Association for Public Policy Analysis & Management

Cost-related medication nonadherence (CRN) can result in hospitalizations and adverse health outcomes. Up to 20% of older adults are estimated to engage in CRN, and the odds are greater for low-income adults and those threatened by hunger, in particular. Low-income older adults may be forced to choose between basic needs, such as food and medication. The Supplemental Nutrition Assistance Program (SNAP) aims to alleviate hunger, but may also help individuals better afford their medications by reducing out-of-pocket food expenditures. Using nationally representative data and Propensity Score Matching (PSM), to the best of our knowledge, this study is the first to rigorously investigate whether participation in SNAP can reduce CRN for adults aged 60 and older. We found that SNAP participants are 3.2 percentage points less likely to engage in CRN than eligible non-participants, and this effect is more than twice as large for those threatened by hunger.

Previous studies have found that SNAP participants have significantly higher odds of CRN than non-participants. These findings, however, may be a spurious association resulting from the grouping of eligible and ineligible (e.g., higher income) adults. This illustrates the challenge with using techniques that do not take into account program eligibility or factors that may influence program participation. In contrast, using PSM we created a matched comparison group of SNAP-eligible older adults, similar to SNAP-participating older adults, in order to parse out the program effects on CRN. Although PSM cannot control for unobserved confounders, our results remained robust to several sensitivity checks, including experimenting with alternative comparison groups that are likely to be more similar to the treatment group in terms of unobservables.

This study fits well into the conference theme as it demonstrates how existing data can be used to examine important policy questions without conducting experiments or primary data collection, which can be expensive and time-consuming efforts. The results of our study provide valuable policy guidance to policymakers and practitioners at a fairly low cost. While SNAP’s primary goal is to alleviate food insecurity, this study emphasizes the importance of quantifying intended and unintended consequences of public assistance programs which is critical for not only determining their effectiveness, but also their value to participants and society as a whole. Furthermore, this study can motivate the design of additional studies that use detailed administrative data to quantify the impact of SNAP benefit amounts on CRN, rely on longitudinal data or primary data collected through surveys, and which consider other populations of interest such as households with children.

Approximately only 42 percent of eligible older adults currently participate in SNAP. This study provides additional impetus to increase SNAP participation among older eligible adults. In particular, health systems and payers have a vested interest in connecting low-income older adults with food assistance programs given (1) the direct connection between food insecurity and poor health outcomes, (2) and (2) now strong evidence that SNAP participation can help improve medication adherence among older adults which can result in significant cost savings for health systems and insurers.