Highlight
In patients recently hospitalized for heart failure, a 90-day food supplementation program was operationally successful: delivery completion reached 93.6%, adherence was moderate to good, and retention was 96.0%.
Despite high feasibility and acceptability, food supplementation with either medically tailored meals or fresh produce was not associated with a reduction in the primary outcome of heart failure readmission or emergency department visits over 90 days compared with usual care.
Fresh produce was rated more favorably than medically tailored meals, suggesting that patient preference and autonomy may matter when designing nutrition-based post-discharge interventions.
Conditioning food delivery on healthcare engagement, such as clinic attendance and medication fills, did not improve heart failure event risk among participants assigned to food supplementation.
Background
Heart failure remains one of the most burdensome chronic cardiovascular syndromes worldwide, marked by frequent hospitalization, high symptom burden, impaired quality of life, and substantial healthcare spending. The period immediately following discharge after a heart failure hospitalization is especially vulnerable. Patients often face persistent congestion, medication changes, functional limitations, dietary uncertainty, and competing social needs. Not surprisingly, this transition window has been a target for interventions aimed at reducing readmissions and improving recovery.
Diet occupies a central place in heart failure self-management. Patients are routinely advised to limit sodium intake, improve overall dietary quality, and maintain nutritional adequacy. Yet translating dietary advice into daily practice is difficult, particularly for those with food insecurity, financial constraints, limited mobility, poor access to healthy foods, or low caregiving support. In this context, “food-as-medicine” strategies have drawn growing attention. These include medically tailored meals, produce prescriptions, and other structured food support programs designed to align nutritional support with clinical needs.
Observational data suggest that poor diet quality and food insecurity are associated with worse heart failure outcomes. However, evidence from randomized trials remains limited, especially in recently hospitalized patients who may stand to benefit most from practical nutrition support. The present trial by Pandey and colleagues addresses that gap by testing whether direct food supplementation after discharge is feasible and whether it might improve short-term clinical outcomes.
Study Design
Trial overview
This study was an open-label, factorial randomized clinical trial conducted at 2 hospitals in Dallas, Texas, between April 2024 and October 2025. The investigators enrolled adults hospitalized for heart failure within 14 days of discharge and followed them for 12 weeks.
Population
A total of 150 participants were enrolled. Median age was 59.5 years, 60.7% were male, median baseline left ventricular ejection fraction was 35%, and baseline Kansas City Cardiomyopathy Questionnaire-Clinical Summary Score (KCCQ-CSS) was 56.6, indicating substantial symptom burden and impaired health status. Importantly, 52.7% had food insecurity, underscoring the social and nutritional vulnerability of the cohort.
Key exclusions included prior heart transplant, left ventricular assist device, inotropic support at discharge, current enrollment in meal delivery programs, and inability to receive home deliveries. Of the 150 enrolled participants, 2 were withdrawn due to clinical deterioration, 1 died, and 6 were lost to follow-up; all were included in the intention-to-treat analysis.
Randomization and interventions
Participants were randomized in a 1:1:1 fashion to one of three groups: medically tailored meals, fresh produce, or usual care. Participants assigned to food supplementation underwent a secondary 1:1 randomization to conditional or unconditional delivery. The conditional strategy linked delivery to healthcare engagement, specifically clinic attendance and medication fills, whereas the unconditional strategy did not.
This design allowed the investigators to address two practical questions. First, can food supplementation be delivered consistently and accepted by patients after a heart failure hospitalization? Second, does making delivery contingent on engagement with care improve outcomes or implementation?
Outcomes
The main implementation outcomes were delivery completion, adherence, and acceptability. The primary clinical outcome was heart failure readmission or emergency department visit for heart failure over 90 days. Secondary outcomes included a hierarchical composite analyzed using a win ratio, prioritizing all-cause death, total heart failure hospitalizations or emergency department visits, and at least a 10-point improvement in KCCQ-CSS.
The choice of a hierarchical composite is notable. Such measures can capture clinically meaningful changes across several domains, especially in modest-sized trials where individual hard events may be infrequent. However, they also require careful interpretation, because favorable movement in lower-priority components may drive the result even when top-level event reduction is absent.
Key Findings
Implementation outcomes: the intervention clearly worked logistically
The trial strongly supports the feasibility of post-discharge food supplementation in this patient population. Delivery completion was 93.6%, a high rate for a home-based intervention in a socially and medically complex cohort. Retention was similarly strong at 96.0%.
Self-reported adherence was also encouraging. Participants receiving medically tailored meals reported consuming the provided food on a mean of 4.7 days per week, while those receiving fresh produce reported adherence on 5.5 days per week. These figures suggest that participants did not merely receive the intervention; many incorporated it into their weekly routine.
Acceptability was high overall, but fresh produce outperformed medically tailored meals on patient rating. The Net Promoter Score was 8.6 for fresh produce versus 7.3 for medically tailored meals, a statistically significant difference with P = .02. This distinction may be clinically informative. Fresh produce may offer more flexibility, cultural fit, or household sharing potential than prepared meals, all of which can influence patient satisfaction and sustained engagement.
Primary clinical outcome: no detectable reduction in 90-day HF utilization
The primary clinical endpoint did not differ significantly between food supplementation and usual care. There were 23 heart failure readmission or emergency department events among 100 participants assigned to food supplementation and 9 events among 50 participants assigned to usual care. The adjusted rate ratio was 1.09, with a 95% confidence interval of 0.49 to 2.43, and P = .83.
This estimate is centered near the null and accompanied by a wide confidence interval. In practical terms, the trial does not provide evidence that food supplementation reduced short-term heart failure utilization, but it also does not exclude potentially meaningful benefit or harm. Rather, it indicates uncertainty in a small exploratory trial.
Secondary hierarchical composite: signal favoring food supplementation
The hierarchical composite favored food supplementation over usual care, with a win ratio of 1.21 (95% CI, 1.14-1.29; P < .001). At face value, this suggests an overall advantage across the ordered domains of death, heart failure events, and health status improvement.
However, interpretation requires restraint. Because the primary endpoint was neutral, the positive win-ratio result should be viewed as hypothesis-generating rather than practice-changing. It is plausible that quality-of-life gains, captured by improvement in KCCQ-CSS, contributed materially to the composite. That would still be important, especially in heart failure, where symptoms and daily functioning are central outcomes. But it is not equivalent to showing fewer readmissions.
Conditional versus unconditional delivery: no advantage to tying food support to healthcare engagement
Among participants who received food supplementation, conditional delivery was not associated with a lower risk of heart failure events compared with unconditional delivery. This is a useful operational finding. Programs that require proof of appointment attendance or medication fills may be administratively more complex and may unintentionally disadvantage patients facing transportation, scheduling, or pharmacy barriers. In this trial, adding such conditions did not appear to improve clinical risk.
Safety and tolerability
The provided abstract does not report a distinct safety signal attributable to either food intervention. Given the nature of the intervention, major adverse effects would not be expected in most patients, although nutritional interventions in heart failure always warrant attention to sodium content, fluid-sensitive foods, renal function, diabetes comorbidity, and patient-specific dietary restrictions.
Clinical Interpretation
This trial is best understood as a positive implementation study with neutral findings on the prespecified primary clinical endpoint. That distinction matters. In healthcare delivery science, feasibility is not a trivial result. Many promising social care interventions fail because they cannot be executed reliably in real-world patients. Here, the investigators showed that recently hospitalized patients with heart failure can be reached, supplied, and retained in a structured nutrition program over 90 days.
Why then was the primary endpoint neutral? Several explanations are plausible.
First, the trial was relatively small and likely underpowered for event-based outcomes. With only 150 participants and a short follow-up period, even a clinically meaningful reduction in heart failure events could have gone undetected.
Second, heart failure readmissions are influenced by many factors beyond diet quality over 12 weeks, including congestion management, renal function, background guideline-directed medical therapy, adherence to medications, outpatient follow-up, socioeconomic instability, and disease severity. Nutrition support may be necessary but insufficient on its own to shift short-term utilization.
Third, the intervention may exert benefits more readily on patient-centered outcomes, such as energy, symptom stability, food security, or quality of life, than on acute healthcare use. The favorable hierarchical composite raises that possibility, though it needs confirmation in larger trials with more granular reporting.
Fourth, the type of food support may matter. Fresh produce was better received than medically tailored meals. Acceptability can influence real-world effectiveness. Patients may engage more consistently with interventions that respect preferences, culture, household dynamics, and autonomy in meal preparation.
Strengths of the Trial
The study has several notable strengths. It used randomization rather than observational comparison, reducing confounding. It focused on a clinically relevant transition period immediately after hospitalization. It enrolled a population with substantial symptom burden and a high prevalence of food insecurity, making the research socially and clinically meaningful. The factorial design was efficient, allowing assessment not only of food supplementation versus usual care but also of conditional versus unconditional delivery. Finally, the inclusion of implementation measures makes the findings directly useful for health systems considering similar programs.
Limitations and Cautions
The trial also has important limitations. The open-label design was unavoidable but introduces the possibility of reporting bias for subjective outcomes such as adherence and acceptability. The sample size and event count were limited, constraining precision. Follow-up lasted only 90 days, which may be too short to detect the full cardiovascular effects of improved diet quality. The study was conducted at 2 hospitals in Dallas, Texas, which may limit generalizability to other health systems, rural settings, or populations with different cultural food practices.
Another key limitation is interpretive tension between the neutral primary endpoint and the strongly positive hierarchical composite. Without detailed component-level reporting in the abstract, clinicians should avoid overreading the composite result. Hierarchical analyses can be informative, but they do not override a neutral prespecified primary utilization endpoint.
It is also unclear from the abstract how the nutritional content differed quantitatively from participants’ baseline intake, how household food sharing may have diluted intervention exposure, and whether benefits varied by food insecurity status, ejection fraction phenotype, renal disease, diabetes, or baseline KCCQ-CSS. These subgroup and mechanism questions are highly relevant for future work.
How This Fits With Current Heart Failure Care
Contemporary heart failure guidelines emphasize multidimensional management: evidence-based pharmacotherapy, decongestion, device therapy when appropriate, self-care education, rehabilitation, and close outpatient follow-up. Nutrition counseling is widely recommended, but prescriptive evidence for specific food delivery interventions has lagged behind.
This trial does not justify concluding that meal or produce delivery reduces short-term heart failure readmissions. It does, however, strengthen the argument that nutrition support can be integrated into post-discharge care pathways, particularly for socially at-risk patients. For clinicians and health systems, the message is practical: these programs are implementable and acceptable, but expectations regarding short-term utilization reduction should remain modest until larger trials are completed.
From a policy standpoint, the findings are relevant to ongoing interest in reimbursing food-as-medicine programs. Payers and health systems increasingly ask whether medically tailored meals or produce support can lower hospital use. This study suggests feasibility alone should not be equated with near-term event reduction. Future reimbursement models may need to value quality of life, food security, and patient engagement alongside utilization outcomes.
Research Implications
Several next-step questions emerge. Larger multicenter trials are needed to estimate effects on hospitalization and mortality with adequate power. Longer follow-up may better capture changes in disease trajectory. Comparative work should explore which nutritional model works best for which patient: medically tailored meals, produce boxes, hybrid programs, or food plus coaching.
There is also a need to identify effect modifiers. Patients with documented food insecurity, poor baseline diet quality, frailty, or frequent admissions may derive greater benefit than lower-risk groups. Embedded mechanistic measures, such as sodium intake, biomarkers, weight trajectories, congestion scores, and medication adherence, could clarify why a program does or does not affect outcomes. Finally, implementation science should continue to evaluate whether flexible, low-burden delivery models outperform more conditional or administratively complex strategies.
Conclusion
Pandey and colleagues provide important randomized evidence at the intersection of cardiovascular care and social nutrition support. In adults recently hospitalized for heart failure, 90-day food supplementation with medically tailored meals or fresh produce was highly feasible, well accepted, and associated with excellent delivery completion and retention. Fresh produce was more acceptable than medically tailored meals, and conditioning food delivery on healthcare engagement did not improve outcomes.
At the same time, the intervention was not associated with improvement in the primary clinical outcome of heart failure readmission or emergency department visits over 90 days versus usual care. A favorable hierarchical composite suggests the possibility of broader patient-centered benefit, but this finding should be considered exploratory in light of the neutral primary endpoint.
For clinicians, the trial supports food supplementation as a realistic supportive service after heart failure hospitalization, particularly in populations with high food insecurity. For researchers and policymakers, it sets a clear agenda: move beyond feasibility to larger, adequately powered studies that determine whether food-as-medicine can improve the outcomes that matter most in heart failure.
Funding and Trial Registration
ClinicalTrials.gov identifier: NCT06115369.
The funding source is not reported in the abstract provided here. Readers should consult the full JAMA Cardiology article for complete funding, conflict-of-interest, and supplementary methodological details.
References
Pandey A, Keshvani N, Coellar JD, Jain AK, Segar MW, Bustillo-Rubio M, Rizvi SK, Peterson ED. Food Supplementation in Patients Hospitalized for Heart Failure: A Randomized Clinical Trial. JAMA Cardiology. 2026;11(6):564-573. PMID: 41949846.
Heidenreich PA, Bozkurt B, Aguilar D, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure. Journal of the American College of Cardiology. 2022;79(17):e263-e421.
McDonagh TA, Metra M, Adamo M, et al. 2023 Focused Update of the 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. European Heart Journal. 2023;44(37):3627-3639.