Under Medicare Part D, seniors can choose to receive drug coverage through either a stand-alone Prescription Drug Plan (PDP) or through Medicare Advantage (MAPDP), which bundles together Parts A, B and D. Theory predicts that MAPDPs, whose payoff function involves total health care costs, have a stronger incentive than PDPs to decrease the formulary generosity of drugs whose use signals that the patient also has high non-drug costs, during open enrollment periods when such potential enrollees can be discouraged from enrolling and current enrollees can be encouraged to disenroll (the selection hypothesis). After open enrollment, MAPDPs have stronger incentives to design formularies that optimize medical management by generously covering drugs whose use reduces hospital costs (Chandra, Gruber and McKnight, forthcoming) (the medical management hypothesis). The objective of this R03 application is to test whether MAPDP and PDP formulary designs during 2006-2010 are systematically different from each other in ways predicted by theory, within the confines of rules governing all Part D plan design.
Our aims are:
Specific Aim 1 : To test whether MAPDP formularies are less generous than PDP formularies during open enrollment for drugs whose use signals that a patient also has high non-drug costs (Selection hypothesis).
Specific Aim 2 : To test whether MAPDP formularies are more generous than PDP formularies after open enrollment for drugs whose use can lower non-drug costs. (Medical management hypothesis).
Specific Aim 3 : To study factors that determine the magnitude of selection, the magnitude of medical management and how the net effect evolves over time. Our basic estimation strategy is a """"""""differences in differences"""""""" approach using the fact that we expect MAPDP plans will exhibit the selection incentive in their generosity towards certain drugs during open enrollment relative to rest of the plan year, relative to PDP plans and exhibit the medical management incentive towards certain drugs during the rest of the year relative to the open enrollment period, relative to PDP plans. We plan to use several data sets that contain information on drug plan design of MAPDPs and PDPs, as well as information on the connection between hospital costs and use of certain drugs. The measures of plan generosity towards a drug include knowing whether it is on the formulary, and if so, knowing its co- payment structure, need for prior authorization, step therapy and quantity limits on refills. While there are some guidelines that all plans must follow regarding plan design, they are allowed considerable flexibility in design. We anticipate the results of our research will have a meaningful impact on public policy and research related to Medicare Part D by showing how the flexibility built into the design of the market is being used by MAPDP vs. PDP plans.
This proposal investigates whether Medicare drug coverage delivered through a comprehensive program (Medicare Advantage) takes into account the spillovers and connections between drug and non-drug costs in the design of their drug formularies relative to stand-alone drug coverage.
