Chris van Tulleken, a doctor and broadcaster, argues in Ultra-Processed People that ultra-processed food is not simply food with too much sugar or salt, but a distinct kind of product, designed to be eaten quickly and in excess, and that the processing itself causes harm. The term comes from the NOVA system, devised by researchers in Brazil, which sorts food by how much it has been processed. More cautious is the UK’s Scientific Advisory Committee on Nutrition, which in 2023 called the link between these foods and poor health concerning, but said it was unclear whether the harm comes from processing or from the calories, saturated fat, salt and sugar most of them contain. Both sides agree that diets heavy in such food are bad for health. They disagree about why, and about whether “ultra-processed” is a useful category at all.
The case against ultra-processed food begins with a striking pattern. A 2024 review combining studies of nearly ten million people linked eating more of it to worse results on 32 of the 45 health measures examined, with the most convincing evidence for deaths from heart disease, type 2 diabetes and anxiety. The pattern is consistent, but it is a correlation. People who eat the most ultra-processed food also tend to have less money, less time and less access to fresh food, and each of those could harm health on its own. A link this broad tells us there is something to explain; it does not yet tell us what.
The second difficulty is the label itself. NOVA can put fizzy drinks, crisps and instant noodles in the same group as supermarket wholegrain bread, fortified breakfast cereal, fruit yoghurt and oat milk. When French food specialists were asked to sort the same foods into NOVA’s groups, they disagreed sharply about roughly a quarter of them, even when given the ingredient list. And when a study of more than 260,000 adults in seven European countries broke the category down, sugary and artificially sweetened drinks and animal-based products such as processed meats were linked to later illness, while ultra-processed breads and cereals were not. A category whose members point in different directions cannot simply be condemned as a whole.
What correlations cannot show, an experiment can. In 2019 Kevin Hall’s team at the US National Institutes of Health had 20 adults live in a research clinic, eating an ultra-processed diet for two weeks and an unprocessed diet for two, in random order. The meals offered were matched for calories, sugar, fat, fibre and salt, and everyone could eat as much or as little as they liked. On the ultra-processed diet they chose to eat about 500 more calories a day, enough to show up as a small gain in weight within the fortnight. Because the same people ate both diets under the same conditions, this is evidence of cause, not just correlation. But it was 20 people, for two weeks, in a clinic, eating one set of menus.
A longer and more realistic trial followed in 2025. In a UK study, 55 adults with overweight or obesity spent eight weeks at home on each of two diets, and both diets met the UK’s official healthy-eating guidelines: one built from minimally processed food, the other from ultra-processed equivalents. Participants lost weight on both. They lost about twice as much on the minimally processed diet, 2.06 per cent of their body weight against 1.05 per cent, although independent experts noted that the order in which people ate the diets affected the results. Two findings sit side by side here. Ultra-processed food did not automatically lead to weight gain, which cuts against the strongest version of van Tulleken’s claim. Yet with the nutrients matched, the less processed diet still did better, which cuts against the idea that nutrients are the whole story.
So something about how these diets are made seems to matter, but what? One explanation is the processing itself: van Tulleken points to additives and to food designed to be hard to stop eating. A rival is that ultra-processed food tends to be softer and more energy-dense, packing more calories into each bite, so people eat it faster and take in more before they feel full. Hall’s 2019 trial hints at the second: to match the fibre, his team added some of it to drinks on the ultra-processed diet, leaving the solid food softer and denser in calories. Both explanations fit the trials so far. Telling them apart needs experiments that change one feature, such as energy density, while keeping the rest the same.
The answer matters because each mistake has a different cost, and different people pay it. If “avoid ultra-processed food” becomes the rule, families with little money or time lose cheap, quick staples such as sliced wholegrain bread, baked beans and fortified cereal, and anyone anxious about eating “clean” gains a new source of guilt about ordinary food. If the rule looks only at nutrients, the gain goes to manufacturers: a soft, energy-dense snack can be relabelled “low sugar” or “high protein” while still encouraging people to eat quickly and too much.
The evidence supports a narrower claim than either side’s headline. The trials suggest that something about how many ultra-processed diets are made leads people to eat more than they need, so “processing is irrelevant” does not survive them. But a category that holds both sugary drinks and wholegrain bread is too broad to condemn outright. The better question is not “is this food ultra-processed?” but “which features, in which foods, make people overeat?” That answer is still being worked out, and until it is, the most honest verdict is a provisional one.