Topics in Medicare Part D: Eligibility, Participation, and Impacts

Abstract

Part D expanded Medicare in 2006 by creating a prescription drug benefit offered through private insurance plans. Key features of this benefit include the Low Income Subsidy (LIS), which provides discounts for plan premium and cost sharing for qualifying beneficiaries, and enrollment regulations, which provide economic incentives to encourage enrollment regardless of health status. Enrollment is voluntary among non-Medicaid duals; consequently, enrollment decisions are critical to program efficiency. This study addresses fundamental aspects of participation in Part D to inform policy design related to insurance access and impacts. Low LIS participation is a major policy concern; however, published estimates of eligibility and take-up are unreliable due to the lack of unified data capturing both dimensions. The objective of Aim 1 was to develop better estimates of whom, among likely eligible beneficiaries, enrolled, and to identify characteristics of non-enrollees. I used an unpublished supplement to the Medicare Current Beneficiary Survey in conjunction with an income imputation procedure, to determine eligibility. Findings suggest that published literature may misrepresent the population unreached by the LIS. In particular, this study brings attention to the high rate of likely LIS eligible beneficiaries obtaining unsubsidized Part D. Aim 2 expanded upon this to estimate the impact of the LIS on drug spending among eligible Part D enrollees using generalized linear regression models. Results indicated that LIS was associated with $534 greater Part D spending compared to non-LIS enrollees. Greater spending was associated with more drug fills and use of higher cost fills. Aim 3 assessed the efficiency of Part D design in offering incentives for timely enrollment by examining the role of health shocks on late enrollment among non-duals. The study focused on beneficiaries who did not obtain Part D or have other creditable coverage following their initial enrollment opportunity using a 5% sample of the Medicare population. Results indicated that a health shock was associated with a greater likelihood of late Part D enrollment suggesting that policy mechanisms alone were not effective in inducing participation. Most beneficiaries remained without coverage, thus penalizing the initial decision to decline Part D may deter later enrollment.