Healthcare Cost Differences in the 1990s: The Influence of Metropolitan Area Marketplace Dynamics

1
Healthcare Cost Differences in the 1990s The
Influence of Metropolitan Area Marketplace
Dynamics

• Merton D. Finkler
• Lawrence University
• August 14, 2003

2
Scope of Study

• How much variation in healthcare cost levels and
  growth rates exists across MSAs?
• Are Medicare payments to providers cost-shifted
  onto private payers?
• Do differences in demographic structure help
  explain differences in cost?
• Do differences in purchaser and provider market
  power help explain differences in cost?
• Sponsor Cobalt Corporation Milwaukee, WI

3
Motivation

• Proprietary reports suggest significant
  differences in cost across MSAs
• Local policy makers suggest low Medicare payments
  drive higher commercial payment
• Payment for health care features different
  national and local incidence implications

4
Cost Indicators

• No comprehensive cost indicator at MSA level
  Evidence proprietary claims data and Medicare
• Comprehensive indicators exist for Medicare
  recipients and HMO enrollees
• Hospital indicators can be compared by MSA
• Cost of serving FEHBP enrollees can be compared
• Focus on expenditures limited attempts to
  separate P from Q

5
Cost Shifting

• Focus of Nov 2002 HCFO Conference
• Common Claim Low Medicare pay implies high
  private pay
• Morrissey No shifting unless relative bargaining
  power change exists or unexploited power exists
• Cutler Evidence of cost shift in the 1980s and
  reduced resource use in the 1990s

6
Who Bears the Burden?

• Argument parallels the incidence of the property
  tax (except as tax on labor)
• National Level Cost of health care is part of
  labor compensation, and labor bears most of the
  burden
• Local Level Cost of health care distinguishes
  MSAs ability to attract and retain labor thus,
  borne locally

7
Data

• 20 large MSAs in the Central USA
• Initial focus on Milwaukee (and 5 close MSAs)
• Add 14 other MSAs Madison,WI 13 with
  population greater than 600K and within 750 miles
• HMO data InterStudy U of MN
• Hospitals American Hospital Association
• Demographics Area Resource File, Census
• Physicians Area Resource File
• FEHBP Blue Cross Blue Shield Intermediary

8
Key Variables

• HMO Premium PMPM
• Non-Governmental Payments to Hospitals per
  Non-Elder
• Medicare Payments (A and B) per Enrollee
• AAPCC through 1997
• Old to Young Working Age Population
• Population 45- 64 / Population 20-34
• Competitiveness Herfindahl for hospitals of
  HMOsHMO Penetration

9
Health Care Costs

• HMO premium PMPM
• 2000 range - 123 (DES) to 178 (MSP)
• 1990 2000 growth 33 (MEM) to 97 (MKE)
• Non-Governmental Payment per Non-Elder
• 2000 range - 587 (KC) -1,165 (IND)
• 1990 2000 growth - 18 (DAY) to 161 (LOU)
• FEHBP PPO - PMPM
• 2000 range - 114 (DAY) - 228 (MKE)

10
Table 1
11
Medicare Payment Levels

• Total Medicare Payments 2000 PEPM
• 347 (FTW) to 559 (PIT) 464 (USA)
• 1990 2000 Growth 23 (DET) to 58 (FTW) 69
  - (USA)
• Medicare Part A 2000
• 195 (FTW) to 353 (PIT) 263 (USA)
• 1990 2000 Growth 24 (DES) to 57 (COL)
• 66 - USA
• Medicare Part B 2000
• 140 (MAD) to 206 (PIT) 200.87 (USA)
• 1990 2000 Growth 11 (DET) to 96 (MEM)
• 74 - USA

12
Table 2
13
Metropolitan Demographics

• Per Capita Income
• 2000 - 26,877 (FTW) to 32,540 (CHI)
  28,738(USA)
• 1990 2000 growth all but St. Louis (45-56)
  USA 50
• Old/Young Ratio
• 2000 84 (MEM) to 135 (PIT) 105 (USA)
• 1990 2000 Growth - 23 (MEM) to 81 (MAD)
• USA 44

14
Table 3
15
Medical Care Providers

• The of Hospitals declined 14 out of 20
• Commercial Admissions Share 2000
• 37 (PIT) to 58 (MAD)
• Herfindahl Index for Commercial Admits
• 2000 416 (CHI) to 4265 (FTW)
• Growth 1990 2000 -4 (GRA) to 288 (CLE)
• Physicians per 1,000 residents
• 2000 1.6 (FTW,GRA) to 3.9 (MAD)
• Growth 1990 2000 0 (CIN) to 24 (DAY)
• Specialists 2000 1.0 (FTW) to 2.6 (MAD)

16
Table 4
17
HMO Characteristics

• HMO Penetration Rate
• 2000 11 (MEM) to 61 (MAD)
• 1990 2000 Growth 50 (MSP) to 705 (IND)
• HMO Competitiveness
• 2000 1.03 (OMA) to 7.27 (MAD)
• 1990 2000 Growth 50 (MSP) to 1992 (IND)
• Capitation - Specialist Revenue 2000
• 2000 0 (OMA,DAY) to 67 (MAD)

18
Table 5
19
Table 6Non-Governmental Payments to Hospitals
20
Implications of Regression

• Commercial payments per NE increased 51/year
• HMO competition reduced hospital payment
• Hospital payments related to MDs/1000
• Medicare payments do not influence commercial
  payments
• Age structure of population negatively influences
  commercial payment level
• Hospital concentration is negatively but
  insignificantly related to commercial payment
• Practice style (admissions/1000) matters

21
Table 7HMO Premium per Member per Month
22
Implications of Regression

• HMO PMPM rose 6.13 per year
• PMPM negatively related to hospital concentration
  level
• HMO penetration rate positively influences PMPM
  (possible reverse causality)
• HMO competitiveness measure does not influence
  PMPM
• Medicare payment levels do not affect PMPM
• Old/Young ratio does not affect PMPM

23
Conclusions

• Indianapolis, Madison, Milwaukee, and Omaha
  deliver relatively expensive commercial
  healthcare
• Akron, Cincinnati, Grand Rapids, and Pittsburgh
  deliver relatively cheap commercial healthcare
• Medicare cost shifting non-existent in the
  aggregate for either specification
• Age structure plays a limited role in explaining
  hospital payments or HMO premiums
• Relative bargaining power seems to matter for
  hospital payments

24
Future Directions

• Increase the number of MSAs analyzed
• Investigate bargaining power e.g., MD group
  practices membership
• Investigate reverse causality (HMO PMPM) through
  evaluation of enrollee age structure
• Differentiate effects of hospital concentration
  scale and contracting economies vs. bargaining
  power