Dr. C. Michael Gibson & Dr. Ambreen Mohamed Discuss: FOOD-HF Randomized Trial

Fresh from AHA 2025, this SnackableHealth session probes a deceptively simple question with system-level ramifications: can “food as medicine” measurably improve short-term outcomes for recently hospitalized heart-failure patients? In FOOD-HF, investigators randomized 150 post-discharge patients into three groups: usual care; medically tailored meals (14 frozen, heart-healthy meals delivered weekly); or fresh produce + pantry staples with recipes to enable patient choice. A second, orthogonal experiment tested conditional delivery (food provided when patients kept clinic visits and refilled medications) versus unconditional delivery, targeting the behavior-change hypothesis head-on. Follow-up lasted 90 days. The primary endpoint was HF readmissions/ED visits; a hierarchical composite secondary endpoint blended mortality, HF events, and clinically meaningful health status change using KCCQ.

Headline result: the primary endpoint was neutral—food supplementation (of either type) did not reduce 90-day readmissions/ED visits versus usual care, likely constrained by low event rates in this short window. But the strategic kicker sits in the composite: patients receiving food support achieved significantly greater quality-of-life gains, with many surpassing the ≥10-point KCCQ threshold that clinicians recognize as a tangible, patient-felt improvement. In other words, no hard-event win at 90 days, but a clear patient-reported benefit—a signal payers and regulators increasingly respect.

Design nuances matter. Patients preferred fresh produce/pantry over prescriptive frozen meals—autonomy and cultural fit trumped convenience. And conditional delivery trended toward better performance on the hierarchical endpoint (behavioral nudge for the win), though—again—short duration and low event counts tempered statistical separation on the primary. Acceptability and implementation readouts were strong, underscoring operational feasibility for county and safety-net settings where food insecurity is prevalent and discharge diets (DASH/low-sodium) are hard to maintain.

So what’s the enterprise takeaway? First, don’t oversell a 90-day hard-event reduction—FOOD-HF doesn’t show it. Second, if your north star includes patient experience and functional status, the quality-of-life lift is real and clinically meaningful. Third, when scaling, favor produce-forward models with behavior-linked conditionality—they were more acceptable and directionally stronger. Finally, the path to hard-event deltas likely requires larger, longer studies, targeted to higher-risk phenotypes (e.g., recurrent utilizers, severe food insecurity, low KCCQ at baseline) and paired with medication optimization and sodium-budget coaching.

Bottom line: FOOD-HF reframes “food as medicine” from a feel-good perk into a patient-centered, implementable intervention that moves the KCCQ needle—and hints that smart incentives can align behavior with better health, even if 90-day readmissions don’t budge (yet). That’s actionable intel for health systems piloting SDoH bundles, and a green light for payers to recognize QoL gains while we build the evidence stack for hard outcomes.