[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.205.176.107. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
Figure 1.
Proportion of Variance in Medicare Spending at the Hospital and Area Levels
Proportion of Variance in Medicare Spending at the Hospital and Area Levels

Data were obtained using 2008 Medicare claims. The unit of analysis was hospitals. COPD indicates chronic obstructive pulmonary disease; HRR, hospital referral region.

Figure 2.
Ranges of Hospital Spending in HRRs by Median HRR Spending
Ranges of Hospital Spending in HRRs by Median HRR Spending

Data were obtained using 2008 Medicare claims. The unit of analysis is hospitals. HRR indicates hospital referral region; orange line, median value.

1.
Bach  PB.  A map to bad policy: hospital efficiency measures in the Dartmouth Atlas. N Engl J Med. 2010;362(7):569-573.
PubMedArticle
2.
Congressional Budget Office. Budget options, volume 1: health care.http://www.cbo.gov/doc.cfm?index=9925. Published December 1, 2008. Accessed February 15, 2009.
3.
Institute of Medicine. Variation in Health Care Spending: Target Decision Making, Not Geography. Washington, DC: National Academy of Sciences; 2013.
4.
Zhang  Y, Baik  SH, Fendrick  AM, Baicker  K.  Comparing local and regional variation in health care spending. N Engl J Med. 2012;367(18):1724-1731.
PubMedArticle
5.
Newhouse  JP, Garber  AM.  Geographic variation in health care spending in the United States: insights from an Institute of Medicine report. JAMA. 2013;310(12):1227-1228.
PubMedArticle
6.
Gottlieb  D, Zhou  W, Song  Y, Andrews  KG, Skinner J, Sutherland J. A standardized method for adjusting Medicare expenditures for regional differences in prices. In: Technical Report: The Dartmouth Institute for Health Policy and Clinical Practice. Lebanon, NH: Center for Health Policy Research, Dartmouth College; 2010.
7.
The Dartmouth Institute for Health Policy and Clinical Practice. Data by region. The Dartmouth Atlas of Health Care website. http://www.dartmouthatlas.org/data/region/. Accessed July 16, 2014.
Research Letter
June 2015

Hospital and Regional Variation in Medicare Payment for Inpatient Episodes of Care

Author Affiliations
  • 1RAND Corporation, Boston, Massachusetts
  • 2Division of Health Policy and Management, University of Minnesota School of Public Health, Minneapolis
  • 3Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts
JAMA Intern Med. 2015;175(6):1056-1057. doi:10.1001/jamainternmed.2015.0674

Health care spending varies widely between geographic regions, but there is disagreement regarding the appropriate policy response.1 Regional policies include reducing Medicare payment rates in high-spending regions,2 limiting the supply of health care facilities using certificate-of-need criteria, and implementing care-improvement collaboratives. The Institute of Medicine opposed regional policies in favor of hospital- and health care professional–focused policies, such as bundled payments, accountable-care organizations, and value-based payments.3 Their concern was that substantial variation in Medicare spending occurs within geographic regions35 and high-performing hospitals and health care professionals in low-performing regions would be unfairly penalized by regional policies. To further inform this debate, we compared the amount of spending variation that occurs between regions vs between hospitals.

Methods

Using claims for all Medicare fee-for-service beneficiaries from January 1, 2008, through December 31, 2008, we summed Medicare payments for hospital and post–acute care services (skilled nursing facilities, home health, and inpatient rehabilitation) for episodes spanning from inpatient admission to 30 days after discharge. Physician payments—a small proportion of both the total episode payments and the variation in episode payments—were not included. Payments were standardized to remove Medicare’s geographic adjustments.6 Ten common conditions were included in the analysis (Figure 1), identified using 27 Medicare Severity Diagnosis Related Groups (MS-DRGs). We normalized episode payments within each MS-DRG to a mean (SD) of 0 (1) to facilitate comparisons across MS-DRGs. Each episode was linked to the hospital with the initial admission and the hospital referral region (HRR).7 We then estimated the average normalized episode payment across all episodes for each originating hospital. Our sample included 3086 hospitals that provided at least 50 episodes across the 10 conditions in 2008; the median value for a hospital was 515 episodes. We excluded HRRs with fewer than 3 hospitals and hospitals with fewer than 20 episodes of a particular condition. Using hospitals as the unit of analysis, we estimated a hierarchical regression model with HRR random intercepts to estimate episode-spending variance components at the hospital and HRR levels. Institutional review board approval was obtained from RAND Corporation.

Results

There were between 3 and 72 hospitals per HRR. Hospital referral regions were associated with between 7.3% (gastrointestinal tract bleeding) and 29.4% (joint replacement of lower extremity) variation in spending, with considerable differences across conditions (Figure 1). Conditions that more frequently involved post–acute care had a greater proportion of HRR-level spending variation.

Figure 2 shows the range of hospital episode spending in each HRR. Hospital episode spending was calculated as the average of normalized Medicare spending across episodes for the 27 MS-DRGs in the sample. Hospital referral regions are ranked from lowest to highest median spending along the x-axis. Each bar represents the range between the lowest- and highest-spending hospital in an HRR, with the median denoted by an orange line. There is a great deal of overlap between HRRs in the range of hospital spending. For example, the lowest-spending HRR contains a hospital with higher spending than the lowest-spending hospital in the highest-spending HRR.

Discussion

Variation in Medicare spending for episodes of acute and post–acute care was driven by both hospital-based and regional factors. Total Medicare spending was also affected by the volume of episodes. While hospital-level factors accounted for most of the variation in episode spending, region also played an important role, particularly for episodes that commonly involve post–acute care. Post–acute care is particularly important given that it accounts for the largest share of variation in per capita Medicare payments.3

These results are consistent with a prior study4 of per capita spending in areas within HRRs and inform the continuing debate regarding provider- vs region-level policies to address variation in spending.3 Because higher- and lower-spending hospitals are located within regions, region-level policies, such as global reductions in payments, would be expected to penalize lower-spending hospitals. However, our results suggest that there is an important regional component driving cost variation for conditions with more post–acute care spending that should be considered in any payment reform.

Back to top
Article Information

Corresponding Author: Peter S. Hussey, PhD, RAND Corporation, 20 Park Plaza, Ste 920, Boston, MA 02116 (hussey@rand.org).

Published Online: April 13, 2015. doi:10.1001/jamainternmed.2015.0674.

Author Contributions: Dr Hussey had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Hussey, Mehrotra.

Acquisition, analysis, or interpretation of data: Hussey, Huckfeldt, Hirshman.

Drafting of the manuscript: Hussey, Hirshman.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Hussey, Huckfeldt, Hirshman.

Obtained funding: Hussey, Mehrotra.

Study supervision: Hussey, Mehrotra.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by the Institute of Medicine.

Role of the Funder/Sponsor: The Institute of Medicine provided input on the design and conduct of the study, but was not involved in the collection, analysis, and interpretation of the data, and in the preparation, review, or approval of the manuscript.

Additional Contributions: Mark Totten, MS, RAND Corporation, provided assistance with data preparation. He was not compensated.

References
1.
Bach  PB.  A map to bad policy: hospital efficiency measures in the Dartmouth Atlas. N Engl J Med. 2010;362(7):569-573.
PubMedArticle
2.
Congressional Budget Office. Budget options, volume 1: health care.http://www.cbo.gov/doc.cfm?index=9925. Published December 1, 2008. Accessed February 15, 2009.
3.
Institute of Medicine. Variation in Health Care Spending: Target Decision Making, Not Geography. Washington, DC: National Academy of Sciences; 2013.
4.
Zhang  Y, Baik  SH, Fendrick  AM, Baicker  K.  Comparing local and regional variation in health care spending. N Engl J Med. 2012;367(18):1724-1731.
PubMedArticle
5.
Newhouse  JP, Garber  AM.  Geographic variation in health care spending in the United States: insights from an Institute of Medicine report. JAMA. 2013;310(12):1227-1228.
PubMedArticle
6.
Gottlieb  D, Zhou  W, Song  Y, Andrews  KG, Skinner J, Sutherland J. A standardized method for adjusting Medicare expenditures for regional differences in prices. In: Technical Report: The Dartmouth Institute for Health Policy and Clinical Practice. Lebanon, NH: Center for Health Policy Research, Dartmouth College; 2010.
7.
The Dartmouth Institute for Health Policy and Clinical Practice. Data by region. The Dartmouth Atlas of Health Care website. http://www.dartmouthatlas.org/data/region/. Accessed July 16, 2014.
×