The meal in the clamshell

Adapted from a newsletter first sent May 8, 2026.

In the spring of 2026 a report estimated that a national rollout of medically tailored meals could avert 2.61 million hospitalizations and save $23.7 billion a year. It’s the kind of number that makes a CFO sit up, and after years of pilots and “promising preliminary data,” I understood why everyone was quoting it.

I kept thinking about something smaller.

I have watched patients hold a meal in a clamshell container they were never going to eat. The science was fine and the macros were fine. But the food tasted like nothing they grew up with. Or it arrived at four in the afternoon on a day when chemo had left them too nauseous to think about food. Or the sodium was so low they assumed something was broken and reached for the takeout menu. Or nobody called to explain why.

A qualitative study around the same time looked at why patients leave these programs and landed in the same place. It wasn’t willpower. It was fit: cultural fit, sensory fit, trust fit.

That changes the question. For about a decade we asked whether medically tailored meals work. Now I think we have to ask whether this meal works for this person, in this week of treatment, in this household, with this food history. No actuarial model answers that. It gets answered at the refrigerator door.

Every big savings estimate rests on a quiet assumption, which is that the meal gets eaten. I don’t know how many of those 2.61 million avoided hospitalizations depend on a fifteen minute phone call nobody put in a budget. I’d like to know.

Previous
Previous

Where the medications ran out