Skip to main content

Socioeconomics and Part D Performance Scores

By January 30, 2014Commentary

CMS is big on performance measures.  So when Part D plans came into existence to provide prescription drug coverage for Medicare beneficiaries, measures covering four areas–customer service, turnover, access and drug pricing and safety–were promulgated and the plans are rated on a five-star system.  The scores are used for bonus calculations for Medicare Advantage plans; low scores may lead to being dropped as a Part D contractor and the public can see the scores and presumably factor them into its buying decisions.  The plans would obviously like to know that they are being judged on an all-things-being equal basis; that, for example, the unique characteristics of their membership isn’t affecting their scores in a negative way compared to other plans.  Research published in Health Affairs suggests this may not be true. (HA Article)

The authors looked at the socioeconomic characteristics of plans’ enrollees to see if there was a correlation with performance scores.  Some Part D measures are case mix adjusted, which might pick up socioeconomic differences, but the medication adherence subset of measures are not.  The researchers therefor focused on performance on that subset.  Data from 2012 was used, including a plan’s performance on medication adherence and the plan’s membership characteristics in regard to income, education and minority status.  The authors first note that there is substantial divergence among plans with respect to the socioeconomic status of their overall membership.  They then find that performance on the adherence measures is correlated with socioeconomic status.  Minority status, low income and not having a high school diploma were all significantly negatively associated with higher adherence scores.  So plans with higher percents of these beneficiaries were very likely to have lower performance scores and at risk for the untoward consequences of these lower scores.  Almost a third of performance variation could be accounted for solely by these characteristics.  Some may argue that no policy changes are needed because the plans should be incented to improve care for this population, but others feel it is unfair to penalize a plan for behavior that may be beyond its control.


Leave a comment