Prospects for Future Health and Health Care Spending Among the Elderly

Jay Bhattacharya, MD, PhD

for Dana Goldman and the RAND group on medical care expenditure

forecasting

A6654C-2 3/06

Medicare Is Central to the U.S. Health Policy Debate

• Largest insurer by far

• Spent US$300 billion in 2005

• Accounts for almost 20% of allmedical spending

• 13% of government spending

We need to know what Medicare’s future looks like

A6654C-3 3/06

Standard Forecasting Approach

Number of people

Amount each person

spendsX = Total

spending

Contains assumptions aboutprices, treatment, medical technology

A6654C-4 3/06

Standard Forecasting Approach

Number of people

Amount each person

spendsX = Total

spending

Contains assumptions aboutprices, treatment, medical technology

A6654C-5 3/06

Government Does a Good Job Forecasting the Number of Medicare Beneficiaries…

0

5,000

10,000

15,000

20,000

25,000

30,000

1983 1986 1989 1995

Low estimate in 1983High estimate in 1983Beneficiaries

(in thousands)

Year1992 1998

Actual

A6654C-6 3/06

Standard Forecasting Approach

Number of people

Amount each person

spendsX = Total

spending

Contains assumptions aboutprices, treatment, medical technology

A6654C-7 3/06

…But Not Medicare SpendingMedicare Part A spending (hospital insurance)

50

70

90

110

130

150

1980 1982 1984 1986 1988 1990 1992 1994Year

Billionsof

dollars(2005)

High estimate1980 Low

estimate1980

A6654C-8 3/06

…But Not Medicare SpendingMedicare Part A spending (hospital insurance)

50

70

90

110

130

150

1980 1982 1984 1986 1988 1990 1992 1994Year

Billionsof

dollars(2005) Actual

High estimate1980 Low

estimate1980

A6654C-9 3/06

…But Not Medicare SpendingMedicare Part A spending (hospital insurance)

50

70

90

110

130

150

1980 1982 1984 1986 1988 1990 1992 1994Year

Billionsof

dollars(2005) Actual

High estimate1980 Low

estimate1980

A6654C-10 3/06

…But Not Medicare SpendingMedicare Part A spending (hospital insurance)

50

70

90

110

130

150

1980 1982 1984 1986 1988 1990 1992 1994Year

Billionsof

dollars(2005)

ActualHigh estimate1980 Low

estimate1980

A6654C-11 3/06

What Explains the Increase in Total Medical Spending over the Last Fifty Years?

A6654C-12 3/06

What Explains the Increase in Total Medical Spending over the Last Fifty Years?

Aging

A6654C-13 3/06

What Explains the Increase in Total Medical Spending over the Last Fifty Years?

Income

Aging

A6654C-14 3/06

What Explains the Increase in Total Medical Spending over the Last Fifty Years?

Insurance

Income

Aging

A6654C-15 3/06

What Explains the Increase in Total Medical Spending over the Last Fifty Years?

Prices

Insurance

Income

Aging

A6654C-16 3/06

What Explains the Increase in Total Medical Spending over the Last Fifty Years?

Technology

Prices

Insurance

Income

Aging

A6654C-17 3/06

The Problem Is Understanding the Effects of Medical Technology

The Left Ventricular Assist Device

A6654C-18 3/06

Is There a Better Way to Forecast Total Medical Spending?

• Centers for Medicare and Medicaid Services requested this study

• Goal: Open the “black-box” of future medical technology

• Focus on the elderly

A6654C-19 3/06

Our Approach

• Identify emerging technologies

• Build demographic-economic model

• Simulate effects of technology on:– Spending– Functional status– Disease

A6654C-20 3/06

Identified Emerging Technologies UsingMethods Pioneered at RAND and UCLA

• Step 1: Reviewed vast literature on emerging technologies– Devices, drugs, treatments, clinical practices– 21,400 articles screened

• Step 2: Convened panels of private sector and academic experts

– Cardiovascular disease– Neurological disorders– Cancer / biology of aging– Geriatricians and social scientists

• Identified 34 key emerging technologies

A6654C-21 3/06

Example:Intraventricular Defibrillators

Used to treat patients with life-threatening arrhythmias: shocks heart to restore natural rhythm.

Target: 50% of patients with heart failure50% of patients post heart attack20% of patients with cardiomyopathy

Likelihood: 30% in 10 yrs30-40% in 20 yrs

Impact: No effect on hospitalizations Life expectancy increases 6-10 months

Cost: $35,000 - $40,000 per case

~3.5 millionin 2004

A6654C-22 3/06

Our Approach

• Identify emerging technologies

• Build demographic-economic model

• Simulate effects of technology on:– Spending– Functional status– Disease

A6654C-23 3/06

Etc.

New 65 year-olds in 2007

Deceased

Survivors

Deceased

Survivors

2007 costs

Health & functional

status 2007

Health & functional

status, 2006

Our Model Tracks Individuals Over Time

Survivors

New 65 year-olds in 2006

100,000 Medicare

beneficiaries(age 65+)in 2005

2006 costs2005 costs

Deceased

A6654C-24 3/06

Our Forecast of Population Growth Matches Vital Statistics

Year

CensusPopulation

65+(in millions)

3035

404550

556065

7075

1998 2003 2008 2013 2018 2023 2028

Model

70.8 M70.3 M

A6654C-25 3/06

Predicted Prevalence of Disease inthe Medicare Population

2003 2008 2013 2018 2023 2028Year

199805

1015202530354045

% of +65

population

Heart

Stroke

DiabetesLung

Disability

A6654C-26 3/06

300

400

500

600

700

800

2000 2005 2010 2015 2020 2025 2030

Year

Billionsof

dollars(2005)

Forecast of Real Health Care Spending by Elderly

Assumes no change in technology

A6654C-27 3/06

Our Approach

• Identify emerging technologies

• Build demographic-economic model

• Simulate effects of technology on:– Spending– Functional status– Disease

A6654C-28 3/06

Our Panel Predicted a Dramatic Expansion of ICD Use

0

100,000

200,000

300,000

400,000

500,000

600,000

2005 2010 2015 2020 2025 2030

550,000

Year

Numberof

procedures

A6654C-29 3/06

Our Panel Predicted a Dramatic Expansion of ICD Use

0

100,000

200,000

300,000

400,000

500,000

600,000

2005 2010 2015 2020 2025 2030

550,000

Year

Numberof

procedures

550,000 $35,000 ~$19Bx =

A6654C-30 3/06

Will Add About $30 Billion Annually toHealth Spending in Steady State

350400450500550600650700750800850

2005 2008 2011 2014 2017 2020 2023 2026 2029Year

Billionsof

dollars(2005)

830 B800 BCurrent medical practice

With ICD expansion

Total health care spending by elderly

A6654C-31 3/06

But Will Not Substantially Improve Functional Status of Elderly Population

0

5

10

15

20

25

30

2000 2005 2010 2015 2020 2025 2030

Year

Current medical practice

% of elderlywith any

functional impairment

With ICD expansion

A6654C-32 3/06

We Evaluated the Technologies Most Likelyto Enter Clinical Practice

103,0003.7Implantable cardio-defibrillators

Technology

Cost per additional life-year

Increase in medical spending* (%)

*Increase in 2030 health care spending relative to status quo without the technology.

A6654C-33 3/06

Technology Will Put Substantial Pressure on Medical Spending by the Elderly

30,00070.4Anti-aging compound (unhealthy)

1,400,000 2.3Pacemaker for atrial fibrillation500,0002.3Left ventricular assist devices500,0008.0Antiangiogenesis (cancer)103,0003.7Implantable cardio-defibrillators

62,0000.5Telomerase inhibitors (cancer)

22,0000.4Treatment of acute stroke18,0000.4 Cancer vaccines

9,00013.8Anti-aging compound (healthy)Technology

Cost per additional life-year

Increase in medical spending* (%)

*Increase in 2030 health care spending relative to status quo without the technology.

A6654C-34 3/06

No Magic Bullet for Medicare

• Much attention has been focused on the emerging demographic crisis

• But medical technology is perhaps more important– Just one technology could increase elderly

health care spending by 14% to 70% annually– But it could also generate substantial social

value

• Our model makes the tradeoffs explicit

A6654C-35 3/06

What Can We Do About Technology?

• “Upstream” rationing– Only allow cost-effective treatments into practice– Britain, Australia are examples– No precedent in U.S. system

• “Downstream” rationing– Markets (prices) determine who gets services– Insurers play a key role

– Source of contention in HMOs– Medicare not even allowed to consider costs when

making coverage decisions

• Policy alternatives– Let Medicare explicitly take costs into consideration– Change the R&D incentives– Buy out innovators

A6654C-36 3/06

Disseminating this Research• Subject of special issue of Health Affairs

• Featured at an Alliance for Health Reform forumin Washington D.C.

• Extensive media coverage, including featured story in the Money section of USA Today

• Key component of NIH-funded conference on biomedical research, attended by NIH Director and heads of many of the Institutes

• Technical advisors to the Medicare Trustees recommended they consider adopting this approach

A6654C-37 3/06

What’s Next?• Expanding the model to add younger cohorts

– Funded by a 5-year grant from National Institute on Aging

• Developing a version for Europe – Funded by Pfizer

A6654C-38 3/06