Geography debate

A new report from the Institute of Medicine may finally put to rest the idea that policymakers should adopt a geographically based value index to set Medicare reimbursements, a proposal favored by providers and members of Congress in regions that have high-quality, low-cost care. But that doesn't mean the debate on geographic variation in healthcare spending and utilization is over.

Academics and policy analysts have examined the concept of geographic variation for four decades, since Dr. John Wennberg—who founded the Dartmouth Atlas of Health Care that studies the patterns of medical resources and utilization—published a paper in 1973 reporting significant variations in the use of healthcare services across regions. Four years ago, members of Congress asked HHS Secretary Kathleen Sebelius to commission two IOM studies on geographic payments under Medicare. The second study, Variation in Health Care Spending: Target Decision Making, Not Geography, was released last week.

After analyzing Medicare and commercial payer data, a committee of 19 experts at the IOM concluded in the 178-page report that geographic variation in healthcare spending is real, but recommended that Congress not adopt a geographic value index because decisions about care are made at the physician and organizational levels, not the regional level.

Among the committee's most notable findings was that variation in total Medicare spending across geographic areas is driven largely by the use of post-acute care, including home health, skilled nursing, rehabilitation, long-term care and hospice services. According to the study, if there were no variation in post-acute-care spending, the variation in total Medicare spending across hospital referral regions would drop by 73%. In contrast, the committee found that variation in spending in the commercial insurance market is based on differences in price markups by providers, not on differences in the use of healthcare services.

“To the degree that we're worried about geographic variation in Medicare spending, we really have to start with what to do about post-acute care,” said Joseph Newhouse, a Harvard professor of health policy and management who served as the committee's chairman. “We didn't make recommendations about what to do, but that's clearly where a lot of the geographic variation is coming from.”

Still, the banner headline of the report, according to committee member Gail Wilensky, was the panel's conclusion that applying a geographic value index for Medicare reimbursement would not help address geographic variation. The researchers reiterated their observation from a March interim report that said an index based on regions would not likely encourage more efficient behavior among providers, and, consequently, not improve the overall value of care.

“There is a lot of variation within any geographic unit, no matter how small you get,” said Wilensky, a senior fellow at Project HOPE in Bethesda, Md. “And if you were to do something like that, what you'd end up doing would be both inefficient and unfair.”

As Wilensky explained, that's because if payments to physicians or hospitals in high-spending regions were reduced, it would be unfair to low-spending, high-value providers practicing in those same areas. “It's a reminder that geographic units don't make decisions about care,” she added. “If you want to change how these organizations or clinicians respond, you need to change the incentives or reward structures they face—and not by the geographic unit.”

One of the five recommendations in the report is that the CMS test new payment reforms that provide incentives for the clinical and financial integration of healthcare systems. Along with the recommendation that Congress not adopt a geographic value index, the other recommendations include suggestions that Congress should urge—and provide necessary resources for—the CMS to make Medicare and Medicaid data easier to access for research purposes; the CMS should conduct ongoing evaluations of the effect of those new payment models; and Congress should give the CMS flexibility to hasten the transition from Medicare fee for service to new payment structures if evaluations show that the new models are effective.

Dr. Elliott Fisher and Jonathan Skinner, Dartmouth professors who have studied geographic variation and co-wrote a blog post last week about the IOM's findings, agreed with all five of the committee's policy recommendations. But they caution that the report's subtitle might leave the impression that geography doesn't matter, and that it could “cast doubt on the promise” of regional initiatives to improve quality and efficiency in healthcare.

“We may not like the value index,” Skinner said in an interview. “But we think local solutions are important. What we say is sometimes working collectively in a local community is the way to fix a problem.”

In their blog, Fisher and Skinner cited the work of the late Nobel-prize winning economist Elinor Ostrom, who found that the best way to tackle complex social problems is through collaboration among a diverse group of stakeholders. To prove their point, they offered examples of regional multistakeholder initiatives in Colorado, Ohio and Georgia that began with a focus on quality and then expanded to both improve health and lower costs.

“The rationale for a geographic focus on healthcare is strong: The factors that determine population health are largely local, rooted in the environmental, social, economic and behavioral determinants of health,” Fisher and Skinner wrote. “Many of the factors that influence healthcare quality and costs are also local, including local supply, pricing behavior and the relative emphasis of providers on profit.”

Now that the IOM committee has finished its work, it's up to Congress and the CMS to determine next steps, Newhouse said.

He pointed to the need to address dramatic regional variations in post-acute care, and offered a few reasons why there is wide variation in these services. “Particularly for home health, it's a very ill-defined benefit as to what exactly should trigger a home health episode and how intense it should be,” he said. “A medical textbook doesn't say how many baths a person should have, in general. And it's dependent on social and family support.” There also could be issues of fraud and abuse, he added.

James Reschovsky, a senior fellow at the Center for Studying Health System Change, who co-authored a May 2012 Health Affairs article that examined the largest contributors to geographic variations in Medicare services, said “the two things that disproportionately drove the most geographic variations were the variations in home health and durable medical equipment. Those are the two areas that historically there have been the most amount of fraud and abuse taking place.”

Overall, experts agree that the final word from the IOM last week sends a strong signal that Congress won't adopt a geographical value index in Medicare. Reschovsky said it's a concept that didn't have a lot of traction to begin with and has even less now.

Skinner echoed that sentiment. “I think a value index is dead,” said. “I think they put the nails in the coffin on that.”

Follow Jessica Zigmond on Twitter: @MHjzigmond