A medically tailored meal program designed with Navajo communities reduced short-term hospital and emergency visits among adults with heart failure, according to a randomized clinical trial.
The program did more than provide heart-healthy meals. It incorporated traditional Diné foods and recipes, worked with Navajo dietitians, farmers and community health workers and supplied appliances when participants lacked reliable ways to store or heat food.
Among 206 adults receiving care at two Indian Health Service sites, 40.6% of those assigned to the meal program had a hospitalization or emergency department visit within 90 days, compared with 57% of those receiving usual dietary guidance.
The results, published in JAMA Internal Medicine, suggest that nutrition support may improve heart-failure care when it addresses not only what people are advised to eat, but whether food is available, culturally meaningful and practical to prepare.
“This study shows the power of centering communities and leveraging their assets to advance holistic health and well-being,” said Dr. Lauren Eberly, the study’s lead author and an assistant professor of cardiovascular medicine at the University of Pennsylvania. “By integrating traditional Indigenous foods with evidence-based nutrition for heart failure, we were able to significantly improve outcomes in a population that has long faced structural barriers to cardiovascular health.”
The trial, known as MUTTON-HF, enrolled adults with heart failure who had experienced a hospitalization or emergency visit during the previous year. Participants were randomly assigned to receive usual care or the meal program for eight weeks, then followed for 90 days.
People in the intervention group received 14 frozen meals each week, enough for two meals a day. The meals followed sodium-restricted heart-health standards while incorporating traditional Navajo foods and recipes.
A local Navajo dietitian and culinary expert helped develop the menu. Researchers also partnered with Navajo farmers and ranchers to source meat and produce and with Tocabe, a Native-run kitchen, to prepare the meals.
The program purchased each meal for $12, not including the additional costs of developing, distributing and delivering the food.
Getting the meals to participants required a delivery system designed for the realities of rural Navajo Nation. Many participants did not have conventional street addresses, so meals were distributed through community pickup sites, smaller food hubs and home delivery supported by public health nurses and community health representatives.
Some participants also lacked reliable household infrastructure. The program provided items such as microwaves, freezers, propane-powered refrigerators and stoves when needed.
Those supports are central to understanding the results. The study tested a comprehensive, community-designed food program, not the isolated effect of a particular traditional ingredient or recipe.
“Food heals in many ways: through good nutrition, culture, identity, and connection,” said Dr. Sonya Shin, the study’s senior author and an associate professor of medicine at Mass General Brigham. “This intervention demonstrates that culturally grounded approaches are not only respectful but also clinically effective.”
Hospitalizations accounted for much of the difference between the groups.
About 12.3% of participants receiving the meals were hospitalized during the 90-day period, compared with 26% in the usual-care group. Heart failure-related hospitalizations occurred in 3.8% of the meal group and 13% of the usual-care group.
Emergency visits alone were also less common in the meal group, but that difference was not statistically conclusive. The primary finding therefore reflects the combined outcome of having at least one hospitalization or emergency visit, with hospitalizations appearing to drive most of the reduction.
The study did not determine exactly why the program reduced medical visits.
Possible contributors include lower sodium intake, more reliable food access, better overall diet quality, less financial pressure, stronger cultural connection or greater support from the delivery network. Several parts of the program may have worked together.
Food insecurity was common at the beginning of the study. More than 60% of participants in each group had low or very low food security, and many also lacked reliable running water, electricity or basic kitchen equipment.
Compared with usual care, the meal program produced a greater reduction in food insecurity and a small increase in fruit and vegetable intake. Participants also showed a greater improvement in the social-limitation portion of a heart-failure quality-of-life questionnaire, which measures how symptoms affect social activities.
The overall quality-of-life score moved in a favorable direction but did not differ clearly between groups. That distinction is important because the news release describes broad quality-of-life improvement more strongly than the complete results support.
Participants receiving meals also had lower weight and systolic blood pressure after eight weeks compared with the control group.
The average difference in weight change between groups was about 6 pounds, while systolic blood pressure differed by about 7 millimeters of mercury. Other measures, including cholesterol, blood sugar and several heart and kidney biomarkers, did not show clear differences.
Weight change in heart failure should not automatically be interpreted as conventional fat loss. It can also reflect changes in fluid retention, food intake or illness. The study did not establish which explanation accounted for the difference.
The trial was pragmatic and open-label, meaning participants and clinicians knew who received the meals. However, data analysts and staff extracting outcomes from medical records were kept unaware of group assignments.
The main analysis included participants according to their original assignments, regardless of whether they ate every meal or completed every survey. Four people were not available for the full 90-day follow-up, leaving primary-outcome information for 202 participants.
The study’s relatively short duration remains an important limitation. Meals were provided for 60 days and hospital use was measured over 90 days.
The trial does not show whether the benefits continued after meal delivery ended, whether the program reduced deaths or whether the cost of producing and distributing the meals would be offset by lower medical spending.
The results also come from two Indian Health Service sites in rural Navajo Nation. They should not automatically be generalized to every Navajo community, other Tribal nations or all adults with heart failure.
Culturally tailored programs cannot simply be copied from one community to another. Food traditions, infrastructure, health systems and community priorities differ.
The study’s strength is that the intervention was built through community participation rather than adapted after the fact. Diné foods and community knowledge were treated as assets within medical care, not as decorative additions to a standard meal plan.
“This is a powerful example of how health systems can partner with communities to co-create solutions that are both effective and sustainable,” Eberly said. “Programs like this can help reshape how we think about chronic disease management.”
The trial provides early evidence that medically tailored meals may reduce acute health care use when the program addresses the full environment in which eating takes place.
The result should not be reduced to the message that traditional foods prevented hospitalization. Instead, the findings support a broader conclusion: Nutrition care may work better when it combines medical guidance with food access, cultural relevance and the practical resources needed to make the program usable.
The trial was supported by the American Heart Association Healthcare by Food Initiative, the National Heart, Lung, and Blood Institute and an Indian Health Service Innovations Award. The American Heart Association provided mentorship on the study’s design and conduct but had no role in collecting, managing or analyzing the data, preparing or approving the manuscript or deciding to submit it for publication.
One author reported receiving payment from the University of Pennsylvania to produce the medically tailored meals used in the study. That author also reported outside payment for meal production and grants from the Dunn Family Charitable Foundation.