Geographic Variation in Health Spending

By September 14, 2009November 2nd, 2009Commentary

Many promoters of health reform have pointed to significant geographic variation in health spending as evidence of waste that if addressed could reduce unnecessary care and costs, with the savings being used to fund access expansion.  The Dartmouth Atlas researchers have been the founding theorists of the geographic variation research and currently hypothesize that the higher spending areas don’t appear to have better health outcomes, so much of their excess spending is wasteful, and that some of the causes of the variation are related to level of supply of hospital beds and specialists and to physician practice preferences which are not necessarily consistent with patients’ wishes or evidence-based medicine.

Richard Cooper is a researcher who has taken a contrary position, which he summarizes in a recent Journal of the American Medical Association commentary.  (JAMA Commentary) Based on his analysis, he believes that geographic variation in spending is real, but has more to do with income and other socio-demographic factors.  He points out that most of the Dartmouth Atlas work relies on Medicare spending and that Medicare spending is not necessarily reflective of total health spending in a particular region.   High-income persons tend to use more health care services than average, but low-income individuals tend to use two to three times the average.  Large urban areas may have a mix of both these subsets, which influence both supply of resources and their utilization.

Cooper’s work is interesting and a useful reminder of the complexity of this issue.  It may be dangerous that so many people now assume that the geographic variation represents waste, which when eliminated will solve our health spending problem.  What is striking about this topic is the failure of researchers to incorporate more data relating to commercial populations.  The largest private insurers all have enormous databases which could be analyzed and combined with Medicare and Medicaid spending to obtain a more complete picture of geographic variation.  Even more surprising is the paucity of studies looking at variation by individual physician.  Physicians order around 90% of medical services.  For two decades or more private health plans have been profiling the practice patterns of doctors to assist them in only delivering appropriate care.  A comprehensive analysis of the amount of individual physician practice variation, within and across geographic regions, would likely not only help more fully explain the causes of variation, but would likely present an opportunity to correct it where it is unjustified.

Leave a Reply