Geographic Variation in Spending–Again

This debate over the extent to which there is geographic variation in spending and what it means seems to never end.  And it is unclear that variation for one payer, i.e. Medicare, is the same as variation for other payers.  In the latest published research, the authors primarily examined methods for determining health status and how that might account for spending variation.    (Med. Care Article)    The authors worked with a database of 1.6 million Medicare beneficiaries from about 60 dispersed communities.  They examined the common approach of looking at end-of-life spending and whether beneficiaries in the last year of life really have the same health status across all regions, which is the Dartmouth Atlas approach.  They also examined whether many case mix adjusters are biased due to physician coding behaviors and whether using different adjusters and adjusters from the current or a prior year would improve the analysis of geographic spending variation.  The researchers first adjusted spending to neutralize regional and site of care reimbursement differences.  They then analyzed end-of-life beneficiary spending and found that there are in fact significant regional differences in the health of beneficiaries in the year in which they died, undermining use of that approach to assess geographic spending differences.  Similarly, they modified the HCC case mix adjuster to limit the effect of conditions where coding behavior was most likely to be discretionary.  They found only a very limited effect of variation in coding behavior in relation to variation in spending.  Finally, when they used current year diagnoses for the HCC adjuster, instead of prior year, far more of the geographic spending variation was explained by health status.   It makes sense that a beneficiary’s health has likely worsened in the year they die.  So if this research is right, policymakers who think billions of dollars can be saved by trying to make all providers behave like providers in “low-cost” regions are attacking the wrong problem.  It is not providers that are the issue, but the underlying causes of worse patient health.