Food as Medicine? Follow the Evidence
April 1, 2024 6 minutes read
By Ellen Feldman, MD
Altru Health System, Grand Forks, ND
Summary Points
- Over the course of a year, 400 patients with uncontrolled type 2 diabetes and food insecurity took part in this randomized clinical trial aimed at improving glycemic control.
- The participants were divided into one of two groups: one receiving intensive interventions, including ingredients enough for 10 nutritious meals weekly for the household, alongside comprehensive support and educational resources, while the other was placed on a waitlist for these same interventions.
- At close of the study, the intervention group demonstrated greater involvement in preventive healthcare practices compared to the control group, but there was no statistically significant disparity observed in the reduction of hemoglobin A1c levels between the two groups.
SYNOPSIS: In this randomized, controlled study of more than 400 individuals with uncontrolled type 2 diabetes and food insecurity, an intensive intervention offering healthy groceries and educational efforts failed to significantly affect glycemic control but resulted in heightened engagement with preventive healthcare services in the intervention group compared to the control.
SOURCE: Doyle J, Alsan M, Skelley N, et al. Effect of an intensive food-as-medicine program on health and health care use: A randomized clinical trial. JAMA Intern Med 2024;184:154-163.
As Hippocrates said, “Let food be thy medicine and medicine be thy food.” Hippocrates timeless adage, dating back to 400 BC, has never seemed more pertinent.1 In an era marked by a surge in diet-related chronic diseases, such as obesity, type 2 diabetes, and hypertension, the quest to harness diet as a tool for improving health is gaining momentum.2 This growing concern has sparked a heightened interest in leveraging dietary modifications to enhance health outcomes, both within private sectors and public health initiatives.2,3
A notable example is the expansion of “Food is Medicine” programs under federally funded health plans, as endorsed by the 2022 White House Conference on Hunger, Nutrition, and Health.4 This momentum has led to an uptick in initiatives that offer meals or healthy ingredients as part of a strategy for bolstering health or managing specific health issues.2-4 Observational studies have highlighted a link between these programs and an improvement in dietary habits, as well as a reduction in food insecurity.
However, the evidence connecting these initiatives to objective evidence of health benefits have been more variable.5,6
Acknowledging the gap in concrete evidence, Doyle et al embarked on a carefully designed randomized clinical trial. Their study evaluated an intensive, clinic-based, food-as-medicine program that provided patients healthy ingredients sufficient for 10 meals weekly for a year for the entire household, along with dietician and nurse consultation and health coaching.
The patients participating in this study were encouraged to chose from a diverse array of fruits, vegetables, and entrees and were provided with recipes and cooking instructions during weekly clinic visits. Additionally, they had access to dietician consultation and preventive care from nursing during these visits, and health coaches maintained involvement and communication through periodic phone calls.
Eligibility for participation in the study included adults enrolled in one healthcare system, with a diagnosis of type 2 diabetes and a hemoglobin A1c (HbA1c) level of at least 8%, self-reported food insecurity, and residency within one of the two designated geographic areas — one urban and one rural. After stratification by location and level of HbA1c, participants were randomly assigned into either an active intervention group or waitlist control group for six months.
Participants in the control group received standard care during the initial six months and were offered the active intervention at the end of this period.
Outcomes were measured at six- and 12-month marks. It is noteworthy that this program, conducted from April 2019 to September 2022, overlapped with the time of the COVID-19 pandemic period, prompting protocol modification for both groups.
Table 1 illustrates statistically significant differences in only a few of the outcomes, including:
- mean number of visits to the program clinic for ingredient pickup and preventive appointments at six and 12 months;
- mean number of encounters with a dietician at six months; and
- self-reported improvement in diet quality at six months.
Table 1. Statistically Significant Differences Between Active Intervention and Control Food Groups |
||
| Treatment Group Mean (Standard Deviation) n = 170 |
Control Group Mean (Standard Deviation) n = 179 |
|
Mean number visits to program clinic (six months) |
13.0 (6.43) |
0.72 (1.20) |
Mean number visits to program clinic (12 months) |
23.40 (6.90) |
10.64 (6.90) |
Mean encounters with dietician (six months) |
2.66 (1.75) |
0.62 (1.26) |
Self-report diet improvement (six months) |
93.3% reported improvement |
77.2% reported improvement |
Non-statistically significant differences were found in some measures of program engagement, such as the mean increase in vegetable consumption at six months, the number of patients abstaining from fast food consumption, and reduction in sweetened beverage consumption at both six and 12 months.
Additionally, biometric measures, such as HbA1c and lipid panels, did not differ significantly between the groups, nor did the frequency of hospital visits or emergency department visits show any association.
Commentary
The findings from this Doyle et al study, a randomized clinical trial employing a waitlist control, demonstrate that participation in an intensive food-as-medicine program led to improvements in self-reported diet quality and increased engagement with healthcare providers.
However, despite these positive outcomes, there was no significant enhancement in glycemic control linked to this intervention, as evidenced by a similar reduction in HbA1c levels in the active and control groups.
Interestingly, both the control and intervention groups experienced a noteworthy decrease in glycemic control measures over time, although the difference between the groups was not statistically significant. Had there not been a control group, it is likely that this improved measure may have been attributed to the active intervention.
The study does not include enough information to determine a mechanism propelling the improvement in HbA1c in the control group. One possibility is that the standard care received by the control group was as effective as the active and intense intervention in reducing mean HbA1c. Another possibility is that being on the waitlist for the active intervention may have motivated behavioral changes for some participants.
Understanding the effect of standard or usual care is crucial for contextualizing the study’s findings. Since all the participants were part of a single healthcare system, knowing the specifics of standard care within that system would aid in generalizing the results. The effectiveness of the control intervention (usual care) may vary in healthcare systems with less intensive standard care protocols.
Furthermore, a more detailed description of what constitutes a “healthy diet” would provide a clearer understanding. It is not clear what entrees were offered during the intervention, nor was a full and complete description of the meal plan provided. Additionally, there was no objective documentation regarding meals eaten outside the 10 provided weekly. One notable limitation of this study is the potential influence of conducting research during the COVID-19 pandemic. While efforts were made to mitigate this effect, the global health crisis could have influenced the results in unforeseen ways, emphasizing the need for cautious interpretation.
In her insightful editorial commentary on the research, Dr. Deborah Grady effectively captures the essence of this study’s critical insight, emphasizing the importance of rigorous methodology: “… investigators should randomize interventions to avoid falsely concluding that pre- vs. post-intervention differences are due to the intervention.”7 Although the effectiveness of food-as-medicine programs in influencing health outcomes remains under investigation, the necessity of conducting well-designed trials with control groups is unequivocal. This approach remains indispensable for accurately assessing the effect of such interventions.
References
- Smith R. “Let food be thy medicine...” BMJ 2004;328:0.
- Downer S, Berkowitz SA. Food is medicine: Actions to integrate food and nutrition into healthcare. BMJ 2020;369:m2482.
- Food is Medicine: Current initiative. The Rockefeller Foundation. https://www.rockefellerfoundation.org/initiative/food-is-medicine/#:~:text=Food%20is%20Medicine%20programs%2C%20including,and%20treat%20diet%2Drelated%20diseases
- Mozaffarian D. The White House Conference on Hunger, Nutrition, and Health — A new national strategy. N Engl J Med 2022;387:2014-2017.
- Go AS, Tan TC, Horiuchi KM, et al; KP NOURISH Study Investigators. Effect of medically tailored meals on clinical outcomes in recently hospitalized high-risk adults. Med Care 2022;60:750-758.
- Bhat S, Coyle DH, Trieu K, et al. Healthy food prescription programs and their impact on dietary behavior and cardiometabolic risk factors: A systematic review and meta-analysis. Adv Nutr 2021;12:1944-1956.
- Grady D. Food for thought — Include controls in policy evaluations. JAMA Intern Med 2024;184:163.
In this randomized, controlled study of more than 400 individuals with uncontrolled type 2 diabetes and food insecurity, an intensive intervention offering healthy groceries and educational efforts failed to significantly affect glycemic control but resulted in heightened engagement with preventive healthcare services in the intervention group compared to the control.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content