Medically tailored groceries improved over time, but not beyond pantry support
A pilot randomized controlled trial tested whether medically tailored groceries and food resource coaching could move the needle for adults living with type 2 diabetes, dyslipidemias, and/or hypertension in a safety-net setting. The study enrolled 210 participants through a co-located food pantry and assigned them to usual pantry support, medically tailored groceries alone, or medically tailored groceries plus coaching for four months. The intervention was built around a straightforward clinical premise: if patients with cardiometabolic disease can reliably obtain nutritious food, the daily burden of glucose, lipid, and blood pressure management should become easier to carry.
It did not translate into a clear between-group separation in this pilot.
Across the cohort, follow-up completion reached 81%, a reasonable retention level for a food-access intervention operating in a resource-constrained environment. Self-reported diet quality, food security, and perceived health improved over time in all groups, each with P < 0.001. Baseline-adjusted analyses, however, did not identify clear differences between the intervention arms and usual pantry support for any of the exploratory outcomes. The pattern is clinically important because it suggests that the pantry itself may already confer a meaningful benefit, narrowing the apparent room for incremental gain from the tailored grocery model during the four-month window.
Blood pressure trended downward, yet statistical separation never solidified
Blood pressure was available from electronic health records for 166 participants, offering a more objective look at one of the most relevant downstream clinical measures. Descriptively, larger blood pressure reductions appeared in the intervention groups. Even so, confidence intervals were wide and the adjusted between-group comparisons did not reach statistical significance. In practical terms, the signal moved in the expected direction, but the study did not establish that medically tailored groceries or coaching outperformed usual pantry support on blood pressure during this early pilot.
Food access interventions often work through several linked mechanisms. Better availability of fresh produce, lower reliance on ultra-processed foods, and more regular meal patterns can reduce sodium load, improve glycemic stability, and support weight and vascular control. Coaching may add a second layer by helping families navigate budgets, storage, cooking, and food selection under real-world constraints. The absence of a clear differential effect in this trial does not negate those mechanisms; it shows how difficult it can be to detect added benefit when the control condition already includes meaningful pantry assistance and the intervention period is relatively short.
A stronger design signal may depend on measurement, dose, and comparison group
The study also included additional exploratory self-reported measures recommended by the funder and detailed in the supplemental materials, including nutrition security, wellbeing measured by EQ-5D-5L, loneliness, adaptive capacity, and the Cooking and Food Provisioning Action Scale. Those outcomes extend beyond diet quality alone and acknowledge that food insecurity affects time, stress, household routines, and psychological strain. Yet even with that broader lens, baseline-adjusted comparisons did not show a clear advantage for either intervention arm over pantry support.
The practical challenge is familiar to clinicians who manage cardiometabolic disease in under-resourced settings. When a food pantry already improves access to staples, adding medically tailored groceries may not produce a large enough incremental effect to register on self-report scales or routine blood pressure data over four months. Dose matters too. A short pilot may be too brief to capture the full physiologic consequences of a better food environment, especially for outcomes such as blood pressure that fluctuate with adherence, stress, medication changes, and the rest of a patient’s social context. Standardized clinical measurement would also matter in a larger study, especially for an outcome as variable as blood pressure.
The pilot still clarifies where the field needs to go next
The trial was registered on ClinicalTrials.gov before recruitment began, with the protocol submitted on February 3, 2024. That pre-registration strengthens confidence in the basic design and reduces the risk of post hoc outcome selection. Funding came from an American Heart Association Food is Medicine Grant, and the funder’s peer review led investigators to add the MTG-only arm and a $15 completion bonus to encourage follow-up data collection. Those modifications matter operationally because they reflect the reality of implementation research: retention, adherence, and participant burden often determine whether a social intervention can be evaluated at all.
The authors’ own conclusion points toward the next study question without overstating this pilot’s results: further studies with standardized clinical measurement are warranted. A larger trial would need enough power to distinguish true treatment effects from the background improvement expected when food support is available to everyone. It would also need to determine whether coaching adds value beyond groceries alone, and whether any benefit emerges most clearly in subgroups defined by baseline food insecurity, medication burden, or disease severity.
For now, the study offers a restrained but useful message. Medically tailored groceries and coaching were feasible to deliver through a community food pantry, participants improved over time, and the intervention arms showed descriptively favorable blood pressure trends. The trial did not prove superiority over usual pantry support, and the wide confidence intervals leave the clinical effect size unresolved. In a field where enthusiasm often runs ahead of evidence, that uncertainty is not a footnote; it is the central result that should shape the next, better-measured trial.