A year of online nutrition classes that barely touched on alcohol. Older adults with memory problems ended up drinking about a fifth less

Drinking too much is one of the risk factors for dementia that people can change, and it matters more once memory has started to slip: people with mild cognitive impairment are already at higher risk, and alcohol adds to it. Yet few programmes for them tackle drinking head on. A team at the University Hospital Erlangen asked a side question of a trial they were already running: if older people with memory problems take a long nutrition course that is not about alcohol, does their drinking change anyway? The answer was published in BMC Public Health on 3 October 2026.

The trial, BrainFit-Nutrition, screened 1,111 people aged 60 and over across Germany in 2022 and enrolled 270 living at home with mild cognitive impairment. Each was randomised to one of two computer brain-training programmes and one of two nutrition courses, all done from home. Both courses ran as 12 online group sessions of 90 minutes every two weeks. One taught a whole-food plant-based diet; the other followed the German Nutrition Society's standard advice for a mixed diet. Alcohol came up in a single session on oils, nuts, seeds and drinks, where both courses advised not drinking at all, warned about alcohol's effects on the brain, and added that anyone who did drink should keep it moderate and preferably choose red wine. This analysis followed the 189 participants who drank at the start, average age 71, about half of them women.

A year later, average drinking had fallen from 1.61 to 1.28 standard drinks a day (a German standard drink is 10 g of alcohol), about a fifth less. The fall was similar with both courses: from 1.52 to 1.27 drinks a day with the plant-based course and from 1.70 to 1.29 with the standard one, and the difference between them was not significant. Men cut back by roughly two-fifths to half a drink a day on either course. Women on the standard course cut back by about a third of a drink, while women on the plant-based course barely changed. Among the 84 people who did cut back, the biggest cuts came from those who drank most at the start, those who were older, those who went to more sessions and, perhaps surprisingly, those with poorer memory scores.

There is a big caveat. Every participant got one nutrition course or the other, so there was no group without counselling, and the trial was designed to test thinking skills, not drinking. That means the analysis cannot show that the classes caused the drop; some of it may reflect people drinking less as they age or as their memory worsens, being in a study, or simply telling interviewers what they hoped to hear, since drinking was self-reported and not checked with blood tests. Participants were also far better educated than most people their age. And on average, people were still drinking more than one standard drink a day a year later. The authors call the work exploratory and want larger studies that include a group with no counselling.