INTRODUCTION

Medicare premiums and cost-sharing can pose a substantial financial burden on beneficiaries, half of whom have incomes below $26,200.1 Beneficiaries with low incomes and assets may qualify for assistance with these costs through full Medicaid, the Medicare Savings Programs (MSPs, which are limited Medicaid benefits), and the Part D Low-Income Subsidy (LIS).2,3 However, a substantial proportion of individuals who qualify for these programs are not enrolled.3,4,5

These programs are linked in ways that may be unknown to Medicare beneficiaries. All beneficiaries who are eligible for full or limited Medicaid qualify for the LIS, which eliminates premiums and lowers cost-sharing in the Medicare Part D program. Moreover, beneficiaries who receive Medicaid are automatically enrolled in the LIS. Because auto-enrollment is the main pathway into the LIS,2 there is concern that low take-up of Medicaid could contribute to low LIS take-up. Foregoing the LIS may have adverse consequences for patients: CMS has estimated that this program reduces average annual out-of-pocket prescription drug spending by $4000 and receiving the LIS has been linked to improved medication adherence.3,6

In this study, we examined take-up of the LIS and Medicaid among Medicare beneficiaries who qualified for both programs. We go beyond prior analyses that report average enrollment by program by examining whether LIS take-up mirrors Medicaid enrollment at income levels where individuals qualify for limited Medicaid benefits that have low take-up rates,5 highlighting opportunities for policy reforms to increase participation in both programs.

METHODS

We analyzed Medicare beneficiaries in the 2008–16 waves of the Health and Retirement Study (surveyed biennially) and linked administrative indicators of Medicaid and LIS enrollment. We restricted our sample to financial respondents (i.e., individuals most knowledgeable about a household’s finances) who qualified for Medicaid and the LIS based on their income and assets.

In primary analyses, we examined LIS and Medicaid take-up as a function of income relative to the federal eligibility threshold for the Qualified Medicare Beneficiary MSP (100% of FPL). Beneficiaries whose income is below this threshold qualify for assistance with out-of-pocket costs in Medicare Parts A and B (provided they have low assets) while those with incomes of 101–135% of FPL only qualify for Medicaid assistance with the Part B premium. In secondary analyses, we assessed take-up above and below state-specific income eligibility thresholds for full Medicaid. Finally, we estimated adjusted differences in LIS take-up above and below these thresholds, controlling for beneficiaries’ demographic and socioeconomic characteristics, health status, disability, Medicare Advantage enrollment, and state and year fixed effects. We adjusted for survey weights to produce estimates that are nationally representative of the community-dwelling Medicare population.

RESULTS

LIS-eligible Medicare beneficiaries qualify for different Medicaid benefits depending on their income and assets (Table 1). Consistent with prior research,5 we found lower Medicaid take-up among beneficiaries who qualified for more limited Medicaid benefits based on income. LIS take-up closely tracked Medicaid enrollment as a function of income and was lowest among beneficiaries with incomes of 101–135% of FPL (Fig. 1). In regression analyses, we estimated that LIS take-up was 24.0 percentage points lower (P < 0.001) among beneficiaries whose income exceeded the eligibility threshold for the Qualified Medicare Beneficiary MSP and 13.7 percentage points lower (P = 0.001) among those whose income exceeded state eligibility thresholds for full Medicaid.

Table 1 Characteristics of Medicaid and Part D Low-Income Subsidies Available to Low-Income Medicare Beneficiaries
Figure. 1
figure 1

Enrollment in Medicaid and the LIS by income. Graphs display proportion of Medicare beneficiaries enrolled in full or partial Medicaid (top panel) and Part D LIS (bottom panel) by income relative to income eligibility threshold for the Qualified Medicare Beneficiary MSP (100% of the Federal Poverty Level). Incomes within ± 4 percentage points of this threshold were excluded. Observations grouped into 20 equally sized income bins; mean enrollment by program is plotted within each bin. Solid lines are estimates; dashed lines are 95% confidence intervals. Medicare beneficiaries in biennial 2008–2016 waves of the Health and Retirement Study (HRS) linked to contemporaneous indicators of Medicaid and LIS enrollment from the Medicare Master Beneficiary Summary File. The study sample included HRS respondents who met eligibility criteria for Medicaid and LIS based on income and assets and excluded respondents in Connecticut, the District of Columbia, Maine, and Indiana (on or after 2014), as these states disregarded income to increase eligibility for the MSPs above the LIS income threshold. Study sample of 7897 HRS observations represented 29,113,139 person-years in community-dwelling Medicare population. Individuals receiving full or limited Medicaid for ≥ 1 month of study year identified from dual eligibility codes in Medicare Master Beneficiary Summary file. Assistance with out-of-pocket costs in Medicare Parts A and B provided via full Medicaid or Qualified Medicare Beneficiary MSP. Individuals must meet income and asset tests to qualify for these programs (Table 1). §Assistance with Medicare Part B premium provided by Specified Low-Income Medicare Beneficiary and Qualifying Individual MSPs. Individuals must meet income and asset tests to qualify (Table 1). ||Individuals receiving full Part D Low-Income Subsidy (LIS) for ≥ 1 month of study year identified from low-income cost share codes in Medicare Master Beneficiary Summary File.

DISCUSSION

Among LIS-eligible Medicare beneficiaries, we observed markedly lower LIS enrollment—corresponding to lower Medicaid take-up—among individuals who qualify for limited Medicaid benefits based on their income. Because Medicaid enrollees automatically receive the LIS, our findings suggest that low enrollment in limited Medicaid contributes to low take-up of the LIS.

Our findings underscore Medicaid’s importance in facilitating take-up of the LIS. Prior research linked low MSP enrollment to limited program awareness and the complexity of applying for these benefits.2 To broaden MSP and LIS participation, states could consider simplifying or automating MSP enrollment, such as waiving asset tests or presumptively enrolling recipients of other means-tested assistance (e.g., food stamps and home heating assistance) in MSPs that cover the Part B premium. Such policies would come at a modest cost to states (in 2020, the Part B premium was $144.60/month) and could be partially offset by savings to State Pharmaceutical Assistance Programs to the extent low-income Medicare beneficiaries use these programs in lieu of the federally funded LIS.2