To Improve Health, Improve Food-as-Medicine Programs

A clinic-based food-as-medicine program did not improve glycemic control among 349 participants with uncontrolled type 2 diabetes and food insecurity who were evaluated in a new randomized clinical trial, according to JAMA Internal Medicine.¹

The participants were randomized to an immediate-treatment group (n = 179) and a waitlist control group (n = 170). Their primary outcome was hemoglobin A_1C (HbA_1C) at 6 months. In the treatment group, participants received up to 10 healthy meal choices/week curated by a dietician, diabetes education, nurse evaluations, and health coaching for 1 year; in the control group, usual care, food bank locations, and an intervention invitation after 6 months were provided. The mean (SD) patient age was 55.4 (11.2) years, and 53.6% were female patients.

At baseline, mean HbA_1C was10.29% (2.02%). Eighty-seven percent reported any diabetes medication use, with metformin and insulin (50.4% each) being the most common. The control group had a slightly shorter time between baseline and 6-month lab test results: 299 (108) vs 306 (114) days.

Body representing healthy food and fitness

Image Credit: Roman - stock.adobe.com

Total clinic visits at 6 and 12 months, respectively, were 13.00 (6.43) and 23.40 (12.20) for the treatment group and 0.72 (1.20) and 10.64 (6.90) for the control group—an “as intended” result, the study authors noted, to demonstrate engagement.

Both groups reported year-over-year improvement in their diet—93.3% of the treatment group and 77.2% of the control group—however, specific measures of diet improvement for the previous week do not demonstrate significant differences, respectively:

• Times ate fruit: 5.56 (3.34) vs 5.24 (3.40)
• Times ate dark green vegetables: 6.29 (20.70) vs 4.01 (2.90)
• Times drank sweetened beverages: 2.78 (3.45) vs 3.41 (4.50)
• Reported never eating fast food: 43.3% vs 36.3%
• Diet improvement index score: 0.39 (0.79) vs 0 (1.00)

Further, although more of the treatment group correctly answered diabetes-related questions, 3 measures of health did not indicate significant adjusted mean differences between the groups:

• Healthy attitudes: 0.13 (95% CI, –0.07 to 0.34)
• Diabetes self-efficacy: –0.03 (95% CI, –0.25 to 0.20)
• Exercise and smoking: 0.05 (95% CI, –0.18 to 0.28)

In addition, while significant HbA_1C reductions were seen in both groups at 6 months, the overall between-group differences were an adjusted –0.10 (95% CI, –0.46 to 0.25; P = .57) and an unadjusted –0.03 (95% CI, –0.42 to 0.35; P = .86), both deemed nonsignificant. Delving deeper, the reductions were a mean 1.3 (2.3) percentage points in the control group and 1.5 (2.2) percentage points in the treatment group—and these results stabilized between the 6- and 12-month measures.

The treatment group had also gained weight compared with the control group, at 6 months, but by 12 months, this difference was no longer significant. Results on differences in blood pressure indicators, hospitalization, emergency department use, outpatient visits, and total claims also were not significantly different.

“The inability of the program to improve glycemic control was despite the treatment group's increased engagement in health care, a priority for patients with greater social disadvantage,” the authors wrote. “Food-as-medicine programs differ in their design, and one explanation for the lack of health improvement in glycemic control is that providing healthy ingredients still may leave participants with the obstacle of preparing the meal.”

With these findings, the investigators stress the need for future evaluation of food-as-medicine programs to include a control group for accurate measure of their effectiveness. An accompanying editorial concurs, noting, “Some might consider it unethical not to provide healthy food to patients with uncontrolled type 2 diabetes and food insecurity. If this trial had been uncontrolled, the intervention would have been found effective. Whenever possible, investigators should randomize interventions to avoid falsely concluding that pre- vs postintervention differences are due to the intervention.”²

References

1. Doyle J, Alsan M, Skelley N, Lu Y, Cawley J. Effect of an intensive food-as-medicine program on health and health care use: a randomized clinical trial. JAMA Intern Med. Published online December 26, 2023. doi:10.1001/jamainternmed.2023.6670
2. Grady D. Food for thought—include controls in policy evaluations. JAMA Intern Med. Published online December 26, 2023. doi:10.1001/jamainternmed.2023.6659