Association for Public Policy Analysis & Management

Background and Significance: This research examined the impact of Supplemental Nutrition Assistance Program (SNAP) participation on the health status of older adults.  Extant studies on the topic have found that among food insecure older adults, SNAP participants had similar or poorer self-reported health and displayed greater risks of hospitalization and mortality relative to nonparticipants.  These studies, however, used non-randomized cross-sectional designs that are susceptible to selection effects and do not take into account changes in the nutritional needs of program participants.  Thus, any presumed relationship that is revealed between SNAP participation and poor health-related outcomes may be due to older adults with poorer health enrolling in the program.  As such, this study used event history modeling to examine the effects of SNAP participation on the health status and risk of hospitalization of older adults.   

Methods: This investigation used longitudinal data from the 2002-2012 Waves of the Rand version of the Health and Retirement Study to estimate separate discrete-time hazard models of poor self-reported health and overnight hospital stays.  The study incorporated both contemporaneous and lagged program participation variables to distinguish the effects of recent enrollment versus persistent enrollment in SNAP.  The final sample to be used for analysis included 6,657 individuals ages 60 and older who were followed for up to 6 waves (i.e., 12 years) to see whether they reported fair or poor health or an overnight hospital stay during the study period. 

Results: The findings revealed that among food insecure older adults, contemporaneous SNAP participants were 3.83 (95% confidence interval: 2.68-5.49) times more likely to report fair or poor health and 2.75 (95% confidence interval: 2.08-3.62) times more likely to report an overnight hospital stay relative to non-participants.  Among food secure older adults, contemporaneous SNAP participants were 29% (95% confidence interval: 11-50%) more likely to report fair or poor health and 19% (95% confidence interval: 4-37%) more likely to report an overnight hospital stay relative to non-participants.   While these findings are consistent with prior research indicating that program participants report poorer health and display greater risk of hospitalization relative to non-participants, incorporating lagged participation variables revealed more favorable program effects.  Among food insecure older adults, lagged participation actually decreased the likelihood of an overnight hospital stay by 46% (95% confidence interval: 10-67%).  Among food secure older adults, lagged participation decreased the likelihood of reporting fair or poor health by 67% (95% confidence interval: 37-61%) and also decreased the likelihood of reporting an overnight hospital stay by 18% (95% confidence interval: 2-32%). 

Conclusions and Implications: While correlating contemporaneous SNAP participation with health outcomes may indicate that program participants fare worse than non-participants in the way of self-reported health and hospitalization risk, this may reflect the simple fact that older adults with poorer health are more likely to enroll in the program.  By using a longitudinal design that examined the lagged effects of SNAP participation, this study provides evidence that program participation may indeed be beneficial in maintaining the health of older adults.