What Can We Learn From Geographic Variation in Health Costs in the US?

Last year I commented on a website that has county-level data on health outcomes for the entire US. A couple of years I commented on the spatial variability in the US in life expectancy at the county level. Another set of spatial analyses has been ongoing for a couple of decades--the Dartmouth Atlas of Health Care. It has focused especially on regional differences in the cost of health care, based on Medicare and Medicaid data, and this spatial variation led people to raise important policy questions. In particular, since the assumption is that Medicare and Medicaid patients are getting essentially the same care anywhere in the country, why are some regions of the country able to deliver that care at lower cost than other regions? Can we use these geographic variations to figure out how to lower health care costs in the higher cost regions since, after all, Americans pay more per person for health care than anyone else on earth.

The National Institute of Medicine created a committee to try to answer this question, and their report was just made available today at the National Academies Press website. It is titled "Variation in Health Care Spending." It turns out that it wasn't easy to track down data to answer this question, since Medicare and Medicaid services are provided in the context of a wider commercial health market. But the committee pulled together a lot of data and concluded that:

(1) Geographic variation in spending and utilization is real, and not an artifact reflecting random noise; and

(2) Variation in spending in the commercial insurance market is due mainly to differences in price markups by providers rather than to differences in the utilization of health care services.

Overall, then, the committee could not determine anything within the regional differences that could lead to policies that could lower costs in some places without harming the care provided. The lack of conclusiveness of the findings, however, is bound to lead to more research and more policy analysis. The fact that key regional differences really do exist is too striking a finding for this new report to put the issue to rest.