A weekly bag of beans, grains and fresh produce from a hospital food pantry: the more bags people collected over nearly three years, the more their weight and blood pressure came down

'Food is medicine' programmes hand out healthy food to people who struggle to afford it, in the hope of treating or preventing illnesses such as high blood pressure, obesity and type 2 diabetes. They reliably improve what people eat and how secure they feel about food. Whether they improve the numbers a doctor checks has been less clear: many trials ran for only weeks or months and found little change in blood pressure or blood sugar. A team at Massachusetts General Hospital and Harvard Medical School looked at a programme that has run for years, and published the results in Preventing Chronic Disease, a journal of the US Centers for Disease Control and Prevention, on 1 October 2026.

The pantry sits inside the hospital's community health centre in Revere, Massachusetts, and gives out weekly plant-based grocery packages: fresh fruit and vegetables, canned beans or nut butter, and whole grains. A dietitian first planned the packages around the 2020 US Dietary Guidelines, with enough food for three meals a day for a week for each family member. Most people who use it have low incomes, public health insurance and a first language other than English. The researchers went back through medical records for adults who picked up at least 26 packages between December 2019 and April 2025 and who had check-up measurements both before and after they started. That left 68 people, average age 57, seven in ten of them women. On average they started out in the obesity range (body mass index 31.7), with an HbA1c of 6.9%, in the diabetes range, and a systolic blood pressure of 130 mm Hg.

On average they collected about 98 packages over 2.7 years, roughly one every week and a half. The team then asked whether people who collected more packages changed more than people who collected fewer, comparing the trend in their measurements before and after they started. They did. For every 100 packages, weight came down by about 2.6 kg more and body mass index by about 1 point more, and diastolic blood pressure (the lower number) by about 3 mm Hg more. Among people whose blood pressure was high at the start, the drops were larger: about 6 mm Hg for each number per 100 packages. Systolic blood pressure across the whole group did not change significantly, and neither did HbA1c, the long-term blood sugar measure, which matches what several earlier trials found.

This is a small look back through health records, not a trial. There was no group of similar people who did not get the food, so the study cannot say the groceries caused the changes. People who kept coming back every week may also have been the ones taking their medicines more regularly or getting more help from the health centre's community workers, and the authors note that both could affect the results. Blood pressure readings came from routine clinic visits, which vary from one day and one clinic to the next, and the blood pressure links grew stronger (and for systolic pressure in people who started high, only appeared) once the team adjusted for the different phases of the COVID-19 pandemic. The authors call plant-based programmes like this one 'promising', and say future ones should run for longer and be tested in randomised trials.