Yet More on Geography and Costs

Research published in Health Services Research and funded by the Center for Studying Health System Change examines yet again the extent to which there may be unwarranted variation in the care of Medicare beneficiaries.  The researchers started by dividing 1.6 million Medicare patients in 60 local health care markets treated by over 5500 physicians into a high cost quartile and a remaining lower cost 75% of enrollees.  On an unadjusted basis, the top 5% of Medicare beneficiaries account for 43% of spending and the top 25% for 85% of total cost to the program.  The research sought to identify factors associated with this high-cost 25% of beneficiaries, which it should be noted was identified as the 25% predicted to be high-cost in the claim years under study, which were 2005-06.  (CSHSC Study)

The factors included beneficiary characteristics, physician and practice characteristics, and local market considerations.  Out of all these sub-factors, the health characteristics of the patient were most strongly related to spending; few factors relating to either physicians or the market had much of an association.  Some that were related included having as a usual physician a specialist, reporting inadequate visit time with the usual physician, having a for-profit provider and a greater supply of medical specialists in the market.

The study is notable for the efforts it went to to adjust and carefully sort through factors and for the fact that it looked at all spending, Medicare as well as supplemental insurance and out-of-pocket expenses.  The researchers also adjusted for geographic variation in prices and input costs.  As might be expected, some characteristics of those in the high-cost quartile were that they were twice as likely to be dual eligibles and 27% were institutionalized as opposed to only one-tenth of a percent in the lower 75% spending group.  Notably, the high 25% had $48,000 in average medical costs versus $7000 for low cost beneficiaries and accounted for 69% of total spending, even after all the adjustments.

As the researchers noted, it is important to identify factors associated with more spending so that any policy fixes do what they are intended to do.  If variation is not related to physician practice patterns, then lowering fees in high-cost regions won’t have an impact.  The authors suggest that the best thing to do appears to be to focus care management strategies on this high-cost quartile.