Geographic Variation in Health Outcomes

Geographic variation seems to be a never-ending source of research analysis, probably because of a hope that somehow if everyplace in the US looked like the highest quality, lowest cost area, we would have a much better health system.  The latest report comes UnitedHealth Group Division Optum.   (Optum Institute Report)   The report looks at both fee-for-service Medicare and commercial data in the usual 306 hospital referral regions.  The findings are very interesting.  Dividing areas into deciles, there is substantial variation in commercial spending, with the highest decile spending 50% more per capita than the lowest.  No particular large geographic area stands out as high or low cost and the variation is driven by both utilization and prices.  Similarly, for Medicare spending per capita varies almost twofold across deciles.  Given Medicare’s fixed prices, the variation is driven more by utilization, health behaviors, practice patterns and chronic conditions.  There is almost no correlation between Medicare and commercial variation or spending patterns, so this is likely not a geographic issue.  If relative income is taken into account, variation patterns change, suggesting that some greater spending is driven by larger incomes.  Similarly, there is significant variation in quality, as reflected in avoidable hospitalizations, avoidable ER use, readmissions, guideline-driven care and medication adherence.  Unlike raw spending, avoidable hospitalization rates tend to be correlated in the commercial and Medicare populations, but the same is not true for readmissions, suggesting different causes are at work.  Some quality measures show wider geographic area patterns, such as Southern areas having lower medication adherence rates.

A very important finding is that there is little correlation in the commercial or Medicare populations between level of spending and overall quality as measured by these researchers.  High spending doesn’t seem to buy better quality and less spending doesn’t seem to mean lower quality.  In general, areas with better population health tend to have lower spending and better quality and the strongest factor that appears linked to better population health, not surprisingly, is better health behaviors.  Poor health behaviors were most strongly linked to higher per capita spending and lower quality.  Stronger economies and social systems also appeared linked to quality and population health.  One clear message is that if you want better health outcomes, you need to focus on how to drive better individual health behaviors.  As we say repeatedly, the only way to do that is hold people responsible for the consequences of their actions.  You engage in bad health behaviors; you pay the extra cost, not the general public.