Association between Supplemental Nutrition Assistance Program participation and cost-related medication nonadherence among older adults with diabetes

Importance: Understanding if the association of social programs with health care access and utilization, especially among older adults with costly chronic medical conditions, can help in improving strategies for self-management of disease. Objective: To examine whether participation in the Supplemental Nutrition Assistance Program (SNAP) is associated with a reduced likelihood of low-income older adults with diabetes (aged >/=65 years) needing to forgo medications because of cost. Design, Setting, and Participants: This repeated cross-sectional, population-based study included 1302 seniors who participated in the National Health Interview Survey from 2013 through 2016. Individuals in the study were diagnosed with diabetes or borderline diabetes, were eligible to receive SNAP benefits, were prescribed medications, and incurred more than zero US dollars in out-of-pocket medical expenses in the past year. The data analysis was performed from October 2017 to April 2018. Exposures: Self-reported participation in SNAP. Main Outcomes and Measures: Cost-related medication nonadherence derived from responses to whether in the past year, older adults with diabetes delayed refilling a prescription, took less medication, and skipped medication doses because of cost. To estimate the association between participation in SNAP and cost-related medication nonadherence, we used 2-stage, regression-adjusted propensity score matching, conditional on sociodemographic and health and health care-related characteristics of individuals. Estimated propensity scores were used to create matched groups of participants in SNAP and eligible nonparticipants. After matching, a fully adjusted weighted model that included all covariates plus food security status was used to estimate the association between SNAP and cost-related medication nonadherence in the matched sample. Results: The final analytic sample before matching included 1385 older adults (448 [32.3%] men, 769 [55.5%] non-Hispanic white, and 628 [45.3%] aged >/=75 years), with 503 of them participating in SNAP (36.3%) and 178 reporting cost-related medication nonadherence (12.9%) in the past year. After matching, 1302 older adults were retained (434 [33.3%] men, 716 [55.0%] non-Hispanic white, and 581 [44.6%] aged >/=75 years); treatment and comparison groups were similar for all characteristics. Participants in SNAP had a moderate decrease in cost-related medication nonadherence compared with eligible nonparticipants (5.3 percentage point reduction; 95% CI, 0.5-10.0 percentage point reduction; P = .03). Similar reductions were observed for subgroups that had prescription drug coverage (5.8 percentage point reduction; 95% CI, 0.6-11.0) and less than $500 in out-of-pocket medical costs in the previous year (6.4 percentage point reduction; 95% CI, 0.8-11.9), but not for older adults lacking prescription coverage or those with higher medical costs. Results remained robust to several sensitivity analyses. Conclusions and Relevance: The findings suggest that participation in SNAP may help improve adherence to treatment regimens among older adults with diabetes. Connecting these individuals with SNAP may be a feasible strategy for improving health outcomes.