Adjusting for Occupational Adjusting for Occupational Mix Differences in the Mix Differences in the

Medicare Hospital Wage Medicare Hospital Wage IndexIndex

North Carolina Rural Health Research & Policy Analysis CenterNorth Carolina Rural Health Research & Policy Analysis CenterCecil G. Sheps Center for Health Services ResearchCecil G. Sheps Center for Health Services Research

University of North Carolina at Chapel HillUniversity of North Carolina at Chapel HillSeptember, 2004September, 2004

This work was funded by the Federal Office of Rural Health PolicThis work was funded by the Federal Office of Rural Health Policy, y, Cooperative Agreement Number 6U1CRH00027Cooperative Agreement Number 6U1CRH00027--04 04

10/19/200410/19/2004 22

The occupational mix adjustment: The occupational mix adjustment: What is it?What is it?

A new computation added to the PPS hospital A new computation added to the PPS hospital hourly wage surveyhourly wage survey

Objective is to obtain Objective is to obtain marketmarket--levellevel average average hourly wages (hourly wages (AHWAHWmktmkt) that reflect differences ) that reflect differences across markets in the across markets in the priceprice of labor, but not in of labor, but not in the the mix of skillsmix of skills hired.hired.

10/19/200410/19/2004 33

When and for whom is it effective?When and for whom is it effective?

Alters the wage index in federal fiscal year 2005Alters the wage index in federal fiscal year 2005Affects all Medicare prospective rates that use Affects all Medicare prospective rates that use the wage index to adjust standard rates per the wage index to adjust standard rates per service, including:service, including:–– inpatient hospitalinpatient hospital–– outpatient hospitaloutpatient hospital–– inpatient rehabinpatient rehab–– skilled nursing facilitiesskilled nursing facilities–– home health agencieshome health agencies

10/19/200410/19/2004 44

Why is it needed?Why is it needed?Some markets (especially stateSome markets (especially state--wide rural wide rural markets) have belowmarkets) have below--average hourly wages average hourly wages bothbothbecause the price of labor is below national because the price of labor is below national average average andand because hospitals in these markets because hospitals in these markets tend to use a less skilled labor mixtend to use a less skilled labor mix

–– In theory, differences in skill mix should reflect In theory, differences in skill mix should reflect differences in service mix, which in turn reflect differences in service mix, which in turn reflect differences in casedifferences in case--mix acuity. DRG weights already mix acuity. DRG weights already adjust payments for acuity; therefore, any differences adjust payments for acuity; therefore, any differences in labor costs caused by skill mix differences should in labor costs caused by skill mix differences should be removed from the wage index.be removed from the wage index.

10/19/200410/19/2004 55

Why now?Why now?BIPA (2000) mandated data collection by BIPA (2000) mandated data collection by job category to make an occupational mix job category to make an occupational mix adjustment possibleadjustment possible–– Congress was responding to pressure from rural Congress was responding to pressure from rural

advocacy groups. Rural hospitals are smaller and advocacy groups. Rural hospitals are smaller and tend to hire a less expensive skill mix, and tend to hire a less expensive skill mix, and conventional wisdom has been that standardizing for conventional wisdom has been that standardizing for occupation mix would increase relative wages in rural occupation mix would increase relative wages in rural markets.markets.

Before BIPA, CMS had delayed Before BIPA, CMS had delayed implementation due to concerns about implementation due to concerns about data quality, cost and validitydata quality, cost and validity

10/19/200410/19/2004 66

How does it work?How does it work?The adjustment is first computed for each The adjustment is first computed for each hospital within a market, to standardize the hospital within a market, to standardize the hospital’s average hourly wage to reflect what it hospital’s average hourly wage to reflect what it would be if that hospital hired the same mix of would be if that hospital hired the same mix of occupations as the national averageoccupations as the national averageAdjusted wages from each hospital are then Adjusted wages from each hospital are then used to compute a new marketused to compute a new market--level average level average hourly wage (hourly wage (AHWAHWmktmkt), from which the wage ), from which the wage index is computedindex is computed

So, there is no individual hospital effect So, there is no individual hospital effect –– only a only a market effectmarket effect

10/19/200410/19/2004 77

Mechanics:Mechanics:Occupational mix adjustment applies only Occupational mix adjustment applies only to specific types of to specific types of clinicalclinical staffstaff–– Nationally, covered occupations account for 48% of Nationally, covered occupations account for 48% of

all hospital paid hours (most are in nursing)all hospital paid hours (most are in nursing)

For each occupation group:For each occupation group:–– Occupational mix adjustment derives from the Occupational mix adjustment derives from the

difference between a hospital’s proportion of paid difference between a hospital’s proportion of paid hours by job category, and the national average hours by job category, and the national average proportion of paid hours by job categoryproportion of paid hours by job category

–– National hospital data come from the Bureau of Labor National hospital data come from the Bureau of Labor Statistics (BLS)Statistics (BLS)

10/19/200410/19/2004 88

Covered Occupation Categories:Covered Occupation Categories:

ClassificationClassificationBLS national percent BLS national percent

paid hours paid hours NursingNursing 37.2737.27 %%Physical TherapyPhysical Therapy 1.36 %1.36 %Occupational TherapyOccupational Therapy 0.46 %0.46 %Respiratory TherapyRespiratory Therapy 1.94 %1.94 %PharmacyPharmacy 2.11 %2.11 %DietaryDietary 0.84 %0.84 %Medical & Clinical LabMedical & Clinical Lab 4.10 %4.10 %

Total, all adjusted categoriesTotal, all adjusted categories 48.08 %48.08 %

10/19/200410/19/2004 99

Mechanics (Mechanics (continued):continued):

Adjustment factor is dollarAdjustment factor is dollar--weighted weighted –– This means that differences in highThis means that differences in high--wage wage

groups count more towards the final groups count more towards the final adjustment to the wage index than differences adjustment to the wage index than differences in lowin low--wage groupswage groups

–– DollarDollar--weighting is accomplished using BLS weighting is accomplished using BLS national hourly wage by category national hourly wage by category –– NOTNOT the the hospitals’ own hourly wages for that particular hospitals’ own hourly wages for that particular groupgroup

10/19/200410/19/2004 1010

Mechanics (Mechanics (continued):continued):

Adjustment to the wage index is Adjustment to the wage index is market market basedbased. . Therefore…Therefore…–– The effect on the wage index depends on the The effect on the wage index depends on the

weighted average of all of the occupation mix weighted average of all of the occupation mix adjustments for hospitals in each market. Large adjustments for hospitals in each market. Large hospitals have more influence than small oneshospitals have more influence than small ones

Example: Example: –– A small MSA that has one very large, sophisticated A small MSA that has one very large, sophisticated

hospital with a negative adjustment factor, plus five hospital with a negative adjustment factor, plus five small hospitals with positive adjustment factors, might small hospitals with positive adjustment factors, might still have a negative adjustment to its wage indexstill have a negative adjustment to its wage index

10/19/200410/19/2004 1111

Data Sources:Data Sources:Hospital data on paid hours by occupation Hospital data on paid hours by occupation category will be collected every three category will be collected every three years (mandated)years (mandated)–– Hospitals can choose to use either…Hospitals can choose to use either…

Payroll and contract hours from the same year as the cost Payroll and contract hours from the same year as the cost report wage survey, orreport wage survey, orA special 4A special 4--week data collection, that must be conducted week data collection, that must be conducted during the first two months of the yearduring the first two months of the year

National data for standard occupation mix National data for standard occupation mix and pay rates come from BLS surveysand pay rates come from BLS surveys

10/19/200410/19/2004 1212

HospitalHospital--level computation level computation

Adjustment factorAdjustment factoris the ratio of the following two sums computed is the ratio of the following two sums computed across all covered job categories:across all covered job categories:

ΣΣ ((National Hrly Mix National Hrly Mix ×× National Hrly Wage)National Hrly Wage)ΣΣ (Hospital Hrly Mix (Hospital Hrly Mix ×× National Hrly Wage)National Hrly Wage)

If a hospital has belowIf a hospital has below--average skill mix, thenaverage skill mix, then⇒⇒ numerator > denominatornumerator > denominator⇒⇒ occupational mix adjustment factor > 1.00occupational mix adjustment factor > 1.00⇒⇒ adjusted hourly wage > actual hourly wageadjusted hourly wage > actual hourly wage

10/19/200410/19/2004 1313

HospitalHospital--level computation level computation (continued)(continued)

Wages for covered occupation groups are Wages for covered occupation groups are multiplied by the computed adjustment factors multiplied by the computed adjustment factors Wages for nonWages for non--covered groups are left alonecovered groups are left aloneNew hourly wage is computed by dividing the New hourly wage is computed by dividing the adjusted plus unadjusted wages by total paid adjusted plus unadjusted wages by total paid hourshoursHospitalHospital’’s new occupationals new occupational--mix adjusted hourly mix adjusted hourly wage is used to compute the marketwage is used to compute the market’’s new s new occupationaloccupational--mix adjusted mix adjusted AHWAHWmktmkt –– there is no there is no individual hospital effectindividual hospital effect

10/19/200410/19/2004 1414

Easiest way to think of the occupationEasiest way to think of the occupation--adjusted hourly wage:adjusted hourly wage:

For each hospital For each hospital ——Adjustment Factors

×Covered group

portions(48% of wages)

1.00×

Non-covered groupportion

(52% of wages)

+ = Occupation-adjusted wages

÷Total paid hours

=

Adjusted hourly wage

10/19/200410/19/2004 1515

MarketMarket--level computation level computation ::

New New AHWAHWmktmkt is computed by summing across all is computed by summing across all hospitals within the labor markethospitals within the labor marketΣΣ (adjusted wages)(adjusted wages)

ΣΣ (actual paid hours)(actual paid hours)

So the occupationSo the occupation--adjusted wage index value for adjusted wage index value for a given market will be highera given market will be higher ONLY ONLY if weighted if weighted average adjustment that includes average adjustment that includes ALLALL of the of the hospitals in the labor market is greater than onehospitals in the labor market is greater than one

10/19/200410/19/2004 1616

How big are the adjustments?How big are the adjustments?

Hospitals’ adjustment factors can be Hospitals’ adjustment factors can be substantialsubstantial–– range across hospitals for all covered groups:range across hospitals for all covered groups:

0.93 to 1.290.93 to 1.29Because the factor only applies to the covered Because the factor only applies to the covered groups, adjustment to hospitals’ overall hourly groups, adjustment to hospitals’ overall hourly wage is lesswage is less–– estimated range across hospitals: estimated range across hospitals: 0.96 to 1.140.96 to 1.14

When averaged within markets, the effect is When averaged within markets, the effect is even smallereven smaller–– estimated range across markets: estimated range across markets: 0.98 to 1.090.98 to 1.09

10/19/200410/19/2004 1717

““Transition” period for Transition” period for implementationimplementation

CMS chose to minimize initial impact by CMS chose to minimize initial impact by implementing only 10% of the adjustment implementing only 10% of the adjustment in FY 2005in FY 2005Although this is called a “transition”, CMS Although this is called a “transition”, CMS did not publish a schedule indicating when did not publish a schedule indicating when or if the adjustments would increase from or if the adjustments would increase from 10% to 100%10% to 100%

10/19/200410/19/2004 1818

.9 1 1.1 1.2 1.3adjustment

shaded boxes identify 25th to 75th percentiles of the distributions

Distribution of adjustment factors and hourly wage effectsacross hospitals and labor markets

hospital adjustment factors hospital hourly wage effectmarket-level AHW effect market AHW effect @ 10%

10/19/200410/19/2004 1919

.99 .995 1 1.005 1.01 .99 .995 1 1.005 1.01

metro non-metro

adjustmentshaded boxes identify 25th to 75th percentiles of the distributions

10%-transition market wage effects by type of labor marketurban (metropolitan) vs. rural (state non-metropolitan)

10/19/200410/19/2004 2020

What is the impact on payments?What is the impact on payments?Very small, for nowVery small, for nowDifficult to identify, because the adjustment was Difficult to identify, because the adjustment was implemented simultaneously with several other implemented simultaneously with several other wage index changes:wage index changes:–– New metropolitan areas New metropolitan areas –– New “laborNew “labor--related share” of hospital paymentsrelated share” of hospital payments–– New market reclassification rulesNew market reclassification rules

Wage index affects only a portion of each PPS Wage index affects only a portion of each PPS rate rate –– The “laborThe “labor--related share” ranges from 62% to 76% related share” ranges from 62% to 76%

of various PPS ratesof various PPS rates

10/19/200410/19/2004 2121

From CMS’ initial predictions of the From CMS’ initial predictions of the impact of the adjustment on ruralimpact of the adjustment on rural--urban urban

distribution of payments: distribution of payments:

MarketMarket--level effect on the wage index is level effect on the wage index is not not what was expectedwhat was expected–– Rural marketsRural markets

31 (63%) have an 31 (63%) have an INCREASE in wage indexin wage index18 (37%) have a 18 (37%) have a DECREASE DECREASE in wage indexin wage index

–– Urban markets:Urban markets:158 (49%) have an 158 (49%) have an INCREASE in wage indexin wage index166 (51%) have a 166 (51%) have a DECREASE DECREASE in wage indexin wage index

10/19/200410/19/2004 2222

What happened?What happened?Hospital data may be inaccurateHospital data may be inaccurateThe rural differentials in occupation mix for The rural differentials in occupation mix for covered categories may be smaller than covered categories may be smaller than expected (especially in nursing, which expected (especially in nursing, which dominates the calculation)dominates the calculation)Critical access hospitals are no longer included Critical access hospitals are no longer included in PPS samplein PPS sample–– This eliminated the rural facilities that are most likely This eliminated the rural facilities that are most likely

to have the lowest skill mix, and therefore the highest to have the lowest skill mix, and therefore the highest occupation mix adjustment factorsoccupation mix adjustment factors