Voucher for Healthy Foods and Diabetes Control: A Randomized Clinical Trial.

Importance: Programs to improve access to healthy foods for people with diabetes exist in some jurisdictions but not others, and their effects on disease control are not established. Objective: To measure the effect of a monthly grocery store voucher that allows access to healthy foods on diabetes control in patients experiencing food insecurity. Design, setting, and participants: This randomized parallel-arm clinical trial was conducted from March 21, 2023, to April 1, 2025, at 7 sites in Toronto, Canada. Patients were randomly allocated to a monthly voucher or usual access to healthy foods. Eligibility required hemoglobin A1c (HbA1c) levels from 6.0% to 11% and food insecurity. Data analysis was conducted in May 2025. Intervention: A $65 monthly voucher ($85 for those in households with 6 or more individuals) for 6 months. Main Outcomes and Measures: The primary outcome was change in HbA1c levels from baseline to 6 months. Secondary outcomes included levels of β-carotene and ascorbic acid, self-reported fruit and vegetable consumption, financial security, food security, and general health. Results: Of 1061 patients approached, 390 were enrolled (mean [SD] age, 60 [13] years; 191 [49%] men, 198 [51%] women); 81 (21%) identified as Black, 93 (24%) identified as South Asian, and 105 (27%) identified as White. Overall, 194 (50%) were randomly allocated to receive the voucher and 196 (50%) to control. The nonsignificant mean difference between voucher and control in the change in HbA1c was -0.18% (95% CI, -0.41 to 0.05; P = .13; mean difference adjusted for baseline HbA1c, age, sex, household size, and condition, -0.10%; 95% CI, -0.33 to 0.12; P = .35). The voucher improved self-reported vegetable consumption (44.7% vs 21.5% 2 or more times per day), fruit consumption (42.6% vs 22.7% 2 or more times per day), food insecurity (risk difference, -0.10; 95% CI, -0.18 to -0.02), and general health (odds ratio, 1.6; 95% CI, 1.1-2.3) but not financial insecurity (risk difference, -0.07; 95% CI, -0.16 to 0.02) or β-carotene level (mean difference, 0.01 µmol/L; 95% CI, -0.08 to 0.06) and ascorbic acid level (mean difference, -0.18 µmol/L; 95% CI, -4.3 to 4.7). Conclusions and Relevance: In this randomized clinical trial, a monthly food voucher did not have an effect on HbA1c control or objective measures of diet quality but the voucher did increase self-reported vegetable and fruit consumption. Trial Registration: ClinicalTrials.gov Identifier: NCT05776420.