Weekly medically tailored groceries lowered HbA1c by 0.40 points in a randomized diabetes trial

Six months of weekly medically tailored grocery deliveries lowered HbA1c by an additional 0.40 percentage points compared with usual care in 460 Medicaid-insured adults with type 2 diabetes. The intervention also improved food security and nutrition security, but it did not significantly change body mass index or hypertension. The randomized trial was published online in Circulation on July 21, 2026. 1

The useful takeaway is not a new diet to copy. It is that supplying healthier food, with enough structure to make it usable, can be part of diabetes care when food access is part of the problem.

The trial enrolled adults in Southern California who had Medicaid insurance and type 2 diabetes. To qualify, participants needed at least two HbA1c measurements of 7.5% or higher during the previous year. The average baseline HbA1c was 9.40%; 58% reported food insecurity, 64.8% were women, and 85.2% identified as Hispanic. 2

Participants were randomly assigned for six months to usual care (153 people), lower-dose medically tailored groceries (153), or higher-dose medically tailored groceries (154). The intervention consisted of weekly home deliveries of healthy produce, scaled to household size, plus matched recipes and telenutrition counseling. The lower-dose deliveries were valued at about $100 to $170 per month; the higher-dose deliveries were about $135 to $210 per month. 1

This was a bundled food-access intervention, not a trial of a named eating pattern. It did not prescribe a carbohydrate limit, a protein target, or a particular menu for everyone. That distinction matters when translating the result to ordinary grocery shopping.

The control group improved too, which is why the trial's most relevant number is the treatment difference of 0.40 points rather than the 0.66-point within-group change. The effect was modest, but the confidence interval did not cross zero. Lower- and higher-dose grocery deliveries produced similar HbA1c reductions, so the study did not identify a clearly superior spending level. 1

The program also improved the conditions needed to follow a healthier diet. Among intervention participants, 83.3% said they ate most or all of the food provided. Only 21.5% engaged in telenutrition counseling, which suggests that the grocery supply itself, the recipes, delivery convenience, or relief from food insecurity may have carried more of the program's practical effect than counseling attendance alone. The trial was not designed to distinguish those components, so that remains an inference rather than a finding. 2

Randomization supports a causal estimate for the complete program in this population. It does not show that buying produce by itself will lower HbA1c by 0.40 points. Weekly delivery, household-scaled quantities, recipes, counseling availability, and improved food security arrived together, so the study cannot assign the benefit to one of them.

The participants also started with poorly controlled diabetes and a high level of food insecurity. A food-secure adult without diabetes should not assume the same HbA1c effect. The study measured six-month glycemic and security outcomes, not diabetes complications, long-term weight change, or cardiovascular events. The lack of a significant BMI change is another reason to read this as a glucose-and-access result rather than a weight-loss program.

The funding and disclosures deserve to stay in view. The study was supported by Kaiser Permanente organizations, while Food Smart provided the groceries and telehealth nutrition counseling at no cost. Jason H.Y. Wu reported research funding from several produce organizations, and Dariush Mozaffarian reported research funding, advisory and consulting relationships, and royalties; the other authors reported no conflicts. These disclosures do not erase a randomized result, but they make independent replication useful. 1

If you have type 2 diabetes and food access is unreliable, ask your diabetes clinician, insurer, or community health program today whether medically tailored groceries or a produce-prescription program is available. Treat it as an addition to ordinary diabetes care, not a reason to change medication or glucose monitoring without clinical advice.

For dietitians, the actionable point is equally concrete: assess food access before adding another diet target. A produce-forward grocery plan may be a sensible translation of this trial, but the evidence supports the delivered, supported program in adults with poorly controlled type 2 diabetes, not a universal promise that any grocery box will improve blood sugar.

Read the journal article or its PubMed record.