Vegetables, calcium and fizzy drinks tracked depression in 1,039 young South Africans, until income came into the sums

Most of what we know about diet and depression comes from Europe, North America and Australia. A study published in Depression and Anxiety on 29 September 2026 looks somewhere the evidence is thin: young adults in South Africa, a country whose diet has been shifting from mostly plant foods toward refined sugar, fat and ultra-processed food. It is part of an international project that wants lifestyle risks for depression and anxiety added to the Global Burden of Disease study, the reference that health ministries use to set priorities.

The data come from the African-PREDICT cohort in the North West Province: 1,039 apparently healthy people aged 20 to 30 (average 24.6), about half Black and half White, about half women, recruited between 2013 and 2017. Trained fieldworkers took three separate 24-hour food recalls from each person, one of them on a weekend day, using food models and photo books for portion sizes. Depression was measured with the PHQ-9 questionnaire; a score of 10 or more, the usual marker of moderate to severe symptoms, counted. At the start, 29.5% met that mark. They were more often Black (62% against 43% of the others) and more often on a low income (58% against 31%).

At first glance diet lined up with mood. People with high scores ate less fruit, vegetables, whole grains, fibre and calcium and drank more sugary drinks. Allowing for how much people ate in total, each extra 100 g of vegetables a day went with 26% lower odds of depression, each extra gram of calcium with 52% lower odds, and each 50 g of sugary drink with about 5% higher odds. But when the team also allowed for age, sex, race and socioeconomic status, those links weakened, and after correcting for the 13 food factors tested at once, none was left. About five years later, 551 people were assessed again. Of the 403 without depression at the start, 50 had developed it. Only one link survived every correction, and only in one of four models: each extra 10 g of milk a day (about two teaspoons) went with 6% lower risk. Ultra-processed food, about 46% of everyone's energy intake, showed no link at all, and more than 90% reported eating no legumes, nuts or seeds, so those foods couldn't be tested.

This is an observational study, so it can't show that any food causes or prevents depression, and it may run the other way: people who are depressed often eat differently. Diet was recorded only once, years before the second check; 42% of the starting group didn't come back; the PHQ-9 is a questionnaire, not a clinical interview; and everyone came from one area and was screened as healthy. The authors themselves call one result, sugary drinks seeming to protect over time, probably a fluke. The useful lesson is the one they stress: in this group, low income and the poorer diet that often comes with it were tangled together, and food advice for mental health is unlikely to work if it ignores the circumstances people eat in.