, the FFS CAHPS sampleincludes people who might have, if giventhe opportunity, chosen to be in an M�Cplan.

Despite these caveats, however, theCAHPS surveys do provide an importantinsight: A large proportion of allbeneficiaries are quite satisfied withMedicare and with their own healthinsurance coverage. Unpublished datafrom both the M�C and Medicare FFSCAHPS and the disenrollment surveyindicate in general there is a relativelyhigh level of satisfaction with Medicareregardless of the plan model in whichbeneficiaries are enrolled. A largeproportion of all beneficiaries rate theirhealth care and Medicare a “10,” on ascale of “10” on composite measuresconstructed by CMS. Beneficiaries inpoorer health, however, give Medicarelower ratings overall (Bernard et al. 2003).The data also suggest that M�C

beneficiaries with health problems may beless satisfied than beneficiaries enrolled inthe traditional FFS Medicare program:The disparities between the satisfactionratings of people in fair or poor health andthe ratings of people in excellent or verygood health were greater for thoseenrolled in M�C than for those enrolledin the traditional FFS Medicare program(Table 5-4). There are also differencesacross individual measures included in thecomposite ratings.24

Changes that Medicarebeneficiaries would like to see ininsurance offerings

Beneficiary and advocacy organizations’concerns about available insuranceoptions can be divided into fourcategories: the adequacy and cost ofbenefits and coverage, the stability ofplans and plan offerings, the complexityof the options available, and the equity inchoices across markets.

• Benefits and costs. The singlegreatest concern among beneficiaryadvocates is coverage of prescriptiondrugs. Major beneficiaryorganizations have called for theaddition of prescription drugcoverage to the basic Medicarepackage (AARP 2002). Someadvocates also believe that theaddition of a drug benefit underMedicare would help to stabilize theM�C market, because Medicarepayments to plans for coveredbenefits would relieve the plans fromat least some portion of the rapidlyincreasing costs of prescription drugs.More generally, advocates areconcerned about increases in out-of-pocket costs incurred bybeneficiaries, both for uncoveredservices and for premiums—


22 These surveys are very large, and can be used to compare enrollees’ reports about their health plans and health care experiences at the plan level, as well as withinand across states and metropolitan statistical areas. A survey of beneficiaries who disenroll from M�C plans is also conducted each year. In addition, information onbeneficiaries’ views about their health plans and insurance coverage is collected in the Medicare Current Beneficiary Survey.

23 The overall plan ratings have generally been higher in the Northeast and lower in the Pacific and Northwest regions (Goldstein et al. 2001).

24 Analyses supplied to MedPAC by CMS indicate that among the 42 states with managed care and DC, M�C enrollees gave higher percentages of positive responsesthan FFS beneficiaries for 2 of the 6 indicators: “Good Communication” and “Flu Shot.” For two other indicators, “Care Quickly” and “Rate Health Care,” neithergroup had a notably higher percentage of positive responses. Generally, FFS received higher percentages of positive responses than M�C for the “Needed Care”composite and “Rate Medicare” indicator (Bennett 2003).

particularly for the higher premiumscharged by M�C plans offeringsupplemental benefits.

• Stability. Beneficiaries have growingconcerns about the stability of M�Coptions. Plan withdrawals over thepast four years have causedfrustration and anger among affectedbeneficiaries, and some beneficiariesare reportedly seeking alternativeprescription drug coverage orreverting to Medigap coverage ratherthan enrolling in an M�C plan(Stuber et al. 2002; Young andMittler 2002). Some advocates haveproposed regulatory changes topromote greater stability, includingrequiring that plan/provider contractslast throughout the calendar year andbe finalized prior to the openenrollment period, and requiringplans that wish to participate in M�Cto commit to the program for a fixedperiod (e.g., three years) (Stuber et al.2002).25 Some advocates also believe

that the instability of the M�Cprogram militates against provisionsthat would restrict beneficiaries fromswitching among plans over thecourse of a year.26

• Complexity. Changes in M�Cavailability, benefits, and premiumcosts, and the introduction of newplan options such as private FFSplans, have made the choice ofinsurance options more complicated.Researchers as well as advocacygroups report that beneficiaries canfind it extremely difficult to sort outtheir options (Young and Mittler2002; Barents 1999; Stuber et al.2002). Specific conditions and limitsof prescription drug coverage offeredby M�C plans can be especiallycomplicated and difficult tosummarize in ways that are useful tobeneficiaries.27 Some advocates havecalled for expanded education andoutreach programs to help

beneficiaries understand their choices(AARP 2002). Greaterstandardization of M�C products tomake it easier for beneficiaries tocompare plan benefits and costs hasalso been proposed. One majorbeneficiary organization supports theuse of standard definitions for allservices covered by plans (AARP2002).

• Equity. The geographic variations inhealth care and insurance costs thatunderlie the Medicare FFS systemaffect the insurance choices availableto Medicare beneficiaries (seeMedPAC’s March 2002 Report toCongress). There are significantgeographic differences in theMedicare�Choice options availableto beneficiaries, as well as largevariations in the richness ofsupplemental offerings and thepremiums charged for these options.These variations are intertwined withcost differences for Medigap policies.Advocates view the variations asinherently unfair and as a threat to theunderlying principles of equityembodied in Medicare. Someadvocates believe that some of thevariations, or at least some of theirnegative effects in terms of equity,could be reduced throughstandardizing benefits and throughrisk-adjusting payments to reduceadverse selection (Dallek et al. 2002).

Health plansFor health plans and insurers, theMedicare market presents bothopportunities and frustrations. Insurersseek a dynamic environment in which abroad range of private options can meetthe needs of a diverse population andwhere there are opportunities for profit.Plans as well as Medigap insurers believe


Consumer Assessment of Health Plans ratings ofMedicare FFS and Medicare�Choice plans

Percent of beneficiaries surveyed givinganswer, by perceived health status

Excellent or very good Fair or poor

Rate their health care a “10”

Medicare FFS 54.4* 43.7M�C 58.9* 44.2

Rate Medicare a “10”

Medicare FFS 49.8 44.6*M�C 49.9 36.5*

Note: FFS (fee-for service), M�C (Medicare�Choice)* Statistically significant differences (p � .05) between Medicare FFS and M�C.

Source: Consumer Assessment of Health Plans Surveys from CMS, Bernard et al. 2002.

T A B L E5-4

25 The effect of requiring plans to make multiple-year commitments is a topic of debate. Some analysts believe that, rather than providing stability for beneficiaries, theseprovisions might deter plans from entering into contracts with Medicare.

26 The lock-in provisions that were partially implemented in 2002 were delayed until 2005 in legislative provisions included in the Public Health Security and BioterrorismResponse Act of 2002 (Public Law 107-188).

27 A 2002 report issued by the HHS Office of Inspector General found that “The information that HMOs provide to beneficiaries about certain elements of the drug benefitis inconsistent, incomplete, and misleading” (Department of Health and Human Services 2002).

that they can better serve beneficiaries ifthere is a “level playing field” whereMedigap products can compete with otherproducts, including M�C plans, that arecurrently subject to different regulationsgoverning underwriting, guaranteed issueand renewal, community rating, andflexibility in benefit design.

Plans’ perspective

From the industry’s perspective, threebasic problems impede the developmentof more successful Medicare markets:Medicare payment levels, administrativeand regulatory requirements, and limitsimposed by Medicare on health plans’ability to design and market new,“flexible” products.

Medicare payment levels Industryrepresentatives maintain that inadequatefunding is the biggest problem facing theM�C program. From their perspective,Medicare payments have to be sufficientto maintain their physician networks. Inparticular, plans believe that the statutoryupdate amount, which has effectivelylimited plans to 2 percent increases, hasfailed to keep up with the increasing costof providing Medicare and non-Medicareservices. Payment updates to M�C plansin the past two years have been far lowerthan the increases in premiums for healthplans in large employer-based markets(American Association of Health Plans[AAHP] 2002). One major industry grouphas called for Congress to change thepayment system to one that pays Medicareplans the higher of 100 percent of localFFS costs or the current M�C rates28

(Blue Cross and Blue Shield Association2002).

Administrative and regulatoryrequirements imposed by MedicarePlan representatives believe that some ofthe data reporting and compliancerequirements imposed by Medicare areexcessively complicated and expensive,and divert funds from patient care. They

also report that some of the instructionsfor complying with these requirements areunclear or contradictory. Plans, despitetheir appreciation of CMS’s recentsimplification efforts (see below), stillhave some concerns about the operationof the M�C risk-adjustment system (seeAppendix A), which they believe isresource intensive and can, because of aneed to correct errors, lead to delays inpayments to plans (AAHP 2002).

Limits imposed by Medicare onplans’ ability to offer “flexible”products Plans and insurers want to beable to market more varied insuranceproducts, including products that lookmore like those available to the workinginsured population. The managed careindustry has recommended expanding therange of choices for beneficiaries bymaking cost contracts a permanent part ofMedicare and allowing M�C plans tovary benefits and premiums withinsegments of service areas (AAHP 2002).Other industry representatives have urgedthe Congress to develop options toincrease participation of PPOs in M�C asa major policy objective (Health InsuranceAssociation of America 2002). SomeMedigap insurers would like to see thestandard packages modernized, or havemore flexibility in offering nonstandardpackages. There is widespread agreementin the insurance industry that any majorrestructuring of the standardized benefitforms should, however, wait until theprescription drug issue and broad reformof Medicare benefits is settled.

CMS policy changes toencourage plan participation

CMS has already taken action to addressperceived problems in M�C markets, inconjunction with a major project beingdirected by the Secretary of Health andHuman Services’ Advisory Committee onRegulatory Reform. Organizationalchanges at CMS, including the creation ofa new Center for Beneficiary Choices,

consolidate oversight responsibilities,which should improve communicationwith plans. CMS has also reduced thenumber of mandatory quality assessmentactivities that participating plans mustconduct, and revised the processes fordeeming plans to be in compliance with avariety of regulatory requirements.

The agency has also made significantchanges designed to reduce theadministrative burden associated with riskadjustment.29 Data collection for riskadjustment across multiple sites of carebegan in October 2000, but M�C plansargued that CMS’s requirements forcollecting and submitting the data weretoo burdensome. In response, theSecretary suspended collection of datafrom ambulatory sites in May 2001 anddirected CMS to investigate alternatives.CMS worked with M�C plans, tradeorganizations, and physicians to develop amultiple-site model to address plans’concerns. CMS announced a preliminaryversion of the model on March 29, 2002.Plans began to collect diagnosis data fromphysician office and hospital outpatientsources in July 2002 and began submittingthe data in October 2002. CMS willannounce the final version of the modelby March 28, 2003, and will begin usingthe model on January 1, 2004.

As described earlier, CMS has alsoinitiated a new demonstration program,focused on PPOs, to foster competition inthe M�C program. Medicare’s PPOdemonstration could be attractive toinsurers for several reasons:

• In some areas, the demonstration willpay more than the M�C paymentrates. The demonstration will pay themaximum of the M�C rates or 99percent of the per capita MedicareFFS spending in a county. Almostone-fourth of the beneficiaries whowill have a PPO demo plan availablelive in counties where higher rateswould be paid.


28 MedPAC has recommended that M�C rates be set equal to 100 percent of local FFS costs (MedPAC March 2002).

29 For 2004, CMS must begin using a risk adjustment system based on a model that uses data from hospital inpatient and ambulatory settings. Also, CMS is required toapply such a model to 30 percent of payments in 2004, and the agency must increase this percentage annually until it reaches 100 percent in 2007.

• Though M�C CCPs may not setpremiums and cost-sharing for thebasic benefit package above a capactuarially set at the national averagefor all Medicare FFS beneficiaries,PPOs in the demonstration will notbe limited by this cap. Benefitconsultants have stated that lifting thecap will allow plans to compete withMedigap for those beneficiaries whoare willing to buy a higher-pricedproduct.

• The demonstration allows fornegotiated risk-sharing between theplan and Medicare. Details of therisk-sharing arrangements have notbeen released, but apparently not allof the demonstration plans areavailing themselves of the option.

When supply anddemand meet in themarketplace

Medicare beneficiaries’ demand forbenefits beyond those found in thetraditional FFS Medicare program hasbeen filled by a broad spectrum ofoptions, with varying degrees of success.Some options, such as M�C plans,primarily replace Medicare FFS whileenhancing some benefits. Other options,such as Medigap, employer-sponsoredsupplemental, Medicaid, and VAprograms, are designed only tosupplement Medicare. Access to thesevarious options depends on beneficiaries’circumstances and geographic locations.

The supply of health insurance options forMedicare beneficiaries is influenced bythe overall health care marketplace. Thestage of the underwriting cycle andinsurance company circumstancesinfluence the supply of plans and thepremiums they charge. Also, theeconomic and regulatory environmentinfluences employers’ willingness toprovide retiree benefits. Finally, the natureof local markets and the balance of powerbetween plans and providers drive plandecisions to enter and remain in local

markets. In this section, we look at theinterplay of supply and demand in themarketplace.

Beneficiary demandThe demand for more comprehensivebenefits is clear: In 2000 only 9 percent ofbeneficiaries in the community had justtraditional FFS Medicare. But the marketmay be changing in the future. Whileemployer-sponsored coverage was held by32 percent of beneficiaries in 2000, manycompanies are cutting back onpostretirement health coverage andeliminating it for new employees. Costpressures will likely fuel the demand forless-expensive options for employers oroptions that retirees can afford on theirown.

Despite the popularity of Medigapcoverage—27 percent of beneficiaries hadMedigap in 2000—it may be becomingless affordable for many beneficiaries,particularly when prescription drugs arepart of the plan. Even those Medigapplans that include a drug benefit do notprovide comprehensive drug coverage.

Medicaid provided additional coveragefor 12 percent of beneficiaries living inthe community in 2000. That coveragemay change to some extent if statebudgets come under increasing pressure.States have taken a variety of steps tolimit Medicaid spending, including cuttingback on prescription drug benefits,increasing cost sharing, and tighteningeligibility criteria (Smith et al. 2003). Asurvey conducted by the NationalConference of State Legislatures in late2002 found that 16 states reported theywould consider eligibility reductions forthe elderly as a means of reducing theirMedicaid costs in 2003 (Bureau ofNational Affairs 2002).

All of this potential increase in demandfor more comprehensive benefits mayrepresent an opportunity for M�C plans,other alternatives to FFS, and Medigapinsurers. But those opportunities may belimited by marketplace realities.

Health plan willingness tosupply coverageHealth plans will only enter the Medicaremarket under certain conditions.Medicare�Choice plans and otheralternatives to Medicare FFS, forexample, need payments that exceed theircosts. If payments are set to equal thosefor FFS Medicare, then the other plansmust lower their costs of care below thoseof Medicare by an amount sufficient tooffset their administrative and marketingcosts, plus their profits. They can do so bybeing more efficient (through utilizationcontrols or disease management programs,for example), receiving discounts fromproviders, enrolling healthierbeneficiaries, or using some combinationof these actions. (If risk-adjustedpayments are fully implemented andaccurately capture the cost of caring forenrollees, then enrollee health statuswould not matter.) Alternatively, planscan enter areas where payments are setabove FFS Medicare costs.

On the other hand, in the currentenvironment M�C plans do not competeagainst only Medicare FFS. Instead, theycompete against a combination ofMedicare FFS and Medigap. To besuccessful, they have to deliver the samecombined set of benefits for less. Thischallenge raises the possibility of nothaving to undercut Medicare FFS costsbut being about equal to Medicare andless than Medigap for the additionalbenefits. In the past, M�C plans tried tokeep premiums low or at zero becausethey did not think that beneficiaries werewilling to pay a premium (or thought thatthose who were willing to do so were badrisks). Sometimes M�C plans leftmarkets rather than adding premiums.Now that premiums for M�C have beenincreasing, it appears that somebeneficiaries are willing to pay for theirproduct and may be comparing M�Cplans and the combination of FFS andMedigap more carefully than they mayhave in the past. For example, althoughplans in many markets have increasedpremiums, and decreased the value of


additional benefits offered in the last twoyears, enrollment in those plans has notdecreased precipitously.

Plans that supplement the basic Medicarebenefit package take their lead from FFSMedicare. Medigap and employer “wrap-around” plans usually depend on theMedicare programs’ coverage decisions inorder to determine coverage for costsharing. The levels of cost sharing underFFS Medicare determine plan cost-sharingliability and thus the cost of the plans tobeneficiaries. Also, the plans—includingM�C plans that supplement the basicMedicare package—are greatly affectedby regulations that determine how thesupplements must interact with Medicare.This is especially true of regulations onplan marketing and rules on how the plansmay integrate employer-sponsoredcontributions.

For network-based alternative plans, amajor constraint on supply is thefeasibility of putting together a network.In some areas of the country, particularlyrural areas, it is very difficult to recruitproviders because they are in a monopolyposition and have no interest in dealingwith managed care organizations. In theM�C program, this has resulted in veryfew MCOs entering rural areas. InCalifornia, for example, the overallpenetration rate of HMOs is very high andthe M�C penetration rate is 35 percent,but the participation rate in countiesoutside MSAs is only 1.1 percent (Goldand Lake 2002). The only M�C choicesin many rural areas are non-networkprivate FFS plans which so far have verylimited membership. Network formationwill also be crucial to success inexpanding the Medicare Select program.

National marketplacedynamicsThe Medicare alternative and supplementmarkets are only a small part of the largerhealth insurance marketplace. As such,they are not immune to larger-scale trendsin the overall market. In recent years theM�C program has reflected some ofthose trends, including the underwritingcycle, the move to larger and looser

networks with less utilization control, andprovider pushback and the decline in full-capitation and other models of risk sharingwith providers.

The underwriting cycle is a term oftenused by health policy analysts to describethe tendency of commercial insurancepremiums to rise at a rate lower than costincreases when the market is profitable asinsurers compete to increase market share,and then to rise at a higher rate as insurerstry to repair profit margins and ridthemselves of money-losing lines ofbusiness. This tendency has been reflectedin the M�C market as plans used M�Cto grow market share in the mid-1990s, inthe anticipation of higher M�C profits,and then pulled back beginning in 1999(Grossman et al. 2002).

In reaction to the anti-HMO backlash ofthe mid-1990s and changes in state laws,plans started to move to less-restrictivenetworks and less emphasis on utilizationcontrols in commercial plans. Thisbroadening of networks and lessening ofcontrols moved into M�C plans as well,which further restricted plans’ ability tomanage underlying care and costs. In acompetitive market, if costs and premiumsrise in reaction to fewer restrictions, morerestrictive plans may begin to once againlook attractive, which may carry over intothe M�C market as well.

Another larger-scale trend has beenprovider pushback against contract termsproposed by network plans. Providers inmany markets have consolidated,increasing their market power and makingit difficult to form desirable networkswithout them. They have also movedaway from accepting risk from plans andmoved to a more FFS-like relationship. InCalifornia, some plans relied heavily onthe capitated risk model, and pushbackfrom providers has caused turbulence andwithdrawal from some markets (Gold andLake 2002).

Taken together, these larger trends reveala dynamic M�C marketplace in whichplans enter and exit just as they do inother managed care markets. This entryand exit by health plans can cause

instability for Medicare beneficiaries andconcern among policymakers, but it is partof the reality of competition. The Medigapmarket has not been particularly volatileover the last few years, however. Instead,Medigap enrollment and premiums havegrown modestly.

Importance of local marketsWhatever the national trends, localmarkets are where beneficiaries maketheir choices, where health care isdelivered, and where insurance plans haveto compete.

Beneficiaries’ real choices are limited towhat they perceive as acceptable andaffordable. Conceptions of acceptableinsurance products will vary along withbeneficiary expectations in different areasof the country. For example, localemployers’ provision of health insurancewill have an effect on their retirees’choices when they become eligible forMedicare. If a beneficiary was in an HMOwhen employed, belonging to an HMO asa Medicare beneficiary may be anobvious, and perhaps a preferable, choice.For a beneficiary with no experience withmanaged care and an attachment to aparticular physician, an HMO may not bean obvious choice.

Beneficiaries’ ability to afford differentchoices may also depend on theiremployment history, as well as on theirincome in retirement. If, for example,beneficiaries have an option thatsubsidizes their expenses, such asemployer-sponsored wrap-aroundsupplemental insurance or Medicaid, theirdemand for HMO options or Medigapmay be lower than without such support.Affordability is a key determinant. In low-income areas, the demand for pricierproducts may be low, unless thepremiums are subsidized by formeremployers.

Health care providers operate in localmarkets as well. They frequently drawcustomers from specific geographic areasand sociodemographic groups. At thesame time, they may have existingrelationships with other providers that


influence expected practice patterns.These relationships can also influencewhich insurance arrangements areconsidered acceptable in local markets. Insome areas capitated contracts withinsurance plans are much more routinethan in others. Therefore, products thatdepend on capitation may only succeed incertain areas.

Health insurers are also sensitive to localmarket conditions because of theregulatory environment. Most insurance isregulated at the state level. Plans judgesome states to be more conducive tocertain forms of insurance than others.State rating rules may also greatly affectthe competition between plan types.Because Medicaid differs by state, plansthat interact with Medicaid also differ bystate. At a more local level, plans react tobeneficiary preferences and provider

characteristics that differ by local area.For example, if there is a monopoly localprovider of a service, such as a large localhospital, plans will be constrained in theircontracts with that provider in ways theywould not be if competing providers wereavailable or not be able to contract at all.Plans also react to the presence of otherplans. Some researchers have found thatlarger numbers of M�C plans competingis correlated with greater value forbeneficiaries at the same cost to theprogram (Pizer and Frakt 2002).

To understand the choices available toMedicare beneficiaries, the individualfeatures of local markets and how theyrelate to competition and market dynamicsmust be examined. MedPAC plans todraw on existing research on local healthcare markets and conduct some case

studies of actual markets to comprehendthe Medigap and supplementalmarketplace. We hope to use the casestudies to clarify what happens in markets,and then draw some conclusions aboutwhere particular kinds of choices might bemade available to Medicare beneficiaries.We might also learn that M�C paymentrates, Medigap rating rules, or stateassistance programs may have unintendedconsequences for insurance competition insome areas. Designing a national programflexible enough to support different kindsof choices in different kinds of localmarkets will be difficult and raise issuesof equity as well. Nevertheless, it will benecessary to address these difficulties ifthe goal is to foster increased choice forMedicare beneficiaries. �


References

American Association of Health Plans. The Medicare�Choice Program: what you needto know, AAHP Report. Washington (DC), AAHP. May 2002, Vol. 3, No. 1.

AARP. The policy book: AARP public policies 2002. Washington (DC), AARP. 2002.

Barents Group, LLC. How Medicare HMO withdrawals affect beneficiary benefits, costs,and continuity of care. Menlo Park (CA), The Henry J. Kaiser Family Foundation.November 1999.

Bennet, S, Centers for Medicare & Medicaid Services, Center for Beneficiary Choices.Written communication to MedPAC. Baltimore (MD). January 2, 2003.

Bernard S, Uhrig J, Tornatore D, et al. Comparing Medicare experiences with fee-for-service versus managed care plans: does health status make a difference? Unpublishedmaterial. In press 2003.

Blue Cross and Blue Shield Association. Congress must act to increase funding and putthe “choice” back in Medicare+Choice. Available athttp://bcbshealthissues.com/proactive/newsroom/release.vtml?id=27923.

Buchmueller TC. The health plan choices of retirees under managed competition, HealthServices Research. December 2000, Vol. 35, No. 5, p. 949–976.

Bureau of National Affairs. BNA’s Healthcare Daily Report. December 19, 2002.

Centers for Medicare & Medicaid Services. The characteristics and perceptions of theMedicare population: data from the 1998 Medicare current beneficiary survey. Baltimore(MD), CMS. 2002. Available at http://www.hcfa.gov/surveys/mcbs/ PubCNP98.htm.

Centers for Medicare & Medicaid Services. Solicitation for proposals for Medicarepreferred provider organization (PPO) demonstrations in the Medicare�Choice program,No. CMS–40420N. Baltimore (MD), CMS. April 12, 2002.

Chollet D, Kirk A, Mathematica Policy Research, Inc. Medigap insurance markets:structure, change, and implications for Medicare. Washington (DC), Department ofHealth and Human Services, Office of the Assistant Secretary for Planning andEvaluation. December 13, 2001.

Congressional Budget Office. Cost estimate: H.R. 4939, Veterans Medicare Payment Actof 2002. Washington (DC), CBO. November 1, 2001.

Dallek G, Dennington A, Biles B, Center for Health Services Research and Policy, TheGeorge Washington University Medical Center. Field report: geographic inequity inMedicare�Choice benefits. Findings from seven communities. New York (NY), TheCommonwealth Fund. September 2002.

Dallek G, Edwards C, Center for Health Services Research and Policy, The GeorgeWashington University Medical Center. Restoring choice to Medicare�Choice: theimportance of standardizing health plan benefit packages. Washington (DC), TheCommonwealth Fund. October 2001.


Department of Health and Human Services. HHS expands health plan options inMedicare�Choice, HHS News. August 27, 2002. Available athttp://www.hhs.gov/news/press/2002pres/ 20020827.html.

Department of Veterans Affairs. Federal benefits for veterans and dependents (2002edition). Washington (DC), VA. 2002.

Ellwood M, Quinn B. Background information on dual eligibles in MSIS FY 1999.Cambridge (MA), Mathematica Policy Research, Inc. February 28, 2002.

General Accounting Office. Low-income Medicare beneficiaries: further outreach andadministrative simplification could increase enrollment, No. HEHS-99-61. Washington(DC), GAO. April 1999.

General Accounting Office. Medigap insurance: plans are widely available but havelimited benefits and may have high costs, No. GAO-01-941. Washington (DC), GAO.July 2001.

General Accounting Office. Retiree health benefits: employer-sponsored benefits may bevulnerable to further erosion, No. GAO-01-374. Washington (DC), GAO. May 2001.

Gold M. Trends reflect fewer choices, monitoring Medicare�Choice: fast facts.Washington (DC), Mathematica Policy Research, Inc. September 20, 2000, No. 4.

Gold M, Justh N. How salient is choice to Medicare beneficiaries? MonitoringMedicare�Choice: fast facts. Washington (DC), Mathematica Policy Research, Inc.January 2001, No. 5.

Gold M, Lake T. Medicare�Choice in California: lessons and insights, No. PR02-61.Menlo Park (CA), The Henry J. Kaiser Family Foundation. September 2002.

Gold M, Mittler J. The structure of supplemental insurance for Medicare beneficiaries:Medicare�Choice—operational insights. Washington (DC), Mathematica PolicyResearch, Inc. June 2001, No. 3.

Gold M, Sinclair S, Cahill M, et al. Medicare beneficiaries and health plan choice, 2000.Washington (DC), Mathematica Policy Research, Inc. January 2001.

Goldstein E, Cleary PD, Langwell KM, et al. Medicare Managed Care CAHPS®: a toolfor performance improvement, Health Care Financing Review. Spring 2001, Vol. 22, No.3, p. 101–107.

Grossman JM, Strunk BC, Hurley RE. Reversal of fortune: Medicare�Choice collideswith market forces, Issue Brief No.52. Washington (DC), Center for Studying HealthSystem Change. May 2002.

Health Care Financing Administration. Memo on intermediaries and carriers, transmittalAB-01-52. Baltimore (MD), HCFA. April 10, 2001.

Health Insurance Association of America, HIAA outlines its 2002 policy priorities.Washington (DC), press release. February 8, 2002. Available athttp://www.hiaa.org/search/content.cfm?ContentID=17648.


Henry J. Kaiser Family Foundation, Health Research and Educational Trust, TheCommonwealth Fund. Erosion of private health insurance coverage for retirees. MenloPark (CA), Henry J. Kaiser Family Foundation. April 2002.

Henry J. Kaiser Family Foundation/Hewitt Associates. The current state of retiree healthbenefits: findings from the Kaiser/Hewitt 2002 retiree survey. Washington (DC), Henry J.Kaiser Family Foundation/Hewitt Associates. December 5, 2002.

Hewitt Associates, LLC. Salaried employee benefits provided by major U.S. employers,2001–2002. Lincolnshire (IL), Hewitt Associates. October 2001.

Institute of Medicine. Crossing the quality chasm: a new health system for the twenty-first century. Washington (DC), National Academy Press. 2001.

Lake T, Rosenbach M. Tracking performance of the Medicare�Choice program: resultsfrom the 1998 and 1999 Medicare CAHPS survey. Princeton (NJ), Mathematica PolicyResearch, Inc. January 18, 2001.

Laschober MA, Topoleski C. A profile of QMB-eligible and SLMB-eligible Medicarebeneficiaries. Baltimore (MD), Barents Group LLC. April 7, 1999.

Lee AJ, Garfinkle SA, Khandker RZ, Norton ED. Impact of Medicare SELECT on costand utilization in 11 states, Health Care Financing Review. Fall 1997, Vol. 19, No.3, p.19–40.

Medicare Payment Advisory Commission. Report to the Congress: assessing Medicarebenefits. Washington (DC), MedPAC. June 2002.


Mercer WM. Mercer/Foster Higgins national survey of employer-sponsored health plans2000. New York (NY), William M. Mercer, Inc. 2002.

National Academy of Social Insurance. Structuring Medicare choice: final report of theStudy Panel on Capitation and Choice. Washington (DC), NASI. April 1998.

Nemore PB, National Senior Citizens Law Center. State Medicaid buy-in programs:variations in policy and practice. Menlo Park (CA), Henry J. Kaiser Family Foundation.December 1999.

Office of Inspector General, Department of Health and Human Services. Medicare HMOprescription drug benefits: information available to beneficiaries on dollar limits, No.OEI-03-00-00430. Washington (DC), HHS. May 2002.

Pezzin LE, Kasper JD. Medicaid enrollment among elderly Medicare beneficiaries:individual determinants, effects of state policy, and impact on service use, HealthServices Research. August 2002, Vol. 37, No. 4, p. 827–847.

Pizer SD, Frakt AB. Payment policy and competition in the Medicare�Choice program,Health Care Financing Review. Fall 2002, Vol. 24, No. 1, p. 83–94.

Politi E. Medicare and TRICARE, Center for Medicare Education Issue Brief.Washington (DC), Center for Medicare Education. 2002, Vol. 3, No. 2. Available athttp://www.medicareed.org/pdfs/papers73.pdf.


Quayle, James R, Office of Personnel Management, Office of the Actuary. Electroniccommunication to MedPAC. Washington (DC), January 10, 2003.

Simmons M. Veterans of the drug price war, The Washington Post. March 5, 2002, p.HE03.

Sing M, Brown R, Hill SC. The consequences of paying Medicare managed plans theircosts, Inquiry. Summer 1998, Vol. 35, No. 2, p. 210–222.

Smith V, Gifford K, Ramesh R, Health Management Associates, and Wachino V, KaiserCommission on Medicaid and the Uninsured. Medicaid spending growth: a 50-stateupdate for fiscal year 2003. Washington (DC), The Henry J. Kaiser Family Foundation.January 2003.

Strombom BA, Buchmeuller TC, Feldstein PJ. Switching costs, price sensitivity andhealth plan choice, Journal of Health Economics. January 2002, Vol. 21, No. 1, p.89–116.

Stuber J, Dallek G, Edwards C, et al., Center for Health Services Research and Policy,The George Washington University Medical Center. Instability and inequity inMedicare+Choice: the impact on Medicare beneficiaries—findings from seven casestudies. Washington (DC), The Commonwealth Fund. January 2002.

Super N. Medigap: prevalence, premiums, and opportunities, NHPF, Issue Brief No. 782.Washington (DC), National Health Policy Forum. September 9, 2002.

Watson Wyatt Worldwide. Retiree health benefits: time to resuscitate? Executivesummary. Watson Wyatt Worldwide. 2002. Available athttp://www.watsonwyatt.com/research/printable.asp?id�W-599.

Young C, Mittler J. Medicare�Choice: views from the field, Operational Insights.Washington (DC), Mathematica Policy Research, Inc. March 2002, No. 7.

Young DA, president, Health Insurance Association of America. Written statementsubmitted in testimony before the Subcommittee on Health, U.S. House ofRepresentatives Committee on Ways and Means. March 14, 2002.

Zarabozo C, Centers for Medicare & Medicaid Services. Telephone conversation with S.Harrison, MedPAC. Washington (DC). January 8, 2003.

Zarabozo C, Centers for Medicare & Medicaid Services. Telephone conversation with S.Harrison, MedPAC. Washington (DC). November 20, 2002.

Zaslavsky AM, Landon BE, Beaulieu ND, Cleary PD. How consumer assessments ofmanaged care vary within regions and among markets, Inquiry. Summer 2000, Vol. 37, p.146–161.


How Medicare pays for services:an overview

A P P E N D I X A

range of health care, including facilityservices—provided in hospital inpatientand outpatient departments, ambulatorycare centers, and skilled nursing facilities,for example—and professional servicesfurnished by physicians, therapists, andother practitioners.

In the traditional fee-for-service (FFS)program, Medicare sets prospectivelydetermined payment amounts (rates)providers will receive for most coveredproducts and services, and providers agreeto accept them as payment in full.1 Thus,in most instances, providers’ payments arebased on predetermined rates and areunaffected by their costs or postedcharges. When beneficiaries use services,providers submit bills to Medicare’s fiscalagents, who pay the predetermined ratesminus beneficiaries’ cost-sharingliabilities, such as deductibles andcoinsurance. Providers then collect theremaining amounts from beneficiaries.2

In the Medicare�Choice (M�C)program, Medicare sets the county-specific monthly capitation payment ratesthat M�C organizations will receive for

A P P E N D I X

How Medicare pays forservices: an overview

AMedicare’s 40 million beneficiaries usethousands of different health care productsand services furnished by over 1 millionproviders in hundreds of marketsnationwide. Medicare pays for theseservices using 15 payment systems thatare generally organized by deliverysetting. These payment systems sharecommon goals, and most have similardesign elements that are tailored toaccommodate the products Medicare isbuying in each setting, the characteristicsof the providers that produce them, theextent to which the same product may befurnished in different settings, and themarket circumstances that affectproviders’ costs. In this appendix, wedescribe the key features of these paymentsystems.

Medicare was enacted to improve accessto care by reducing the financial burdensfaced by elderly people (and later,disabled people) in obtaining medicallynecessary acute care services. To achievethis objective, Medicare helps itsbeneficiaries pay for covered products andservices in 15 different health caresettings. These settings encompass the full


enrolled beneficiaries. M�C plans mayoffer beneficiaries additional benefits notcovered in the traditional program andcharge additional premiums if the totalcost of all covered benefits exceedsMedicare’s capitation payment rates.M�C plans, however, acceptresponsibility for contracting with andpaying health care providers and suppliersfor the products and services they furnishto enrolled beneficiaries.

Recent legislation—the Balanced BudgetAct of 1997 (BBA), the Balanced BudgetRefinement Act of 1999 (BBRA), and theMedicare, Medicaid, and SCHIP BenefitsImprovement and Protection Act of 2000(BIPA)—fundamentally changed the wayMedicare pays for many products andservices. These laws required the Centersfor Medicare & Medicaid Services(CMS)3 to develop and adopt newprospective payment systems (PPSs) forservices furnished by skilled nursingfacilities, hospital outpatient departments,home health agencies, rehabilitationfacilities, long-term care hospitals, andpsychiatric facilities. The legislation alsorequired CMS to change the method for

1 Medicare pays for some services—those furnished by long-term care hospitals and psychiatric facilities, for example—based on a provider’s incurred allowable costs. Inthese instances, providers receive interim payments, usually reflecting their unit costs in the preceding year; discrepancies between interim payments and allowable costsare resolved (settled) annually after the end of the provider’s cost reporting period.

2 Most beneficiaries have secondary insurance; in this case, Medicare’s fiscal agents generally bill the secondary payer directly for the beneficiary’s liability.

3 CMS was formerly known as the Health Care Financing Administration.

making prospective capitation paymentsto health care organizations under theM�C program. In addition, CMS hasmodified its PPSs for hospital inpatientacute care, physician services, andambulance services, and proposedchanging its payment methods for durablemedical equipment.

In this appendix, we describe the 15 majorpayment systems Medicare uses to payproviders for products and services theyfurnish to Medicare beneficiaries. Webegin with an overview of key structuralelements that are present—explicitly orimplicitly—in virtually all prospectivepayment systems. This overview isfollowed by six sections that describe thepayment systems, grouped as follows:

• inpatient acute care in short-termhospitals and psychiatric facilities;

• ambulatory care furnished byphysicians, hospital outpatientdepartments, ambulatory surgicalcenters, and clinical laboratories;

• post-acute care furnished by skillednursing facilities, home healthagencies, inpatient rehabilitationfacilities, and long-term carehospitals;

• dialysis services furnished inoutpatient centers and hospice care;

• ambulance services and productsfurnished by durable medicalequipment suppliers; and

• services furnished by private healthplans under the M�C program.

Key structural elements ofMedicare’s prospectivepayment systems

Medicare’s payment policies and methodsare often seen as extremely complex, aperception strengthened by the myriadpolicy changes enacted in recentlegislation. Even without these changes,however, Medicare’s size and scope—

encompassing a full range of health careproducts and services from many differenttypes of providers in hundreds of marketsnationwide—would make its paymentmethods seem complicated. Furthercomplexity stems from the current mix ofpayment systems, in which traditionalpayment methods based on providers’costs and charges have not yet been fullyreplaced by prospectively determinedpayment rates.

Nevertheless, Medicare’s paymentsystems reflect common goals andproblems that are addressed using ahandful of similar structural elements.Focusing on the goals and structuralelements helps make the payment systemsand related policy issues moreunderstandable.

As discussed in previous MedPACreports, Medicare’s prospective paymentsystems are intended to support itsprincipal policy objective—promotingbeneficiaries’ access to high-quality carein the most appropriate clinical settingwithout imposing undue financial burdenson beneficiaries or taxpayers. To achievethis objective, Medicare’s paymentsystems must set payment rates that areconsistent with efficient providers’ short-run costs of producing services. That is,payment rates must accurately reflectpredictable cost variations amongproducts and services, including variationsthat result from patient characteristics andlocal market factors that are beyondproviders’ control.

To set and maintain accurate paymentrates for many products and services—even in a single setting—is a difficult task.At a minimum, policymakers need thefollowing conditions (Table A-1, p. 222):

• The products and services Medicareis buying must be well defined.

• The relative costliness of eachproduct or service compared with thatof the average service unit must bemeasurable.

• Production processes used byproviders must be understood well

enough to identify the major inputsthat contribute to efficient providers’unit costs.

• Patient or beneficiary characteristicsand market circumstances that mayaffect providers’ costs must beknown and measurable.

• A payment update method must bedeveloped to adjust payment ratesannually, consistent with changes ininput prices and other factors thatmay affect efficient providers’ costsover time.

Defining the products andservices Medicare is buying The products Medicare buys in eachsetting are defined by the unit of paymentand a compatible classification system.The unit of payment may be an individualservice (a physician office visit, forexample), a day of care (care in a skillednursing facility), an episode of care (ahospital stay), or a month of service (as inthe M�C program). Generally, the unit ofpayment should match the unit of serviceand the way providers think aboutdelivering care in the setting.

Consistent with the unit of payment, theclassification system identifies distinctservices, types of patient care products, orpatients who are expected to requiredifferent amounts of resources. In someMedicare payment systems—the hospitalinpatient PPS, for example—theclassification categories reflect differentclinical problems and treatment strategiesas indicated by diagnoses and procedures.In others, such as those for physician,hospital outpatient, or ambulatory surgicalservices, the categories reflect differentprocedures or evaluation and managementservices. In all payment systems, theclassification categories define theproducts for which Medicare will pay.

Setting relative values Relative values measure the expectedcostliness of a unit in each classificationcategory compared with the expectedaverage costliness of all units. Categories

220 How Med i ca r e pay s f o r s e r v i c e s : a n o v e r v i ew

that require more resources than averagehave higher relative values, and those thatrequire fewer resources have lower ones.Relative values are often referred to ascase-mix weights.

Setting a national basepayment rate The base payment rate represents theamount Medicare would pay for anaverage unit of service in a market withnational average input prices, if no otherpayment adjustments applied. The basepayment rate in each setting should reflectthe costs the payment rates are intended tocover—operating costs alone or operatingand capital costs together.4 Base paymentamounts per unit are sometimes calledconversion factors as in the physician feeschedule or the hospital outpatient PPS.

In some of Medicare’s paymentsystems—those for hospital inpatient,outpatient, skilled nursing, or home healthservices, for instance—the Congress hasrequired CMS to set national basepayment amounts to reflect nationalaverage historical costs for the affectedproviders. In general, average historicalunit costs in the base year have beenupdated to the first payment year bytaking into account industry-wide inflationand changes in case mix during theintervening years. In some instances,however (the physician fee schedule, forexample) measures of providers’historical costs are simply not available.In these cases, the initial base paymentamounts often have been set so that totalprojected payments in the first year underany new payment system would equaltotal projected payments under thepreceding system for the same year.

Adjusting for local marketconditions Input prices differ among markets acrossthe nation and these differences generallyaffect efficient providers’ costs in

predictable ways. Consequently,Medicare’s payment rates in each marketshould be adjusted to reflect the localprice level. To make these adjustments,policymakers must have one or moremeasures of geographic variation in inputprices—such as the area wage index in thehospital inpatient acute care PPS or thegeographic practice cost indexes in thephysician fee schedule. Policymakers alsomust know what proportions of providers’unit costs are affected by variations ininput prices. This information is used todetermine how much of the national basepayment rate should be adjusted by thegeographic input price factor for eachmarket area. Most Medicare paymentsystems use a version of the hospital wageindex.

Other adjustments Most payment systems have otheradjustments related to unusualcharacteristics of patients, servicesfurnished, providers, or market areas inwhich providers operate. In manyinstances, these adjustments are intendedto account for factors that mightsubstantially alter the resources needed toprovide services. In other cases, theyreflect policymakers’ decisions to supportcertain activities, such as providinggraduate medical education, serving adisproportionate share of low-incomepatients, or furnishing services to ruralbeneficiaries. Some payment systems,such as the acute inpatient hospital PPS,have more adjustments than others.

Updating payment rates Payment rates for most settings must beupdated annually to reflect changes intechnology, practice patterns, and marketconditions. Thus CMS must developmethods and data sources to be used inupdating the base payment amount, theproduct classification system, and therelative values. Other payment

adjustments also may need periodicrevision as conditions change. In mostpayment systems, the national basepayment rate is updated annually to reflectthe forecasted increase in an industry-specific national input-price index called amarket basket (MB) index. The MBindex, developed by CMS, tracks nationalaverage price levels for labor and otherinputs, weighted to reflect the relativeimportance of each input category in thespecific industry.5 This update affects allpayment rates equally, so it does not affectthe distribution of payments amongproduct categories or across providers.

Updating the relative values affects thedistribution of payments among productsand services, and among providersaccording to their case or service mixes.In some payment systems, such as thosefor acute inpatient hospital care andinpatient rehabilitation services, relativevalues are updated annually. In othersystems, such as the physician feeschedule and the skilled nursing facilityand home health PPSs, the relative valuesare updated less frequently.

The configuration of these elements varieswidely among Medicare’s paymentsystems, reflecting differences in thenature of the services Medicare is buying,the characteristics of the providers thatproduce them, and how market conditionsaffect providers’ costs. In addition,Medicare’s payment systems ofteninclude provisions designed to offset orweaken providers’ financial incentives toshift beneficiaries’ care among settings.These financial incentives reflect fixed-price payment for bundles of services—providers can lower their costs andincrease profits by shifting the provisionof some services to another setting wherethey would be paid for in a differentpayment system. These incentives alsomay arise because Medicare sets paymentrates separately for each setting and may


4 Operating costs consist of expenses for room, board, routine and special care, and ancillary services, such as laboratory tests, therapy, and imaging. Capital costs, suchas rent, interest, and depreciation, are included in the payment rates in some payment systems (such as the skilled nursing facility PPS) or excluded and paid separately.

5 For physician services, CMS uses the Medicare Economic Index (MEI), a weighted average of price changes for inputs used to provide care. These include physiciantime and effort, wage rates for nonphysician employees, and office expenses. The MEI is similar conceptually to the market basket index, except that it includes anadjustment for productivity growth.

pay different amounts for the sameservice, depending on the setting in whichit is furnished.

The remainder of this appendix describeshow the key elements are combined foreach of the 15 payment systems Medicareuses to pay providers for services theyfurnish to its beneficiaries.

Acute inpatient services

This section describes Medicare’spayment methods for acute inpatient carefurnished to beneficiaries in:

• short-term general hospitals.

• specialty psychiatric facilities.

Payment for acute careservices in short-termgeneral hospitals Each year, about one of every fiveMedicare beneficiaries enrolled in thetraditional program has one or moreinpatient stays in a short-term acute care


Summary of Medicare’s current payment systems by setting

Acute inpatient care Ambulatory care Post-acute care

Payment Hospital Ambulatory Skilled Homesystem Acute care Psychiatric outpatient surgical Outpatient nursing healthdescription hospitals facilities Physicians departments centers laboratories facilities agencies

Fiscal year began 1984 1983 1992 2000 1982 1984 1998 2001Basis of payment

Product definitionUnit of payment

Product classification system

Policies defining product boundaries

Product relative valuesComponents of relative values

Source of relativevalues

Base payment rate/conversion factorComponents of base amount

Source of base amount

T A B L EA-1

Prospective

Discharge

509 DRGs

72-hour ruleshort-staytransfers; high-cost outliers

Single value foreach DRG

Hospitals’ billedcharges

Facility costswith limit

Discharge

None

None

None

None

Prospective

Service

7,000�

HCPCS codes

Differentials bysetting, multipleor atypicalservices

Physician work;practiceexpenses;liabilityinsurance

Expertjudgement;practiceexpense data;premium survey

Prospective

Service

HCPCSgrouped in 570APCs

High-costoutliers; multipleservice discount

Single value foreach APC

Median ofestimatedservice costs

Prospective

Procedure

HCPCS in 8proceduregroups

Multiple servicediscount

Single amountfor each group

Median ofestimatedservice costs

Prospective

Test

1,100�

HCPCS codes

None

Combined withbase amount

None

Prospective

Day

44 RUG–IIIgroups

None

Therapyservices;nursing care

Staff-time studies

Prospective

60-day episode

80 HHRGs

Fewer than 5visits; high-costoutliers

Single value foreach HHRG

Estimated meancost per HHRG

continued on next page

Labor-related;nonlabor; capital

Updatedproviders’ 1982costs

Current per unitoperating costs

Facility’s annualcost report

Singleconversionfactor (for sumof relativevalues)

Projectedspending underprecedingmethod

Labor-related;other

Updated 1996OPD chargesadjusted to costs

Labor-related;other

1986 survey ofASCs’ costs andcharges

Carrier-specificrates with limit

Updated 1983lab charges

Therapy; nursingcare; routinecare

Targetaggregatespending

Labor-related,other

Spending inprecedingsystem

hospital.6 They receive care in more than4,800 facilities that contract withMedicare to provide services and agree toaccept the program’s predeterminedpayment rates as payment in full.7

Payments for inpatient care (about $94billion in 2001) account for the largestcomponent—about 34 percent—ofMedicare spending. These payments alsoprovide the largest single source of

hospitals’ revenues—about 23 percent ofoverall revenues.

From its inception in 1966 until 1983,Medicare paid hospitals for inpatientservices based on their incurred costs.This payment method gave providers littleincentive to produce services efficiently.Because they were costly and relativelyeasy to distinguish, episodes of hospital

inpatient care (stays) were the first to beconverted to prospectively determinedpayment, beginning in fiscal year (FY)1984. The hospital PPS is a maturesystem, but it nevertheless needs frequentadjustments to keep up with changes intechnology, practice patterns, and marketconditions that affect the amount and mixof resources hospitals use to furnishinpatient care. The inpatient PPS pays


6 The Medicare inpatient hospital benefit covers beneficiaries for 90 days of care per illness episode, with a 60-day lifetime reserve. Illness episodes begin whenbeneficiaries are admitted for care and end after they have been out of the hospital or a skilled nursing facility for 60 consecutive days. Beneficiaries are liable for adeductible of $840 for the first hospital stay in an episode. Daily copayments—currently $210—are imposed beginning on the 61st day.

7 Except for convenience items or services not covered by Medicare, providers are not permitted to charge beneficiaries more than the predetermined payment rate.Medicare pays the predetermined rate minus any beneficiary liability, such as a deductible or copayment; the provider then collects the remaining amount from thebeneficiary or a supplemental insurer.


Acute inpatient care Ambulatory care Post-acute care

Payment Hospital Ambulatory Skilled Homesystem Acute care Psychiatric outpatient surgical Outpatient nursing healthdescription hospitals facilities Physicians departments centers laboratories facilities agencies

Adjustments for local market conditionsLabor input prices

Other input prices

Other paymentadjustments

Payment updatemethod

Payments for capitalcosts

Other policies

T A B L EA-1

Hospital wageindex (HWIr)

COLA

Low-incomepatients (DSH);GME programs

Rise in hospitalmarket basketindex

Separateprospectiverates

Higher rates inlarge urbanareas; policiesfor ruralproviders

None

None

None

Rise in TEFRAmarket basketindex

Separate costpass-through

National limitadjusted toreflect localmarket wagelevel

SeparateGPCIs: work,practiceexpenses, PLI

None

Reduced ratesfor nonphysicianpractitioners

SGR formula

Included inpayment rate

10 percent add-on for healthprofessionalshortage areas(HPSAs)


None

None



Newtechnologypass-through;transitionalcorridors


None

None

Rise in CPI–U


None

None

None

None

Rise in CPI–U


National limit �median ofcarriers’ rates

Hospital wageindex (HWIu)

None

None

Rise in SNFmarket basketindex


None


None

None

Rise in homehealth marketbasket index


10 percentadd-on for ruralbeneficiaries


Note: APC (ambulatory payment classification), ASC (ambulatory surgery center), BLS (Bureau of Labor Statistics), CAH (critical access hospital), CMG (case-mix group), COLA(cost of living adjustment applied in Alaska and Hawaii), CPI–U (consumer price index for all urban consumers), CPT (Current Procedural Terminology), DRG (diagnosisrelated group), DSH (disproportionate share), GME (graduate medical education), FFS (fee-for-service), GPCI (geographic practice cost index), HCPCS (Healthcare CommonProcedure Coding System), HHRG (home health resource group), HWIr (hospital wage index with geographic reclassifications), HWIu (hospital wage index withoutgeographic reclassifications), LTC (long-term care), OPD (outpatient department), PE (practice expense), PLI (professional liability insurance), RUG–III (resource utilizationgroup, version III), SGR (sustainable growth rate), SNF (skilled nursing facility), TEFRA (Tax Equity and Fiscal Responsibility Act of 1982).

hospitals predetermined per-dischargerates that are based primarily on twofactors:

• the patient’s condition and relatedtreatment strategy, and

• market conditions in the facility’slocation.

Using information about patients’diagnoses, procedures, age, and dischargedestination reported on hospitals’ claims,Medicare assigns discharges to diagnosis

related groups (DRGs), which grouppatients who have similar clinicalproblems and are expected to requiresimilar amounts of hospital resources.Each DRG has a national relative weightthat reflects the expected relativecostliness of inpatient treatment for apatient in that group compared with thatfor the average Medicare patient. Groupsexpected to require more resources thanaverage have higher weights, and thoseexpected to require fewer resources havelower ones.

The payment rates for DRGs in each localmarket are determined by adjusting anational average base payment amount(the amount that would be paid for anaverage patient in a facility located in anaverage market) to reflect the input-pricelevel in the local market, and thenmultiplying the adjusted local amount bythe relative weight for each DRG.Payment rates also are increased forfacilities that operate approved physician(resident) training programs, those that



Services forPost-acute care special populations Other services

Payment Inpatient Long-term Outpatient Durablesystem rehabilitation care dialysis Hospice Ambulance medical Medicare+Choicedescription facilities hospitals care services services equipment plans

Fiscal year began 2002 2003 1982 1983 2002 1986 1998Basis of payment

Product definitionUnit of payment

Product classification system

Policies defining product boundaries

Product relative valuesComponents of relative values

Source of relativevalues

Base payment rate/conversion factorComponents of base amount

Source of base amount

T A B L EA-1

Prospective

Discharge

385 CMGs

Short-stayoutliers/deaths;transfers; high-costoutliers

Single value foreach CMG


Prospective

Discharge

499 LTC–DRGs

Short-stay outliers;transfers; high-costoutliers

Single value foreach LTC–DRG


Prospective

Dialysis treatment

None

None

None

None

Prospective

Day

4 care typegroups

Beneficiary givesup curativetreatment

Combined withbase amounts

None

Prospective

Trip

14 HCPCS within9 service levels,2 CPT codes

Base rateand mileage

Single valuefor eachservice level

Negotiatedrulemaking

Prospective

Item

HCPCS within 6equipmentcategories

None

Combined withbase amounts

None

Prospective

Month

Beneficiaries’demographics andhealth risk

All-inclusive capitationpayment rate

One value for eachenrollee category

FFS bills 1990–1995


Labor-related;other

Projectedspending underpreceding method

Labor related;other


Labor-related; other

1977–1979 costreports

Labor-related;other

Cost data fromMedicaredemonstration

Single conversionfactor for baserate


Single amount

Allowed chargesin 1986–1987

Updated 2001 rate;blended national/county rate

Historical FFSspending in countyand nation

treat a disproportionate share of low-income patients, and for other factors.

Because the inpatient PPS accounts for alarge share of Medicare spending, it facesongoing scrutiny, often leading totechnical and policy improvements. ThePPS payment rates are intended to coverthe costs that reasonably efficientproviders would incur in furnishing highquality care, thereby rewarding thosewhose costs fall below the payment rates.However, financial performance under thePPS differs substantially among certaingroups of hospitals. Some of thesedifferences represent intended effects of

policies adopted by the Congress. In otherinstances, they may reflect unintendedresults of inaccurate or inappropriatepayment adjustments, and failures toaddress factors that affect efficientproviders’ costs in certain circumstances.

Defining the hospital inpatientacute care products MedicarebuysUnder the inpatient PPS, Medicare setsper-discharge payment rates for distincttreatment episodes represented by 508DRGs, which are based on patients’clinical conditions and treatment

strategies.8 Clinical conditions aredescribed by patients’ dischargediagnoses, including the principaldiagnosis—the main problem requiringinpatient care—and up to eight secondarydiagnoses indicating other conditions thatwere present at admission (comorbidities)or developed during the hospital stay(complications). The treatment strategy—surgical or medical treatment—isdescribed by the presence or absence ofup to six procedures performed during thestay. Age, sex, and dischargedestination—for example, home, anotherPPS hospital, or a skilled nursing


8 Although the federal DRG classification system includes 527 categories, 19 are no longer used for Medicare payment.


Services forPost-acute care special populations Other services

Payment Inpatient Long-term Outpatient Durablesystem rehabilitation care dialysis Hospice Ambulance medical Medicare+Choicedescription facilities hospitals care services services equipment plans

Adjustments for local market conditionsLabor input prices

Other input prices

Other payment adjustments

Payment update method

Payments for capital costs

Other policies

T A B L EA-1


None

Low-incomepatients

Rise in modifiedTEFRA marketbasket index

Included inprospective rates

Higher rates inrural areas


COLA

None

Rise in modifiedTEFRA marketbasket index

Included inprospective rates

None

40% 1986HWI� 60%1980 BLS wageindex

None

Higher rates forhospital-basedfacilities

No routine update


Exceptions; extrapayments forsome tests anddrugs

Hospice wageindex

None

None



Annual paymentper beneficiarycapped

PE GPCI forphysician feeschedule

None

Rural and low-volume add-ons

Rise in CPI–U


Qualifying CAHson cost-basedreimbursement

Carrier-specificrates with limit

None

Product-specificnational limits

Rise in CPI–U


None

Hospital wage index(HWIu); GPCIs

None

Floor rates

Rise in aggregate FFSspending; 2 percentminimum

Included in payment rate

None

Note: APC (ambulatory payment classification), ASC (ambulatory surgery center), BLS (Bureau of Labor Statistics), CAH (critical access hospital), CMG (case-mix group), COLA(cost of living adjustment applied in Alaska and Hawaii), CPI–U (consumer price index for all urban consumers), CPT (Current Procedural Terminology), DRG (diagnosisrelated group), DSH (disproportionate share), GME (graduate medical education), FFS (fee-for-service), GPCI (geographic practice cost index), HCPCS (Healthcare CommonProcedure Coding System), HHRG (home health resource group), HWIr (hospital wage index with geographic reclassifications), HWIu (hospital wage index withoutgeographic reclassifications), LTC (long-term care), OPD (outpatient department), PE (practice expense), PLI (professional liability insurance), RUG–III (resource utilizationgroup, version III), SGR (sustainable growth rate), SNF (skilled nursing facility), TEFRA (Tax Equity and Fiscal Responsibility Act of 1982).

facility—are also occasionally used todistinguish groups of patients who areexpected to use different amounts ofresources.

The DRG definitions have a tree-likestructure. Based on the principaldiagnosis, cases are first assigned to 1 of25 major diagnostic categories (MDCs),reflecting the affected organ system (suchas the digestive system) or the etiology ofthe condition (such as burns or significanttrauma). Within each MDC, cases aresubdivided into those with and thosewithout operating room or othersignificant procedures. Each of thesebroad groups is then further divided; thesurgical group by type of procedure andthe medical group by specific type ofcondition as indicated by the principaldiagnosis. Finally, medical and surgicalsubgroups are often subdivided further toform DRGs distinguished by the presenceor absence of comorbidities orcomplications indicated by specificsecondary diagnoses.9

CMS annually reviews the DRGdefinitions to ensure that they continue toinclude cases with clinically similarconditions requiring comparable amountsof inpatient resources. When the reviewshows that clinically similar cases within aDRG consume atypical quantities ofresources, CMS often reassigns them to adifferent DRG with comparable resourceuse; less often, CMS creates a newDRG.10

In return for receiving Medicare’spredetermined payments, hospitals areexpected to furnish a reasonably well-defined bundle of inpatient services foreach DRG. Facing fixed payment rates,

however, providers have financialincentives to reduce their inpatient costsby moving some normally includedservices to another setting—such as anoutpatient department or a skilled nursingfacility—and bill those servicesseparately. To counter these financialincentives, Medicare has adopted policiesthat help strengthen the boundaries of theinpatient service bundles associated withthe DRGs. Thus, patients must stayovernight before their discharges qualifyfor payment under the inpatient PPS.Related outpatient department servicesthat were delivered in the three daysbefore admission are included in thepayment for the inpatient stay and maynot be separately billed (the 72-hour rule).Similarly, payments for services may bereduced when patients are transferred toanother hospital after a stay that is morethan one day shorter than the nationalaverage stay for the DRG. The samepayment reductions apply for certainDRGs when patients are transferred topost-acute care facilities, such asrehabilitation or skilled nursing facilities,or discharged to receive clinically relatedhome health care that begins within threedays.

Setting the payment ratesMedicare sets separate per-dischargeoperating and capital payment rates,which are intended to cover the operatingand capital costs that efficient facilitieswould be expected to incur in furnishingcovered inpatient services.11 Operatingpayment rates cover costs for labor andsupplies; capital payment rates cover costsfor depreciation, interest, rent, and certainproperty-related expenses for insuranceand taxes.

Medicare sets operating and capitalpayment rates using similar methods andfactors. In general, CMS sets nationalpayment rates for all types of cases bymultiplying a base payment amount by therelative weight for each DRG. The DRGpayment rates are then adjusted to reflectthe local level of input prices in eachmarket area. Finally, operating and capitalpayment rates are adjusted to account forcertain hospital- and case-specific factors.

The base payment amounts Medicaresets two separate operating base paymentamounts (known as standardized paymentamounts): one for large urban areas—metropolitan statistical areas (MSAs) witha population of one million or more—andone for all other urban and rural areas.12

These base payment amounts representwhat a hospital located in these areaswould be paid for operating expenses foran average Medicare patient (before anyadjustments). The base operating amountsper discharge for FY 2003 are $4,251 forlarge urban areas and $4,184 for otherareas.

Capital payments have only recently beenmade fully prospective, having completeda 10-year phase-in during FY 2001.13 Thebase capital rate for discharges fromhospitals in large urban areas for FY 2003is $419; it is $407 for hospitals located inother areas.

The diagnosis related group relativeweights Medicare assigns a weight toeach DRG reflecting the average relativecostliness of cases in that group comparedwith that for the average Medicare case.The same DRG weights are used to setoperating and capital payment rates. CMSrecalibrates the DRG weights annually


9 These groups are sometimes divided further to form DRGs for pediatric patients (under age 17); a few DRGs are also distinguished by patient sex or dischargedestination.

10 For example, CMS established a new DRG when it found that tracheostomy patients were substantially more costly than others in the same DRGs.

11 Certain costs are excluded from the inpatient PPS and paid separately, such as direct costs of operating graduate medical education programs, organ acquisition costs,and bad debts related to beneficiaries’ nonpayment of their cost-sharing liabilities (deductibles and copayments).

12 Hospitals in Puerto Rico receive a 50/50 blend of the federal base payment amount and a Puerto Rico-specific rate.

13 New hospitals are exempt from prospective payment for capital costs for two years. During this period, they are paid 85 percent of their Medicare-allowable capitalcosts.

based on average standardized billedcharges for all PPS cases in each DRG inthe most recent Medicare bill file.14

Adjustment for market conditionsMedicare’s base operating and capitalpayment rates are adjusted to reflect theexpected impact of differences in localmarket prices for labor and other inputs.The base operating payment is adjusted byan area wage index; in Alaska andHawaii, a cost of living adjustment(COLA) is also applied. The area wageindex is intended to measure differencesin hospital wage rates among labormarkets; it compares the average hourlywage for hospital workers in each MSA orstatewide rural area relative to thenationwide average.15 The wage index isapplied to the labor-related portion of thestandardized payment amount—71percent of the total—which reflectsCMS’s estimate of the portion ofoperating costs affected by local wagerates and fringe benefits. The wage indexis revised each year based on wage datareported by PPS hospitals on their annualMedicare cost reports. The COLA reflectsthe higher costs of supplies and othernonlabor resources in Alaska and Hawaii;it increases the nonlabor portion of PPSoperating payments—29 percent of thetotal—for hospitals in these states by asmuch as 25 percent.

The federal rate for capital payments isadjusted to reflect local market conditionsusing a geographic adjustment factor(which is based on the area wage index)and, for Alaska and Hawaii, the sameCOLA as used for operating payments.

Other adjustments Payment rates alsomay be adjusted to reflect higher costs of

care in hospitals that operate approvedresident training programs, revenue lossesassociated with treating low-incomepatients, and the financial burden ofexceptionally high-cost cases. Theseadjustments are intended to preserveaccess to care for Medicare beneficiariesby protecting hospitals that face certaincost or revenue pressures.16 Medicare alsomakes special payments to several groupsof hospitals.17 Most of these specialpayment provisions are designed to helprural hospitals, although some urbanfacilities also may qualify.

Indirect medical education paymentsTeaching hospitals receive add-onpayments to reflect the additional(indirect) costs of patient care associatedwith operating approved physiciantraining programs. The size of the indirectmedical education (IME) adjustmentapplied to DRG payments depends on thehospital’s teaching intensity, as measuredby the number of residents per bed. In2001, approximately 1,100 hospitalsreceived IME payments.

Disproportionate share paymentsHospitals that treat a disproportionateshare (DSH) of low-income patientsreceive additional payments that areintended to partially offset their revenuelosses from furnishing uncompensatedcare. The DSH adjustment is based onnine different formulas and depends onurban or rural location, number of acutecare beds, and other characteristics. Theamount of the adjustment—the percentagefrom the applicable formula multiplied bythe hospital’s total DRG payments—depends on the hospital’s low-incomepatient share. A hospital’s low-incomepatient share is the sum of the percentage

of its Medicare inpatient days furnished topatients eligible for Supplemental SecurityIncome benefits and the percentage of itstotal acute inpatient days furnished toMedicaid patients. No DSH payments aremade unless a hospital’s low-incomepatient share exceeds 15 percent.

Until 2001, small urban hospitals—thosewith fewer than 100 beds—and most ruralproviders had to meet substantially higherminimum low-income patient shares toqualify for DSH payments. The BIPAreduced the qualifying thresholds forsmall urban and rural providers to thesame level applied for larger urbanhospitals. In 2001, these policy changesexpanded eligibility for DSH paymentsfrom about 1,800 hospitals to about 2,800hospitals; about 800 of the newly eligiblefacilities were in rural areas.

Outlier payments In general, hospitalsare expected to offset losses on somecases (in which costs exceed the paymentrate) with gains on others (in which costsare below payments). Some cases,however, are extraordinarily costly,producing losses that may be too large tooffset. Hospitals facing fixed paymentrates have strong financial incentives toavoid patients who may be likely torequire extraordinary care. To promoteaccess to high-quality inpatient care forseriously ill beneficiaries, Medicaremakes extra payments for these so-calledoutlier cases, in addition to the usualoperating and capital DRG payments.Outlier cases are identified by comparingtheir costs to a DRG-specific thresholdthat is the sum of the hospital’s DRGpayment for the case (both operating andcapital), any IME and DSH payments, anda fixed loss amount. For instance, in 2003


14 Hospitals’ billed charges are standardized to improve comparability. This involves adjusting charges to remove differences associated with variations in local marketprices for inputs and those related to the size and intensity of hospitals’ resident training activities.

15 A hospital may request geographic reclassification to an adjacent market area for the standardized payment amount, the wage index (and capital geographicadjustment factor), or both. To qualify, a hospital must demonstrate that it is located within 15 miles of the border of the adjacent area. It also must show that its hourlywages are above average for its market area (above 106 percent for rural hospitals and 108 percent for urban hospitals) and comparable to the average in the areato which it seeks reclassification (at least 82 percent for rural hospitals and 84 percent for urban hospitals).

16 Medicare also reimburses acute-care hospitals for bad debts resulting from beneficiaries’ nonpayment of deductibles and copayments after providers have madereasonable efforts to collect the unpaid amounts. The BBA reduced these payments, but the BIPA added some back. As a result, Medicare paid 70 percent of allowablebad debts in FY 2000.

17 These special payment provisions are discussed in greater detail in MedPAC’s June 2001 Report to the Congress: Medicare Payment Advisory Commission. Report tothe Congress: Medicare payment policy. Washington (DC), MedPAC. June 2001.

the threshold is set at the hospital’s DRGpayment plus any IME and DSHpayments plus $33,560—the nationalfixed loss amount—adjusted to reflectinput price levels in the hospital’s localmarket. Medicare pays 80 percent ofhospitals’ costs above their fixed lossthresholds. Costs for individual cases areestimated by reducing the hospital’scovered charges for the case by its overallMedicare cost-to-charge ratio from itsmost recent settled annual cost report.IME and DSH adjustments are not appliedto outlier payments. Outlier payments arefunded by offsetting reductions in theoperating base payment amounts (5.1percent) and the capital federal rate (5.3percent).

Transfer policy Medicare reduces DRGpayments when the patient is transferredto another PPS hospital, or in someinstances to a post-acute care setting.When a patient is transferred to anotherPPS hospital, the transferring facility ispaid a per diem amount for each daybefore the transfer occurs, up to amaximum of the full DRG payment.18

The hospital receiving a transferredpatient is paid as if the case had not beentransferred.19 Beginning in FY 1999,discharges in 10 DRGs are treated astransfers if patients are sent to a long-termcare hospital or a rehabilitation,psychiatric, or skilled nursing facility, orthey receive clinically related home healthcare. This policy is intended to strengthenthe boundaries of the hospital inpatientservice bundle by reducing providers’financial incentives to unbundle servicesnormally furnished during the hospitalinpatient stay. The 10 affected DRGs wereselected by the Secretary of HHS based ontheir high volume and disproportionatelyhigh likelihood of post-acute care use. TheSecretary was authorized to expand the set

of DRGs to which this policy appliesbeginning in FY 2001, but has not yetdone so.

Payment updates Both the operatingand capital payment rates are updatedannually. The operating update is set bythe Congress in law; the annual capitalupdate is determined by the Secretary ofHHS. In recommending annual updates,the Commission and CMS useframeworks that take into accountprojected changes in input prices, scienceand technology, productivity, and otherfactors expected to affect efficienthospitals’ costs.

Recommended and statutory updates forthe operating and capital payment ratesare generally expressed relative to theprojected increase in the hospital MBindex, which measures changes innational average prices for inputshospitals purchase to produce services. Anupdate usually would be expressed then asbeing equal to MB or MB minus 0.5percentage points, for example.

Payment for specialtypsychiatric facilities Medicare beneficiaries with mentalillnesses or alcohol- and drug-relatedproblems are frequently treated inspecialty psychiatric facilities, eitherfreestanding hospitals or specializedhospital-based units. These hospitalsgenerally furnish short-term acute care. Tobe admitted to a specialty facility, patientsgenerally must be considered a risk tothemselves or others.20 Payments topsychiatric facilities (almost $3 billion in2001) represent only a small part of totalMedicare spending (about 1 percent), butthe program accounts for about 30 percentof psychiatric facilities’ annual revenues.

Psychiatric facilities are paid forfurnishing care to Medicare beneficiariesunder cost growth limits established in theTax Equity and Fiscal Responsibility Actof 1982 (TEFRA); payments are based ontheir incurred average operating costs perdischarge, subject to an annually adjustedfacility-specific limit (see text box).

The Congress required CMS to developand implement a per diem PPS to replacethe earlier payment methods; CMS plansto implement the new system in 2003.

As is the case for stays in short-term acutecare hospitals, beneficiaries treated inspecialty psychiatric facilities areresponsible for a deductible—$840 in2003—for the first admission during aspell of illness, and for a copayment—$210 per day—for the 61st through 90thdays. Beneficiaries treated for psychiatricconditions in specialty facilities also arecovered for 90 days of care per illnessepisode, with a 60-day lifetime reserve.21

Over their lifetimes, however,beneficiaries are limited to 190 days oftreatment in freestanding psychiatrichospitals.

Ambulatory care

Medicare beneficiaries receive ambulatorycare services from a variety ofpractitioners in several settings. The mostcommon ambulatory services are:

• physician services.

• outpatient hospital care.

• ambulatory surgical care.

• outpatient laboratory services.

These physicians and providers furnish awide range of services, including some


18 The per diem rate is the hospital’s DRG payment rate divided by the national average length of stay for the same DRG. Generally, hospitals receive twice the per diemrate for the first day and the per diem rate for each additional day up to the full DRG rate. Hospitals may also receive outlier payments calculated using a loss thresholdprorated to reflect the length of stay.

19 If the patient is discharged to yet another PPS hospital, the transfer payment rules again apply.

20 Beneficiaries are also treated for psychiatric or alcohol- and drug-related conditions in regular beds in acute care hospitals; in these instances providers are paid underthe acute care inpatient PPS.

21 Beneficiaries are liable for a higher copayment for each lifetime reserve day—$420 per day in 2003.


Payment for facilities exempt from the prospective payment system for acute care hospitals

From Medicare’s inception until1983, all hospitals were paidbased on their Medicare-

allowable incurred costs. In the TaxEquity and Fiscal Responsibility Act of1982 (TEFRA), the Congress setfacility-specific limits on hospitals’operating costs per discharge, withpenalties and rewards based on whethertheir costs were above or below thefacility-specific limit or target. In 1984,short-term general acute care hospitalsmoved to the inpatient prospectivepayment system (PPS), but theCongress excluded other classes offacilities because the types of casesthey treated and the relationshipsbetween case characteristics andefficient providers’ costs were not wellunderstood.

Five classes of facilities were paidunder TEFRA between 1983 and2002—cancer hospitals, children’shospitals, long-term care hospitals, andrehabilitation and psychiatric facilities(hospitals and units). From 1983 to1998, each provider was paid anoperating amount for each discharge,equal to the lesser of its currentoperating costs or a facility-specifictarget amount. The facility-specifictarget amount (limit) for each providerwas based on its operating costs per

discharge during its base year, updatedfor inflation using a TEFRA marketbasket index which measures changesin the prices of goods and services thatspecialty facilities must buy to produceinpatient care. These facilities werepaid for capital costs based on theirMedicare-allowable incurred expensesfrom 1983 to 1998. From 1998 to 2002,facilities were paid 85 percent ofallowable capital costs.

Because facilities’ operating targetswere based on their own historicalcosts, TEFRA payments often variedsubstantially among facilities. Inaddition, new providers often enteredthe Medicare program with higher coststhan older providers had, giving newproviders higher targets and creatingpayment inequities. The Congressrequired the Centers for Medicare &Medicaid Services (CMS) to designPPSs for the three largest classes offacilities—rehabilitation facilities,long-term hospitals, and psychiatricfacilities. Rehabilitation facilities beganpayment under a PPS in January 2002;long-term care hospitals beganpayment under a PPS in October 2002.Speciality psychiatric facilities willcontinue to be paid under TEFRA untila PPS for this group is implemented.

To reduce inequities in target amounts,the Balanced Budget Act of 1997(BBA) established temporary nationalcaps on facilities’ target amounts from1998 through 2002 for three providergroups: long-term care hospitals andrehabilitation and psychiatric facilities.(Cancer and children’s hospitalscontinued to be paid under the oldTEFRA method during this period.) Inaddition, the BBA temporarily reducedcapial payments for all TEFRAfacilities to 85 percent of theirallowable capital costs. Beginning in2003, these provisions expired andfacilities returned to the old TEFRApayment method.

However, the BBA also establishedtwo permanent features in the TEFRApayment system. One is paymentlimitations for new specialty facilitiesexcluded from the acute care hospitalPPS on or after October 1, 1997. Theother is revised incentive payments forfacilities with costs below their targetsand relief payments for facilities withcosts above their targets.

Facilities’ operating targets are updatedaccording to a TEFRA market basket.The market basket index for FY 2003 is3.5 percent. �

common to more than one setting. Forexample, beneficiaries may receiveidentical services in physicians’ officesand hospital outpatient departments.Outpatient laboratory services helpphysicians in offices and outpatientdepartments to diagnose, treat, andmonitor patients’ illnesses or conditions.Some ambulatory surgeries can beperformed in physicians’ offices,outpatient departments, or ambulatorysurgical centers.

Payment for physicianservices Physician services include office visits,surgical procedures, and a broad range ofother diagnostic and therapeutic services.These services are furnished in allsettings, including physicians’ offices,hospitals, ambulatory surgical centers,skilled nursing facilities and other post-acute care settings, hospices, outpatientdialysis facilities, clinical laboratories, andbeneficiaries’ homes. Medicare payments

to physicians (about $56 billion in 2001)account for about 20 percent of totalspending.

The Medicare physician payment systemwas implemented in 1992. To makepredetermined payments for physicianservices, Medicare uses a fee schedulewith payment rates for more than 7,000services. Many services have twopayment rates: a higher rate for servicesprovided in nonfacility settings, such asphysicians’ offices, and a lower rate for

those furnished in facilities, such ashospitals. Rates are lower for servicesfurnished in facilities because physicians’practice costs are generally lower. Also,when a service is provided in a facility,Medicare pays both the facility and thephysician.

Each service has a weight—called arelative value unit—that measures therelative costliness of three types ofresources used to provide physicianservices: physician work, practiceexpenses, and expenses for professionalliability insurance (PLI). Payment ratesfor services in each local market aredetermined by adjusting each relativeweight to reflect the input-price level inthat market, and then multiplying the totalof the adjusted weights by a dollar amountcalled the fee schedule’s conversionfactor. Payment rates for physicians’services are adjusted further when theyare:

• furnished by practitioners other thanphysicians.

• furnished in Health ProfessionalShortage Areas (HPSAs).

• provided by a physician who has notagreed to accept Medicare’s paymentrate as payment in full.

• atypical (for example, the service isassisting the primary surgeon ratherthan serving as the primary surgeonperforming a surgical procedure).

Payments are updated every yearaccording to a formula called thesustainable growth rate (SGR) system,which is intended to keep spendinggrowth consistent with growth in thenational economy.

The physician fee schedule was adoptedmore than 10 years ago, but efforts toimprove it continue. For example, CMS isworking with the physician community torefine the relative weights for practiceexpenses.

Defining the physician servicesthat Medicare buysUnder the physician fee schedule, the unitof payment is the individual service, suchas an office visit or a diagnosticprocedure. These products, however,range from narrow services (an injection)to broader bundles of services associatedwith surgical procedures, which includethe surgery and related preoperative andpostoperative visits. All services—surgical and nonsurgical—are classifiedand reported to CMS according to theHealthcare Common Procedure CodingSystem (HCPCS), which contains codesfor more than 7,000 distinct services.

Setting the payment ratesUnder the fee schedule, payment rates arecalculated by adding three relativeweights and multiplying the sum by theconversion factor. The weights reflect therelative costliness of the inputs used toprovide physician services: physicianwork, practice expenses, and PLIexpenses. The relative weights forphysician work are based on physicians’assessments of the relative levels of time,effort, skill, and stress associated witheach service. The relative weights forpractice expense are based on theexpenses physicians incur when they rentoffice space, buy supplies and equipment,and hire nonphysician clinical andadministrative staff. The PLI relativeweights are based on the premiumsphysicians pay for professional liabilityinsurance.

In calculating payment rates, each of thethree relative weights is adjusted to reflectthe price level for related inputs in thelocal market where the service isfurnished. Three geographic practice costindexes are used for this purpose. The feeschedule payment amount is thendetermined by summing the adjustedweights and multiplying the total by thefee schedule conversion factor.

Payments under the physician feeschedule also may be adjusted to reflectother factors. First, payments aredecreased if services are furnished bycertain nonphysician practitioners.

Services provided by physician assistantsand nurse practitioners are paid at 85percent of physicians’ fees, and nursemidwives’ services are paid at 65 percent.

Second, payments are adjusted accordingto so-called payment modifiers that appearon claims for payment to show whetherthe service provided was atypical. Forexample, physicians use a modifier to billfor a service when they serve as assistantsurgeons. Payment for an assistantsurgeon is 16 percent of the fee scheduleamount for a surgical procedure. Othermodifiers apply to multiple surgicalprocedures performed for the same patienton the same day, preoperative orpostoperative management withoutsurgical care, and bilateral surgery.

Third, under the Medicare incentivepayment program, physicians receivebonus payments when they provideservices in HPSAs. These payments areintended to attract more physicians toHPSAs. The bonus increases payments tothese physicians by 10 percent (excludingbeneficiary coinsurance).

Fourth, payments are adjusted downwardwhen services are furnished by physicianswho are not in Medicare’s participatingphysician and supplier program. Paymentrates for services provided bynonparticipating physicians are 95 percentof the fee schedule payment rate.

The fee schedule’s relative weights areupdated at least every five years; HCPCScodes and the conversion factor areupdated annually. The update of relativeweights includes a review of changes inmedical practice, coding changes, newdata, and the addition of new services. Incompleting its review, CMS receivesadvice from a group of physicians andother professionals sponsored by theAmerican Medical Association andphysician specialty societies.

The annual updates for the conversionfactor are made according to the SGRsystem. If actual spending is less than thetarget, the update is greater than thechange in input prices for physicianservices. If actual spending is greater than


the target, the update is less than thechange in input prices.

Payment for outpatienthospital care Medicare beneficiaries receive a widerange of services in hospital outpatientdepartments, from injections to surgicalprocedures requiring general anesthesia.Spending for these services is growingrapidly, largely because of changes intechnology and medical practice that havefostered new services and encouragedshifts in care from inpatient to ambulatorycare settings. Outpatient hospital careaccounted for about 7 percent of totalMedicare spending in 2001, or about $16billion.22

Medicare originally paid hospitals foroutpatient care based on their allowableincurred costs. The BBA almostcompletely eliminated such cost-basedpayment by requiring CMS to developand adopt an outpatient PPS, which wasimplemented in August 2000.

In requiring the outpatient PPS, theCongress also reduced beneficiarycopayments for outpatient hospital care.When the BBA was enacted, copaymentsaccounted for about 50 percent of totalMedicare payments to hospitals foroutpatient care. Under the new paymentsystem, beneficiaries’ share of totalpayments will slowly decline. MedPAChas recommended that the Congressaccelerate the reduction in thesecopayments.

Like the payment system for physicianservices, the new outpatient PPS is a feeschedule. It sets payment rates forindividual services based on a set ofrelative weights, a conversion factor, andan adjustment for geographic differencesin input prices. The PPS also includes anoutlier adjustment for extraordinarilyhigh-cost services and so-called pass-through payments for certain newtechnologies that are used as inputs in thedelivery of services.

Because of uncertainty about the effects ofthe new system, certain types of hospitalsare at least partially protected fromfinancial losses. Cancer and children’shospitals are permanently held harmlessfrom losses; small rural hospitals are heldharmless through 2003. Other hospitalsthat experience losses are eligible forpartially offsetting payment adjustmentsthrough 2003.

Defining the outpatient hospitalproducts that Medicare buysMedicare pays for outpatient servicesbased on the individual service orprocedure provided, as identified by anHCPCS code. CMS classified procedures,evaluation and management services, anddrugs and devices furnished in outpatientdepartments into about 570 ambulatorypayment classifications (APCs). TheseAPCs group items and services that areclinically similar and use comparableamounts of resources. More than 300 ofthe APCs identify drugs or devices used inconjunction with a procedure. In addition,some new services are assigned to certain“new technology” APCs based only onsimilarity of resource use. CMS chose toestablish new technology APCs becausesome services were too new to berepresented in the data used to develop theoutpatient PPS. Services remain in theseAPCs for two to three years while CMScollects the clinical and cost datanecessary to refine and update the APCclassification system. Additional servicesmay be placed in the new technologyAPCs after review by CMS.

Within each APC, CMS bundles integralservices and items with the primaryservice. For example, the bundle for asurgical procedure includes operating andrecovery room services, mostpharmaceuticals, anesthesia, and surgicaland medical supplies. In deciding whichservices to bundle and which to payseparately, CMS considered commentsfrom hospitals, hospital suppliers, andothers. For example, in response to publiccomments, CMS separated corneal tissue

acquisition, maintenance, and distributionfrom services requiring corneal tissue.CMS also pays separately for blood, bloodproducts, and plasma-based andrecombinant therapies.

Unlike all other services included in theoutpatient PPS—for which the unit ofpayment is the service or procedureprovided—partial hospitalizations forpsychiatric services are paid on a per diembasis. These intensive outpatientpsychiatric services may be provided by ahospital outpatient department or by acommunity mental health center, and theper diem payment rate represents theexpected facility costs for a day of care.

Setting the payment ratesPayment rates in the outpatient PPS areintended to cover hospitals’ operating andcapital costs for the facility services theyfurnish; professional services (physicians’services provided to individual patients,for example) are paid separately.Outpatient payment rates are determinedby multiplying the relative weight for anAPC by a conversion factor. Except forthe new technology APCs, each APC hasa relative weight that is based on themedian cost of services in that APC.Services are assigned to a new technologyAPC based on their expected cost. Newtechnology APCs range from $0–$50 to$5,000–$6,000, with an additionalcategory at $19,500–$20,500; the relativeweights are set at the midpoint of theseranges.

The conversion factor translates therelative weights into dollar paymentamounts. The initial conversion factor wasset so that projected total payments—including beneficiaries’ copayments—would equal the estimated amount thatwould have been spent under the oldpayment methods, after correcting forsome anomalies in statutory formulas.

To account for geographic differences ininput prices, the labor portion of theconversion factor (60 percent) is adjustedby the hospital wage index.


22 Total spending on all hospital outpatient services (those covered by the outpatient PPS as well as those paid under separate fee schedules or based on costs) accountedfor $18.4 billion in 2001.

The outpatient PPS includes fouradditional payment adjustments: pass-through payments for new technology;outlier payments for high-cost services;hold-harmless payments for cancer,children’s, and small rural hospitals; andtransitional corridor payments that help tolimit hospitals’ financial losses under thePPS.

In addition to the new technology APCs,the pass-through payments are a secondway that the outpatient PPS accounts fornew technologies. Unlike the newtechnology APCs, however, pass-throughpayments are not payments for individualservices. Instead, they are payments forcertain new technology items—drugs,biologicals, and implantable devices—thatare used in the delivery of services. Bysupplementing the payments forindividual services, pass-throughpayments are meant to help ensurebeneficiaries’ access to new technologiesthat are not well represented in data thatCMS uses to set the PPS payment rates.For drugs and biologicals, the paymentsare based on average wholesale prices.For devices, the payments are based oneach hospital’s costs (as determined byadjusting its charges using a cost-to-charge ratio). By law, total pass-throughpayments are limited to 2.5 percent oftotal payments under the outpatient PPS,and the conversion factor is reduced by2.5 percent to finance them. If CMSprojects that pass-through payments willexceed this limit during a year, the agencyis required to reduce all pass-throughpayments in that year by a uniformpercentage to meet the limit. However,CMS did not maintain budget neutralityfrom August 2000 to April 2002.

Outlier payments are made for individualservices or procedures that haveextraordinarily high costs, compared withthe payment rates for their APC group. In2003, outliers are defined as services withestimated costs that exceed a thresholdequal to 2.75 times the PPS payment rate.Hospitals will be reimbursed for 45percent of the difference between thethreshold and the estimated cost of theservice in 2002. Aggregate outlier

payments are limited to 2 percent of totalpayments; outlier payments are financedby reducing the conversion factor by 2percent.

The BBRA mandated that cancerhospitals and outpatient departments ofsmall rural hospitals (100 or fewer beds)be held harmless from financial lossesunder the PPS. This protection ispermanent for cancer hospitals; smallrural hospitals are protected until 2003. Inaddition, the BIPA extended permanenthold-harmless protection to children’shospitals. These hospitals will be paidaccording to the PPS payment rates, but iftheir PPS payments are lower than thosethey would have received under previouspolicies, they will receive extra paymentsto make up the difference.

To smooth the way to the outpatient PPS,the Congress mandated transitionalcorridor payments in the BBRA that willcontinue through 2003. The amount ofthese payments depends on the differencebetween a hospital’s PPS payments andwhat it would have received under theprevious payment policy. Corridorpayments are intended to make up a highproportion of hospitals’ small losses but adeclining proportion of larger losses. Forexample, in 2000 and 2001, corridorpayments made up 80 percent of lossesthat were less than 10 percent of what thehospital would have received underprevious policy, but only 70 percent oflosses in the 10 to 20 percent range. In2002 and 2003, the transitional corridorpayments make up declining proportionsof hospitals’ revenue losses under thePPS.

The APC groups and their relativeweights are reviewed and revisedannually. The review considers changes inmedical practice, changes in technology,the addition of new services, new costdata, and other relevant information. CMSconsults with a panel of outside experts aspart of this review.

CMS also annually updates theconversion factor by the projectedincrease in hospital market basket indexunless the Congress stipulates otherwise.

Payment for care providedby ambulatory surgicalcentersSince 1982, Medicare has paid for thefacility costs of surgical proceduresprovided in freestanding or hospitalowned and operated ambulatory surgicalcenters (ASCs). ASCs are distinctfacilities that furnish only ambulatorysurgery; the most common procedures arecataract removal and lens replacement,other eye procedures, and colonoscopy.Payments to ASCs (about $1.6 billion in2001) account for less than 1 percent oftotal Medicare spending.

Medicare pays for surgery-related facilityservices provided in ASCs—such asoperative nursing, recovery care,anesthetics, drugs, and other supplies—using a simple fee schedule. (Medicarepays for the related physician services—surgery and anesthesia—under thephysician fee schedule.) The ASC feeschedule sets payment rates for only nineprocedure groups. The payment rates areadjusted to reflect geographic differencesin market input prices. Medicare mustrevise the payment rates at five-yearintervals based on a survey of ASCs’ costsand charges. Between revisions, the ratesare to be updated annually using theconsumer price index for all urbanconsumers (CPI–U).

Defining the care that Medicarebuys from ambulatory surgicalcentersThe unit of payment in the ASC paymentsystem is the individual surgicalprocedure. Each of the 2,300 proceduresapproved for payment in an ASC isclassified into one of nine paymentgroups.

Approved procedures generally arelimited to those that are provided inhospital inpatient settings and can also beperformed safely in outpatient facilities.Procedures frequently performed inphysicians’ offices are specificallyexcluded from ASC coverage. ASC-approved procedures usually require lessthan 90 minutes of operating room time


and less than 4 hours of recovery roomtime.

Setting the payment ratesTo set ASC payment rates, CMS mustsurvey a sample of ASCs every five yearsto collect data on their costs and chargesfor individual procedures. After auditingthe survey data, CMS adjusts ASCs’charges to reflect costs using cost-to-charge ratios. CMS sets the nationalpayment rate for each of the nine paymentgroups equal to the estimated median costof procedures in that group. To accountfor geographic differences in market inputprices, CMS adjusts the labor portion ofthe rate using the hospital wage index forthe ASC’s location. The labor portion ofthe rate is currently 34.45 percent.23 ASCpayment rates also are adjusted whenmultiple surgical procedures areperformed during the same operativesession. In this case, the ASC receives fullpayment only for the procedure with thehighest payment rate; payments for theother procedures are reduced to one-halfof their usual rates.

Between revisions to the payment system,the ASC payment rates are to be updatedannually based on the CPI–U. The BBAlimited those updates to the CPI–U minus2 percentage points (but not less thanzero) from FY 1998 through FY 2002.CMS also is required by law to updateevery two years the list of proceduresperformed in ASCs that are eligible forMedicare payment.

Payment for outpatientlaboratory services Clinical laboratory tests help physiciansdiagnose, treat, and monitor patients’illnesses and conditions. Beneficiariesmay receive tests during a hospital stay ora visit to a physician’s office or outpatientdepartment. Medicare pays hospitals fortests furnished during a hospital stay aspart of the bundled inpatient payment. Incontrast, Medicare pays the labs directlybased on a fee schedule for tests

performed in an outpatient setting. Threemain types of labs serve these ambulatorypatients: hospital-based labs; independentlabs, which usually serve a region; andphysician office labs, which generallyperform only relatively simple tests.Although Medicare payments account forabout 30 percent of laboratories’ revenues,laboratory payments account for about 2percent of total Medicare spending.

Medicare uses a simple PPS (feeschedule) established in 1984. Paymentrates were initially set separately for morethan 1,100 tests in each carrier’sgeographic market, based on what locallabs charged in 1983; since then, the rateshave been updated periodically forinflation. PPS payment rates are alsolimited by national service-specificmaximums that affect almost all labclaims.

Defining the laboratory productsMedicare buysMedicare sets payment rates for more than1,100 HCPCS codes used in billing forlaboratory services. Although in theorythere is a separate code for each service, inpractice a single HCPCS code mayidentify more than one testing method fora given substance or more than onesubstance analyzed by a single method.Panel tests, which are tests commonlyordered together, have their own HCPCScodes as well.

Setting the payment ratesThe fee schedule payment rates representthe total payment to laboratories;beneficiary copayments are not required.CMS assigns payment amounts for alllaboratory HCPCS codes in each carriermarket based upon 1983 charges from thelaboratories in that market. Medicarepayments were set at the 60th percentileof prevailing charges for freestandinglaboratories and the 62nd percentile forhospital-based laboratories in each area. In1987, fees for outpatient services inhospital laboratories, other than thoseperformed in sole community hospitals,

were reduced to the 60th percentile ofprevailing charges. Fee schedule amountsdiffer from carrier to carrier in someinstances, but no separate geographicadjustment is provided. Beginning in1986, the Congress established upperlimits on laboratory payment rates, callednational limitation amounts (NLAs).NLAs are based on the median of allcarrier rates for each test. The NLAs havebeen repeatedly reduced and currently areset at 74 percent of the median of all localfee schedule amounts for each procedure.Because so many of the carrier paymentrates are constrained by the NLAs, mostlaboratory services are paid the samenational rate.

When newly developed tests are used bylaboratories, CMS either assigns paymentrates based on their similarity to existingtests or requires carriers to independentlyset the rates for the first year of use.Carriers must research and set their ownpayment amounts. They may obtain costdata from manufacturers, receive paymentdata from other carriers, or perform theirown analyses.

Post-acute care

Many beneficiaries receive post-acutecare from one of four types of providers:

• skilled nursing facilities

• home health agencies

• inpatient rehabilitation facilities

• long-term care hospitals

Most patients use this care immediatelyfollowing an acute hospital stay.

Payment for skilled nursingfacility services Beneficiaries who need short-term skilledcare (nursing or rehabilitation services) onan inpatient basis following a hospital stayof at least three days are eligible to receivecovered services in skilled nursing


23 The labor-related portion of the rate was determined by calculating the average percentage of facility costs attributable to labor expenses for the 90 facilities includedin the 1986 cost survey. The 1994 cost survey—which has not been used to update payment rates—showed that 37.66 percent of facility costs were related to laborexpenses.

facilities (SNFs). SNFs can be hospital-based units or freestanding facilities.About 1.4 million beneficiaries use SNFcare in a year, but Medicare’s paymentsfor these services account for only about10 percent of freestanding nursingfacilities’ revenues; they make up lessthan 2 percent of hospitals’ revenues.Similarly, payments to SNFs ($15.3billion in 2001) represent only about 6.5percent of total Medicare spending.

Medicare adopted a new PPS for SNFservices on July 1, 1998. Throughout mostof the 1980s and 1990s, however, SNFswere paid on the basis of their costs,subject to limits on their per diem routinecosts (room, board, and routine nursingcare); no limits were applied for ancillaryservices (such as drugs and therapy).Under the PPS, SNFs are paid apredetermined rate for each day of care.The per diem rates are based primarily onthe patient’s service needs and marketconditions in the facility’s location.Patients are assigned to 44 groups, eachcontaining patients with similar serviceneeds who are expected to require similaramounts of resources. The daily rate foreach group is the sum of threecomponents:

• a fixed amount for routine services(such as room and board, linens, andadministrative services),

• a variable amount reflecting theintensity of nursing care patients areexpected to require, and

• a variable amount for the expectedintensity of therapy services.

The rates are computed separately forurban and rural areas, and a portion of thetotal rate is adjusted to reflect marketconditions in each SNF’s location.

The SNF PPS has problems characterizingand classifying patient days, therebyraising questions about its ability togenerate payments that accurately reflectefficient providers’ costs of furnishingcare. Partly in response to this problem,

the Congress temporarily increasedpayments to SNFs. Two of the threepayment increases expired at the end ofFY 2002.

The skilled nursing facilityproduct Medicare buysMedicare sets daily payment rates for 44resource utilization groups, version III(RUG–III), which are intended to grouppatients with similar expected serviceneeds. Patients’ expected service needsare determined by periodic assessments oftheir condition, including their needs forintensive physical, occupational, or speechtherapy; special treatments (such as tubefeeding); and their functional status (theirability to manage unassisted ordinarydaily activities, such as eating, bathing,and dressing).

Setting the payment ratesThe PPS rates are expected to cover alloperating and capital costs that efficientfacilities would be expected to incur infurnishing covered SNF services. Each ofthe 44 RUG–III groups has a daily ratecomprising a fixed routine amount plus anursing component and a therapycomponent. The nursing component iscalculated by multiplying a base rate fornursing by a national relative weight thatreflects the intensity of nursing care thatpatients in each RUG–III category areexpected to receive. For groups thatrequire intensive therapy, the therapycomponent is calculated by multiplying abase rate for therapy by a national relativeweight that reflects the expected intensityof therapy; a fixed rate is used for groupsreceiving routine therapy. Rates are setseparately for urban and rural SNFs.

The rates are adjusted to account fordifferences in input prices among localmarkets. The labor-related portion of thedaily payment rate—75 percent for FY2002—is multiplied by the hospital wageindex in the SNF’s location, and the resultis added to the nonlabor portion. Rates areupdated annually, based on the projectedincrease in the SNF market basket index,

a measure of the national average pricelevel for the goods and services SNFspurchase to provide care.

The initial payment rates in 1998 were setto reflect the projected amount that SNFsreceived in 1995, updated for inflation.24

The Congress subsequently increased thepayment rates temporarily in severalways:

• The BBRA increased rates for all 44RUG–III groups by 4 percent for carefurnished from April 2000 throughSeptember 2002.

• The BIPA increased the base rate forthe nursing component by 16.66percent for care furnished from April2001 through September 2002.

• The BBRA and BIPA increased ratesfor 14 rehabilitation groups by 6.7percent and those for 12 complexcare groups by 20 percent. Theseincreases were intended to give CMStime to refine the RUG–IIIclassification system, and they expirewhen CMS adopts that refinement.

Payment for home healthcare services Beneficiaries who are generally confinedto their homes and need skilled care (froma nurse, physical therapist, or speechtherapist) on a part-time or intermittentbasis are eligible to receive certainmedical services at home. Coveredservices, delivered by home healthagencies (HHAs) in visits to beneficiaries’homes, include:

• skilled nursing care

• physical, occupational, and speechtherapy

• medical social work

• home health aide services

Beneficiaries are not required to make anycopayments for these services.


24 By law, this projection excluded costs of SNFs that were exempt from Medicare’s routine cost limits or that had so-called atypical exceptions in 1995. The projectionincluded only 50 percent of the difference between the average costs of hospital-based and freestanding facilities.

About 2.2 million beneficiaries used homehealth care in 2001. Medicare’s paymentsto HHAs were about $9 to $10 billion25 in2000, accounting for around 6 percent oftotal Medicare spending but a large shareof HHAs’ total revenues.

Until October 2000, HHAs generally werepaid on the basis of their incurred averagecosts per visit, subject to annuallyadjusted limits. In October 2000, CMSadopted a new PPS in which HHAs arepaid a predetermined rate for each 60-dayepisode of home health care. The paymentrates are based on patients’ conditions andservice use, and they are adjusted toreflect the level of market input prices inthe geographical area where services aredelivered. If fewer than 5 visits aredelivered during a 60-day episode, theHHA is paid per visit by visit type, ratherthan by the episode payment method.Adjustments for several other specialcircumstances, such as high-cost outliers,can also modify the payment. Paymentrates also are increased for patients inrural areas.

Setting rates for Medicare home healthservices has always been complicated bythe lack of a clear definition of the benefit.The benefit was originally intended forshort-term, posthospital recovery care forbeneficiaries who could not leave theirhomes, but changes to eligibility criteriahave expanded the benefit. Beneficiarieswho have no preceding hospital stay andare capable of spending significant timeoutside their homes are now eligible toreceive covered services furnished in anunlimited number of home care episodes.

The home health productsMedicare buysMedicare purchases home health servicesin 60-day episodes. For each episode ofcare, the payment amount is intended tocover what an efficient provider wouldhave to spend in furnishing visits,supplies, outpatient therapy, and patientassessments. The severity of a patient’scondition changes the expected amount ofresources—chiefly the number and type ofvisits—required for high-quality care. To

capture differences in expected resourceuse, patients receiving 5 or more visits areassigned to 1 of 80 home health resourcegroups (HHRGs) based on diagnosis,functional capacity, and service use.

Setting the payment ratesThe HHRGs range from groups ofrelatively uncomplicated patients to thosecontaining patients who have severemedical conditions, severe functionallimitations, and a need for extensivetherapy. Each HHRG has a nationalrelative weight reflecting the averagerelative costliness of patients in that groupcompared with the average Medicarehome health patient. The payment ratesfor HHRGs in each local market aredetermined by adjusting a nationalaverage base amount—the amount thatwould be paid for a typical home healthpatient residing in an average market—toreflect the input-price level in the localmarket and then multiplying the adjustedlocal amount by the relative weight foreach HHRG.

The initial national average base paymentamount for a typical home health episodein 2001 was set so that projected spendingwould equal the amount that would havebeen spent under the previous paymentsystem. This amount was reducedbeginning in 2003 to account for certainpreviously deferred payment reductions.Further, because providers receivepayments on a per-visit basis for patientswho are furnished fewer than 5 visits in 60days, the base amount was adjusted toreflect this policy. It was also reduced 5percent to account for anticipated high-cost outlier payments. For FY 2003, thenational average payment rates forHHRGs range from $1,000 to $6,000.

To capture local market conditions, theper-episode payment rate is divided intolabor and nonlabor portions; the laborportion—77 percent—is adjusted by aversion of the hospital wage index toaccount for geographic differences in themarket prices for labor-related inputs tohome health services. For most services

provided in facilities, the location of thefacility determines the local areaadjustment that applies. For home healthservices, however, the local areaadjustment is determined by thebeneficiary’s residence. The total paymentis the sum of the adjusted labor portionand the nonlabor portion.

Payment rates are temporarily increasedby 10 percent for care delivered tobeneficiaries who live in rural areas. Thisis intended to compensate for potentiallyhigher visit costs in rural areas related tolow patient volume and long distancesbetween patients.

When a patient’s episode of care involvesan unusually large number or a costly mixof visits, the HHA may be eligible for anoutlier payment. To be eligible, imputedepisode costs must exceed the paymentrate by 13 percent or more. Episode costsare imputed by multiplying the estimatednational average per visit costs by type ofvisit—adjusted to reflect local inputprices—by the number of visits by typeduring the episode. When these estimatedcosts exceed the outlier threshold, theHHA receives a payment equal to 80percent of the difference in addition to theepisode payment.

The base rate is updated annually. Theupdate is based on the projected change inthe home health market basket index,which measures changes in the prices ofgoods and services home health agenciesmust buy to produce care.

Payment for inpatientservices in rehabilitationfacilities After an illness, injury, or surgical care,some patients need intensive inpatientrehabilitation services, such as physical,occupational, or speech therapy.Relatively few beneficiaries use intensiverehabilitation therapy because they mustbe able to tolerate and benefit from threehours of therapy per day to be eligible fortreatment in an inpatient rehabilitationsetting. Among those who qualify, manyare admitted to inpatient rehabilitation


25 Estimates vary.

facilities (IRFs), which may befreestanding hospitals or specialized,hospital-based units. Others may receivecare in a SNF, especially in markets thatlack IRFs or have few rehabilitation beds.Although payments to IRFs (about $4.2billion in 2001) represent only a small partof total Medicare spending (about 1percent), Medicare accounts for a largeshare of IRF revenues.

Until January 1, 2002, Medicare paidIRFs (under TEFRA) on the basis of theirincurred average costs per discharge,subject to annually adjusted facility-specific limits. Beginning in January2002, IRFs are paid predetermined per-discharge rates based primarily on thepatient’s condition (diagnoses, functionaland cognitive statuses, and age) andmarket conditions in the facility’slocation. Discharges are assigned to case-mix categories containing patients withsimilar clinical problems that are expectedto require similar amounts of resources.Each case-mix category has a nationalrelative weight reflecting the expectedrelative costliness of treatment for apatient in that category compared withthat for the average Medicare inpatientrehabilitation patient. The payment ratesfor case-mix categories in each localmarket are determined by adjusting anational average base payment amount toreflect the input-price level in the localmarket, and then multiplying the adjustedlocal amount by the relative weight foreach case-mix group. Payment rates alsoare increased for facilities located in ruralareas and those that treat adisproportionate share of low-incomepatients.

Defining the inpatientrehabilitation products Medicare buysUnder the inpatient rehabilitation PPS,Medicare sets payment rates for 385intensive rehabilitation products—calledcase-mix groups (CMGs)—defined bytypes of treatment episodes. Patients areassigned to 380 of these treatmentcategories based on the primary reason forintensive rehabilitation care (for example,a stroke or burn); their age and levels offunctional and cognitive impairments; and

the types of comorbidities present duringthe stay. The other five categories are forpatients discharged before the fourthday—short-stay outliers—and for thosefew who die in a facility. Further, IRFsmay receive only partial payment forother patients who do not receive a fullcourse of intensive therapy because theyare discharged to another facility and thelength of stay is less than that typicallyprovided to patients with the samecondition.

Setting the payment ratesThe PPS payment rates are intended tocover all operating and capital costs thatefficient facilities would be expected toincur in furnishing covered rehabilitationservices. The initial payment level (baserate) for a typical discharge—$12,193 forFY 2003—is intended to reflect theprojected amount providers would havebeen expected to receive per dischargeunder the previous payment system(TEFRA) in 2003. Because providers willreceive additional payments under thePPS for extraordinarily costly patients(high-cost outliers), the projected amountis reduced (3 percent) to maintain thesame expected total spending. Further,reflecting its experience with similarfinancial incentives under other discharge-based PPSs, CMS decreased the base rate(by 1.16 percent) in the expectation thatproviders would lower their costs byreducing lengths of stay compared withthose under TEFRA.

The base rate is adjusted to account fordifferences in input prices among markets.The labor-related portion of the basepayment amount—72 percent—ismultiplied by a version of the hospitalwage index, and the result is added to thenonlabor portion. The adjusted rate foreach market is multiplied by the relativeweights for all CMGs to create local PPSpayment rates. Payment rates areincreased for IRFs located in rural marketsand for those that treat low-incomepatients. Rural facilities’ payment ratesare increased by 19 percent to compensatefor their tendencies to have fewer cases,longer lengths of stay, and higher averagecosts per case. An IRF also is eligible to

receive higher payment rates if it serves atleast one low-income patient. Thepayment adjustment for each facility isbased on its low-income patient share,which is the sum of two proportions: theproportion of total inpatient daysfurnished to beneficiaries eligible forSupplemental Security Income benefitsand the proportion of total patient daysfurnished to Medicaid patients. Afteradjustments for local market conditions,rural location, and type of treatmentcategory, the CMG payment rates rangefrom $3,819 to $58,590.

Finally, IRFs receive additional paymentsfor high-cost outliers when their costsexceed a fixed-loss threshold. An IRF hasa threshold for each CMG equal to itsregular payment rate plus a national fixed-loss amount ($11,211) adjusted by thewage index for the IRF’s market. Forhigh-cost outliers, IRFs receive theirregular payment rates plus 80 percent oftheir costs above the fixed-loss threshold.

Both the base rate and relative weights areupdated annually. The base rate is updatedusing the TEFRA market basket index(used for facilities originally excludedfrom the acute care hospital PPS)expanded to reflect changes in the price ofcapital. The relative weights are updatedbased on changes in national averagecharges per discharge for each CMG.

Payment for servicesfurnished in long-term care hospitals Patients with clinically complex problems,such as multiple acute or chronicconditions, may need hospital care forrelatively extended periods of time. Someare admitted to long-term care hospitals(LTCHs). Other patients—especially inthe many markets without thesehospitals—may be cared for in acute carehospitals or SNFs. Payments to LTCHs(about $2 billion in 2001) represent only asmall part of total Medicare spending (lessthan 1 percent); however, Medicareaccounts for a substantial proportion ofthese hospitals’ revenues.

Beginning in October 2002, LTCHs arepaid predetermined per-discharge rates


based primarily on the patient’s diagnosisand market conditions in the facility’slocation.26 Before then, LTCHs were paidfor furnishing care to Medicarebeneficiaries under TEFRA.

Discharges are assigned to case-mixcategories containing patients with similarclinical problems that are expected torequire similar amounts of resources.Each case-mix category has a nationalrelative weight reflecting the expectedrelative costliness of treatment for apatient in that category compared withthat for the average Medicare LTCHpatient. The payment rates for case-mixcategories in each local market aredetermined by adjusting a nationalaverage base payment amount to reflectthe input-price level in the local market,and then multiplying the adjusted localamount by the relative weight for eachcase-mix group. Payment rates also areincreased for hospitals located in Alaskaand Hawaii and for cases that areextraordinarily costly. Payment rates areadjusted for patients who have very shortstays and for those who are transferred toan acute care hospital, an inpatientrehabilitation facility, or a skilled nursingfacility for a specified amount of time,followed by readmission to the sameLTCH.

Defining the long-term carehospital products Medicare buysUnder the PPS for care in LTCHs,Medicare sets payment rates for 499 typesof treatment episodes. These episodes arecalled long-term care diagnosis relatedgroups (LTC–DRGs). The groupingsystem for episodes is the same one usedfor the acute care hospital PPS. Patientsare assigned to these treatment categoriesbased on the discharge diagnosis,including the principal diagnosis; up toeight secondary diagnoses; up to sixprocedures performed; age; sex; and

discharge status. LTCHs may receivepartial payments for patients who do notreceive a full course of treatment.

Setting the payment ratesThe PPS payment rates are intended tocover all operating and capital costs thatefficient LTCHs would be expected toincur in furnishing covered acute long-term care services. The initial paymentlevel (base rate) for a typical discharge—$34,956 for FY 2003—is intended toreflect the projected amount providerswould have been expected to receive perdischarge under the previous paymentsystem in FY 2003. Because providerswill receive additional payments under thePPS for extraordinarily costly patients(high-cost outliers), the projected amountis reduced (8 percent) to maintain thesame expected total spending. Further,reflecting its experience with similarfinancial incentives under other discharge-based PPSs, CMS decreased the base rate(by 0.34 percent) in the expectation thatproviders would lower their costs byreducing lengths of stay compared withthose under the old payment system.

The base rate is adjusted to account fordifferences in input prices among markets.This adjustment is being phased in overfive years. The labor-related portion of thebase payment amount—73 percent—ismultiplied by a version of the hospitalwage index and the result is added to thenonlabor portion.27 For LTCHs in Alaskaand Hawaii, the nonlabor portion isadjusted by a COLA and added to thelabor-related portion.28 The adjusted ratefor each market is multiplied by therelative weights for all LTC–DRGs tocreate local PPS payment rates.

Relative weights for the LTC–DRGsdiffer from the acute care hospital DRGweights. Medicare assigns a weight toeach LTC–DRG reflecting the average

relative costliness of cases in the groupcompared with that for the averageMedicare case. LTC–DRGs with fewerthan 25 cases in 2001 have been groupedinto 5 categories based on their averagecharges; relative weights for these 5 case-mix groups have been determined basedon the average charges for theLTC–DRGs in each of these 5 groups.

LTCHs are paid adjusted PPS rates forpatients who do not receive a full courseof treatment. Short-stay outliers aredefined as cases with a length of stay upto and including five-sixths of thegeometric average length of stay for theLTC–DRG. For short-stay outliers,LTCHs are paid the least of:

• 120 percent of the cost of the case,

• 120 percent of the LTC–DRGspecific per diem amount multipliedby the length of stay for that case, or

• the full LTC–DRG payment.

LTCHs are paid adjusted PPS rates forpatients who are extraordinarily costly.High-cost outlier cases are identified bycomparing their costs to a LTC–DRG-specific threshold that reflects the DRGpayment for the case plus a fixed lossamount. For example, in 2003 thethreshold is set at the LTC–DRG paymentplus $24,450—the national fixed lossamount—adjusted to reflect the inputprice levels in the local market. Medicarepays 80 percent of the LTCHs’ costsabove their fixed loss thresholds. High-cost outlier payments are funded byoffsetting reductions in the base paymentamount (8 percent).

LTCHs receive one payment for patientswho are transferred from the LTCH toanother facility for a specified period oftime and return to the LTCH—so-called“interrupted stays.” Interrupted stays aredefined as those cases in which an LTCH


26 LTCHs began receiving payments under the new PPS at the beginning of their FY 2003 cost reporting periods. During a five-year transition period, they are paid ablend of the PPS rate and their updated facility-specific rate. For example, in the first year of PPS, payments will be made up of 20 percent PPS rates and 80 percentfacility-specific rates; in the second year, payments will be made up of 40 percent PPS rates and 60 percent facility-specific rates.

27 The wage index used to adjust LTCH payments is calculated from wage data reported by acute care hospitals without the effects of geographic reclassification.

28 The COLA reflects the higher costs of supplies and other nonlabor resources in Alaska and Hawaii; it increases the nonlabor portion of the payment by as much as 25percent.

patient is discharged to an inpatient acutecare hospital, an IRF, or a SNF for aspecified period followed by readmissionto the same LTCH. The specified periodof time for an interrupted stay is 9 dayswhen the patient is discharged to an acutecare hospital, 27 days for discharge to anIRF, and 45 days for discharge to a SNF.PPS payment is based on the LTC–DRGassigned to the case at discharge to theother facility.

Finally, Medicare has established policiesto discourage transfers between LTCHsand other providers, followed byreadmissions to the LTCH, when theLTCH and any of the other providers arelocated in the same facility or on the samecampus (colocated). Medicare’s concernabout such transfers is that they may occuras a result of financial instead of clinicalconsiderations. Within a cost reportingperiod, Medicare treats transfers to co-located acute care hospitals followed byreadmissions to the same LTCHs above athreshold of 5 percent of all cases as ifthey were one LTCH discharge forpayment purposes. Until the threshold isexceeded, Medicare treats each case as adischarge. A separate 5 percent thresholdapplies to cases transferred to colocatedSNFs, IRFs, and psychiatric facilities.

Services for specialpopulations

Many Medicare beneficiaries have specialneeds resulting from end-stage renaldisease (ESRD) or a terminal illness.These beneficiaries may receive servicesin two specialized settings:

• outpatient dialysis facilities

• hospices

Payment for outpatientdialysis services Individuals with ESRD—irreversible lossof kidney function—require either dialysisor kidney transplantation to survive. In1972, the Social Security Act extended allMedicare Part A and Part B benefits toindividuals with ESRD who are entitled to

receive Social Security benefits. Thisentitlement is nearly universal, covering93 percent of all people with ESRD in theUnited States. Total Medicare spendingfor these beneficiaries has outstrippedexpectations—reaching about $15 billionin 2001—primarily because ofunanticipated growth in the ESRDpopulation. The 350,000 enrolled ESRDbeneficiaries in 2001 accounted for 0.8percent of total Medicare enrollment,compared with only 0.1 percent ofenrollment in 1974. This enrollmentgrowth reflects population aging andimprovements in clinical knowledge andtechnique that have enabled successfultreatment of older patients and those withcoexisting illnesses who might not havebeen treated 30 years ago.

Because of the scarcity of kidneysavailable for transplantation, most peoplewith ESRD receive dialysis treatmentsthree times per week in either freestandingor hospital-based facilities. Medicarespending for outpatient dialysis andinjectable drugs administered duringdialysis (about $6.7 billion in 2001)accounts for 2 percent of total programexpenditures but is a predominant share ofrevenues for dialysis facilities. Medicarepays dialysis facilities a predeterminedamount for each dialysis treatment theyfurnish, using a payment system firstimplemented in 1983. The prospectivepayment—called the composite rate—isintended to cover the bundle of services,tests, drugs, and supplies routinelyrequired for dialysis treatment and isadjusted only to account for differences inlocal input prices.

Even though technological advances havechanged the provision of dialysis caresince the composite rate was established,CMS has not modified the unit ofpayment. Although CMS has occasionallychanged the dialysis bundle, it has notused explicit criteria to determine whichservices should be included.Consequently, the composite ratecurrently excludes several new injectabledrugs and clinical laboratory tests thathave diffused widely into medical practiceover the past decade; providers are paid

for these services based on their incurredcosts. The BIPA required the Secretary ofHHS to:

• include in the composite rate by July2002 diagnostic laboratory tests anddrugs that were routinely used infurnishing dialysis care but that werebeing billed separately.

• recommend to the Congress in astudy whether the composite rateshould be updated annually orperiodically.

A draft of this study is currently beingreviewed within CMS.

Defining the dialysis productsMedicare buysMedicare covers two methods ofdialysis—hemodialysis and peritonealdialysis. In hemodialysis, a patient’s bloodis cycled through a dialysis machine,which filters out body waste. About 90percent of all dialysis patients undergohemodialysis three times per week indialysis facilities. Peritoneal dialysis usesthe membrane lining the peritoneal cavityto filter excess waste products, which arethen drained from the abdomen. Patientsundergo peritoneal dialysis five to seventimes per week in their homes.

The unit of payment is the dialysistreatment. The composite rate paymentsystem differs from Medicare’s otherprospective payment systems because ituses only one product category to definethe service bundle Medicare is buying.Although different equipment, supplies,and labor are needed for hemodialysis andperitoneal dialysis, the current systemdoes not differentiate payment based ondialysis method.

Providers may separately bill Medicarefor certain injectable medications,including erythropoietin and vitamin Danalogues, and laboratory tests that are notincluded in the composite rate bundle. TheCongress has set the payment forerythropoietin at $10 per 1,000 unitswhether it is administered intravenouslyor subcutaneously in dialysis facilities orin patients’ homes. Providers receive 95


percent of the average wholesale price forseparately billable injectable medicationsother than erythropoietin administeredduring in-center treatments. Finally,providers furnishing laboratory servicesoutside the composite rate bundle are paidaccording to the laboratory fee schedule.

Setting the payment ratesThe composite rate is intended to cover alloperating and capital costs that efficientproviders would incur in furnishingdialysis treatment episodes in dialysisfacilities or in patients’ homes. The basepayment rate is $131 for hospital-basedfacilities and $127 for freestandingfacilities in 2002. Medicare caps itspayments to facilities at an amount equalto three dialysis sessions per week,although dialysis may be given morefrequently.

The labor-related portion of the compositerate—40 percent in 2002—is adjusted forlocal market differences in input pricesusing a wage index created in 1987. Thiswage index blends 60 percent of a wageindex based on 1980 Bureau of LaborStatistics hospital wage data with 40percent of the fiscal year 1986 PPShospital wage index. Both componentwage indexes use labor markets based on1980 definitions for MSAs and statewiderural areas. The blended wage index islimited by a floor and a ceiling; areas thathave blended index values lower than 90percent of the national average are raisedto the 90 percent level (the wage index“floor”), while those with blended indexvalues higher than 130 percent of thenational average are lowered to the 130percent level (the “ceiling”). Thus, theminimum payment is $121 and themaximum is $144 per dialysis treatmentin 2002.

A dialysis facility may apply for anexception to its composite rate whendialysis costs exceed the base paymentrate. The four circumstances that mayjustify a payment exception are: (1)serving an atypical patient mix, (2)furnishing services to patients who areusing fewer than three dialysis sessionsper week, (3) serving an isolated area in

which the facility is essential to ensurebeneficiaries’ access to care, or (4)extraordinary circumstances, such asfurnishing dialysis in an area affected bynatural disaster.

Dialysis facilities are reimbursed for baddebt that results when, after a good faitheffort, they are unable to collectbeneficiaries’ 20 percent coinsuranceamounts for dialysis services.

Payment for hospiceservicesTerminally ill beneficiaries (certified tohave a projected life expectancy of sixmonths or less) may elect to receivehospice care, which aims to help thesepatients continue to live as normally aspossible and remain in their homes.Therefore, the hospice benefit covers awide array of services, including:

• physician services.

• skilled nursing services.

• counseling (dietary, spiritual,bereavement, and other counselingservices).

• medical social services.

• drugs and biologicals for pain controland symptom management.

• physical, occupational, and speechtherapy.

• home health aide and homemakerservices.

• inpatient respite care.

To be eligible for hospice services,beneficiaries must give up other coveredservices related to curative treatment ofthe terminal condition, although Medicarestill pays for unrelated care. Twentypercent of Medicare beneficiaries whodied in 1998 used hospice care. Paymentsto hospices (almost $3.4 billion in 2001)represent a small part of total Medicarespending (about 1 percent), althoughMedicare makes up a large share ofhospice revenues.

Medicare pays hospices for each day abeneficiary is eligible and under hospicecare, regardless of the amount of servicesfurnished on any given day. Per diempayment rates are based on a fee schedulewith separate rates for four broadcategories of care. The rate for each day isadjusted to reflect local market conditions.

Defining the hospice productsMedicare buys and settingpayment ratesFor hospice services, Medicare setspredetermined daily payment ratesaccording to a fee schedule for four broadcategories of care: routine home care,continuous home care, inpatient respitecare, or general inpatient care. Patients areassigned to these categories based on thetype of care they actually receive eachday.

The daily payment rates representpayment in full for all costs that hospicesincur in furnishing services identified inpatients’ care plans. The initial paymentlevel (base rate) per category is adjustedto account for differences in wage ratesamong markets. The labor-related portionof the base payment amount—69 percentfor routine and continuous home care, and54 percent and 64 percent for inpatientrespite care and general inpatient care,respectively—is adjusted by the hospicewage index for the location in which careis furnished, and the result is added to thenonlabor portion. The base rates areupdated annually by the projected increasein the acute care hospital MB index.

A hospice’s annual aggregate paymentsare limited by a capped amount ($17,391for FY 2003) multiplied by the number ofbeneficiaries newly enrolled during theyear. The capped amount is updatedannually by the consumer price index forall urban consumers, U.S. city average(CPI–U).

Other services

Medicare also pays for other services andproducts used by beneficiaries in thetraditional fee-for-service program,including:


• ambulance services.

• durable medical equipment.

Payment for ambulanceservicesMedicare pays for both emergency andnonemergency ambulance services,including ground, water, and air serviceswhen the use of other means oftransportation to health care serviceswould be harmful to beneficiaries’ health.Ambulance staff provide a range ofservices to stabilize and treat patients intransit.

Ambulance providers are either facilitybased (hospital, skilled nursing facility, orhome health agency) or freestandingsuppliers, a distinction integral to past andcurrent payment. Until April 1, 2002,Medicare based payments for ambulanceservices on providers’ reported costs(facility-based providers) and charges(freestanding providers). Providers werepaid a base rate, which covered the costsof services and supplies, and a mileagepayment. This approach provided fewincentives for cost containment and oftenresulted in payment and coveragedisparities among similar providers. Theseissues, together with increasedexpenditures, led the Congress to mandatein the BBA that CMS implement aprospective fee schedule.

Several issues delayed fee scheduleadoption, including how to adjust forhigher costs incurred by low-volumeproviders, how to ensure that aggregatepayments to ambulances were notreduced, and whether to require additionalcoding to document the medical necessityof services. After a fairly extensive rule-making period, including the formationand guidance of an advisory committeefor CMS, the final rule was issued inFebruary 2002 for implementationeffective April 1, 2002.

Defining the ambulance productMedicare buysUnder the new prospective fee schedule,14 HCPCS codes are used to distinguishthe level of services provided, supplies

and equipment used, and mileage.Ambulance suppliers may also bill twoCPT codes for electrocardiograms.Ambulance-administered drugs areconsidered supplies and are notreimbursed separately. Payments arereduced when a beneficiary dies beforethe ambulance arrives at the scene.

CMS adopted nine transport service levelsfrom the National Emergency MedicalServices Training Blueprint as revised bythe Department of Transportation. Whendifferent from the Blueprint, state andlocal laws preempt the Blueprint forvehicle staffing and clinical certificationrequirements. CMS assigned relativevalue units for seven of the service levels(ground only) through negotiated rule-making with the advisory committee.

Setting the payment ratesThe new fee schedule establishes paymentamounts that, for ground or waterservices, are the product of a nationallyuniform relative value for the service, ageographic area adjustment factor, and anationally uniform conversion factor. Theconversion factor is based on fourestimates for 2002 through 2006:spending levels (both program andbeneficiary), inflation, the mix of servicelevels performed, and the increase inMedicare enrollment. If these behavioraland other assumptions prove different,CMS will adjust the conversion factorprospectively, in order to keep the totalamount of payments in the system equalto the level prior to fee scheduleimplementation. For air services, the basepayment is the product of an unadjustednationally uniform value for the serviceand a geographic adjustment; there is noconversion factor or relative value unit.

The geographic adjustment accounts forvarying costs of conducting business indifferent regions of the country, and isequal to the geographic practice expenseindex for the Medicare physician feeschedule. The geographic areas are thoseused for the physician fee schedule,selected by location of the patient whenput on the ambulance. The geographicindex applies to 70 percent of the base rate

for ground services and 50 percent of thebase rate for air services; it does not applyto the mileage payment rate.

A separately calculated payment is madefor mileage to account for costsattributable to use of the ambulancevehicle. To reflect cost differences,mileage rates vary between ground and airtransport and also distinguish betweenfixed wing and rotary wing (helicopter)transport. For rural ground trips, CMSprovides a 50 percent add-on to themileage rate for the first 17 miles and a 25percent add-on for miles 18 through 50, asestablished in BIPA. Rural air tripsreceive a 50 percent add-on to the baserate and to all of the miles from the time apatient is placed on board.

The ambulance fee schedule will bephased in through a five-year transitionperiod of blended payments. For Aprilthrough December 2002, providers werepaid 80 percent by the former method and20 percent according to the new feeschedule. For 2003, the percentage haschanged to 60 percent by the formermethod and 40 percent according to thenew fee schedule. By 2006 payments willbe 100 percent according to the new feeschedule.

The conversion factor, the air ambulancebase rates, and the mileage rates will beupdated annually based on the rise inCPI–U. However, during the rule-makingprocess, BIPA mandated an update of 4.7percent for services furnished betweenJuly 1, 2001, and December 31, 2001, 2percentage points higher than the CPI–U.

The fee schedule applies to all entitiesproviding services and they must acceptthe fee schedule amount as payment infull. Critical access hospitals that have noother ambulance service provider orsupplier within a 35-mile driving distanceare the sole exception; they receive cost-based reimbursement.

Payment for durablemedical equipment When medical equipment is needed totreat a beneficiary’s illness or injury at


home, it is covered under the durablemedical equipment (DME) benefit.Medicare spent about $7 billion on DMEin 2001, about 3 percent of fee-for-serviceprogram spending.

Wheelchairs and respirators are typical ofthe equipment Medicare pays for underthis benefit. To be covered, the equipmentmust:

• withstand repeated use.

• serve a primarily medical purpose.

• generally not be useful to a personwithout an illness or injury.

Thus, disposable supplies such asbandages or incontinence pads, orotherwise useful equipment such as ahumidifier, would not be covered underthis benefit.

Medicare also covers prosthetics,orthotics, and some medications under itsDME benefit. Covered prostheticsgenerally are artificial limbs; orthoticsinclude orthopedic braces and somesupportive garments. Medication that isnecessary to the function performed bydurable equipment is also covered underthis benefit—for example, heparinadministered in a home dialysis system,albuterol in a nebulizer, or chemotherapydrugs in an infusion pump.

Medicare has paid DME suppliers using afee schedule since 1986. Under the feeschedule, covered items are classified intoproduct groups within six major classes.The payment amount for each productgroup is a weighted average of local andregional prices, updated annually by theCPI–U. Suppliers are generally paid eithera monthly rate for rentals or a lump sumfor purchased items. Medicare also coversthe cost of repairs, maintenance, delivery,and supplies necessary to use purchasedequipment. Beneficiaries are responsiblefor a 20 percent copayment.

The durable medical equipmentMedicare buysDME payments include a monthly rentalfee or a lump-sum purchase fee. Under the

DME fee schedule, Medicare sets pricesfor equipment by category and productgroup. Equipment is assigned to one of sixcategories based on its nature—whether ornot it is inexpensive, needs frequentservice, or is a rental item subject to anexplicitly limited period of use. The sixDME categories are:

• inexpensive or routinely purchasedequipment.

• items requiring frequent andsubstantial servicing.

• customized items.

• prosthetic and orthotic devices.

• capped rental items.

• oxygen and oxygen equipment.

Within the 6 categories, equipment isfurther categorized into about 2,000product groups. Examples of productgroups are high-strength, lightweightwheelchairs and rental portable oxygensystems. All items within the sameproduct group have the same paymentrate.

The central issue in DME payment policyis the frequent failure of Medicare’spayments to reflect current market prices.It is difficult for CMS to price DME in away that is consistent with the marketbecause the product definitions are toobroad. Each product code has only onepayment rate, but one product code can beused for many different items withvarying prices in the retail market. Also,changing Medicare’s payment rates in anyway other than simple updating has beencumbersome.

The BBA gave Medicare the authority toapply a so-called test of inherentreasonability to some items that havewell-developed retail markets; this allowsCMS some price-setting flexibility. CMShas also conducted a competitive biddingdemonstration to test the effects ofcompetition on prices for certain DMEitems. In three phases of thedemonstration, competitive bidding

lowered prices for selected DME items 17percent, 21 percent, and 22 percent.

Setting the payment ratesTo ensure beneficiaries’ access to neededDME, the fee schedule must coverefficient suppliers’ costs of furnishingequipment for rental or purchase.Generally, the current fees are an averageof the allowed charges from 1986 and1987, adjusted by the CPI–U to accountfor inflation.

Over time, the inflation-adjusted priceshave failed to reflect changes in medicalequipment technology and other factorsthat have caused market retail prices todiverge from Medicare’s payment rates.Recent legislation established twoalternatives to the inflation adjustment.One is that Medicare can adjust prices byas much as 15 percent in one year forDME that is frequently purchased byother payers. To make the priceadjustment, CMS would use an inherentreasonableness test based on a survey ofmarket prices. The other is that Medicarecan freeze some prices or put a limit onthe amount of the annual increase.

Medicare uses different methods amongthe six broad equipment categories forcapturing variations in prices due to localmarket conditions. In some instances,Medicare sets a separate fee schedule foreach state based on local allowed chargesin 1986–87. In other cases, Medicare uses10 regional fee schedules in which theprices in each region are based on anaverage of allowed charges in theconstituent states. Both the state andregional schedules are subject to floorsand ceilings to limit the variability inprices across the country. A third methodis an item-by-item determination by thecarrier. Rental payments are subject to anational payment limit. The applicable feeschedule is determined by the location ofbeneficiaries’ residences rather than thelocation of the DME provider. Allprogram payments are reduced by the 20percent coinsurance paid by beneficiaries.


Medicare�Choice plans

Medicare beneficiaries may choose toreceive their Medicare benefits from aprivate plan participating in theMedicare�Choice program rather thanfrom the traditional program. Under someM�C plans, beneficiaries may receivebenefits beyond those offered undertraditional Medicare and may payadditional premiums. Medicare pays plansa capitated rate for the 12 percent ofbeneficiaries currently enrolled. Thesepayments amounted to $37 billion in2002, 15 percent of total Medicarespending.

Medicare payment rates for M�C plansare based on enrolled beneficiaries’characteristics and the counties in whichthey live. Medicare uses beneficiaries’characteristics—primarily age and sex—to develop a measure of their expectedrelative risk for covered health spending.The payment rate for a plan enrolling abeneficiary is then calculated using thebase rate for the beneficiary’s county ofresidence, adjusted for the beneficiary’sexpected relative health risk. The base ratefor each county is based on its historicaverage per capita spending in thetraditional Medicare program, local levelsof input prices, and the health riskcharacteristics of its Medicare population.

In response to concerns that plans couldnot survive in areas with low paymentrates (because of historically low percapita Medicare spending), the Congressset floors to raise the lowest rates.

Many analysts have been concerned thatthe current risk adjusters, based mostly ondemographic variables, do not account forpredictable differences in spending forcovered services among beneficiaries.More accurate risk adjusters are beingphased in.

Defining theMedicare�Choice productsMedicare buysUnder the M�C program, Medicare buysmonthly insurance coverage for its

beneficiaries from private plans. Thecoverage must include all Medicarebenefits, except that plans may limitenrollees’ choices of providers morenarrowly than under the traditional fee-for-service program.

Medicare’s payment rates for a month ofcoverage are based on beneficiaries’counties of residence and on their relativeexpected cost, as predicted bydemographic and diagnostic healthfactors. The county-level rates aredetermined administratively, based onstatutory formulas. The 2003 rate for acounty is the highest of three values:

• a floor rate of $548 for counties inmetropolitan areas with 250,000 ormore people, or $495 for all othercounties;

• the county’s 2001 rate increased by 2percent; or

• a 50/50 blend of an input price-adjusted national average rate and anupdated historical rate based on thecounty’s 1997 payment rate.

All blended rates are adjusted by a budgetneutrality factor that constrains nationalpayments. For 2003, budget neutralitycould not be achieved; thus, the blendedrates were not applicable.

Medicare currently calculates abeneficiary’s relative expected cost—ascompared with the average expected costfor all Medicare beneficiaries—based onseven factors:

• age,

• sex,

• whether the beneficiary has ESRD,

• whether the beneficiary is alsocovered by Medicaid,

• whether the beneficiary isinstitutionalized,

• whether the beneficiary (or spouse) iscurrently covered as an active workerunder an employer-sponsored plan,and

• a health risk factor currently based ondiagnoses assigned when thebeneficiary used certain Medicare-covered services during the precedingyear.

Setting the payment ratesThe original theory behind settingpayment rates for private plans was thatthe rates should be based on how much itwould cost the traditional Medicareprogram to provide coverage for thosewho enrolled in the plans. Before theBBA, rates were set at 95 percent of theexpected cost of providing coverage underthe traditional Medicare program.Medicare would thus save 5 percent of theexpected spending on behalf of abeneficiary when the beneficiary enrolledin a private plan.

The theory raised several concerns inpractice, however. Beneficiaries’ spendingin the traditional Medicare program variessubstantially across counties; per capitaspending in the highest county was three-and-a-half times that for the lowestcounty. Therefore, the payment rates forprivate plans were three-and-a-half timeshigher in some counties than in others. Asa result of low payment rates and otherfactors, few beneficiaries in lower-spending areas had private plans availableto them, while most beneficiaries inhigher-spending counties had plans withextra benefits available. The BBAchanged the rate-setting to the approachdescribed earlier in an effort to reduce ratevariation across the country and enticeprivate plans into serving more counties.

The three county rates are updatedannually. The floor rates are updated bythe national average growth in per capitaspending in the traditional Medicareprogram. The county’s prior-year rates areincreased by 2 percent, thus serving as aminimum update of 2 percent. Finally, theblended rates are recalculated andadjusted by a percentage constrained bybudget neutrality. In most years, theblended rates have not been applicablebecause of the budget-neutralityconstraint. �


An introduction to how Medicaremakes coverage decisions

A P P E N D I X B

particular service for a beneficiary. Inaddition to developing coverage decisionsthrough Medicare’s contractors, CMSimplements policies through the national

A P P E N D I X

An introduction to howMedicare makes coveragedecisions

BMedicare covers items and services thatare included in a Medicare benefitcategory, are not statutorily excluded, andare reasonable and necessary based onsection 1862(a)(1)(A) of the SocialSecurity Act. Although the statute setsforth the broad categories of benefitscovered by Medicare, neither the statutenor regulations provide an all-inclusivelist of the specific items and services thatare reasonable and necessary forbeneficiaries’ medical care. The Centersfor Medicare & Medicaid Services (CMS)and the contractors who review, process,and adjudicate Medicare claims—including the fiscal intermediaries (FIs)for Part A services, carriers for certainPart B services, and durable medicalequipment regional contractors(DMERCs)—determine whether servicesare reasonable and necessary, and,therefore, covered under Medicare.

There are several ways for services to becovered under Medicare. The vastmajority of explicit coverage decisions aredeveloped by Medicare’s contractors.These decisions, referred to as localmedical review policies (LMRPs), applyonly to specific services provided in thecontractor’s regional jurisdiction.Contractors also can make individualdecisions about the coverage of a


coverage decision (NCD) process. NCDsare national policies on the coverage ofspecific medical services. Both the localand the national coverage processesexplicitly consider whether services meetMedicare’s statutory requirements for“reasonable and necessary” care.

The NCD and LMRP processes are notthe only means by which Medicare candevelop and implement coverage policies.Policies affecting the coverage of servicesare also published in Medicare’s providermanuals and program memorandums.These policies are developed by CMS;like NCDs, they are binding for allcontractors and apply nationwide. Finally,Medicare’s coding requirements may alsoimplicitly affect the coverage of services.

It is worth noting that the majority ofservices—including those that fall into anexisting payment method or category—donot go through Medicare’s explicitcoverage process. Rather, these servicesare paid through CMS’s prospectivepayment mechanisms. Under Medicare’sprospective payment systems (PPSs),providers serve as the purchaser and makedecisions about which items and serviceswill be furnished in the payment bundle.Broader payment bundles, such as thediagnosis-related groups in the hospital

Statutory limits onMedicare coverage

Title VIII of the Social SecurityAct authorizes Medicarebeneficiaries to obtain healthservices from any institution,agency, or person qualified toparticipate in the Medicareprogram. The statute listscategories of items and serviceseligible for Medicare coverage andspecifies that no payment may bemade for services that are not“reasonable and necessary for thediagnosis or treatment of illness orinjury or to improve thefunctioning of a malformed bodymember” (Social Security Act,Title XVIII, Section 1862(a)(1)(A)). In recent years, Medicare hasalso been statutorily authorized tocover certain preventiveservices—mainly diseasescreenings—through statute. �

inpatient PPS, provide more leeway forproviders to furnish services of theirchoice compared with narrower paymentbundles, such as the ambulatory paymentclassification groups in the hospitaloutpatient PPS. As discussed in Chapter 4,both the hospital inpatient and outpatientPPSs provide additional payment forcertain new technologies.

This appendix summarizes the process bywhich coverage decisions are made in theMedicare program. First, we describe theprocess by which NCDs are made. Thenwe summarize the local coverage decisionmaking process and assess some of thesimilarities and differences between thenational and local coverage decisionmaking process. In the next two sections,we describe examples of coverage policiesmade in CMS’s provider manuals andexplain how Medicare’s coding processmay affect the coverage of new services.Lastly, we describe the current process bywhich coverage decisions can be appealedand the changes to the appeal processmandated by the Medicare, Medicaid, andSCHIP Benefits Improvement andProtection Act of 2000 (BIPA).

The national coveragedetermination process

The NCD process, administered by CMSstaff in the agency’s national office inBaltimore, is usually reserved for thoseitems or services that have the potential toaffect a large number of beneficiaries andthat have the greatest impact on Medicare

(National Health Policy Forum 2001).NCDs cannot vary from region to regionbecause all contractors andMedicare�Choice (M�C) plans arerequired to follow NCDs. The NCDprocess is initiated when CMS receives aformal request from the public. Inaddition, CMS staff can initiate theprocess if they find that: (1) inconsistentlocal coverage policies exist; (2) theservice represents a significant medicaladvance, and no similar service iscurrently covered by Medicare; (3) theservice is the subject of substantialcontroversy; or (4) the potential for rapiddiffusion or overuse exists.

The NCD process is initiated lessfrequently than the local medical reviewprocess. Over the past 30 years, CMS hasmade about 300 national coveragedecisions. By contrast, Medicare’scontractors have made about 9,000 localcoverage decisions during the past decade(Davison 2002). CMS makes relativelyfew NCDs because:

• Most decisions to cover services arenot controversial.

• Most services do not meet the criteria(listed previously) for CMS to initiatean NCD.

• Limited resources may affect CMS’sability to initiate more NCDs.

• Manufacturers and providers of amedical service may be apprehensiveabout requesting an NCD becausethey perceive that the decision could

result in an “all or nothing” scenarioin terms of their ability to obtainMedicare reimbursement.

A negative NCD can be especiallyproblematic for providers of a service forwhich Medicare constitutes a large shareof the market. However, NCDs aresometimes written for a specific clinicalindication of an item or service and can bemodified once new clinical information isavailable. For example, CMS implementedan NCD in 1991 to cover the implantationof an automatic defibrillator for patientswith a documented episode of life-threatening ventricular tachyarrhythmia orcardiac arrest not associated withmyocardial infarction. In 1999, CMSmodified the NCD to include threeadditional clinical indications (CMS1999).1

CMS uses an evidence-based approach toevaluate items and services for coverage.This approach is based on applying thebest available medical evidence accordingto the generally accepted hierarchy ofevidence.2 CMS refers most NCDrequests to outside impartial groups tosupplement the agency’s scientific andmedical expertise. One such expertgroup—the Medicare Coverage AdvisoryCommittee (MCAC)—was chartered bythe Secretary in 1998 to supplement theagency’s clinical expertise and allow forpublic input and participation. TheMCAC, which consists of six medicalspecialty panels and an ExecutiveCommittee, gives CMS its opinion onwhether a specific item or service meetsthe criteria for Medicare coverage.3 The

246 An i n t r odu c t i o n t o how Med i ca r e make s c o v e r age de c i s i o n s

1 The three additional indications are: (1) a documented episode of cardiac arrest due to ventricular fibrillation not due to a transient or reversible cause; (2) ventriculartachyarrhythmia, either spontaneous or induced, not due to a transient or reversible cause; or (3) familial or inherited conditions with a high risk or life-threateningventricular tachyarrythmias such as hypertrophic cardiomyopathy.

2 In reviewing coverage, CMS weighs the medical and scientific evidence in accordance with a fairly standardized hierarchy that ranks the relative authority given tovarious types of studies. This hierarchy of evidence is as follows, ranked with the most authoritative first:(1) Controlled clinical trials published in peer-reviewed medical or scientific journals;(2) Controlled clinical trials completed and accepted for publication in peer-reviewed medical or scientific journals;(3) Assessments initiated by CMS;(4) Evaluations or studies initiated by Medicare contractors; and(5) Case studies published in peer-reviewed medical or scientific journals that present treatment protocols.

3 The six specialty panels are: medical and surgical procedures; drugs, biologics and therapeutics; medical devices; durable medical equipment; laboratory anddiagnostic services; and diagnostic imaging. An Executive Committee—including the chair and vice chairs of each of these committees, a representative at-large, twoindustry representatives, and two consumer representatives—tries to ensure that consistent standards for decision-making are applied across the panels. An issue is firstreviewed and discussed by one of the specialty panels, which develops specific recommendations. The recommendations are then forwarded to the Executive Committeefor review and the preparation of a final recommendation to CMS.

MCAC serves only an advisory role; allfinal decisions are made by CMS. Theagency uses other outside groups,including the Agency for HealthcareResearch and Quality, to performtechnology assessments—independent,systematic analyses of the safety andeffectiveness of medical services.

The process of making most NCDs isrelatively lengthy because of the manysteps involved, which often includeconvening the MCAC and conducting atechnology assessment. For the 10 NCDsmade in fiscal year 2001, the average timefrom the date of the decisionmemorandum (announcing CMS’s intentto implement a decision) to the date ofimplementation was 156 days (Thompson2002).4 Six of the 10 decisions exceededCMS’s self-imposed time frame of 180 to270 days.

National coverage policies are publishedin Medicare’s coverage issues manual. Inaddition, information about both nationaland local coverage decisions is availablethrough the Internet. CMS’s websiteincludes current information about NCDsbeing developed as well as those that havebeen decided and implemented.5 CMS’swebsite also provides a mechanism tosearch through national and localcoverage policies, as well as supplyinglinks to contractors’ websites which postdraft and final LMRPs.

Local medical reviewprocess

Medicare’s contractors are tasked withreviewing claims for services furnished byproviders, physicians, and suppliers andpaying only for those services that meetMedicare’s coverage requirements.

Consequently, contractors play animportant role in protecting the integrityof the Medicare program. LMRPs areadministrative and educational tools toassist providers in submitting correctclaims for payment. They may containinstructions about any or all of thefollowing types of provisions: coding,benefit category, statutory exclusion, ormedical necessity.

LMRPs are developed by eachcontractors’ medical director. Thesepolicies outline how contractors willreview claims to ensure that they meetMedicare coverage requirements. Eachmedical director evaluates the medicalnecessity and reasonableness of servicesfurnished to beneficiaries by providerswithin the contractor’s jurisdiction.Circumstances for which medicaldirectors may develop new or revisedLMRP include:

• certain services demonstrating asignificant risk to the Medicare trustfund, as identified by potentially highcost or high volume of services;

• need for developing uniform LMRPsacross the contractor’s multiplejurisdictions; and

• frequent denials being issued oranticipated for an item or service.

LMRPs must be consistent with nationalguidance that includes decisions andpolicies made through the NCD process orpublished in CMS’s provider manuals orprogram memorandums. Contractors candevelop LMRPs for services not coveredby national guidance. In addition, LMRPscan provide more specific informationabout an NCD. For example, severalcontractors have issued LMRPs about theuse of intravenous iron therapy furnished

to end-stage renal disease (ESRD) patientsto treat iron deficiency anemia.6 TheseLMRPs provide specific instructionsabout the intravenous iron therapy NCDimplemented by CMS in December 2000.Finally, the existence of one or moreLMRPs does not preclude CMS frommaking an NCD. As noted in the previoussection, CMS may consider making anNCD because of varying LMRPs.

The process for developing a LMRPincludes drafting language based on areview of medical literature and thecontractor’s understanding of localpractices. LMRPs must consider and bebased on the strongest evidence available(HCFA 2000). Contractors are required topermit interested parties to submitscientific, evidence-based information andhave open meetings for the purpose ofdiscussing draft LMRPs. Carriers mustestablish carrier advisory committees(CACs) in each state, which provide aforum for information exchange betweencarriers and physicians. CACs meet atleast three times per year and arecomposed of physicians, a beneficiaryrepresentative, and representatives fromother medical organizations (CMS 2002a).

In contrast to NCDs, LMRPs apply onlyin the contractor’s jurisdiction.Consequently, coverage policies varyacross localities because contractors caneach set policies within their specificgeographic jurisdiction. CMS encouragescontractors who operate in two or morestates to develop uniform local coveragepolicies across all jurisdictions to theextent possible. In addition, medicaldirectors from the carriers and FIsparticipate in work groups for specificclinical areas, such as chronic painmanagement, anesthesiology, and clinical


4 The 10 coverage determinations were for: (1) intestinal and multivisceral transplantation; (2) biofeedback for the treatment of urinary incontinence; (3) pelvic floorelectrical stimulation for the treatment of urinary incontinence; (4) ocular photodynamic therapy with verteporfin; (5) cryosurgical salvage therapy for recurrent prostatecancer; (6) positron emission tomography for the diagnosis and treatment of selected oncologic conditions; (7) percutaneous transluminal angioplasty of the carotidartery concurrent with stenting; (8) liver transplantation for patients with hepatocellular carcinoma; (9) coverage of liver transplants in nonapproved centers during theemergency in Houston; and (10) coverage of liver transplants in nonapproved centers during the emergency in Houston (amendment).

5 CMS’s website, which provides information about national and local coverage policies, is available at http://www.cms.hhs.gov/coverage/.

6 Contractors that have implemented LMRPs concerning the use of intravenous iron therapy include First Coast Service Options, Inc. and the Mutual of Omaha InsuranceCompany.

laboratory services. These groups providethe medical directors an opportunity todiscuss issues related to coverage,including issues raised by providers andbeneficiaries.7 In contrast to the localdecisions made by the FIs and carriers, thefour DMERCs are required to create oneset of coverage policies that applynationwide.

Generally, contractors cannot developpolicies to cover experimental orinvestigational services. However,beginning in 1995, Medicare haspermitted the coverage of certain devicesfor which the Food and DrugAdministration (FDA) has granted aninvestigational device exemption (IDE)and the coverage of certain servicesrelated to those devices (HCFA 1995).8

Specifically, contractors can considercovering a device for which the FDA has:(1) granted an IDE; (2) provided aclassification of nonexperimentalinvestigational device, for whichunderlying questions of safety andeffectiveness have been resolved for thatdevice type (i.e., the device falls under“category B”); and (3) required thatclinical trials be conducted, withbeneficiaries participating in the FDA-approved clinical trial. The intent of thisrule was to provide the opportunity forbeneficiaries to gain quicker access to newservices while permitting opportunities forproviders and manufacturers of the serviceto build the body of evidence necessaryfor seeking broader coverage. Medicaredoes not cover investigational devicesgranted an IDE that are classified as“category A”—experimental andinvestigational devices for which absoluterisk of the device type has not beenestablished.

Coverage policiesimplemented in programmanuals

Coverage policies also can beimplemented through policies publishedin Medicare’s program manuals andmemorandums.9 Program manuals,including the Medicare intermediarymanual and the Medicare carrier manual,contain operating instructions, policies,and procedures based on statutes,regulations, and directives. Programmemorandums are another vehicle forCMS to transmit new policies andprocedures that are often but notnecessarily linked to a specific programmanual. Policies published in manuals andmemorandums can set forth when andunder what circumstances services may becovered and paid for by Medicare. Forexample:

• The Medicare intermediary manualprovides coverage information abouthemodialysis treatments furnished toESRD patients. This policy limitsMedicare’s payment for hemodialysisfurnished to beneficiaries with ESRDto a maximum of three treatments perweek even though hemodialysis canbe furnished on a daily basis. Medicaldirectors can make individualcoverage determinations forbeneficiaries who require more thanthree hemodialysis treatments perweek (CMS 2003).

• CMS issued a program memorandumin 2002 about the coverage ofdiagnostic services furnished byqualified audiologists. Thememorandum set forth the specificcircumstances for which diagnosticservices provided to evaluate thesymptoms associated with hearingloss or ear injury would be coveredby Medicare and the qualifications

audiologists need to be consideredqualified by Medicare (CMS 2002b).

These policies are developed by CMSstaff and are binding on all contractors.The number of coverage decisionsimplemented in this manner is unknown.

Medicare’s coding process

CMS’s coding requirements mayimplicitly affect the coverage of newservices. (See Chapter 4 for a relateddiscussion on paying for new technologiesin Medicare’s PPSs.) Medicare’s paymentsystems are organized around standardsets of codes that describe the servicesfurnished by providers to beneficiaries.All services must be appropriately codedfor providers to receive payment fromMedicare. Some providers contend thatdelays in updating codes result in delaysin payments for new services, althoughthere is no clear evidence of problemswith access to these services. Timelycoding updates are especially important inthe outpatient sector, where paymentbundles are small and most servicesrequire a code for providers to be paid.Organizations who assign new outpatientcodes include CMS, the AmericanMedical Association, the Health InsuranceAssociation of America, and the BlueCross Blue Shield Association.

Appeals process

Beneficiaries and providers have theopportunity to appeal the denial ofcoverage for services that contractorsbelieve do not fall within a Medicarebenefit category, are not reasonable andnecessary, or are otherwise excluded bystatute or regulation. Currently, theappeals process for Part A and Part Bservices offers up to five levels forbeneficiaries and providers wishing to


7 Meetings of these clinical work groups are not required to take place in public settings.

8 Manufacturers submit marketing applications for clearance or approval of devices to the FDA. For certain devices, the FDA may require that clinical trials be conductedto obtain clinical information to determine the device’s safety and effectiveness. Generally, for these devices to be shipped lawfully for purposes of conducting the clinicaltrial, the sponsor must obtain an approved IDE from the FDA.

9 This section specifically excludes national coverage decisions published in program memorandums and the coverage issues manual.

appeal a contractor’s initial determinationthat a claim should not be paid, either infull or in part, by Medicare.10

The process begins when contractorsnotify beneficiaries and providers (theappellants) in writing of the reasons thatthey have denied coverage for a service.The appellants may request that theapplicable contractor reconsider or reviewthe denial of coverage (Figure B-1). Ifdissatisfied with the reconsideration of thedenial of coverage, appellants can appealthe decision to Administrative Law Judges(ALJs), who are employed by the SocialSecurity Administration. After the hearingwith an ALJ, cases may be appealed to theDepartmental Appeals Board of theDepartment of Health and HumanServices, the final level of administrativeappeal. Cases may then be appealed to theU.S. Federal District Courts.

As set forth in Figure B-1, the process hasseparate paths for appeals of Part A andPart B claims. Currently, depending onthe type of service that is being appealed,the appeals process differs in terms of:

• the time frames for Medicare to acton an appeal,

• the minimum value amount of aclaim to be appealed to an ALJ,

• the availability of an expeditedreview,

• the use of independent externalreviewers, and

• the right of beneficiaries to continuereceiving a service.

Figure B-1 shows some of the differencesin the time frames for Medicare to actupon an appeal. Appellants have from 60days for Part A services to 6 months forPart B services to file a reconsideration.The minimum value of services that canbe appealed to an ALJ varies for Part Aand Part B services. For inpatient hospitalservices only, appellants can ask for anexpedited review by a QualityImprovement Organization (QIO) for anoncoverage decision. Inpatients cannot

be discharged from the hospital orcharged for additional time in the hospitaluntil the QIO issues a determinationwithin one full working day afterreceiving the request.

Two sections of BIPA call for CMS tomodify the appeals process:

• Section 521 establishes uniformprocesses for handling appeals of PartA and Part B services after beingfurnished to a beneficiary. Forexample, BIPA establishes thatdisputed services must be worth atleast $100 for appellants to appeal toan ALJ, sets forth a 90-day time limitfor the ALJs and the DepartmentalAppeals Board to each make adecision about the case, and allowsappellants to escalate the case to thenext level if this deadline is not met.In addition, Section 521 establishes a

new appeals entity—qualifiedindependent contractors—toreconsider contractors’ initialdeterminations.

• Section 522 clarifies when nationaland local coverage policies can bechallenged by beneficiaries beforereceiving services. Section 522 alsorequires that CMS submit annualreports to the Congress regarding theamount of time the agency took tocomplete and fully implement NCDsfor the previous fiscal year.

CMS has not yet fully implemented thechanges mandated by BIPA. The agencyhas published proposed rules toimplement Sections 521 and 522 and hassubmitted a report to the Congress on thetime required for CMS to complete andfully implement the 10 NCDs made infiscal year 2001 (Thompson 2002).


Medicare's process for appeals of Part A and Part B claims

FIGUREB-1

Note: FI (fiscal intermediary), AIC (minimum amount in controversy), DAB (Departmental Appeals Board).

Source: Centers for Medicare & Medicaid Services. Medicare and Medicaid programs: changes to the Medicare claims appeal procedures, Federal Register. November 15, 2002, Vol. 67, No. 221, p. 69312–69363.

First

Second

Third

Fourth

Fifth

Part A

Initial determination

FI reconsiderationAIC � $0

Administrative law judgeAIC � $100

Departmental Appeals BoardAIC � $0

Federal District CourtAIC � $1,000

60 days to file

60 days to file

60 days to fileDAB may decline review

60 days to file

Part B

Initial determination

Carrier reviewAIC � $0

Carrier hearingAIC � $100

Departmental Appeals BoardAIC � $0

Federal District CourtAIC � $1,000

Administrative law judgeAIC � $500; $100 for homehealth (A/B split)

60 days to file

6 months to file

6 months to file

60 days to fileDAB may decline review

60 days to file

Level of appeal

FIGUREB-1

10 The section focuses on appeals related to Part A and Part B services. Medicare has a separate process for appeals related to M�C services, including an externalreview process and an expedited process for certain types of appeals.

References

Centers for Medicare & Medicaid Services. Intermediary manual, Part 3, Chapter II,Coverage of services. January 2003. Available athttp://www.cms.hhs.gov/manuals/13_int/a3165.asp.

Centers for Medicare & Medicaid Services. Medicare program integrity manual, Chapter13, Local medical review policy. December 2002a. Available athttp://www.cms.gov/manuals/108_pim/pim83c13.asp#Sect1.

Centers for Medicare & Medicaid Services. National coverage determinations (NCDs):implantation of automatic defibrillators. July 1, 1999. Available athttp://www.cms.hhs.gov/ncd/searchdisplay.asp?NCD_ID�110&NCD_vrsn_ num�1.

Centers for Medicare & Medicaid Services. Payment for services furnished byaudiologists (transmittal AB-02-080). June 7, 2002b. Available athttp://www.cms.hhs.gov/manuals/memos/comm_date_dsc.asp.

Davison B, CMS. Local medical review policy 101. Washington (DC), MedicalTechnology Leadership Forum. June 28, 2002.

Health Care Financing Administration. Criteria and procedures for extending coverage tocertain devices and related services, Title 42, Code of Federal Regulations, parts 405 and411. September 19, 1995, Vol. 60, No. 181, p. 48417–48425.

Health Care Financing Administration. Local medical review policy (LMRP)development and format (transmittal AB-00-116). November 24, 2000. Available athttp://www.cms.hhs.gov/manuals/pm_trans/2000/memos/comm_date_dsc.asp.

National Health Policy Forum. Medicare coverage: lessons from the past, questions forthe future. Washington (DC), The George Washington University. 2001.

Thompson TG. Report to Congress on national coverage determinations. Washington(DC), Department of Health and Human Services. June 2002.


Inpatient payments for rural hospitals

A P P E N D I X C

areas (defined as a population above1 million) and those in other urbanand rural areas; and

• raise the cap on most rural hospitals’disproportionate share (DSH)payments.

Implementing a low-volume adjustment

Making Medicare payments approximatean efficient provider’s costs requiresaccounting for factors beyond providers’control that may affect the costs offurnishing services. Patient volume maybe one such factor, particularly in smalland isolated communities where providersfrequently cannot achieve the economiesof scale of their larger counterparts, andthus have higher per case costs. Thecurrent prospective payment system (PPS)rates do not directly account for therelationship between cost and volume,placing low-volume providers at afinancial disadvantage.

The critical access hospital (CAH), solecommunity hospital, and Medicare-

A P P E N D I X

Inpatient payments for ruralhospitals

CAs discussed in Chapter 2A on paymentadequacy and updates for hospitalpayments, MedPAC previously issuedfour recommendations designed toimprove payments for rural hospitals thathave been considered by the Congress butnot yet enacted (MedPAC 2001a,MedPAC 2002). We are reissuing theserecommendations. Chapter 2Asummarizes the four recommendations,their rationales, and their combinedimpact on Medicare inpatient payments.This appendix provides additionalbackground, explanation, and support forthe four recommendations, as well asimpact estimates for each individualpolicy change.

The four recommendations would:

• implement a low-volume adjustmentto the inpatient base rates;

• reevaluate (with an eye towardreducing) the labor share (whichdetermines the portion of the basepayment rate that is adjusted by eacharea’s wage index value);

• eliminate the differential in base ratesbetween hospitals in large urban


dependent hospital programs benefit manysmall and isolated hospitals, even thoughthese programs do not directly address thesmall-scale issue. Eligibility for theseprograms is not well targeted to low-volume hospitals, however, and paymentsare based at least partially on hospital-specific costs, which may reflect poormanagement and other providerinefficiencies. A low-volume adjustmentcould address these issues more directly;for that reason, MedPAC recommendsthat the Congress enact such anadjustment.

Effects of low volume oncosts and financialperformanceTo determine whether low-volumehospitals have higher costs than otherhospitals, we examined the relationshipbetween total (all payer) inpatient volumeand Medicare costs per discharge.1 Ouranalysis showed a statistically significantrelationship between discharge volumeand costs per discharge, after controllingfor cost-related factors in the paymentsystem.2 The volume and cost relationshipis most pronounced for facilities withfewer than 200 discharges per year

1 Although Medicare payments are intended to cover the costs of treating Medicare patients, a hospital’s total volume of service determines its unit costs of production.

2 These factors include case mix as measured by diagnosis related groups, base rate (separate for hospitals in large urban areas and those in other urban and ruralareas), area wage index value, outlier frequency, and teaching intensity.

(Figure C-1), which have per case coststhat are more than 20 percent aboveaverage. The relationship becomesrelatively flat after about 500 discharges.

Low-volume hospitals account for only asmall fraction of acute care facilities; 2percent of hospitals have fewer than 200discharges and 11 percent have fewer than500 discharges. The vast majority of thesefacilities—85 percent—are in ruralcounties.

Hospitals’ financial performance underMedicare’s inpatient PPS is stronglyrelated to inpatient volume: Margins riseas volume increases (Table C-1). Theaggregate inpatient margin is negative forhospitals with 500 or fewer discharges,while hospitals in larger-volume groupshave margins ranging from 5 to 17percent.3 This strongly indicates that low-

254 I n pa t i e n t paymen t s f o r r u r a l h o sp i t a l s

Hospital discharge volume and hospital cost per case, 1997FIGUREC-1

Source: MedPAC analysis of cost report and MedPAR data from CMS.

0 100 200 300 400 500 600 700 800 900 1000

25

10

20

15

5

0

�5

�10

30

35

40Per

cent

vari

ation f

rom

mea

n c

ost

per

case

Total (all payer) discharges

Medicare inpatient margin, by discharge volume, 1999

Percent ofhospitals with

Total discharges Margin negative margin

� 200 �16.4% 66.7%201 to 500 �2.1 50.2501 to 1,000 4.6 39.01,001 to 2,500 5.0 37.72,501 to 5,000 6.5 32.75,001 to 10,000 10.1 24.010,001 to 20,000 12.3 19.4� 20,000 17.4 7.4

Note: The Medicare inpatient margin reflects the change in disproportionate share payments enacted by theMedicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 (BIPA). Analysis based ondata from two-thirds of the hospitals covered by prospective payment in 1999, which includes some that havesince been designated critical access hospitals.

Source: MedPAC analysis of cost report and MedPAR data from CMS.

T A B L EC-1

3 We show the Medicare inpatient margin for this calculation, despite the fact that it overstates hospitals’ financial performance under Medicare in the absolute, because itis inpatient costs that are affected by a hospital’s volume of discharges.

volume providers are disadvantaged byrates based on average volume and thatcurrent programs targeted to ruralproviders do not fully correct for thisproblem.

Access considerations The issue of a low-volume adjustment ismost critical for isolated hospitals, wherethe facility is important for maintainingbeneficiaries’ access to care. Suchfacilities, because of their marketcircumstances, have little ability to growand take advantage of economies of scaleand scope of services realized by largerfacilities. Adjusting payments for a low-volume facility that is near other facilities,on the other hand, is not a priority becausebeneficiaries’ access to care is less likelyto be affected. In fact, the close proximityof other hospitals may be one of theprimary reasons for the hospital’s lowvolume of service.

Low-volume hospitals are more isolatedthan those with higher volume, but mostlow-volume hospitals would not meet the35-mile distance standard used fordesignating sole community hospitals.Just over half of low-volume hospitals aremore than 25 road miles from the nearesthospital, and 86 percent have no potentialcompetitors within 15 miles.

R E C O M M E N D A T I O N 2 A - 2 :

The Congress should enact a low-volume adjustment to the rates usedin the inpatient PPS. This adjustmentshould apply only to hospitals thatare more than 15 miles from anotherfacility offering acute inpatient care.

The Commission believes that a low-volume adjustment would strengthen thecurrent inpatient PPS by aligningpayments better with efficient providers’

costs. The adjustment should reflect thebasic underlying relationship betweenpatient volume and costs per discharge,avoiding cliffs (points in the formulawhere a small change in volume wouldproduce a large change in payment) thatmight provide inappropriate incentives.

To avoid problems with annual volumevariation and to encourage stability in thelevel of the adjustment over time, thevolume adjustment should be set for anindividual facility based on a multiyearaverage volume. The level of theadjustment should be periodicallyreexamined to reflect improvements madein the inpatient PPS that might affect themeasured relationship between volumeand cost.4

To illustrate the financial impact of a low-volume adjustment, we simulated anadjustment that increases payments by upto 25 percent and drops to zero forhospitals with 500 or more discharges.5

This formula, for example, would providea 20 percent increase in payments forhospitals with 100 discharges and a 10percent increase for those with 300discharges. We limited the add-on tohospitals more than 15 miles from thenearest acute care facility. About 10percent of all PPS hospitals would qualify,and about a quarter of these alreadyreceive some assistance from the solecommunity or Medicare-dependentprogram but would benefit more from thelow-volume adjustment. The increase inpayments probably would enable somecritical access hospitals to come back intothe PPS (if these facilities were allowed toreverse their CAH status), because theadjusted base payment rate would betterreflect their underlying cost structure.6

Similarly, many hospitals might decidenot to become CAHs if a low-volumeadjustment were available.

Reevaluating the laborshare used in geographicadjustment

The labor share, which CMS revisesperiodically in updating the market basketindex, is an estimate of the nationalaverage proportion of hospitals’ costsassociated with inputs directly orindirectly affected by local wage levels.The labor share is used to determine theportion of the PPS base payment rate towhich the wage index is applied forgeographically adjusting rates. Forinpatient hospital services, CMS has setthe labor share at 71.1 percent—itsestimate of the share of hospitals’ totalexpenses comprising wages and salaries,fringe benefits, and other labor-relatedcost elements using locally purchasedinputs (Table C-2). For reasons detailed


4 Examples of policy changes that could affect the cost and volume relationship include case-mix refinements, such as all patient refined diagnosis related groups, and anoccupational mix adjustment to the wage index, both of which the Commission has recommended in past reports.

5 The payment adjustment we simulated produces a multiplier that is applied to the PPS base payment rate for a case, similar to the way the indirect medical educationand disproportionate share adjustments are applied. Only hospitals with fewer than 500 discharges would have their payments adjusted. The low-volume adjustmentmultiplier � [1.25 � (0.0005 x d)] if d � 500; otherwise the mulitplier � 1.0, where d � total inpatient acute care discharges.

6 Rural hospitals that have fewer than 15 beds (25 including swing beds) and are located more than 35 miles from the nearest hospital offering similar services (oralternatively have been designated in a comprehensive state plan as a critical access hospital for care in isolated rural areas) can apply to become a critical accesshospital. These hospitals receive full cost-based payment for both inpatient and outpatient services.

Components ofnational labor share

for inpatient care

Category Share

Total labor-related 71.1%

Wages and salaries 50.2Employee benefits 11.2Nonmedical professional fees 2.1Postage 0.3All other labor-intensive 7.3

Note: All other labor-intensive includes businessservices, computer processing, landscape andhorticultural services, building maintenance andrepair, laundry services, auto repair, paymentsto membership organizations, appliance repair,and indirect business taxes.

Source: CMS analysis of hospital data from Medicarecost reports, U.S. Census, Bureau of EconomicAnalysis, and American Hospital Association.

T A B L EC-2

below, MedPAC recommends that theSecretary reevaluate (with a view towardreducing) the labor share for inpatientpayments.

Rationale for reducing thelabor share The input categories included in the laborshare were originally selected in 1983when the hospital inpatient PPS wasadopted. Most of these inputs are stillpurchased largely in local markets.However, some categories such as postageare likely purchased in national marketsand not influenced by local wage levels.Still others (data processing andaccounting services, for instance) mayinclude some inputs that are purchased innational markets and some that are boughtlocally. As a result, the national averagelabor share may be somewhat lower thanthe current estimate of 71.1 percent.

This problem could be addressed byreexamining CMS’s construction of thenational labor share. This would likelyresult in a lower labor share, which wouldreduce the proportion of the national basepayment amount adjusted by the wageindex. Hospitals located in low-wagemarkets (wage index less than 1.0) wouldreceive higher payments, while thoselocated in high-wage markets wouldreceive lower payments. Overall, thispolicy change would transfer paymentsfrom urban to rural hospitals. Some urbanhospitals would benefit, however, becausethey are located in markets with wageindexes below 1.0, and some ruralhospitals would receive reduced paymentsbecause they are located in market areaswith wage indexes above 1.0.

Developments sinceMedPAC’s rural report About a year after our rural report(MedPAC 2001a) was published, CMSrebased the input categories in the hospitalmarket basket, as it does routinely everyfive years. CMS did not alter the inputcategories included in the wagecomponent of the market basket, but itrevised the weight (share of total costs)

for the labor-related inputs based on thelatest data, which resulted in a proposal toraise the labor share from 71.1 percent to72.5 percent.

Around this time, we obtained preliminaryresults from a multivariate analysis of thefactors explaining variation in hospitals’Medicare costs and payments per case.This analysis provided strong evidencethat the current labor share of 71.1 percentoverstates the labor-related share ofnational input costs. However, contrary towhat many observers have assumed, thestudy found that the labor-related share ofexpenses is lower in high-wage markets(most of which are in urban areas) than inlow-wage markets (most of which arerural). This pattern occurs becausehospitals in major metropolitan areasgenerally provide more sophisticatedservices and treat more complex patients,which raises their costs for plants andequipment.

Although CMS remains reluctant to basethe labor share calculation on amultivariate analysis approach, because ofits complexity and the difficulty of usingit to identify a specific point estimate, theagency pulled back its proposal to raisethe labor share pending furtherdevelopmental work.


The Secretary should reevaluate thelabor share used in the wage indexsystem that geographically adjustsrates in the inpatient PPS, with anyresulting change phased in over twoyears.

In the coming year, MedPAC willundertake a follow-up study designed toidentify the best labor share value for thehospital industry as a whole. Because theshare of labor-related expenses variesaccording to the circumstances ofhospitals, the goal will be to identify thevalue that minimizes error (that is, resultsin the smallest possible differencebetween hospitals’ individual labor sharesand the national average).

Eliminating the base rate differential

In Medicare’s inpatient PPS, the operatingbase payment rate for hospitals in largeurban areas (metropolitan areas with morethan 1 million people) is 1.6 percent abovethe payment rate for other hospitals, andthe differential is 3.0 percent for thecapital base rate (comprising about 10percent of the overall rate). Current datado not support this differential, andMedPAC recommends eliminating it.

History of the base rate differential The current payment differential reflectspolicy decisions made more than a decadeago. When the Congress established theinpatient PPS, base payment rates for ruralhospitals were set 20 percent below thosefor urban hospitals, and no distinction wasmade among hospitals in urban areasbased on the population of themetropolitan area. This initial differentialreflected actual cost differences observedin the base data used to establish the PPSrates.

Starting in 1988, the Congress enactedseparate updates for hospitals in largeurban, other urban, and rural areas,effectively creating three separate basepayment rates, while also substantiallyreducing the difference in base ratesbetween rural and urban hospitals.Hospitals in large urban areas receivedhigher updates at the time becauseanalysis showed that the higher costs ofthose hospitals were not fully recognizedby PPS payment policies.

In 1990, the operating base rate for ruralhospitals was 7.0 percent lower than therate for other urban hospitals, while therate for large urban hospitals was 1.6percent higher than the other urban rate(the current differential). The OmnibusBudget Reconciliation Act of 1990 setupdate factors to eliminate the gap inpayment rates between rural and otherurban hospitals by fiscal year 1995, partlybecause analysis showed that rural


hospital costs were 40 percent belowthose for urban hospitals while aggregatepayments were 45 percent lower.

Rationale for eliminating the differentialMedicare margin data provide support foreliminating the current differential.Inpatient margins for rural and other urbanhospitals are substantially lower thanthose of large urban hospitals (Table C-3).This difference in performance is due inlarge part to the higher payment ratesreceived by hospitals that qualify for DSHand indirect medical education (IME)adjustments; such hospitals are muchmore likely to be located in large urbanareas. However, even after removing DSHpayments and the portion of the IMEpayment above the measured costrelationship, hospitals in large urban areasstill have Medicare margins for theremaining payments that are 3.0 to 3.5percentage points higher than those ofother hospitals. The current base ratedifferential accounts for about half of thisdifference in margins.

Statistical analysis also supportseliminating the differential in base rates.When hospitals in large urban areas arecompared with all other hospitals, norelationship between large urban locationand costs per case is apparent aftercontrolling for cost-related paymentadjustments in the inpatient PPS. Wefound that rural hospital costs were about2 percent lower than those of large urbanhospitals, but this analysis was based on1997 data and does not account for the 2percent higher cost growth experiencedannually by rural hospitals between 1997and 2000. If the analysis were run usingmore recent data, the cost differencebetween hospitals in large urban and ruralareas would likely be much smaller, if notnonexistent.

Providing one base rate for all hospitalswould also eliminate the need forgeographic reclassification for the baserate.7 To qualify for base rate

reclassification, a hospital mustdemonstrate that it is close to an area witha higher base rate and that its costs arecloser to the amount it would be paid if itwere reclassified than to the amount underits current classification. In other words, ahospital with costs above its base rate canbe reclassified, whereas a hospital withcosts below its base rate cannot. Thispolicy produces an undesirable incentiveby rewarding high-cost hospitals with ahigher base rate without any otherjustification.


The Congress should raise theinpatient base rate for hospitals inrural and other urban areas to thelevel of the rate for those in largeurban areas, phased in over twoyears.

Raising the cap ondisproportionate sharepayments

Medicare’s disproportionate shareadjustment for hospital inpatient servicesis designed to offset the financial pressureof uncompensated care. However, theCommission has concluded that the

current system has several design flawsand has previously recommended a majorreform of the system. As an interimmeasure, we recommend raising the capon DSH payments that currently applies tomost rural hospitals.

The current disproportionateshare adjustment Medicare distributes DSH paymentsthrough a hospital-specific percentageadd-on to the PPS base rate. The add-onfor each case is determined by a complexformula and each hospital’s share of low-income patients, which is the sum of tworatios—Medicaid patient days as a shareof total patient days, and patient days forMedicare beneficiaries who receiveSupplemental Security Income (SSI) as apercentage of total Medicare patient days.

Problems with the currentsystem and responses todate The Commission has previouslyrecommended policy changes toameliorate two key problems with theexisting DSH payment system (MedPAC2000, MedPAC 2001b):

• The current low-income sharemeasure does not includeuncompensated care, and


Medicare inpatient margin, by location, 2000

Margin including Margin excludingDSH payments DSH payments

Hospital and above-cost and above-costgroup IME payments IME payments

All hospitals 10.9% 1.5%

Large urban areas 15.3 3.2Other urban areas 7.2 �0.4Rural areas 2.6 0.2

Note: DSH (disproportionate share), IME (indirect medical education). Above-cost IME payments are those in excessof MedPAC’s estimate of the relationship between teaching intensity and costs per discharge.


T A B L EC-3

7 This form of geographic reclassification is awarded less frequently than reclassification to obtain a higher wage index, which responds to inaccuracies in the wage indexsystem caused by the use of metropolitan statistical areas (MSAs) to represent health care labor markets.

• The system has separate paymentrates for 10 hospital groups, with theleast favorable rates given to mostrural hospitals and to urban facilitieswith fewer than 100 beds.

The Balanced Budget Refinement Act of1999 (BBRA) mandated that CMS collectdata on uncompensated care from all PPShospitals beginning with fiscal year 2002cost reports, which may pave the way forincluding uncompensated care in thecalculation of hospitals’ low-incomeshares. Then the Medicare, Medicaid, andSCHIP Benefits Improvement andProtection Act of 2000 (BIPA) partiallyimplemented our recommendation byapplying the most liberal current threshold(minimum low-income share needed toqualify for a payment adjustment) to allhospitals. We estimate that this madeabout 840 additional rural hospitals (40percent of all rural facilities) eligible toreceive DSH payments. However, BIPAcaps the DSH add-on that most ruralhospitals can receive at 5.25 percent,while some urban facilities currentlyreceive far higher adjustments.

Since MedPAC’s complete reformpackage probably cannot be implementeduntil at least fiscal year 2005 because ofthe time required to collect and processuncompensated care data, an appropriateinterim step is needed to bridge the gapbetween the BIPA provision and thesystem MedPAC envisions.


The Congress should raise the cap onthe disproportionate share add-on ahospital can receive in the inpatientPPS from 5.25 percent to 10 percent,phased in over two years.

Although there is no right level for thecap, a cap of 10 percent would bring DSHpayments for rural hospitals to roughly themidpoint between the amount that BIPAproduced and the amount implied by theproportion of the care furnished by rural

hospitals to the two largest groups of low-income patients. Rural facilities wereresponsible for 12.8 percent of the careprovided to Medicaid and uncompensatedcare patients nationally in 1999 (FigureC-2), but with the DSH payment rules ineffect at the time, only 3.1 percent ofpayments went to rural providers.8 BIPArules increased rural hospitals’ share ofpayments to 6.9 percent, and raising thecap to 10 percent would lift this share to9.8 percent.

The Congress should not remove the DSHpayment cap altogether now, for tworeasons. First, it would result in somehospitals receiving large increases in theirDSH payments, only to have theirpayments cut again if uncompensated care

is later brought into the system used todistribute payments.

Second, eliminating the cap might resultin unusually large payment increases forsome rural hospitals, and the aggregateincrease in payments would be three timesthat of our recommended approach. Thecurrent DSH distribution formula isgraduated, offering a higher payment ratefor the mostly public, inner-city hospitalswith the largest low-income shares. Thiswas done in an attempt to compensate forthese hospitals’ unusually largeuncompensated care burdens and their lowMedicare penetration (often below 20percent). Applying this formula in ruralareas, where hospitals have much higherMedicare penetration (often above 70percent), could result in windfall-level


Rural hospitals' shares of low-income patient costsand disproportionate share payments

FIGUREC-2

Source: MedPAC analysis of data from the American Hospital Association annual survey of hospitals and impact file data from CMS.

Note: The 5.25 percent cap on the disproportionate share add-on was enacted by the Medicare, Medicaid,and SCHIP Benefits Improvement and Protection Act of 2000 (BIPA) and went into effect on April 1, 2001.Low-income costs for this analysis include Medicaid and uncompensated care.

16

8

4

12.8

3.1

6.9

9.8

0Share of disproportionate share payments:

12

Per

cent

Prior toBIPA

Share oflow-income

costs

5.25percent cap

10percent cap

8 Because uncompensated care data from the Medicare cost reports are not yet available, this analysis is based on data from the American Hospital Association annualsurvey of hospitals.

payment adjustments. If the Congressapproves revamping the DSH paymentsystem to bring uncompensated care intothe low-income share calculation, itshould consider avoiding this problem byapplying a single formula to all hospitalswithout a graduated rate structure.

Impact ofrecommendations

Three of our four recommendations toimprove rural hospital payments call for atwo-year phase-in schedule. To displaythe full impact, Table C-4 shows the one-year increase in inpatient paymentsresulting from each of therecommendations and Table C-5 (p. 260)shows the two-year increase for each. The

combined impact of all four policychanges, reflecting their interactiveeffects, is presented in Chapter 2A.

Implementing a low-volume adjustment(which we are recommending forimmediate implementation) wouldincrease aggregate inpatient payments byless than 0.1 percent. But despite the smalloverall impact, this policy change wouldincrease payments for hospitals withfewer than 200 discharges by about 8percent and for those with 201 to 500discharges by 4 percent. In addition, theaggregate impact might be somewhatlarger if critical access hospitals areallowed to return to the PPS to takeadvantage of the higher paymentsafforded by this policy change.

Although our recommendation that CMSreevaluate the labor share used in thehospital wage index system does notspecify an exact value for the labor share,we simulated an illustrative reduction to68 percent from the current 71.1 percent.CMS would implement this changebudget neutrally, which would increasepayments for rural and other urbanhospitals by 0.2 percent while decreasingpayments for large urban hospitals by thesame amount.

Eliminating the differential in basepayments rate for hospitals in rural andother urban areas would raise paymentsfor hospitals in these areas by 1.2 percent.This increase is less than the 1.6 percentdifferential in base rates under currentpolicy because many of the hospitals paidcost-related rates under the solecommunity hospital and Medicare-dependent programs would not be affectedby the policy change.

Raising the cap on DSH payments to 10percent would increase rural hospitals’payments by 1.2 percent on average.Although urban hospitals with fewer than100 beds would see similar increases,there are so few of these facilities that theincrease for all urban hospitals is less than0.1 percent.

Our recommendations generally providethe largest payment increases to hospitalsthat do not benefit from any of theexisting programs aimed at helping ruralhospitals—the rural referral, solecommunity, and small rural Medicaredependent programs. The only exceptionis the low-volume adjustment that likelywould not benefit such hospitals if theyhave more than 50 beds. Hospitals nothelped by current programs have thelowest Medicare inpatient margins undercurrent policy—3.7 percent for those withfewer than 50 beds and 2.5 percent forthose with more than 50 beds. Raising thecap on DSH payments produces thelargest difference, with hospitals nothelped by any current program receivingan increase of over 2 percent comparedwith less than 1 percent for all other ruralfacilities.


One-year impact on Medicare inpatient payments of four recommendations to improve

payments for rural hospitals

Change in payments for each recommendation

Implement Reduce Eliminate RaiseHospital Baseline low-volume labor share base rate DSH cap togroup margin adjustment to 68 percent differential 10 percent

All hospitals 10.3% * 0.0% 0.3% 0.1%

Urban 11.3 0.0% –* 0.3 *Rural 3.9 * 0.1 0.6 0.6

Large urban 13.6 0.0 –0.1 0.0 *Other urban 7.7 0.0 0.1 0.8 *Rural referral 3.9 0.0 0.1 0.6 0.6Sole community 4.6 0.1 0.1 0.3 0.1Small rural Medicare-

dependent 7.2 0.2 0.2 0.7 0.5Other rural � 50 beds 3.7 0.2 0.2 0.8 1.0Other rural � 50 beds 2.5 * 0.2 0.8 1.1

Major teaching 20.7 0.0 –0.1 0.2 0.0Other teaching 9.6 0.0 * 0.4 *Nonteaching 5.4 * * 0.4 0.2

Note: DSH (disproportionate share). Baseline margin is the actual 2000 margin adjusted to reflect the increase indisproportionate share payments implemented in 2001 and the decrease in indirect medical educationpayments implemented in 2003. Analysis excludes critical access hospitals.* Less than 0.05 percent

Source: MedPAC analysis of impact file and MedPAR data from CMS.

T A B L EC-4


Two-year impact on Medicare inpatient payments of four recommendations to improve

payments for rural hospitals

Change in payments for each recommendation

Implement Reduce Eliminate RaiseHospital Baseline low-volume labor share base rate DSH cap togroup margin adjustment to 68 percent differential 10 percent

All hospitals 10.3% * 0.0% 0.7% 0.2%

Urban 11.3 0.0% –* 0.6 *Rural 3.9 * 0.2 1.2 1.2

Large urban 13.6 0.0 –0.2 0.0 *Other urban 7.7 0.0 0.2 1.5 *Rural referral 3.9 0.0 0.2 1.2 1.2Sole community 4.6 0.1 0.1 0.6 0.3Small rural Medicare-

dependent 7.2 0.2 0.4 1.4 0.9Other rural � 50 beds 3.7 0.2 0.4 1.7 2.1Other rural � 50 beds 2.5 * 0.4 1.6 2.2

Major teaching 20.7 0.0 –0.2 0.3 0.0Other teaching 9.6 0.0 * 0.7 *Nonteaching 5.4 * * 0.9 0.4

Note: DSH (disproportionate share). Baseline margin is the actual 2000 margin adjusted to reflect the increase indisproportionate share payments implemented in 2001 and the decrease in indirect medical educationpayments implemented in 2003. Analysis excludes critical access hospitals.* Less than 0.05 percent

Source: MedPAC analysis of impact file and MedPAR data from CMS.

T A B L EC-5

References

Medicare Payment Advisory Commission. Report to the Congress: Medicare in ruralAmerica. Washington (DC), MedPAC. June 2001a.


Medicare Payment Advisory Commission. Report to the Congress: Medicare paymentpolicy. Washington (DC), MedPAC. March 2001b.



A data book on hospital financial performance

A P P E N D I X D

Table D-3 shows the trend in Medicareinpatient length of stay.

Further tables present data on a number ofmargin measures for prospective paymentsystem (PPS) hospitals, based onMedicare cost report data. This analysisfeatures our overall Medicare margin thatincorporates payments and costs forinpatient and outpatient services, as wellas hospital-based home health, skillednursing, and PPS-exempt units. Marginsfor each of these components and theoverall Medicare margin (that includesgraduate medical education and Medicarebad debt) are presented by hospital group:

Table D-4 shows the trend in Medicareinpatient margins.

Table D-5 shows the distribution ofMedicare inpatient margins for 2000.

Table D-6 shows the trend in Medicareoutpatient margins for 1996 through 2000.

Table D-7 shows the distribution ofMedicare outpatient margins for 2000.

Table D-8 shows the trend in hospital-based Medicare skilled nursing facilitymargins for 1996 through 2000.

Table D-9 shows the trend in hospital-based Medicare home health agencymargins for 1996 through 2000.

A P P E N D I X

A data book on hospitalfinancial performance

DThis appendix provides data on hospitalfinancial performance. Tables in this databook provide variables by hospital groupand are presented for 10 years (1991 to2000) unless otherwise noted below.Tables include data from the Medicarecost reports and the American HospitalAssociation annual survey of hospitals.Medicare cost report data from 2000include imputed values for hospitalswhose 2000 cost reports were notavailable (about 27 percent ofobservations). Hospitals are grouped byseveral attributes, including location(urban and rural), teaching status (majorteaching, other teaching, nonteaching),receipt of disproportionate sharepayments, census region, and ownershipstatus. All measures, with the exception ofdistribution data, are national aggregates,not the averages of individual facilities;this provides an overview of the industryas a whole. Definitions of the variablesincluded in these tables can be found inthe table notes.

The data book starts with case-basedvariables:

Table D-1 shows the trends in hospitalpayments per case, costs per case, andlength of stay.

Table D-2 shows the trend in Medicarecost per discharge.


Table D-10 shows the trend in MedicarePPS-exempt unit margins for 1996through 2000.

Table D-11 shows the trend in the overallMedicare margins for 1996 through 2000.

Table D-12 shows the distribution of theoverall Medicare margins for 2000.

The analysis is then expanded fromMedicare to comparative tables amongpayers. These tables contain aggregatevalues for all community hospitals, whichincludes all PPS hospitals and most PPS-exempt facilities.

Table D-13 shows the trend in payment-to-cost ratio by source of revenue.

Table D-14 shows the trend in gains orlosses by source of revenue.

The appendix concludes with data onhospital total margins. The total marginincludes all patient care services fundedby all payers, plus nonpatient revenue.

Table D-15 shows the trend in hospitaltotal margins.

Table D-16 shows the distribution ofhospital total margins for 2000.

266 A da t a book on ho sp i t a l f i n an c i a l p e r f o rmance

Change in hospital payment, cost, and length of stay indicators, 1991–2001

Medicare Medicare Medicare Medicare Total Costs peroperating Market payments costs per length length adjusted

Year update basket per discharge discharge of stay of stay admission

1991 3.4% 4.4% 6.1% �7.0% �2.7% �1.3% �5.5%1992 3.0% 3.2% 6.2% �4.6% �3.3% �1.6% �5.7%1993 2.7% 3.1% 3.5% �1.2% �5.5% �2.3% �3.4%1994 2.0% 2.6% 3.1% �1.1% �6.0% �3.8% �0.1%1995 2.0% 3.2% 4.9% �1.2% �6.2% �4.3% �0.5%1996 1.5% 2.4% 5.5% �0.5% �5.6% �3.5% �0.4%1997 2.0% 2.0% 0.5% �0.1% �3.6% �1.9% �1.5%1998 0.0% 2.9% �0.1%% �1.7% �2.2% �0.9% �2.3%1999 1.1% 2.5% 0.6% �2.9% �1.3% �1.8% �2.7%2000 1.1% 3.6% 1.2% �2.9% �1.9% �1.9% �2.1%2001 3.4% 4.3% N/A N/A N/A �1.3% �4.7%

Note: N/A � not available. Excludes critical access hospitals.

Source: MedPAC analysis of Medicare cost report data (fourth quarter 2002) from CMS, and data from the American Hospital Association annual survey of hospitals.

T A B L ED-1


Change in Medicare inpatient costs per discharge, 1991–2000

Hospital group 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

All hospitals 7.0% 4.6% 1.2% �1.1% �1.2% �0.5% 0.1% 1.7% 2.9% 2.9%

Urban 6.7 4.4 1.1 �1.5 �1.4 �0.6 0.1 1.6 2.9 2.8Rural 8.7 5.9 2.1 0.8 0.1 0.7 1.4 2.8 3.3 3.4

Large urban 6.1 3.4 1.3 �2.0 �1.5 �0.7 0.2 1.6 2.7 2.6Other urban 7.6 6.1 0.8 �0.6 �1.2 �0.3 �0.1 1.7 3.3 3.1Rural referral 8.7 5.6 2.1 0.2 �0.4 �0.1 1.0 3.2 3.8 3.4Sole community 8.6 4.8 2.5 1.1 1.6 1.2 1.6 2.6 2.4 4.4Small rural Medicare-dependent 9.2 4.7 1.8 1.5 �2.5 3.7 2.4 2.2 1.3 5.0Other rural � 50 beds 6.8 6.3 2.2 2.3 2.1 1.9 1.2 4.5 3.1 3.0Other rural � 50 beds 8.7 7.0 1.5 0.8 �0.3 0.2 1.7 1.9 3.6 2.1

Major teaching 6.9 3.7 2.0 �2.5 �1.1 0.5 �0.2 2.0 3.2 2.6Other teaching 6.8 4.5 0.8 �1.2 �0.8 �0.9 0.3 1.1 2.4 3.1Nonteaching 7.2 4.8 1.1 �0.7 �1.8 �0.9 0.3 2.0 3.3 3.0

Major teachingPublic 7.3 5.6 0.3 �3.5 �1.8 5.3 1.5 0.4 5.3 5.9Private 6.8 3.3 2.3 �2.4 �0.9 �0.5 �0.5 2.4 2.7 1.9

Other teachingPublic 8.6 5.2 0.4 �1.1 �1.9 �2.8 �0.6 4.9 2.9 0.9Private 6.6 4.5 0.9 �1.2 �0.7 �0.8 0.4 0.9 2.4 3.2

NonteachingPublic 9.0 5.6 2.1 0.8 �1.0 0.6 0.7 2.3 2.4 4.1Private 6.8 4.7 0.9 �1.0 �1.9 �1.2 0.2 2.0 3.5 2.8

DSHLarge urban 6.2 3.0 0.9 �2.1 �1.4 �0.5 0.6 1.3 2.9 2.9Other urban 7.9 6.5 0.8 �0.4 �1.4 �0.2 0.1 1.6 3.6 3.7Rural 9.4 7.1 2.3 0.1 �1.4 0.2 1.6 3.6 3.0 3.8

Non-DSH 7.1 4.8 1.5 �0.9 �0.9 �0.7 �0.1 1.9 2.7 2.5

Teaching and DSH 7.0 4.3 0.9 �1.7 �1.0 �0.2 0.2 1.1 3.1 3.1Teaching and non-DSH 6.5 4.5 2.1 �1.4 �0.6 �1.0 �0.1 2.2 1.6 2.4Nonteaching and DSH 7.0 4.8 0.8 �0.8 �2.4 �1.6 0.5 2.2 3.0 3.6Nonteaching and non-DSH 7.4 4.9 1.2 �0.6 �1.3 �0.4 0.1 1.9 3.5 2.6

New England 2.7 4.3 2.6 0.9 �0.5 �1.6 �0.7 0.1 1.2 �0.2Middle Atlantic 6.7 4.7 2.2 �0.7 0.1 �0.9 0.0 0.7 2.1 1.9South Atlantic 6.8 4.6 1.0 �1.8 �2.1 �0.8 0.5 1.4 3.5 4.7East North Central 7.5 5.0 1.0 �0.6 �0.2 �0.4 �0.5 2.2 2.5 2.5East South Central 10.2 7.3 0.1 �3.2 �1.9 1.2 0.8 2.8 3.3 2.6West North Central 6.3 4.9 1.4 0.1 �0.6 2.6 1.4 2.5 3.9 2.3West South Central 8.5 3.9 1.9 �1.6 �3.4 �1.7 0.1 1.5 2.7 4.5Mountain 6.4 5.4 �0.3 0.4 �1.4 0.3 0.7 2.8 2.7 5.4Pacific 6.9 3.0 0.2 �1.7 �1.5 �0.2 0.8 2.8 5.2 4.3

Voluntary 6.9 4.6 1.4 �1.0 �0.9 �0.3 0.0 1.7 2.8 2.6Proprietary 6.2 3.6 �0.7 �3.0 �3.6 �3.8 0.8 1.4 4.1 4.4Urban government 7.9 5.5 0.8 �1.5 �2.0 1.5 �0.3 1.9 3.6 4.3Rural government 9.5 6.3 3.1 2.0 0.1 1.7 1.6 2.7 2.0 3.8

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teaching hospitals have aratio of less than 0.25. Excludes critical access hospitals.


T A B L ED-2


Change in Medicare inpatient length of stay, 1991–2000

Hospital group 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

All hospitals �2.7% �3.3% �5.5% �6.0% �6.2% �5.6% �3.6% �2.2% �1.3% �1.9%

Urban �3.0 �3.4 �5.9 �6.3 �6.6 �5.9 �3.7 �2.2 �1.2 �1.9%Rural �1.3 �3.1 �3.5 �4.3 �4.7 �4.2 �3.2 �2.2 �1.6 �2.3%

Large urban �3.4 �3.8 �5.7 �6.7 �6.4 �5.9 �3.5 �2.1 �1.0 �1.7Other urban �2.3 �2.8 �6.0 �5.8 �6.7 �5.9 �3.8 �2.3 �1.4 �2.1Rural referral �1.9 �3.7 �4.6 �6.3 �6.0 �5.7 �3.5 �1.7 �1.8 �1.6Sole community �1.0 �2.2 �3.0 �2.9 �3.6 �3.5 �3.0 �2.7 �1.4 �2.4Small rural Medicare-dependent �0.5 �2.7 �2.3 �2.0 �3.9 �1.2 �1.3 �2.3 �3.0 �1.8Other rural � 50 beds �2.0 �2.5 �1.7 �3.4 �1.3 �2.9 �2.8 �3.3 0.0 �4.8Other rural � 50 beds �1.1 �3.2 �3.2 �3.6 �5.2 �3.8 �3.6 �2.0 �1.6 �2.6

Major teaching �3.2 �3.5 �5.8 �7.2 �6.7 �6.6 �4.5 �2.4 �0.9 �1.5Other teaching �3.0 �3.4 �6.2 �6.3 �6.3 �6.0 �3.8 �2.5 �1.6 �1.7Nonteaching �2.3 �3.4 �4.9 �5.4 �6.1 �5.0 �3.1 �2.0 �1.2 �2.2

Major teachingPublic �2.2 �3.1 �5.8 �5.6 �6.8 �5.0 �2.4 �4.3 �0.5 �1.6Private �3.4 �3.5 �5.8 �7.5 �6.7 �6.9 �4.9 �2.0 �1.0 �1.5

Other teachingPublic �3.6 �2.0 �7.5 �6.2 �6.8 �7.4 �4.7 �0.5 �2.2 �2.3Private �3.0 �3.5 �6.1 �6.3 �6.3 �5.9 �3.8 �2.7 �1.6 �1.7

NonteachingPublic �0.9 �3.1 �3.5 �3.5 �4.9 �3.9 �2.7 �2.1 �1.4 �1.6Private �2.6 �3.4 �5.1 �5.7 �6.3 �5.3 �3.2 �2.0 �1.1 �2.3

DSHLarge urban �3.6 �3.6 �5.8 �6.3 �6.4 �5.9 �3.5 �2.3 �1.0 �1.2Other urban �2.4 �2.6 �6.1 �5.8 �6.7 �5.9 �3.7 �2.3 �1.3 �1.9Rural �1.2 �2.8 �3.5 �4.4 �5.6 �5.4 �3.8 �2.0 �2.2 �2.2

Non-DSH �2.3 �3.6 �5.1 �6.0 �5.9 �5.3 �3.6 �2.2 �1.3 �2.3

Teaching and DSH �3.2 �3.1 �6.1 �6.6 �6.5 �6.5 �4.0 �2.5 �1.4 �1.5Teaching and non-DSH �2.7 �3.9 �5.9 �6.7 �6.3 �5.7 �4.2 �2.5 �1.4 �2.0Nonteaching and DSH �2.6 �3.3 �5.2 �5.1 �6.5 �5.3 �3.1 �2.0 �1.0 �1.8Nonteaching and non-DSH �2.2 �3.5 �4.6 �5.6 �5.8 �4.9 �3.2 �2.0 �1.3 �2.5

New England �7.8 �4.3 �5.4 �7.5 �8.6 �7.8 �6.4 �3.2 �1.7 �3.3Middle Atlantic �2.8 �2.2 �5.8 �6.3 �6.7 �6.7 �4.9 �3.6 �1.5 �2.3South Atlantic �2.6 �4.2 �5.0 �6.1 �6.6 �5.7 �3.2 �2.1 �1.5 �0.9East North Central �2.8 �3.9 �6.0 �6.5 �5.8 �6.1 �3.5 �1.9 �1.4 �2.1East South Central �0.5 �2.5 �5.4 �6.1 �6.4 �4.4 �3.7 �1.7 �1.5 �2.5West North Central �2.7 �3.8 �5.6 �4.9 �5.0 �3.0 �2.3 �2.0 �1.4 �2.7West South Central �1.3 �3.4 �4.4 �5.4 �6.8 �4.9 �2.7 �1.7 �1.0 �1.4Mountain �3.2 �2.7 �6.7 �5.1 �5.7 �3.8 �1.7 �1.7 �0.6 �1.6Pacific �3.1 �4.8 �6.2 �4.8 �3.2 �3.1 �0.4 �0.1 0.6 0.3

Voluntary �2.9 �3.4 �5.6 �6.3 �6.3 �5.7 �3.9 �2.4 �1.4 �2.0Proprietary �2.5 �3.7 �5.3 �5.8 �6.6 �5.8 �2.6 �1.3 �0.2 �1.2Urban government �2.2 �2.8 �5.8 �5.3 �6.6 �5.3 �3.2 �2.3 �1.1 �1.5Rural government �0.7 �3.0 �2.5 �2.7 �4.0 �3.5 �2.7 �2.1 �1.8 �2.0

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teaching hospitals have aratio of less than 0.25. Excludes critical access hospitals.


T A B L ED-3


Medicare inpatient margins excluding payments for direct graduatemedical education, by hospital group, 1991–2000

Hospital group 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

All hospitals �2.4% �0.9% 1.3% 5.6% 11.1% 16.1% 16.5% 14.4% 12.3% 10.8%

Urban �2.2 �0.8 1.6 6.4 11.8 16.9 17.4 15.6 13.5 12.1Rural �3.7 �1.4 �0.5 0.6 6.1 10.7 10.0 6.3 4.4 2.7

Large urban �1.6 0.4 3.0 8.6 13.9 19.1 19.5 17.9 16.0 14.6Other urban �3.3 �2.9 �0.8 2.7 8.3 13.5 14.3 12.2 9.8 8.2Rural referral �3.7 �1.0 �1.1 0.0 5.8 10.5 10.0 6.3 4.8 3.1Sole community �0.9 2.1 4.1 5.2 8.6 12.9 11.5 8.0 5.8 3.8Small rural Medicare-dependent 1.2 3.3 2.4 �0.6 6.7 10.9 11.8 10.3 8.9 5.9Other rural � 50 beds �5.4 �4.2 �1.2 �0.8 4.5 9.9 9.9 5.6 3.9 2.0Other rural � 50 beds �7.1 �5.7 �3.8 �1.8 4.6 9.3 8.3 3.8 1.3 0.5

Major teaching 6.8 8.7 10.9 16.8 21.5 25.8 25.9 24.9 23.1 22.9Other teaching �2.8 �1.7 0.7 4.8 10.0 15.0 15.5 13.8 12.1 10.2Nonteaching �6.4 �5.0 �3.0 0.6 6.6 11.7 12.4 9.3 6.7 4.9

Major teachingPublic 10.8 11.4 14.4 21.0 26.1 28.5 28.7 27.5 23.9 21.0Private 5.9 8.2 10.1 15.8 20.3 25.2 25.2 24.3 22.9 23.3

Other teachingPublic �1.5 �0.4 1.9 4.9 10.4 14.3 15.9 11.0 10.0 7.9Private �2.9 �1.7 0.7 4.8 10.1 15.0 15.5 14.0 12.3 10.3

NonteachingPublic �6.3 �5.1 �3.5 �2.0 3.9 8.2 7.9 4.8 3.0 0.2Private �6.4 �4.9 �2.9 1.0 7.1 12.3 13.2 10.0 7.3 5.7

DSHLarge urban 2.2 4.6 7.7 13.6 18.5 23.1 22.8 21.5 19.8 18.5Other urban �1.4 �0.9 1.2 4.8 10.7 15.6 16.3 14.1 11.7 9.8Rural �2.7 �1.1 �0.4 0.1 7.3 12.7 12.1 7.3 5.7 3.8

Non-DSH �6.7 �5.4 �3.9 �0.4 5.2 10.8 11.6 9.1 7.0 5.6

Teaching and DSH 3.1 4.7 7.4 12.5 17.3 21.5 21.5 20.4 18.4 17.0Teaching and non-DSH �4.6 �3.2 �1.8 2.2 7.7 14.0 14.6 12.9 11.7 10.7Nonteaching and DSH �4.2 �2.5 �0.1 3.9 10.3 15.6 15.8 12.7 10.3 8.3Nonteaching and non-DSH �8.1 �7.0 �5.3 �2.2 3.5 8.4 9.5 6.3 3.6 2.0

New England �2.1 0.0 1.3 5.3 10.0 17.3 19.3 17.6 15.8 16.8Middle Atlantic 1.1 2.3 4.5 8.9 12.7 18.2 19.1 20.2 19.8 20.0South Atlantic �5.9 �4.3 �2.3 2.7 9.5 14.5 15.5 12.4 10.0 7.0East North Central �5.1 �3.4 �1.2 2.2 7.1 12.1 13.3 10.0 7.8 6.4East South Central �3.7 �4.4 �1.9 4.0 11.2 15.9 15.2 11.9 10.4 8.2West North Central �3.0 �2.7 �1.2 2.4 7.1 10.9 10.3 8.6 6.3 6.3West South Central �4.5 �2.3 �0.6 4.0 11.4 17.6 17.0 15.3 12.1 10.5Mountain 1.7 3.4 6.5 8.4 13.1 17.0 16.6 12.8 9.4 5.8Pacific 1.4 4.3 7.9 13.2 18.9 23.0 21.6 18.4 15.3 12.9

Voluntary �2.4 �1.0 1.0 5.1 10.1 15.2 15.9 13.8 11.9 10.7Proprietary �4.7 �2.4 1.2 7.8 15.5 21.3 20.7 18.6 15.7 13.5Urban government 1.5 2.5 5.3 1.3 16.1 19.6 19.6 17.5 14.8 12.1Rural government �4.6 �3.1 �2.2 �2.7 3.0 7.4 6.6 3.0 1.5 �0.8

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teaching hospitals have aratio of less than 0.25. A margin is calculated as revenue minus costs divided by revenue; margins are based on Medicare-allowed costs. Medicare inpatient marginincludes services covered by the inpatient prospective payment system. 2000 values are imputed for hospitals whose 2000 cost reports were not available (about 27percent of observations). Excludes critical access hospitals.


T A B L ED-4


Distribution of Medicare inpatient margins excluding payments for directgraduate medical education, by hospital group, 2000

PercentPercentile with

Number of negativeHospital group hospitals 10th 25th 50th 75th 90th margins

All hospitals 4,124 �17.2% �5.5% 6.0% 16.7% 26.9% 36.3%

Urban 2,458 �13.7 �2.7 7.8 17.9 27.8 31.2Rural 1,666 �22.4 �9.0 2.9 14.4 24.6 43.7

Large urban 1,405 �12.5 �1.0 10.1 20.8 29.6 27.3Other urban 1,053 �14.9 �5.0 5.0 14.0 22.8 36.4Rural referral 222 �13.2 �6.9 1.9 11.8 19.1 44.6Sole community 515 �21.7 �8.2 4.8 17.2 28.6 39.8Small rural Medicare-dependent 221 �18.9 �6.0 7.4 17.0 27.1 35.7Other rural � 50 beds 354 �29.8 �12.9 2.3 15.6 24.8 46.6Other rural � 50 beds 354 �24.4 �9.9 �0.2 9.2 19.1 50.8

Major teaching 280 5.4 15.0 23.2 31.4 38.9 5.7Other teaching 755 �7.3 1.6 9.4 18.1 27.3 21.9Nonteaching 3,089 �20.4 �8.0 3.4 14.1 23.7 42.5

Major teachingPublic 76 5.9 13.4 22.2 30.5 37.2 3.9Private 204 5.1 15.2 23.8 31.5 39.1 6.4

Other teachingPublic 57 �21.1 �3.8 6.8 14.6 21.6 35.1Private 698 �6.8 1.8 9.5 18.3 27.6 20.8

NonteachingPublic 813 �25.2 �10.5 1.1 13.4 23.6 47.2Private 2,276 �18.3 �6.8 4.0 14.3 23.8 40.9

DSHLarge urban 748 �4.2 6.1 15.9 25.6 34.2 15.2Other urban 579 �9.5 �1.2 8.0 16.2 25.6 27.1Rural 375 �21.0 �6.8 5.5 18.8 28.5 37.6

Non-DSH 2,422 �21.6 �8.8 2.3 12.6 21.7 44.7

Teaching and DSH 683 �2.7 6.4 15.7 25.7 34.5 12.3Teaching and non-DSH 352 �10.8 �1.3 7.7 16.9 25.6 27.6Nonteaching and DSH 1,019 �13.7 �3.7 7.5 18.3 27.2 32.2Nonteaching and non-DSH 2,070 �23.0 �9.9 0.9 11.8 21.2 47.6

New England 172 �17.9 �6.7 8.2 21.2 28.8 33.1Middle Atlantic 467 �8.6 1.8 12.2 24.1 35.4 20.6South Atlantic 598 �14.9 �4.4 4.4 14.0 21.6 37.5East North Central 652 �23.1 �11.2 �0.2 10.8 19.8 50.5East South Central 386 �11.4 �1.0 9.2 19.7 27.8 27.5West North Central 499 �19.9 �9.5 1.5 11.8 20.9 47.9West South Central 569 �16.2 �3.7 8.4 18.8 28.7 32.0Mountain 267 �23.5 �7.4 3.9 16.0 25.0 41.6Pacific 514 �17.1 �3.7 9.0 19.5 29.6 29.4

Voluntary 2,520 �15.4 �5.1 5.7 16.0 26.0 36.3Proprietary 658 �11.5 �0.7 10.1 20.2 29.5 26.4Urban government 312 �17.4 �5.5 7.9 18.7 29.6 34.0Rural government 634 �26.9 �10.8 0.9 13.4 23.9 47.5

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teaching hospitals have aratio of less than 0.25. A margin is calculated as revenue minus costs divided by revenue; margins are based on Medicare-allowed costs. Medicare inpatient marginincludes services covered by the inpatient prospective payment system. Data are imputed for hospitals whose 2000 cost reports were not available (about 27 percent ofobservations). Excludes critical access hospitals.


T A B L ED-5


Medicare outpatient margins excluding payments for direct graduate medical

education, by hospital group, 1996–2000

Hospital group 1996 1997 1998 1999 2000

All hospitals �7.6% �6.7% �16.1% �16.4% �13.7%

Urban �7.9 �6.9 �16.2 �16.5 �13.6Rural �6.4 �5.9 �15.7 �16.1 �13.9

Large urban �8.3 �7.1 �16.9 �16.7 �13.9Other urban �7.4 �6.6 �15.3 �16.2 �13.3Rural referral �5.4 �5.1 �14.4 �14.4 �11.1Sole community �4.3 �2.8 �13.7 �14.3 �12.0Small rural Medicare-dependent �9.6 �8.8 �18.4 �18.3 �17.5Other rural � 50 beds �10.1 �9.4 �18.6 �20.0 �19.9Other rural � 50 beds �7.8 �7.5 �17.8 �18.3 �17.2

Major teaching �10.5 �10.0 �19.3 �18.3 �17.8Other teaching �6.9 �6.4 �15.0 �15.3 �12.0Nonteaching �7.0 �5.7 �15.6 �16.4 �13.2

Major teachingPublic �12.6 �13.1 �20.3 �19.3 �21.8Private �10.0 �9.3 �19.0 �18.0 �16.9

Other teachingPublic �7.9 �7.5 �13.8 �14.5 �15.4Private �6.9 �6.3 �15.1 �15.4 �11.8

NonteachingPublic �7.1 �7.5 �16.4 �16.2 �16.1Private �7.0 �5.3 �15.5 �16.4 �12.6

DSHLarge urban �8.8 �8.0 �17.6 �17.2 �15.5Other urban �7.5 �6.6 �15.7 �16.5 �13.7Rural �5.4 �4.0 �16.0 �16.8 �16.0

Non-DSH �7.2 �6.2 �15.4 �15.9 �12.4

Teaching and DSH �8.8 �8.4 �17.2 �17.0 �15.5Teaching and non-DSH �7.1 �6.4 �15.4 �15.3 �11.5Nonteaching and DSH �6.6 �5.1 �15.9 �16.7 �13.6Nonteaching and non-DSH �7.2 �6.1 �15.4 �16.2 �13.0

New England �7.8 �7.4 �14.9 �14.6 �13.0Middle Atlantic �10.6 �9.2 �18.0 �17.4 �13.0South Atlantic �6.3 �5.3 �13.8 �14.6 �13.1East North Central �7.8 �7.9 �17.4 �17.8 �15.6East South Central �6.6 �6.4 �16.9 �18.4 �16.4West North Central �6.5 �5.6 �14.2 �14.2 �12.7West South Central �7.0 �4.3 �14.8 �16.2 �14.0Mountain �6.1 �4.3 �13.9 �13.7 �11.2Pacific �7.9 �6.6 �18.4 �18.4 �12.6

Voluntary �7.6 �6.6 �16.0 �16.1 �13.0Proprietary �6.5 �4.3 �15.5 �17.8 �14.0Urban government �9.7 �9.7 �17.5 �17.3 �18.1Rural government �6.9 �7.6 �16.8 �16.2 �16.5

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater,while other teaching hospitals have a ratio of less than 0.25. A margin is calculated as revenue minus costs divided by revenue;margins are based on Medicare-allowed costs. 2000 values were imputed for hospitals whose 2000 cost reports were notavailable (about 27 percent of observations). Excludes critical access hospitals.

Source: MedPAC analysis of Medicare cost report data (fourth quarter 2002 for 1996, 1998–2000 data; fourth quarter 1999 for 1997data) from CMS.

T A B L ED-6


Distribution of Medicare outpatient margins excluding payments for directgraduate medical education, by hospital group, 2000



All hospitals 3,897 �31.6% �22.8% �15.3% �8.6% �1.3% 91.7%

Urban 2,346 �31.7 �22.4 �14.4 �7.2 0.3 89.8Rural 1,551 �31.2 �23.5 �16.7 �10.3 �4.1 94.5

Large urban 1,351 �32.0 �22.3 �14.0 �6.8 1.0 88.7Other urban 995 �31.5 �22.5 �14.5 �7.6 �1.0 91.2Rural referral 207 �24.6 �18.0 �11.6 �6.7 1.0 89.4Sole community 488 �27.9 �20.5 �14.3 �8.9 �3.4 93.0Small rural Medicare-dependent 204 �33.6 �27.0 �19.2 �12.8 �8.0 96.6Other rural � 50 beds 331 �36.7 �26.6 �20.3 �14.6 �7.5 97.3Other rural � 50 beds 321 �31.2 �24.3 �18.0 �12.3 �6.0 96.0

Major teaching 253 �34.5 �25.4 �16.9 �9.7 �1.1 92.9Other teaching 718 �29.6 �20.1 �13.0 �6.3 0.9 88.4Nonteaching 2,926 �31.7 �23.2 �15.8 �9.1 �2.1 92.3

Major teachingPublic 55 �51.6 �29.3 �15.8 �8.0 �0.8 92.7Private 198 �31.7 �25.2 �16.9 �9.8 �1.3 92.9

Other teachingPublic 44 �31.0 �21.8 �14.6 �9.2 �1.2 90.9Private 674 �29.5 �20.0 �12.9 �6.2 0.9 88.3

NonteachingPublic 761 �36.1 �25.1 �18.1 �11.0 �5.1 96.3Private 2,165 �30.6 �22.3 �15.0 �8.4 �1.0 90.9

DSHLarge urban 702 �33.7 �23.1 �14.7 �7.3 1.2 88.7Other urban 545 �29.8 �21.6 �14.6 �8.3 �1.2 91.2Rural 338 �32.4 �24.2 �16.7 �10.7 �5.2 95.6

Non-DSH 2,312 �31.4 �22.8 �15.5 �8.6 �1.5 92.1

Teaching and DSH 631 �31.3 �22.2 �14.7 �7.8 �0.8 91.3Teaching and non-DSH 340 �31.0 �20.3 �12.0 �4.6 2.7 86.5Nonteaching and DSH 954 �32.6 �23.6 �15.3 �8.8 �1.2 90.9Nonteaching and non-DSH 1,972 �31.4 �23.0 �16.0 �9.2 �2.5 93.1

New England 172 �25.2 �18.1 �13.8 �8.2 �4.0 95.3Middle Atlantic 452 �29.2 �20.3 �11.0 �1.4 6.4 78.5South Atlantic 586 �29.2 �21.5 �15.3 �9.1 �3.3 94.9East North Central 638 �32.3 �24.5 �16.8 �10.0 �1.9 92.5East South Central 303 �35.6 �24.5 �17.9 �10.6 �5.0 96.0West North Central 444 �29.0 �21.4 �15.6 �9.3 �3.4 94.6West South Central 544 �34.3 �23.8 �15.2 �9.4 �3.3 94.9Mountain 261 �29.2 �21.1 �14.0 �7.5 �1.0 91.2Pacific 497 �37.4 �25.1 �15.8 �8.5 1.0 88.9

Voluntary 2,410 �29.5 �21.1 �14.4 �7.5 0.3 89.7Proprietary 627 �35.2 �25.1 �16.0 �8.8 �3.0 93.5Urban government 272 �39.4 �26.3 �16.9 �8.8 �1.6 92.6Rural government 586 �34.3 �24.6 �18.1 �11.6 �6.7 97.3

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teaching hospitals have aratio of less than 0.25. A margin is calculated as revenue minus costs divided by revenue; margins are based on Medicare-allowed costs. Data are imputed for hospitalswhose 2000 cost reports were not available (about 27 percent of observations). Excludes critical access hospitals.

Source: MedPAC analysis of Medicare cost report and impact file data (fourth quarter 2002) from CMS.

T A B L ED-7


Hospital-based Medicare skilled nursing facilitymargins excluding graduate medical


Hospital group 1996 1997 1998 1999 2000

All hospitals �10.2% �12.0% �25.3% �56.0% �57.3%

Urban �10.2 �12.2 �25.4 �54.9 �56.7Rural �10.6 �11.0 �24.6 �60.8 �60.0







Non-DSH �9.8 �11.4 �25.5 �56.1 �57.3




Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater,while other teaching hospitals have a ratio of less than 0.25. A margin is calculated as revenue minus costs divided by revenue;margins are based on Medicare-allowed costs. 2000 values are imputed for hospitals whose 2000 cost reports were notavailable (about 27 percent of observations). Excludes critical access hospitals.


T A B L ED-8


Hospital-based Medicare home health agencymargins excluding graduate medical


Hospital group 1996 1997 1998 1999 2000

All hospitals �4.4% �4.0% �24.1% �13.2% �10.0%

Urban �4.5 �3.9 �22.3 �11.9 �8.9Rural �4.2 �4.3 �30.0 �17.6 �13.9







Non-DSH �4.5 �4.0 �24.6 �13.4 �10.0






T A B L ED-9


Hospital Medicare PPS-exempt unit marginsexcluding graduate medical education,

by hospital group, 1996–2000

Hospital group 1996 1997 1998 1999 2000

All hospitals 4.3% 2.3% �1.7% �2.3% �0.7%

Urban 3.9 1.8 �2.3 �2.7 �1.0Rural 7.2 5.9 2.2 0.4 1.4

Large urban 4.3 1.7 �3.1 �4.4 �2.3Other urban 3.3 2.0 �1.0 �0.1 1.1Rural referral 10.5 10.1 9.3 5.0 4.5Sole community �0.7 �4.2 �11.9 �7.9 �8.3Small rural Medicare-dependent 7.0 2.4 �6.4 �0.2 0.1Other rural � 50 beds �0.6 1.6 �5.5 �5.3 �5.7Other rural � 50 beds 8.2 7.1 1.8 �0.6 4.7

Major teaching 1.5 �0.3 �8.7 �8.2 �7.3Other teaching 4.2 0.9 �1.4 �3.2 0.2Nonteaching 5.7 4.6 1.2 1.0 1.6

Major teachingPublic �5.3 �0.6 �26.3 �31.3 �24.3Private 3.0 �0.2 �3.1 �1.6 �1.8

Other teachingPublic 7.1 5.1 �5.1 �5.3 �0.3Private 4.0 0.6 �1.1 �3.0 0.2

NonteachingPublic 4.4 3.1 �0.5 0.1 �2.2Private 5.9 4.8 1.4 1.2 2.2

DSHLarge urban 3.9 1.5 �4.2 �6.7 �4.1Other urban 3.9 1.8 �0.7 �0.5 1.1Rural 11.3 13.4 8.6 7.1 7.0

Non-DSH 4.1 2.2 �1.0 �0.3 0.6

Teaching and DSH 2.8 0.5 �5.5 �6.6 �3.8Teaching and non-DSH 4.0 0.4 �0.6 �0.6 0.6Nonteaching and DSH 7.3 5.8 3.5 2.2 2.6Nonteaching and non-DSH 4.1 3.3 �1.2 �0.2 0.5

New England �4.9 �7.3 �2.3 �1.8 �4.3Middle Atlantic 4.9 2.6 �7.3 �6.6 �4.4South Atlantic 3.8 5.6 3.5 2.4 3.3East North Central 3.4 �0.2 �2.3 �1.1 0.8East South Central 4.1 1.6 �0.9 1.3 1.0West North Central 4.4 1.9 �3.4 �4.5 �2.7West South Central 3.2 3.3 0.2 �5.2 �2.1Mountain 9.4 �1.3 �5.6 �1.1 2.1Pacific 10.5 7.2 0.7 �1.9 �0.5

Voluntary 3.8 1.2 �0.8 �0.9 0.0Proprietary 8.1 7.1 1.8 �0.7 4.3Urban government 1.4 2.9 �12.4 �14.3 �11.3Rural government 2.2 2.5 �3.9 �2.7 �4.8

Note: DSH (disproportionate share). PPS (prospective payment system). PPS-exempt units include inpatient psychiatric and rehabilitationservices. Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teachinghospitals have a ratio of less than 0.25. A margin is calculated as revenue minus costs divided by revenue; margins are based onMedicare-allowed costs. 2000 values are imputed for hospitals whose 2000 cost reports were not available (about 27 percent ofobservations). Excludes critical access hospitals.


T A B L ED-10


Overall Medicare margins including payments for direct graduate medical


Hospital group 1996 1997 1998 1999 2000

All hospitals 9.9% 10.4% 6.5% 5.1% 5.0%

Urban 10.7 11.5 7.8 6.4 6.3Rural 5.0 4.1 �1.2 �2.5 �2.9

Large urban 12.3 13.2 9.5 8.4 8.4Other urban 8.2 8.8 5.1 3.3 2.9Rural referral 5.9 5.4 0.0 �1.7 �1.9Sole community 6.1 4.8 �0.7 �1.6 �2.1Small rural Medicare-dependent 3.2 3.3 �0.6 �0.7 �1.5Other rural � 50 beds 2.4 1.7 �4.0 �3.4 �4.3Other rural � 50 beds 4.2 2.9 �3.1 �5.0 �5.6

Major teaching 17.2 19.0 15.3 13.7 14.9Other teaching 9.6 10.1 6.7 5.7 5.0Nonteaching 6.5 6.7 2.0 0.1 �0.2

Major teachingPublic 18.3 19.5 15.1 10.6 11.4Private 16.8 18.9 15.3 14.4 15.7

Other teachingPublic 9.5 11.0 4.4 4.4 2.6Private 9.7 10.1 6.9 5.8 5.2

NonteachingPublic 3.6 2.9 �1.8 �3.2 �4.8Private 7.1 7.4 2.6 0.7 0.6

DSHLarge urban 15.5 16.1 12.4 11.3 11.5Other urban 10.0 10.5 6.7 4.8 4.2Rural 7.5 6.4 0.3 �1.0 �1.7

Non-DSH 5.6 6.2 2.1 0.8 0.7

Teaching and DSH 14.4 15.2 11.7 10.3 10.4Teaching and non-DSH 8.4 9.4 6.0 5.4 5.5Nonteaching and DSH 10.0 9.8 4.9 3.2 2.8Nonteaching and non-DSH 3.8 4.2 �0.5 �2.4 �2.6

New England 10.4 11.7 9.0 7.7 9.0Middle Atlantic 12.0 13.7 11.5 11.7 13.5South Atlantic 9.1 9.7 5.9 4.4 2.4East North Central 6.7 7.3 2.7 0.9 0.9East South Central 10.2 9.4 4.5 3.5 2.8West North Central 5.5 5.5 1.7 0.0 0.9West South Central 10.1 10.1 6.5 3.8 3.4Mountain 10.5 10.5 5.2 2.6 1.0Pacific 15.1 14.4 9.8 7.2 6.4

Voluntary 9.3 10.2 6.3 5.0 5.1Proprietary 13.6 13.0 10.0 8.2 7.7Urban government 12.3 12.6 7.8 5.3 4.5Rural government 2.6 1.4 �3.8 �4.8 �5.9



T A B L ED-11


Distribution of overall Medicare margins including payments for directgraduate medical education, by hospital group, 2000



All hospitals 3,866 �18.8% �9.3% 0.1% 9.1% 17.6% 49.8%

Urban 2,325 �15.9 �6.9 2.5 10.6 19.3 43.1Rural 1,541 �22.4 �12.6 �3.6 5.5 13.9 60.0

Large urban 1,338 �14.7 �5.1 4.5 13.2 21.7 37.7Other urban 987 �16.9 �8.8 �0.1 7.8 14.4 50.4Rural referral 207 �17.9 �10.8 �3.3 4.7 11.5 60.9Sole community 487 �22.2 �12.0 �1.9 7.9 17.3 57.3Small rural Medicare-dependent 203 �20.7 �9.9 0.2 7.7 15.0 49.8Other rural � 50 beds 324 �28.4 �14.4 �4.2 4.8 14.1 59.0Other rural � 50 beds 320 �24.9 �13.8 �6.0 2.1 10.5 70.9

Major teaching 252 �1.4 7.6 15.2 22.2 28.9 11.5Other teaching 715 �9.4 �2.4 4.2 11.0 19.2 33.4Nonteaching 2,899 �21.3 �11.7 �2.4 6.3 14.4 57.2

Major teachingPublic 55 �2.2 1.8 12.3 19.6 26.1 16.4Private 197 �0.8 8.8 16.3 22.6 29.6 10.2

Other teachingPublic 44 �11.0 �5.1 0.7 6.1 12.5 45.5Private 671 �9.3 �2.0 4.6 11.2 19.6 32.6

NonteachingPublic 752 �26.0 �14.7 �4.6 4.6 13.4 63.0Private 2,147 �19.4 �10.9 �1.8 6.9 15.0 55.1

DSHLarge urban 699 �8.2 0.5 9.1 17.8 24.7 23.7Other urban 544 �12.4 �4.2 2.3 9.7 16.4 41.4Rural 336 �22.1 �10.6 �1.3 8.3 17.2 53.9

Non-DSH 2,287 �21.9 �12.5 �3.1 5.5 13.2 59.2

Teaching and DSH 630 �5.8 0.8 8.8 17.4 25.1 22.5Teaching and non-DSH 337 �12.4 �4.8 2.9 11.0 18.5 37.4Nonteaching and DSH 949 �16.1 �7.2 1.5 10.0 18.0 45.3Nonteaching and non-DSH 1,950 �22.5 �13.6 �4.3 4.5 11.9 62.9

New England 171 �18.0 �9.7 1.6 10.8 19.3 44.4Middle Atlantic 448 �9.9 �1.4 7.0 16.4 26.4 30.4South Atlantic 583 �15.8 �8.6 �0.9 8.3 14.4 53.0East North Central 635 �22.7 �13.9 �5.3 3.8 12.1 63.9East South Central 300 �14.5 �4.6 2.5 10.4 17.4 39.7West North Central 443 �19.9 �13.2 �4.2 4.3 11.9 62.5West South Central 538 �19.4 �7.9 1.0 9.0 17.6 46.5Mountain 259 �23.2 �11.7 �1.6 8.4 16.1 54.8Pacific 489 �18.9 �8.0 2.5 12.0 21.1 42.9

Voluntary 2,396 �17.2 �8.8 0.1 9.0 17.8 49.6Proprietary 619 �14.7 �4.8 3.9 12.7 19.9 37.6Urban government 270 �20.4 �11.0 �0.4 9.0 18.2 50.4Rural government 581 �26.6 �14.4 �4.5 4.5 13.6 63.2

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teaching hospitals have aratio of less than 0.25. A margin is calculated as revenue minus costs divided by revenue; margins are based on Medicare-allowed costs. Data are imputed for hospitalswhose 2000 cost reports were not available (about 27 percent of observations). Excludes critical access hospitals.


T A B L ED-12


Hospital payment-to-cost ratios,by source of revenue, 1991–2001

Uncompensated PrivateYear Medicare Medicaid care payers

1991 88.4% 81.6% 19.6% 129.7%1992 88.8 90.9 18.9 131.31993 89.4 93.1 19.5 129.31994 96.9 93.7 19.3 124.41995 99.3 93.8 18.0 123.91996 102.4 94.8 17.3 121.51997 103.6 95.9 14.1 117.61998 102.6 97.9 13.2 113.61999 101.1 96.7 13.2 112.32000 100.2 96.1 12.1 112.52001 99.4 98.0 12.2 113.2

Note: Payment-to-cost ratios indicate the relative degree to which payments from each payer cover the costs oftreating its patients. Operating subsidies from state and local governments are considered payments foruncompensated care, up to the level of each hospital’s uncompensated care costs. Data are for communityhospitals and reflect all types of patient care services. Imputed values are used for missing data (about 35percent of observations),which corrects for underrepresentation of proprietary and public hospitals relative tovoluntary institutions. Most Medicare and Medicaid managed care patients are included in the private payerscategory. The costs allocated to Medicare and Medicaid include CMS’s allowed and nonallowed costs.

Source: MedPAC analysis of data from the American Hospital Association annual survey of hospitals.

T A B L ED-13

Gains or losses as a percent of total hospital costs, by source of revenue, 1991–2000

Othergovernmentpayers and Uncompensated Private Total

Year Medicare Medicaid subsidies care payers Nonpatient gains

1991 �4.4% �2.3% 0.4% �4.8% 11.6% 3.5% 4.0%1992 �4.4 �1.2 0.2 �4.9 11.8 3.3 4.81993 �4.1 �0.9 0.2 �4.8 10.9 3.3 4.41994 �1.2 �0.9 0.2 �4.9 8.7 3.1 5.01995 �0.3 �0.9 �0.1 �5.0 8.5 3.7 6.01996 0.9 �0.7 �0.1 �5.1 7.9 4.3 7.21997 1.4 �0.5 �0.1 �5.2 6.7 4.9 7.21998 1.0 �0.2 0.0 �5.2 5.5 5.1 6.11999 0.4 �0.4 0.1 �5.4 5.2 5.1 4.92000 0.1 �0.4 0.1 �5.3 5.4 5.1 4.8

Note: Gains or losses are the difference between the cost of providing care (or operating a nonpatient service) and the payment received. Operating subsidies from state andlocal governments are considered payments for uncompensated care, up to the level of each hospital’s uncompensated care costs. Subsidies in excess of uncompensatedcare costs are combined with revenue from other government payers. Nonpatient reflects both other operating and nonoperating revenue. Data are for community hospitalsand reflect both inpatient and outpatient services. Imputed values are used for missing data (about 35 percent of observations), which corrects for underrepresentation ofproprietary and public hospitals relative to voluntary institutions. Most Medicare and Medicaid managed care patients are included in the private payers category. Gainsand losses from the sources shown sum to total gains (except due to rounding). The costs allocated to Medicare and Medicaid include CMS’s allowed and nonallowedcosts.

Source: MedPAC analysis of data from the American Hospital Association annual survey of hospitals.

T A B L ED-14


Hospital total margins, by hospital group, 1991–2000

Hospital group 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

All hospitals 4.4% 4.4% 4.4% 5.0% 5.8% 6.1% 6.0% 4.3% 3.8% 3.4%

Urban 4.3 4.3 4.3 4.9 5.6 5.9 5.8 4.2 3.6 3.3Rural 5.2 5.3 5.2 5.6 6.6 7.3 6.9 5.1 4.9 4.4

Large urban 3.7 3.7 3.9 4.3 4.9 5.2 5.2 3.8 3.0 2.7Other urban 5.5 5.2 5.2 6.0 6.9 7.2 7.0 5.0 4.7 4.3Rural referral 6.7 6.9 6.3 6.8 8.4 9.2 9.3 6.9 7.4 6.4Sole community 5.1 5.1 5.1 5.6 5.7 6.5 6.1 4.8 3.9 3.9Small rural Medicare- 3.1 2.4 3.9 3.3 3.9 4.7 4.5 2.3 3.3 2.7

dependentOther rural � 50 beds 2.2 2.3 2.5 2.1 2.8 4.3 2.8 1.1 1.3 1.0Other rural � 50 beds 4.5 4.8 4.7 5.6 6.7 7.1 5.9 4.6 3.3 3.4

Major teaching 3.7 3.4 3.4 3.3 4.0 3.5 4.8 3.3 2.8 1.5Other teaching 4.6 4.5 4.6 5.3 6.3 7.0 6.5 4.2 3.9 4.2Nonteaching 4.8 5.0 4.9 5.9 6.5 7.1 6.3 5.0 4.3 4.1

Major teachingPublic 4.5 4.2 4.5 2.8 3.1 2.8 5.1 4.5 3.2 �0.1Private 3.3 3.0 3.0 3.4 4.3 3.7 4.7 2.9 2.6 2.1

Other teachingPublic 5.4 4.2 4.4 3.8 4.9 6.0 4.2 3.1 2.7 2.9Private 4.6 4.5 4.7 5.5 6.4 7.1 6.6 4.3 4.0 4.3

NonteachingPublic 4.3 4.6 4.2 4.7 5.5 6.0 5.7 4.0 3.1 3.3Private 4.8 5.0 5.1 6.1 6.7 7.3 6.4 5.2 4.6 4.2

DSHLarge urban 3.2 3.4 3.6 3.9 4.4 4.4 4.7 3.2 2.5 1.9Other urban 5.9 5.6 5.5 6.3 6.9 7.2 7.2 5.1 4.8 4.2Rural 7.2 7.5 5.8 6.1 7.2 7.8 6.9 4.9 4.7 4.2

Non-DSH 4.6 4.5 4.6 5.3 6.3 7.1 6.4 4.9 4.4 4.4

Teaching and DSH 4.0 4.0 4.0 4.2 4.8 4.9 5.5 3.6 3.0 2.4Teaching and non-DSH 4.9 4.0 4.5 4.9 6.5 7.1 6.5 4.6 4.6 4.7Nonteaching and DSH 5.1 5.2 5.3 6.3 6.7 7.2 6.3 4.8 4.5 4.0Nonteaching and non-DSH 4.5 4.7 4.6 5.5 6.2 7.1 6.3 5.1 4.2 4.2

New England 2.2 2.2 3.1 2.6 3.0 4.0 5.1 2.7 1.5 1.6Middle Atlantic 1.4 0.9 1.9 2.6 3.0 3.1 3.5 1.6 0.4 0.9South Atlantic 6.0 6.2 5.7 6.6 7.5 8.4 8.0 5.7 6.0 4.2East North Central 4.8 4.8 4.8 5.6 6.3 6.4 6.9 4.7 5.3 3.9East South Central 6.4 5.6 4.9 5.2 6.6 7.3 5.0 3.4 3.2 3.2West North Central 4.9 4.5 4.7 6.6 7.3 7.4 7.8 5.8 4.9 4.9West South Central 5.8 7.4 6.2 6.7 7.4 7.3 6.4 5.6 4.2 3.9Mountain 5.5 5.4 7.0 7.4 7.7 8.2 4.5 5.3 3.7 4.3Pacific 4.7 4.1 4.1 3.6 4.4 4.5 5.3 4.5 3.6 4.4

Voluntary 4.3 4.1 4.1 4.7 5.7 5.8 6.2 4.0 3.1 3.2Proprietary 5.0 6.3 6.9 8.9 8.3 9.8 5.7 6.8 9.0 7.7Urban government 4.6 4.2 4.3 3.5 4.0 4.0 5.1 4.2 3.0 1.2Rural government 4.6 5.0 4.5 4.7 5.8 6.5 5.3 3.7 3.5 3.5

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teaching hospitals have a ratioof less than 0.25. A margin is calculated as revenue minus costs divided by revenue. Total margin includes all patient care services funded by all payers, plus nonpatientrevenue. 2000 values are imputed for hospitals whose 2000 cost reports were not available (about 27 percent of observations). Excludes critical access hospitals.


T A B L ED-15


Distribution of hospital total margins, by hospital group, 2000



All hospitals 4,051 �9.7% �1.9% 2.7% 7.0% 12.0% 32.8%

Urban 2,402 �10.6 �2.2 2.7 7.2 13.0 33.3Rural 1,649 �8.5 �1.6 2.6 6.5 10.8 31.9

Large urban 1,365 �11.1 �3.0 2.1 6.9 13.1 36.9Other urban 1,037 �9.9 �1.1 3.6 7.4 12.5 28.6Rural referral 217 �1.5 2.3 5.8 8.3 14.2 14.7Sole community 506 �8.6 �1.7 2.5 6.5 11.0 34.6Small rural Medicare-dependent 222 �9.9 �3.1 1.8 5.4 9.6 37.8Other rural � 50 beds 352 �10.4 �3.1 1.6 5.6 9.1 36.4Other rural � 50 beds 352 �6.3 �1.0 2.7 6.2 9.8 30.4

Major teaching 269 �10.8 �3.2 0.8 4.6 8.8 41.3Other teaching 724 �7.3 �0.6 3.3 7.1 12.1 27.9Nonteaching 3,058 �10.0 �2.0 2.7 7.1 12.3 33.2

Major teachingPublic 71 �14.8 �5.6 �0.3 3.1 8.8 50.7Private 198 �7.8 �2.2 1.4 4.9 8.7 37.9

Other teachingPublic 50 �6.3 �2.1 1.8 5.4 10.3 34.0Private 674 �7.3 �0.5 3.4 7.1 12.2 27.4

NonteachingPublic 813 �8.6 �1.9 2.2 6.0 9.6 32.3Private 2,245 �10.5 �2.0 2.9 7.5 13.5 33.5

DSHLarge urban 732 �11.6 �3.7 1.2 5.9 12.3 40.2Other urban 567 �9.0 �0.4 3.7 7.5 13.1 26.5Rural 369 �12.4 �3.6 1.9 6.6 10.3 38.8

Non-DSH 2,383 �8.5 �1.5 3.0 7.1 12.0 31.1

Teaching and DSH 665 �9.6 �2.2 1.8 5.8 11.1 34.6Teaching and non-DSH 328 �6.6 0.0 3.8 8.0 12.0 25.3Nonteaching and DSH 1,003 �12.4 �3.3 2.3 7.3 12.8 35.6Nonteaching and non-DSH 2,055 �8.6 �1.6 2.9 7.0 12.1 32.0

New England 171 �7.2 �1.0 2.3 4.6 10.0 30.4Middle Atlantic 466 �11.1 �3.2 0.7 4.0 7.2 42.1South Atlantic 594 �11.0 �1.6 3.2 7.9 14.8 31.3East North Central 648 �6.8 �0.4 3.8 7.4 11.4 26.2East South Central 382 �12.4 �3.6 2.1 5.9 9.6 39.0West North Central 488 �4.7 0.0 3.6 7.3 12.1 24.6West South Central 548 �11.8 �3.9 2.3 7.1 12.6 39.6Mountain 262 �7.9 �1.3 4.1 8.7 14.1 27.9Pacific 492 �10.5 �2.2 2.3 7.6 14.8 33.3

Voluntary 2,479 �8.5 �1.3 2.8 6.5 10.6 31.2Proprietary 638 �15.1 �4.1 3.4 13.8 21.5 37.3Urban government 300 �10.7 �2.1 2.0 5.4 9.3 34.7Rural government 632 �8.7 �2.2 2.0 6.2 9.6 33.4

Note: DSH (disproportionate share). Major teaching hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while other teaching hospitals have aratio of less than 0.25. A margin is calculated as revenue minus costs divided by revenue. Total margin includes all patient care services funded by all payers, plusnonpatient revenue. Data are imputed for hospitals whose 2000 cost reports were not available (about 27 percent of observations). Excludes critical access hospitals.


T A B L ED-16

Commissioners’ voting on recommendations

A P P E N D I XE

In the Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000, the Congress required MedPAC to call forindividual Commissioner votes on each recommendation, and to document the voting record in its report. The information below satisfiesthat mandate.

Chapter 1: Context for Medicare spendingNo recommendations

Chapter 2: Assessing payment adequacy and updating payments in fee-for-service Medicare

Section 2A: Hospital inpatient and outpatient services

2A-1 The Secretary should add 13 DRGs to the post-acute transfer policy in fiscal year 2004 and then evaluate the effects on hospitalsand beneficiaries before proposing further expansions.

Yes: Burke, DeBusk, DeParle, Durenberger, Feezor, Hackbarth, Muller, Nelson, Newhouse, Raphael, Reischauer,Rosenblatt, Rowe, Smith, Stowers

No: WolterNot voting: Wakefield

2A-2 The Congress should enact a low-volume adjustment to the rates used in the inpatient PPS. This adjustment should apply only tohospitals that are more than 15 miles from another facility offering acute inpatient care.

Yes: Burke, DeBusk, DeParle, Durenberger, Feezor, Hackbarth, Muller, Nelson, Newhouse, Raphael, Reischauer,Rosenblatt, Rowe, Smith, Stowers, Wakefield, Wolter

2A-3 The Secretary should reevaluate the labor share used in the wage index system that geographically adjusts rates in the inpatientPPS, with any resulting change phased in over two years.

Yes: Burke, DeBusk, DeParle, Durenberger, Feezor, Hackbarth, Muller, Nelson, Raphael, Reischauer, Rosenblatt, Rowe,Smith, Stowers, Wakefield, Wolter

Not voting: Newhouse

2A-4 The Congress should raise the inpatient base rate for hospitals in rural and other urban areas to the level of the rate for those in largeurban areas, phased in over two years.


A P P E N D I X

Commissioners’ voting on recommendations

E


2A-5 The Congress should raise the cap on the disproportionate share add-on a hospital can receive in the inpatient PPS from 5.25percent to 10 percent, phased in over two years.

Yes: Burke, DeBusk, Durenberger, Feezor, Hackbarth, Muller, Nelson, Newhouse, Reischauer, Rosenblatt, Rowe,Smith, Stowers, Wakefield, Wolter

No: DeParleNot voting: Raphael

2A-6 The Congress should increase payment rates for the inpatient PPS by the rate of increase in the hospital market basket, less0.4 percent, for fiscal year 2004.


2A-7 The Congress should increase payment rates for the outpatient PPS by the rate of increase in the hospital market basket, less0.9 percent, for calendar year 2004.


Section 2B: Physician services

2B The Congress should update payments for physician services by the projected change in input prices, less an adjustment forproductivity growth of 0.9 percent, for 2004.

Yes: Burke, DeBusk, DeParle, Durenberger, Feezor, Hackbarth, Muller, Nelson, Newhouse, Raphael, Reischauer,Rosenblatt, Smith, Stowers, Wakefield, Wolter

Absent: Rowe

Section 2C: Skilled nursing facility services

2C-1 The Secretary should continue a series of nationally representative studies on access to skilled nursing facility services(similar to studies previously conducted by the Department of Health and Human Services’ Office of Inspector General).


Absent: Rowe

2C-2 The Congress should eliminate the update to payment rates for skilled nursing facility services for fiscal year 2004.


Absent: Rowe

2C-3A Consistent with previous MedPAC recommendations, the Secretary should develop a new classification system for care inskilled nursing facilities.

Because it may take time to develop this system, the Secretary should draw on new and existing research to reallocatepayments to achieve a better balance of available resources between the rehabilitation and nonrehabilitation groups.

To allow for immediate reallocation of resources, the Congress should give the Secretary the authority to:

284 Comm i s s i o ne r s ’ v o t i ng on r e commenda t i o n s

• remove some or all of the 6.7 percent payment add-on currently applied to the rehabilitation RUG–III groups.

• reallocate money to the nonrehabilitation RUG–III groups to achieve a better balance of resources among all of theRUG–III groups.

2C-3B If necessary action does not occur within a timely manner, the Congress should provide for a market basket update, less anadjustment for productivity growth of 0.9 percent, for hospital-based skilled nursing facilities to be effective October 1,2003.


Section 2D: Home health services

2D-1 The Secretary should continue a series of nationally representative studies on access to home health services (similar tostudies previously conducted by the Department of Health and Human Services’ Office of Inspector General).


Absent: Rowe

2D-2 The Congress should extend for one year add-on payments at 5 percent for home health services provided to Medicarebeneficiaries who live in rural areas.


Absent: Rowe

2D-3 The Congress should eliminate the update to payment rates for home health services for fiscal year 2004.

Yes: Burke, DeBusk, DeParle, Durenberger, Feezor, Hackbarth, Muller, Nelson, Newhouse, Reischauer, Rosenblatt,Smith, Stowers, Wakefield, Wolter

Not voting: RaphaelAbsent: Rowe

Section 2E: Outpatient dialysis services

2E The Congress should update the composite rate payment by the projected change in input prices, less 0.9 percent, forcalendar year 2004.

Yes: Burke, DeBusk, DeParle, Durenberger, Hackbarth, Muller, Nelson, Newhouse, Raphael, Reischauer,Rosenblatt, Smith, Stowers, Wakefield, Wolter

Not voting: FeezorAbsent: Rowe


Section 2F: Ambulatory surgical center services

2F-1 The Secretary should expedite collection of recent ASC charge and cost data for the purpose of analyzing and revising the ASCpayment system.


Absent: Rowe

2F-2 The Congress should eliminate the update to payment rates for ASC services for fiscal year 2004.

Yes: Burke, DeBusk, Durenberger, Feezor, Hackbarth, Muller, Nelson, Newhouse, Raphael, Reischauer, Rosenblatt,Smith, Stowers, Wakefield, Wolter

Not voting: DeParleAbsent: Rowe

2F-3 Until the Secretary implements a revised ASC payment system, the Congress should ensure that payment rates for ASCprocedures do not exceed hospital outpatient PPS rates for those procedures, after accounting for differences in the bundle ofservices covered.

Yes: Burke, DeBusk, Durenberger, Feezor, Hackbarth, Muller, Nelson, Newhouse, Raphael, Reischauer, Rosenblatt,Smith, Stowers, Wakefield, Wolter

Not voting: DeParleAbsent: Rowe

Chapter 3: Access to care in the Medicare program No recommendations

Chapter 4: Payment for new technologies in Medicare’s prospective payment systems The Secretary should introduce clinical criteria for eligibility of drugs and biologicals to receive pass-through payments under theoutpatient prospective payment system.


Absent: Rowe

Chapter 5: Health insurance choices for Medicare beneficiaries No recommendations

286 Comm i s s i o ne r s ’ v o t i ng on r e commenda t i o n s

Acronyms

AAHP American Association of Health Plans

ACE–PRO Access to Care for the Elderly Project

ADL activity of daily living

AHA American Hospital Association

AHCA American Health Care Association

AHRQ Agency for Healthcare Research and Quality

AIC minimum amount in controversy

ALJ administrative law judge

AMA American Medical Association

APC ambulatory payment classification or ambulatory payment category

ASC ambulatory surgical center

AWP average wholesale price

BBA Balanced Budget Act of 1997

BBRA Balanced Budget Refinement Act of 1999

BIPA Medicare, Medicaid, and SCHIP Benefits Improvement and ProtectionAct of 2000

BLS Bureau of Labor Statistics

CAC carrier advisory committee

CAH critical access hospital

CAHPS Consumer Assessment of Health Plans Survey

CalPERS California Public Employees’ Retirement System

CAT computerized automated tomography

CBO Congressional Budget Office

CC complication or comorbidity

CCP coordinated care plan

CMG case-mix group

CMS Centers for Medicare & Medicaid Services

COLA cost of living adjustment

CPI–U consumer price index for all urban consumers

CPT Current Procedural Terminology

DAB Departmental Appeals Board

DME durable medical equipment

DMERC durable medical equipment regional contractor

DoD Department of Defense

DRG diagnosis related group

DSH disproportionate share

ECI employment cost index

ED emergency department

ESRD end-stage renal disease

FDA Food and Drug Administration


Acronyms

FEHBP Federal Employees Health Benefits Program

FFS fee-for-service

FI fiscal intermediary

FY fiscal year

GAO General Accounting Office

GDP gross domestic product

GI gastrointestinal

GME graduate medical education

GPCI geographic practice cost index

GPO group purchasing organization or Government Printing Office

HCBC home and community-based care

HCFA Health Care Financing Administration

HCPCS Healthcare Common Procedure Coding System

HHA home health agency

HHRG home health resource group

HHS Department of Health and Human Services

HI Hospital Insurance (Medicare Part A)

HMO health maintenance organization

HPSA health professional shortage area

HSC Center for Studying Health System Change

HWI hospital wage index

HWIr hospital wage index with geographic reclassification

HWIu hospital wage index unreclassified

ICD–9–CM International Classification of Diseases, Ninth Revision, ClinicalModification

IDE investigational device exemption

IHS Indian Health Service

IME indirect medical education

IOM Institute of Medicine

IPS interim payment system

IRF inpatient rehabilitation facility

LMRP local medical review policy

LOS length of stay

LTC long-term care

LTC–DRG long-term care diagnosis related group

LTCH long-term care hospital

LUPA low utilization payment adjustment

M+C Medicare+Choice

MB market basket

MCAC Medicare Coverage Advisory Committee

MCBS Medicare Current Beneficiary Survey

MCO managed care organization

MDC major diagnostic category

MedPAC Medicare Payment Advisory Commission

MedPAR Medicare Provider Analysis and Review

290 Ac ronyms

MEI Medicare Economic Index

MEPS Medical Expenditure Panel Survey

MPR Mathematica Policy Research, Inc.

MRI magnetic resonance imaging

MSA metropolitan statistical area

NAIC National Association of Insurance Commissioners

NCD national coverage decision

NCFE National Century Financial Enterprises

NCHS National Center for Health Statistics

NHAMCS National Hospital Ambulatory Medical Care Survey

NHE national health expenditure

NHIS National Hospital Indicators Survey or National Health Interview Survey

NHS National Health Service (United Kingdom)

NICE National Institute for Clinical Excellence (United Kingdom)

NKF National Kidney Foundation

NLA national limitation amount

OACT Office of the Actuary

OASIS Outcomes and Assessment Information Set

OBRA Omnibus Budget Reconciliation Act

OIG Office of Inspector General

OMB Office of Management and Budget

OPD outpatient department

OR operating room

OSCAR Online Survey, Certification, and Reporting system

PAC post-acute care

PACE Program of All-Inclusive Care for the Elderly

PBM pharmaceutical benefit management organization

PE practice expense

PET positron emission tomography

PFFS private fee-for-service

PHI private health insurance

PLI professional liability insurance

POS point-of-service (plan)

PPO preferred provider organization

PPS prospective payment system

ProPAC Prospective Payment Assessment Commission

PTCA percutaneous transluminal coronary angioplasty

QI–1 Qualifying Individuals–1

QIO quality improvement organization

QMB qualified Medicare beneficiary

RUG–III resource utilization group, version III

S/HMO Social Health Maintenance Organization

SCHIP State Children’s Health Insurance Program

SGR sustainable growth rate


SHIP State Health Insurance Assistance Program

SLMB specified low-income Medicare beneficiary

SMI Supplementary Medical Insurance (Medicare Part B)

SMS Socioeconomic Monitoring System

SNF skilled nursing facility

SSI Supplemental Security Income

TEFRA Tax Equity and Fiscal Responsibility Act of 1982

UK United Kingdom

USRDS United States Renal Data System

VA Department of Veterans Affairs

292 Ac ronyms

More about MedPAC


Commission members

Glenn M. Hackbarth, J.D., chairmanIndependent consultantBend, OR

Robert D. Reischauer, Ph.D., vice chairmanThe Urban InstituteWashington, DC

Term expires April 2004

Sheila P. Burke, M.P.A.,R.N., F.A.A.N.Smithsonian InstitutionWashington, DC

Allen D. FeezorCalifornia Public Employees’Retirement SystemSacramento, CA

Ralph W. MullerStockap & AssociatesChicago, IL

Joseph P. Newhouse, Ph.D.Harvard UniversityBoston, MA

Alice Rosenblatt, F.S.A.,M.A.A.A.WellPoint Health NetworksThousand Oaks, CA

John W. Rowe, M.D.Aetna Inc.Hartford, CT


Nancy-Ann DeParle, J.D.JPMorgan PartnersWashington, DC

David F. DurenbergerNational Institute of Health PolicyWashington, DC

Carol RaphaelVisiting Nurse Service of New YorkNew York, NY

Mary K. Wakefield, Ph.D.,R.N., F.A.A.N.Center for Rural HealthUniversity of North DakotaGrand Forks, ND

Nicholas J. Wolter, M.D.Deaconess Billings ClinicBillings, MT


Autry O.V. “Pete” DeBuskDeRoyalPowell, TN

Glenn M. Hackbarth, J.D.

Alan R. Nelson, M.D.American College of Physicians-American Society of Internal MedicineWashington, DC

Robert D. Reischauer, Ph.D.

David A. SmithAFL–CIOWashington, DC

Ray E. Stowers, D.O.Oklahoma State University College of Osteopathic MedicineTulsa, OK

Commissioners’ biographies

Sheila P. Burke, M.P.A., R.N., F.A.A.N., is the Smithsonian Institution’sundersecretary for American Museums and National Programs. Before joining theSmithsonian, she was executive dean and lecturer in public policy at the John F. KennedySchool of Government, Harvard University, Cambridge. From 1986 to 1996, Ms. Burkewas chief of staff for former Senate Majority Leader Bob Dole and was elected Secretaryof the Senate in 1995. She currently serves as a board member of the Kaiser FamilyFoundation, the Kaiser Commission on Medicaid and the Uninsured, the Center forHealth Care Strategies, Inc., the Academy for Health Services Research and HealthPolicy, the American Board of Internal Medicine Foundation, WellPoint HealthNetworks, Chubb Insurance, Community Health Systems, the University of SanFrancisco, and Marymount University. She also sits on the national advisory council atthe Center for State Health Policy and has chaired the National Academy of SocialInsurance’s project on Restructuring Medicare for the Long-Term. Ms. Burke holds aB.S. in nursing from the University of San Francisco and an M.P.A. from HarvardUniversity.

Autry O.V. “Pete” DeBusk is chairman, chief executive officer, and founder ofDeRoyal, a global supplier of medical products and services in the acute care, patientcare, wound care, and original equipment manufacturing markets. Mr. DeBusk formedhis first company in 1970 with a patent he received on an orthopedic product. In 1976 heconsolidated his many product lines into one company, DeRoyal Industries. A member ofseveral community organizations, Mr. DeBusk is also chairman of the Board of Trusteesat Lincoln Memorial University in Harrogate, Tenn., as well as a founder of the AutryO.V. DeBusk facility, Boys and Girls Club, Powell, Tenn. As an innovative leader in themedical industry, he received a prestigious award from Duke University in 2000recognizing his original contributions to orthopedic surgery. He received his B.S. degreefrom Lincoln Memorial University and attended graduate school at the University ofGeorgia.

Nancy-Ann DeParle, J.D., is a senior advisor to JPMorgan Partners, LLC, andadjunct professor at the Wharton School of the University of Pennsylvania. From 1997 to2000, she served as administrator of the Health Care Financing Administration (HCFA),which is now the Centers for Medicare & Medicaid Services. After resigning fromHCFA, Ms. DeParle became a joint fellow of Harvard University’s Institute of Politics atthe Kennedy School of Government and the Interfaculty Health Policy Forum. Beforejoining HCFA, Ms. DeParle was associate director for health and personnel at the WhiteHouse Office of Management and Budget. From 1987 to 1989 she served as theTennessee Commissioner of Human Services. She has also worked as a lawyer in privatepractice in Nashville, Tenn., and Washington, DC. Ms. DeParle received a B.A. degreefrom the University of Tennessee; B.A. and M.A. degrees from Oxford University,where she was a Rhodes Scholar; and a J.D. degree from Harvard Law School.


David F. Durenberger, J.D., is president of Policy Insight, LLC; senior health policyfellow at the University of St. Thomas; and chairman and chief executive officer of theNational Institute of Health Policy. He is also chairman of Citizens for Long Term Care,president of the Medical Technology Leadership Forum, and a member of the KaiserFoundation Commission on Medicaid and the Uninsured. From 1978 to 1995, he servedas the senior U.S. Senator from Minnesota, as well as a member of the Senate FinanceCommittee and chair of its Health Subcommittee. He was a member of the SenateEnvironment Committee as well as the committee now known as the Health, Education,Labor, and Pensions Committee, and chaired the Senate Select Committee onIntelligence. He was vice chairman of the Pepper Commission and a member of theCongressional Bio-Ethics Commission, the Advisory Commission on IntergovernmentalRelations, and the Congressional Advisory Committee to the Office of TechnologyAssessment. Senator Durenberger is a graduate of St. John’s University, received his J.D.degree from the University of Minnesota, and served as an officer in the U.S. Army.

Allen D. Feezor is assistant executive officer, Health Benefit Services, CaliforniaPublic Employees’ Retirement System (CalPERS). Previously, Mr. Feezor was vicepresident for planning, marketing and managed care for University Health Systems ofEast Carolina in Greenville, NC. From 1985 to 1995, he was chief deputy commissionerfor the North Carolina Department of Insurance, where he chaired two national taskforces that pioneered state health insurance and small group reform. He has headed the430,000-member North Carolina Teachers,’ State Employees’ and Retirees’ Health Planand has served as Senior Representative in Washington, DC, for the Blue Cross/BlueShield Association. He was a founding faculty member of the National Academy forState Health Policy and a contributor to two Institute of Medicine studies—one on thefuture of health benefits and another on improving Medicare. He currently serves on theboards of Pacific Business Group on Health and the Integrated Health Association. Mr.Feezor earned his B.A. and M.A. degrees in political science from Duke University.

Glenn M. Hackbarth, J.D., is chairman of the Commission and an independentconsultant living in Bend, Ore. He has experience as a health care executive, governmentofficial, and policy analyst. He was chief executive officer and one of the founders ofHarvard Vanguard Medical Associates, a multispecialty group practice in Boston thatserves as a major teaching affiliate of Harvard Medical School. Harvard Vanguard wascreated from the staff-model delivery system that was the original core of HarvardCommunity Health Plan. Mr. Hackbarth previously served as senior vice president ofHarvard Community Health Plan. From 1981 to 1988, he held positions at theDepartment of Health and Human Services, including deputy administrator of the HealthCare Financing Administration. Mr. Hackbarth received his B.A. from Penn StateUniversity and his M.A. and J.D. degrees from Duke University.

Ralph W. Muller is managing director of Stockap & Associates, a hospital consultingfirm. In 2001 to 2002, he was a visiting fellow at the King’s Fund in London. Until July2001, he was president and chief executive officer of the University of Chicago Hospitalsand Health Systems (UCHHS), a position he held since 1985. Before joining the hospital,he held senior positions with the Commonwealth of Massachusetts, including deputycommissioner of the Department of Public Welfare. Mr. Muller is past chairman of theAssociation of American Medical Colleges, past chairman of the Council of TeachingHospitals and Health Systems, and past vice-chairman of the University Health SystemConsortium. He is past chairman of the National Opinion Research Center, a socialservice research organization. Mr. Muller received his B.A. in economics from SyracuseUniversity and his M.A. in government from Harvard University.

298 Comm i s s i o ne r s ’ b i og raph i e s

Alan R. Nelson, M.D., is an internist-endocrinologist who was in private practice inSalt Lake City until becoming chief executive officer of the American Society of InternalMedicine (ASIM) in 1992. Following the merger of ASIM with the American College ofPhysicians (ACP) in 1998, Dr. Nelson headed the Washington office of ACP-ASIM untilhis semi-retirement in January 2000 and now serves as special advisor to the executivevice president and chief executive officer. He was president of the American MedicalAssociation in 1989 to 1990. Dr. Nelson also serves on the Board of Trustees ofIntermountain HealthCare, a large integrated health system headquartered in Salt LakeCity. A member of the prestigious Institute of Medicine of the National Academy ofSciences (IOM), he serves on the IOM Roundtable on Environmental Health SciencesResearch and Medicine and was chair of the study committee on Rural and EthnicDisparities in Health Care. Dr. Nelson received his M.D. from Northwestern University.

Joseph P. Newhouse, Ph.D., is the John D. MacArthur Professor of Health Policyand Management at Harvard University and director of Harvard’s Division of HealthPolicy Research and Education. At Harvard since 1988, Dr. Newhouse was previously asenior corporate fellow and head of the economics department at RAND. He hasconducted research in health care financing, economics, and policy, and was the principalinvestigator for the RAND Health Insurance Experiment. Recipient of severalprofessional awards, he is a member of the Institute of Medicine, a former chair of theProspective Payment Assessment Commission, and a former member of the PhysicianPayment Review Commission. He is also a past president of the Association for HealthServices Research and the International Health Economics Association and has beenelected to the American Academy of Arts and Sciences. Dr. Newhouse is editor of theJournal of Health Economics. He received a B.A. from Harvard College and a Ph.D. ineconomics from Harvard University.

Carol Raphael is president and chief executive officer of the Visiting Nurse Service(VNS) of New York, the country’s largest voluntary home health care organization. VNSprograms include post-acute and long-term care, family and children services,rehabilitation, hospice, mental health, and public health, as well as a health plan fordually eligible Medicare and Medicaid beneficiaries. Ms. Raphael developed the Centerfor Home Care Policy and Research, which studies the management, cost, quality, andoutcomes of home- and community-based services. Previously, Ms. Raphael served asthe executive deputy commissioner of the Human Resources Administration in charge ofthe Medicaid and public assistance programs in New York City. Ms. Raphael has servedon several Robert Wood Johnson Foundation advisory committees and New York Statepanels, including the New York State Hospital Review and Planning Council, for whichshe chairs the Fiscal Policy Committee. She is on the boards of Excellus, Inc., the StatenIsland University Health System, and the American Foundation for the Blind, and is amember of the Pfizer Hispanic Advisory Board. She has an M.P.A. from HarvardUniversity’s Kennedy School of Government.

Robert D. Reischauer, Ph.D., is vice chairman of the Commission and president ofThe Urban Institute. Previously, he was a senior fellow with the Brookings Institutionand from 1989 to 1995 was the director of the Congressional Budget Office. Dr.Reischauer currently serves on the boards of the Academy of Political Sciences, theCenter on Budget and Policy Priorities, and the Committee for a Responsible FederalBudget. He also serves on the editorial board of Health Affairs, chairs the NationalAcademy of Social Insurance’s project on Restructuring Medicare for the Long-Term,and is a member of the Institute of Medicine and the National Academy of PublicAdministration. Dr. Reischauer received his A.B. degree from Harvard College and hisM.I.A. and Ph.D. from Columbia University.


Alice Rosenblatt, F.S.A., M.A.A.A., is chief actuary and executive vice president ofIntegration Planning and Implementation at WellPoint Health Networks. Before joiningWellPoint in 1996, she was a principal at Coopers & Lybrand LLP (now PNC), whereshe consulted with insurers, health plans, providers, and employers. She is a formersenior vice president and chief actuary of Blue Cross Blue Shield of Massachusetts andBlue Cross of California. Other positions include work for The New England andWilliam M. Mercer, Inc. Ms. Rosenblatt has served on the Board of Governors of theSociety of Actuaries and the American Academy of Actuaries. She previously chaired theAcademy’s federal health committee and work group on risk adjustment. Ms. Rosenblatthas testified on risk adjustment before subcommittees of the Committee on Ways andMeans and the Committee on Commerce of the U.S. House of Representatives. She has aB.S. and an M.A. in mathematics from the City College of New York and the CityUniversity of New York, respectively.

John W. Rowe, M.D., is chairman and chief executive officer of Aetna Inc., one ofthe nation’s largest healthcare insurers. Prior to joining Aetna, Dr. Rowe served aspresident and chief executive officer of Mount Sinai NYU Health. Previously, Dr. Rowewas president of The Mount Sinai Hospital and the Mount Sinai School of Medicine inNew York City, where he currently is a professor of medicine. Before joining MountSinai in 1988, Dr. Rowe was a professor of medicine and the founding director of theDivision on Aging at Harvard Medical School and chief of gerontology at Boston’s BethIsrael Hospital. He has authored over 200 scientific publications, mostly on thephysiology of the aging process, as well as a leading textbook of geriatric medicine. Dr.Rowe was director of the MacArthur Foundation Research Network on Successful Agingand is coauthor, with Robert Kahn, Ph.D., of Successful Aging (Pantheon, 1998). Heserved on the Board of Governors of the American Board of Internal Medicine and aspresident of the Gerontological Society of America, and is a member of the Institute ofMedicine of the National Academy of Sciences.

David A. Smith is senior policy advisor to the president of the AFL–CIO, where hepreviously served as director of the Public Policy Department. Prior to joining theAFL–CIO, he served as senior deputy budget director and as Commissioner of EconomicDevelopment for the City of New York. Mr. Smith spent most of the 1980s inWashington as an aide to Senator Edward M. Kennedy and as a senior economist at theJoint Economic Committee. Mr. Smith has taught economics and public policy at theUniversity of Massachusetts and the New School for Social Research, and is a seniorfellow at the Century Foundation. He is a member of the Board of Directors of PublicCampaign and of the National Bureau of Economic Research, a fellow of the NationalAcademy of Social Insurance, and a member of the Advisory Committee to the Export-Import Bank. Mr. Smith attended Tufts University and received an M.Ed. from HarvardUniversity.

Ray E. Stowers, D.O., is director of the Oklahoma Rural Health Policy and ResearchCenter as well as director of rural health in the Department of Family Medicine at theOklahoma State University College of Osteopathic Medicine. He was in private ruralpractice for 25 years at Family Medicine Clinics, Inc., in Medford, Okla., and is amember of the National Rural Health Association. Dr. Stowers is first vice president ofthe American Osteopathic Association and has served that organization in manycapacities, including several related to physician coding and reimbursement issues. Hehas been on the Physician Payment Review Commission and was a founding member ofthe American Medical Association’s Relative Value Update Committee. Dr. Stowersreceived his B.S. and B.A. degrees from Phillips University in Oklahoma and his D.O.degree from the University of Health Sciences College of Osteopathic Medicine inKansas City, Mo.

300 Comm i s s i o ne r s ’ b i og raph i e s

Mary K. Wakefield, Ph.D., R.N., F.A.A.N., is director and professor, Center forRural Health at the University of North Dakota. Dr. Wakefield has held administrativeand legislative staff positions in the U.S. Senate and served on many public and privatehealth-related advisory boards. From 1997 through 1998, she was on President Clinton’sAdvisory Commission on Consumer Protection and Quality in the Health Care Industry.Dr. Wakefield was a member of the Institute of Medicine’s Committee on Quality HealthCare in America and a fellow of the American Academy of Nursing. In 2000, she wasappointed to the National Advisory Committee on Rural Health, Office of Rural HealthPolicy, Health Resources and Services Administration. Dr. Wakefield received her B.S.in nursing from the University of Mary, Bismarck, N.D., and her M.S. and Ph.D. fromthe University of Texas at Austin.

Nicholas J. Wolter, M.D., is a pulmonary and critical care physician who serves aschief executive officer for Deaconess Billings Clinic (DBC), Billings, Mont. DBC is aregional, not-for-profit medical foundation consisting of a multispecialty group practice,hospital, health maintenance organization, research division, and long-term care facilityserving a vast rural area in the northern Rockies. Dr.Wolter began his Billings Clinicpractice in 1982 and served as medical director of the hospital’s intensive care unit from1987 to 1993. He began his leadership role with the successful merger of the clinic andhospital in 1993. Dr. Wolter is a diplomate of the American Board of Internal Medicineand serves on the boards of many regional and national health care organizations. He hasa B.A. from Carleton College, an M.A. from the University of Michigan, and an M.D.from the University of Michigan Medical School.



Commission staff

Mark E. Miller, Ph.D.Executive director

Lu Zawistowich, Sc.D.Deputy director

Publications director

Sarah Thomas, M.S.

Research directors

Jack Ashby, M.H.A.

Jill Bernstein, Ph.D.

Scott Harrison, Ph.D.

Kevin J. Hayes, Ph.D.

Sally Kaplan, Ph.D.

Karen Milgate, M.P.P.

Julian H. Pettengill, M.A.

Nancy Ray, M.S.

Analysts

Sharon Bee Cheng, M.S.

David V. Glass, M.S.

Timothy F. Greene, M.B.A.

Craig K. Lisk, M.S.

Ann Marshall, M.S.P.H.

Anne Mutti, M.P.A.

Susanne Seagrave, Ph.D.

Joan Sokolovsky, Ph.D.

Mae Thamer, Ph.D.

Ariel Winter, M.P.P.

Chantal Worzala, Ph.D.

Daniel Zabinski, Ph.D.

Research assistant

Sarah Lowery, B.A.

Special assistant to the executive director

Marian Lowe

General counsel

Helaine Fingold, J.D.

Administrative staff

Reda H. Broadnax, B.S., Executive officer

Wylene Carlyle

Diane E. Ellison

Plinie (Ann) Johnson

Cheron McCrae

Rachel Vallieres, B.A.

Cynthia Wilson



M A R C H 2 0 0 3


Re

po

rt to

the

Co

ng

re

ss

: Me

dic

ar

e P

ay

me

nt P

olic

y

|M

ar

ch

20

03




















advocates.







payment policymedicare...march 2003 report to the congress medicare payment policy 601 new jersey...

Documents