Putting costs on the map
Population’s health might trump geography: report
n assessing geographic differences in Medicare spending, it may be wise to assess a population’s health rather than just the quality of care it’s receiving, a study has concluded.
The analysis found that assessing additional and better measures of health reduced the magnitude of unexplained geographic differences in Medicare spending, according to results published in the May 12 New England Journal of Medicine and authored by researchers from George Mason University and the Urban Institute.
The trend in healthcare reform is to make providers more accountable for the care they provide by bundling payments of hospitals and doctors, establishing accountable care organizations, and penalizing hospitals with high readmission rates (May 17, p. 6). However, the NEJM study makes the point that “policies that focus on area-level spending without adequate adjustment for differences in beneficiaries’ health status could inappropriately reward or penalize certain geographic areas,” said Jack Hadley, professor in the College of Health and Human Services’ Health Administration and Policy Department at George Mason and senior author of the study, in an interview.
Without knowing all of the factors that account for geographic differences in Medicare spending, policymakers should resist looking for simple solutions to limit expenditures in high-cost areas, the study stated.
For hospitals in areas classified as “high cost” under Medicare, “such institutions should be encouraging their representatives to argue in Congress that the cost differences across geographic areas aren’t necessarily due to inefficiency, and that they shouldn’t be getting penalized by Medicare,” Hadley said.
That’s not to say that inefficiencies don’t exist, Hadley said. Patients may stay in hospitals when it isn’t necessary, or get additional testing that isn’t needed. “It’s just that the geographic differences in spending may not be the best indicators of where those inefficiencies are.”
Others in the industry who assessed the report’s findings agree that provider behavior,
Iwhile important, doesn’t reflect the entire story of cost variation. “Geographic variation in education levels, household income, insurance status prior to Medicare enrollment, smoking, genetics, etc., are all factors in a complex equation that results in differences in disease burden, adherence, patient preferences, and ultimately, global utilization,” said William Golden, professor of medicine and public health at the University of Arkansas for Medical Sciences.
Previous studies have often identified variation across geographic areas in how much Medicare spends per beneficiary, Hadley stated. What they’ve found is “in the highest-cost areas Medicare spends about 50% more per beneficiary than in the lowest-cost areas, even after adjusting for basic demographics such as age, gender and race, and differences in medical-care practices,” he said.
Regional disparities in Medicare spending was highlighted last year in an article in the New Yorker by surgeon and author Atul Gawande, who found that higher spending areas weren’t necessarily delivering the best care, debunking the claim that patients get what they pay for. In particular, he explored the reasons why Medicare spends twice the national average per enrollee in McAllen, Texas—one of the most expensive healthcare markets in the country—than in other comparable markets, even though the quality of care there has been under scrutiny (Nov. 9, 2009, p. 12).
For the NEJM study, however, researchers took a specific focus on measuring health, using an expanded set of criteria that included both pre-existing conditions and changes in health throughout the year. The study also controlled for patients’ demographic characteristics, family income, supplemental insurance coverage and area-level measures of healthcare supply.
Researchers estimated the differences in Medicare spending between high-and lowcost geographic areas, using data on Medicare spending by 6,725 elderly Medicare patients collected by the Medicare Current Beneficiary Survey from 2000 to 2002. Unadjusted, Medicare spending per beneficiary was 52% higher in the most expensive regions than in the least expensive regions. But after adjusting for demographic and baseline health characteristics and changes in health status, researchers saw the difference in spending between the highest and lowest spending regions shrink to 33%.
“Our study shows that an individual’s health explains almost onethird of the difference in Medicare spending per beneficiary between the highest-and lowest-cost areas, while previous studies have assigned a smaller role to health measures,” Stephen Zuckerman, a senior fellow in the Urban Institute’s Health Policy Center and the lead author of the study, said in a written statement.
The problem is that health status explains only about 40% of these geographic spending variations. “We’re still left with substantial unexplained variations that likely deal with utilization and payment. The bottom line is we just don’t use health services in a consistent way,” said J.B. Silvers, a professor of healthsystems management at Case Western Reserve University in Cleveland.
Silvers: “We just don’t use health services in a consistent way.”
Hadley: “Geographic differences ... may not be best indicators.”