Geographic Variation in Health Care: Changing Policy Directions

Originally published by the Center for Studying Health System Change (HSC) as NIHCR Policy Analysis No. 4, April 2011.

Geographic Variation in Health Care Spending and Use

For more than four decades, researchers have documented wide geographic differences in fee-for-service Medicare spending and service utilization across the United States. An early and influential conclusion drawn from this body of research was that health care is delivered far more efficiently in some parts of the country than in others, implying that geography itself largely determines the type and amount of care patients receive.

With health care spending continuing to outpace economic growth and approaching one-fifth of gross domestic product, some policy makers have seized on reducing unwarranted geographic variation as a mechanism to control costs without sacrificing access or quality. Analysis of Medicare fee-for-service claims data, in particular, has fueled calls for public policies designed to push health care providers and patients in high-cost regions toward the seemingly more efficient practice patterns observed in low-cost areas.

Reassessing the Evidence on Geographic Variation

This Policy Analysis, authored by Jill Bernstein, James D. Reschovsky, and Chapin White, reviews the evidence and key inferences from geographic variation research that have shaped the ongoing policy debate about health care efficiency. The authors find that more recent research, which employs improved data sources and more sophisticated analytical approaches, suggests that unwarranted geographic variation is not as extensive as previously believed.

These updated findings support a growing consensus that payment reforms narrowly targeted by geography would be ineffective at addressing either local or national problems of health care costs and quality. Instead, the authors argue that a more direct and productive strategy for reducing unwarranted variation in health care use and spending is to focus on broader payment reform and oversight mechanisms that can drive greater efficiency across all geographic areas.

From Geographic Targeting to Broader System Reform

While geographic variation research has brought the twin issues of uneven care and spending disparities to the forefront of health policy discussions, the analysis concludes that the broader health care system, rather than geography, should be the primary focus of reform efforts. The health reform law includes multiple pathways to broader system reform, such as the creation of accountable care organizations, bundled payment models, and patient-centered medical homes, all of which may help drive higher efficiency and improved quality of care.

The distinction matters for policy design. If geographic variation in spending and utilization primarily reflects differences in how the health care system is organized and financed rather than inherent regional characteristics, then targeting individual geographic areas for cost reduction is less likely to succeed than implementing reforms that change the incentive structures and delivery models governing care everywhere. The evidence reviewed in this analysis points strongly in that direction, suggesting that the most promising strategies for controlling health care costs and improving quality are those that reshape how care is organized, paid for, and delivered on a national basis rather than penalizing or rewarding providers based on where they practice.

Key Policy Takeaways

Early geographic variation research drew attention to dramatic differences in per-capita Medicare spending across hospital referral regions, leading many to conclude that patients in high-spending areas received more care without better outcomes. These findings were widely cited in policy debates and contributed to proposals for geographically targeted payment adjustments.

However, newer studies that account for differences in patient health status, income, insurance coverage, and prices paid by private insurers alongside Medicare paint a more nuanced picture. When these factors are included in the analysis, the degree of unexplained geographic variation shrinks considerably. Some areas that appeared inefficient under simpler analyses look quite different when a broader set of variables is considered.

The analysis also highlights the risks of using geographic variation in Medicare spending as a proxy for overall health system efficiency. Medicare covers only a portion of the population, and spending patterns in the fee-for-service program may not reflect what is happening in the broader health care market, including the privately insured and managed care populations. As a result, policies designed solely around Medicare geographic variation data could produce unintended consequences.

Rather than zeroing in on geographic targets, the authors recommend pursuing payment reform strategies that create the right incentives for providers and patients regardless of location. Accountable care organizations, which hold groups of providers jointly responsible for the cost and quality of care for a defined patient population, represent one such approach. Bundled payments, which set a single payment for all services related to an episode of care, offer another. Patient-centered medical homes, which emphasize care coordination and chronic disease management through a primary care practice, provide yet another model. Each of these approaches addresses the systemic drivers of inefficiency and variation rather than attempting to legislate efficiency by zip code.

Sources and Further Reading

Bernstein, Jill, James D. Reschovsky, and Chapin White, "Geographic Variation in Health Care: Changing Policy Directions," NIHCR Policy Analysis No. 4 (April 2011).

Dartmouth Atlas of Health Care, Center for the Evaluative Clinical Sciences, Dartmouth Medical School.

Congressional Budget Office, "Geographic Variation in Health Care Spending" (February 2008).

Institute of Medicine, "Variation in Health Care Spending: Target Decision Making, Not Geography" (2013).