Background
ART adherence is essential for favorable HIV outcomes, yet suboptimal adherence persists, particularly among low-income populations. Scarcity theory posits that economic hardship imposes cognitive burdens that undermine health behaviors like adherence, implying that alleviating scarcity may improve adherence. The Supplemental Nutrition Assistance Program (SNAP), the largest US anti-poverty program, has two features that create a unique opportunity to evaluate its effect on ART adherence. First, SNAP benefit disbursement dates are randomly assigned at program enrollment, generating random variation in payment timing akin to a randomized trial. Second, SNAP produces a predictable “SNAP cycle” of decreased financial stress and scarcity after payment, followed by a recurrence at month’s end.
Methods
We enrolled SNAP recipients with HIV receiving care at the University of Pennsylvania Health System who reported missing ≥2 ART doses in the prior 30 days. Participants were followed 4 months with daily ART adherence measured by electronic pill caps (CleverCaps). We applied a Bayesian hierarchical change point model with weakly informative priors and random intercepts for participant and month to estimate adherence by day since SNAP payment, accounting for within-person correlation and monthly clustering. Posterior inference was via Markov Chain Monte Carlo with all Gelman–Rubin R̂ < 1.05 indicating good convergence.
Results
We enrolled 59 participants between 6/2024–8/2025 (median age 56 years, 73% male, 70% with income <100% poverty level, median SNAP benefit $291). Median adherence was 71% (IQR 49–89) based on 5,219 person-days of observation. The model identified a change point in adherence 19 days after SNAP receipt (95% Credible Interval [CrI] 7–26), with a 98% posterior probability that the adherence slope decreased thereafter (Figure shows posterior distributions). After the change point, adherence declined by 0.7%/day (95% CrI –1.6 to –0.1), for a total decrease of 10% by the end of the benefit cycle (95% CrI 3-19).
Conclusions
ART adherence rose substantially in the weeks after each SNAP payment and then fell sharply towards the end of the SNAP benefit cycle. To our knowledge, this is the first study to causally link SNAP payment with ART adherence, showing that monthly variation in financial stress influences daily health behavior. With upcoming SNAP policy changes expected to reduce participation, these results underscore the importance of poverty alleviation for HIV outcomes.
Table or Figure
