A year of nutrition classes, measuring cups included. In 33 older adults with diabetes or prediabetes, blood sugar did not budge, but body fat slipped

Most diet advice for diabetes arrives as a list of things to cut. A team at the University of Brasília tried something softer with older patients at a public primary care unit in Ceilândia, a satellite city of Brasília: a year of group sessions built on the idea of 'eating competence', which asks whether people enjoy food, are willing to try new things, notice their own hunger and fullness, and can plan and put together meals. Their results were published in the journal Nutrients on 30 October.

The clinic's records listed 187 patients aged 60 or over with type 2 diabetes or prediabetes. After phone calls and the exclusions (insulin users, people with diabetes complications or memory problems), 38 enrolled and 33 finished: 19 with diabetes and 14 with prediabetes, 29 of them women, most retired and with little schooling. Each got a kit of measuring cups and spoons at the first meeting, and a dietitian, a nurse and nutrition students met them again every three months. The sessions covered using household measures to build a plate of rice, beans, couscous, meat or eggs; reading labels that say 'diet', 'light' or 'fit'; whole fruit versus juice; never eating starchy food or fruit on its own but with fibre, protein or fat; and eating slowly, away from screens and with company when possible.

A year later the headline measures were flat. Average HbA1c was 6.09% at the start and 6.09% at the end, and BMI barely changed. The total eating competence score rose from 37.9 to 38.8 out of 48, too small to rule out chance, and most people were already above the 'competent eater' cut-off of 32 when they started (28 of 33), which left little room to improve. Food habits reported in a short questionnaire did not shift either. The one clear change was body fat measured by bioimpedance, down from 39.9% to 38.5% on average with weight steady, which suggests a little fat was swapped for lean mass, though the study did not test why.

Within the group, people whose meal-planning and cooking skills ('contextual skills') improved most tended to lose the most body fat, and those who became more open to new foods tended to drop in a nutrition-risk checklist. The authors call these correlations exploratory, and they ran many of them, so some could be chance. There was no comparison group, so nobody can say whether the classes caused the drop in body fat. Medication changes were not recorded, nearly all participants were women, and the starting HbA1c was already close to the prediabetes range, so there was little to bring down. The authors argue that the separate parts of eating competence may say more than the total score, and that the idea deserves a proper controlled trial.