With all of the changes going on in reimbursement and the private insurance market, the increasing enrollment in Medicare Advantage (MA) has received little attention. The Kaiser Family Foundation has a nice article here about not only trends in enrollment but also the implications. Some key points from that article:

It estimated that in 2017, a full one-third of Medicare beneficiaries would be enrolled in Medicare Advantage (19 million patients), and by 2027 it could be more than 4 in 10.

Enrollment is “highly concentrated among a handful of firms” which indicates low market competition.

Penetration varies considerably by state, from a high of 56% (MN) to a low of 1% (AK), but this is not necessarily an urban/rural issue: penetration is 65% around Miami but only 14% near Baltimore.

The article goes on to talk about trends in premiums and out-of-pocket costs, but what it doesn’t touch on is the potential impact on researchers. For those using administrative data to perform observational studies, it is common practice to exclude Medicare Advantage patients because much of their care may not be captured by Medicare billing data. Most understand that excluding this large portion of patients likely introduces some level of bias (or, at a minimum, makes generalizing results to all Medicare beneficiaries impossible), but what’s typically ignored is the huge variation in penetration by geographic location. If MA plans introduce selection bias into a study, then self-selection by 1% of the population in one area compared with 56% of the population in another area has huge implications for the impact and significance of the bias introduced. As far as I know, very little (to any) work has been done to try to quantify the impact this may have. Most studies examining differences between those who are and are not enrolled in MA plans do so on a national level, providing little insight into the regional impact.

It’s clear that more work needs to be done to investigate how the ever-growing (and regionally-varying) penetration of MA plans affects observational research.