[Skip to Content]
[Skip to Content Landing]
Table 1.  Measures of Geographic Variationa
Measures of Geographic Variationa
Table 2.  Measures of Relative Change in Spending Ranka
Measures of Relative Change in Spending Ranka
1.
Wennberg  JE.  Tracking Medicine: A Researcher’s Quest to Understand Health Care. New York, NY: Oxford University Press; 2010.
2.
Skinner  JS, Fisher  ES, Wennberg  JE.  The Efficiency of Medicare. Cambridge, MA: National Bureau of Economic Research; 2001.
3.
Fisher  ES, Wennberg  DE, Stukel  TA, Gottlieb  DJ, Lucas  FL, Pinder  EL.  The implications of regional variations in Medicare spending: part 2: health outcomes and satisfaction with care.  Ann Intern Med. 2003;138(4):288-298.PubMedGoogle ScholarCrossref
4.
Appleby  J, Raleigh  V, Frosini  F, Bevan  G, Gao  H, Lyscom  T.  Variations in Health Care: The Good, the Bad and the Inexplicable. London, England: The King's Fund; 2011.
5.
Burwell  SM.  Setting value-based payment goals—HHS efforts to improve US health care.  N Engl J Med. 2015;372(10):897-899.PubMedGoogle ScholarCrossref
Research Letter
March 2016

Geographic Variation in Medicare Expenditures, 2003-2012

Author Affiliations
  • 1The Dartmouth Institute for Health Policy and Clinical Practice, Lebanon, New Hampshire
JAMA Intern Med. 2016;176(3):405-407. doi:10.1001/jamainternmed.2015.7814

Geographic variation in health care use and expenditures among Medicare beneficiaries has long been studied in the United States.1 Such variation is associated with inefficiency of health care where utilization rates vary across geographic regions.2 This variation is wasteful because higher utilization rates are not associated with better outcomes.3 One might hope that public reporting would decrease such variation. This study sought to determine whether variation in 6 categories of per capita Medicare expenditures, as well as total per capita expenditures, had changed in the 10 years beginning 2003.

Methods

For 306 geographically defined hospital referral regions (HRRs) in each calendar year from January 1, 2003, to December 31, 2012, numbers of beneficiaries of Medicare Part A and Part B fee-for-service plans were obtained from the Dartmouth Atlas Project website (http://www.dartmouthatlas.org), where such values have been publicly posted and available since 2005. Total Part A and Part B per capita Medicare fee-for-service expenditures adjusted for sex, age, race, and price, as well as expenditures in 6 categories (inpatient care in hospital and skilled nursing facilities, physician care, outpatient facility care, home health care, hospice care, and durable medical equipment [DME]), were also obtained from the Dartmouth Atlas Project website. For each year and expenditure category, commonly used measures of geographic variation4 were calculated: the extreme ratio, which is the highest value divided by the lowest value; the interquartile ratio, which is the 75th percentile value divided by the 25th percentile value; the coefficient of variation, which is the ratio of the standard deviation to the mean; and the weighted coefficient of variation, which is the ratio of the population-weighted standard deviation to the population-weighted mean. Furthermore, for each year and expenditure category, each HRR was assigned to a spending quintile. The proportion of HRRs that stayed in the same quintile during the 10-year study period was calculated. Data analysis was conducted from September 1 to October 31, 2015.

Results

Measures of geographic variation were low and generally stable for total care and inpatient care, low but increasing somewhat for physician and outpatient facility care, modest but increasing for home health care, and decreasing from high levels for hospice care and DME expenditures (Table 1). Per capita spending growth was highest for outpatient facility and hospice care and lowest for DME and inpatient care. The large majority of HRRs that were in the lowest or highest spending quintile in 2003 remained in the same quintile in 2012 (Table 2); total, physician, and home health care expenditures were most entrenched in extreme spending quintiles.

Discussion

Despite the ready availability of data showing widespread geographic variation in per capita Medicare expenditures, measures of geographic variation in spending remained flat or increased for all spending categories except hospice care and DME. The highest and lowest spending HRRs remained in the same categories for long periods.

These findings are important for 2 reasons. First, they provide insight into how health care delivery has changed: that measures of geographic variation for hospice care and DME expeditures decreased from relatively high levels and had more movement across spending quintiles suggest that use of these services converged, although hospice services became relatively more common and DME expeditures became relatively less common. In contrast, increasing measures of geographic variation and less movement across spending quintiles for physician and home health care expenditures suggest increasing divergence in opinion about how to use these services most effectively over time. This finding is disconcerting because the opposite effect would be expected if data availability brought about collaborative change.

Second, these findings suggest that policy efforts have not reined in health care spending in areas with high costs. While the Centers for Medicare & Medicaid Services has initiated efforts to hasten uptake of value-based reimbursement,5 seemingly those efforts should focus on HRRs with high costs and high levels of use in order to curtail expenditure growth, reduce geographic variation in per capita spending, and improve the value of health care.

Back to top
Article Information

Corresponding Author: William B. Weeks, MD, PhD, MBA, The Dartmouth Institute for Health Policy and Clinical Practice, 1 Medical Center Dr, Williamson Translational Research Building, Lebanon, NH 03766 (wbw@dartmouth.edu).

Published Online: January 25, 2016. doi:10.1001/jamainternmed.2015.7814.

Author Contributions: Dr Weeks 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.

Conflict of Interest Disclosures: None reported.

References
1.
Wennberg  JE.  Tracking Medicine: A Researcher’s Quest to Understand Health Care. New York, NY: Oxford University Press; 2010.
2.
Skinner  JS, Fisher  ES, Wennberg  JE.  The Efficiency of Medicare. Cambridge, MA: National Bureau of Economic Research; 2001.
3.
Fisher  ES, Wennberg  DE, Stukel  TA, Gottlieb  DJ, Lucas  FL, Pinder  EL.  The implications of regional variations in Medicare spending: part 2: health outcomes and satisfaction with care.  Ann Intern Med. 2003;138(4):288-298.PubMedGoogle ScholarCrossref
4.
Appleby  J, Raleigh  V, Frosini  F, Bevan  G, Gao  H, Lyscom  T.  Variations in Health Care: The Good, the Bad and the Inexplicable. London, England: The King's Fund; 2011.
5.
Burwell  SM.  Setting value-based payment goals—HHS efforts to improve US health care.  N Engl J Med. 2015;372(10):897-899.PubMedGoogle ScholarCrossref
×