Medication Adherence Among Medicare Beneficiaries with Cardiovascular Conditions After the 2024 Inflation Reduction Act Provisions

INTRODUCTION

Older adults with cardiovascular disease often face substantial out-of-pocket costs for essential medications required to manage their chronic conditions. Rising prescription drug prices – driven in part by expensive, brand-name cardiometabolic therapies – have exacerbated these challenges, contributing to cost-related non-adherence and poor outcomes.^1,2 The Inflation Reduction Act (IRA), signed in August 2022, introduced comprehensive reforms to reduce out-of-pocket drug spending for Medicare beneficiaries, including millions with cardiovascular conditions.

Two IRA provisions took effect in January 2024. First, eligibility for full low-income subsidies (LIS) – which limit premiums, deductibles, and co-payments – was expanded to include beneficiaries with incomes between 135–150% of the federal poverty level (FPL). Second, the 5% coinsurance requirement in the catastrophic coverage phase of Medicare Part D was eliminated (effectively capping annual out-of-pocket drug spending at ~$3,300). Using a quasi-experimental difference-in-differences approach, we assessed the impact of these reforms on cost-related medication non-adherence among Medicare beneficiaries with cardiovascular risk factors and conditions.

METHODS

We used National Health Interview Survey (NHIS) data from 2021–2024. We included Medicare beneficiaries ≥65 years with ≥1 cardiovascular risk factor or condition (hypertension, hyperlipidemia, diabetes, obesity, coronary heart disease, angina, myocardial infarction, stroke) who reported taking ≥1 prescription medication in the past year.³

To evaluate the impact of expanding full LIS eligibility, we compared Medicare beneficiaries with household incomes between 135–150% FPL who became newly eligible under the IRA (intervention group) to beneficiaries with incomes ≤135% FPL already eligible for full subsidies and therefore unaffected by the policy (control group). To evaluate the impact of eliminating catastrophic coverage coinsurance, we compared beneficiaries with incomes >150% FPL who were not eligible for LIS (intervention group) with the same control group. Beneficiaries dually enrolled in Medicaid were excluded from our intervention groups as they qualified for pre-existing cost protections. The pre-policy period was 1/1/2021 to 12/31/2023 and the post-policy period was 1/1/2024 to 12/31/2024. The primary outcome was self-reported cost-related medication non-adherence, defined as skipping doses, reducing doses, delaying fills, or foregoing fills due to cost.

We fit survey-weighted linear regression models adjusted for age, sex, race/ethnicity, income, education, employment, and the number of chronic cardiovascular conditions. Our models included an interaction term for group (intervention vs. control) and period (pre- vs. post-IRA provisions), which provided difference-in-differences estimates. We assessed pre-policy period parallel trends visually and through statistical testing. To evaluate the robustness of our findings, we conducted a sensitivity analysis excluding adults with diabetes to distinguish the 2024 provisions from earlier insulin out-of-pocket price caps.⁴

All analyses incorporated NHIS weights to produce nationally representative estimates, accounting for non-response and under-coverage. Missingness was low (<1%), so complete-case analyses were performed. Analyses were conducted in R version 4.4.0. Institutional review board approval was not required as we used publicly available, de-identified data.

RESULTS

The study population included 28,010 Medicare beneficiaries with cardiovascular risk factors and conditions (weighted mean age 74.3 years; 53.9% female). Following implementation of the 2024 IRA reforms, Medicare beneficiaries who became newly eligible for full LIS experienced a 5.5-percentage point reduction (pp) (95% CI −10.9, −0.2) in cost-related medication non-adherence compared with controls already eligible for full subsidies prior to the policy change (Figure 1A). In contrast, Medicare beneficiaries who were subject to elimination of the coinsurance requirement for catastrophic coverage – but not eligible for LIS – experienced a smaller decrease in cost-related medication non-adherence relative to controls that was not statistically significant (adjusted difference-in-differences −1.7pp [95% CI −3.9, 0.5]) (Figure 1B). Analyses excluding beneficiaries with diabetes produced findings similar in direction and magnitude to the primary analysis, although estimates were no longer statistically significant (LIS expansion: adjusted difference-in-differences −4.7pp [95% CI −10.4, 1.0]); catastrophic coverage: −1.8pp [95% CI −4.1, 0.5]). Pre-policy trends were parallel across all analyses.

Figure 1. Cost-Related Prescription Medication Non-Adherence Among Adults with Cardiovascular Risk Factors or Conditions Before and After the Inflation Reduction Act’s 2024 Prescription Drug Provisions.

The plot displays the survey-weighted proportion of National Health Interview Survey participants with cardiovascular risk factors or disease reporting cost-related prescription medication non-adherence by year. Eligible Medicare beneficiaries included adults age ≥65 years who reported taking ≥ 1 prescription medication. Panel A compares Medicare beneficiaries newly eligible for full low-income subsidies (household income 135% to 150% of the FPL) with a control population of Medicare beneficiaries who were already eligible for full LIS prior to IRA implementation (household income ≤135% FPL). Panel B compares Medicare beneficiaries potentially impacted by the elimination of the 5% coinsurance requirement in the catastrophic coverage phase of Medicare Part D (but not eligible for LIS; household income >150% FPL) with the same control group). The pre-policy period was from 2021–2023 and the post-policy period was 2024. Error bars indicate 95% confidence intervals. To evaluate the parallel trends assumption for the difference-in-differences design, we fit survey-weighted linear regression models restricted to the pre-implementation period (i.e. 2021–2023) and included an interaction term between year (as a continuous variable) and group (i.e. Medicare vs. control). A group-by-year interaction p-value > 0.05 indicates that trends prior to policy implementation were not significantly different between the two groups. This group-by-year interaction p-values were 0.82 (Panel A) and 0.87 (Panel B) for cost-related medication non-adherence in the primary analysis, suggesting against differential pre-period trends.

FPL = Federal Poverty Level

DISCUSSION

In this national study of Medicare beneficiaries with cardiovascular risk factors and conditions, expanding eligibility for full LIS to adults with incomes between 135–150% of FPL as part of the IRA’s 2024 provisions was associated with a 5.5 percentage points reduction in cost-related medication non-adherence relative to controls. At the population level, this corresponds to approximately 70,000 fewer socioeconomically vulnerable older adults with cardiovascular conditions experiencing financial barriers to medication adherence. In contrast, eliminating the coinsurance requirement for catastrophic coverage was not associated with a significant reduction in cost-related medication non-adherence.

Many individuals with cardiovascular risk factors and conditions are prescribed costly brand-name drugs, which can lead to financial strain, non-adherence, and poor outcomes.¹ Guideline-directed medical therapy for systolic heart failure includes multiple brand-name agents costing over $3,000 annually, while treatments for cardiac amyloidosis can exceed $15,000 per year.^5,6 Although these therapies reduce morbidity and mortality, access is often limited by steep out-of-pocket costs, particularly among individuals with low incomes.⁷ Our results indicate that subsidies that reduce cost-sharing for prescription medications may mitigate these barriers and improve adherence for lower-income beneficiaries.

While the elimination of coinsurance for catastrophic coverage did not improve medication adherence in the broader Medicare population, additional IRA provisions are expected to further reduce cost-sharing. These include an annual out-of-pocket spending cap of $2,000 for all beneficiaries that took effect in January 2025 and negotiated prices for select high-cost medications (including several cardiovascular therapies), which will begin in 2026.^2,8 At the same time, several Medicare Part D plans have responded by altering their plan design in ways that could increase patient cost-sharing and limit financial relief from the IRA.⁹ Understanding the net effects of these changes will be a critical area of future research.

This study has limitations. First, cost-related medication non-adherence was self-reported, though this measure correlates well with prescription data.¹⁰ Second, because NHIS lacks out-of-pocket medication spending, we could not identify those who directly benefited from the catastrophic coverage cap.

CONCLUSION

Following implementation of the IRA’s 2024 prescription drug provisions, cost-related medication non-adherence significantly decreased among Medicare beneficiaries with cardiovascular risk factors and conditions who were newly eligible for full low-income subsidies. These findings indicate that the IRA’s prescription drug reforms may improve medication access and adherence among Medicare beneficiaries, with important implications for cardiovascular health in the United States.