Adherence patterns in a randomized controlled trial of a client-choice pantry-based food is medicine program

Top Things to Know

Choice and behavioral support may work better together than defaults alone.

High retention in a population receiving safety-net care suggests that a client-choice FIM model integrated with a health system and community food provider can be feasible.

Future FIM trials should distinguish food offered, food selected and food actually consumed, and follow participants longer to determine whether client-choice approaches and coaching produce durable dietary changes and ultimately improve clinical outcomes.

Summary of Conclusion/Findings

This paper reports a 4-month pilot randomized controlled trial of a client-choice, pantry-based FIM program among 210 adults with lower incomes and diabetes, dyslipidemia and/or hypertension recruited from a safety-net health center in Dallas; participants received usual pantry services, medically tailored groceries (MTGs) promoted through default grocery selections, or MTGs plus food resource coaching. Of 330 people who completed screening and were eligible, 210 (64%) enrolled, and 171 (81.4%) of those randomized completed at least three of four visits plus pre- and postintervention surveys; no measured participant characteristics significantly distinguished completers from noncompleters. Participants who preferred Spanish were more likely to enroll than those who preferred English, a finding the authors suggest may reflect bilingual recruitment and the availability of Spanish-speaking staff. The MTG default initially increased selection of foods aligned with the Dietary Approaches to Stop Hypertension (DASH) eating pattern, but this effect weakened over time; by the final visit, MTG alone did not significantly differ from usual care on either nutritional-quality measure. In contrast, participants receiving MTGs plus coaching selected 6.9 percentage points more MTG foods than usual care at the final visit in adjusted analysis, while DASH scores improved over time across all three groups and between-group differences at the final visit were not significant. Findings are exploratory because this was a small pilot at one Dallas safety-net setting, randomization left some baseline imbalances between groups, foods obtained from the pantry were not necessarily consumed, food obtained elsewhere was not captured, and the study did not evaluate clinical outcomes.