Few nutrition ideas have travelled as fast as the warning against ultra-processed foods. Study after study links diets high in such products — soft drinks, packaged snacks, ready meals, reconstituted meats — to obesity, type 2 diabetes, heart disease and earlier death. The headlines are usually stark. The science underneath them is genuinely concerning, but also more uneven and more interesting than a simple good-food, bad-food story allows.
The classification at the centre of the debate is called NOVA, developed by researchers in Brazil. It sorts foods by the extent and purpose of industrial processing, from unprocessed ingredients through to ultra-processed formulations made largely from extracts, additives and cosmetic flavourings. As we have noted in our coverage of food and health, a definition built on processing rather than nutrients was always going to produce some awkward border cases.
What the Strongest Evidence Shows
The most persuasive evidence comes from a controlled feeding trial run by the National Institutes of Health, in which volunteers ate about 500 calories a day more, and gained weight, on an ultra-processed diet compared with an unprocessed one — even though meals were matched for calories, fat, sugar and fibre on paper. People ate the ultra-processed food faster, a clue that texture and energy density, not just ingredients, drive overconsumption.
Large observational studies across several countries consistently find that people deriving most of their energy from ultra-processed products fare worse on many health measures. These studies cannot prove causation on their own — heavy consumers differ in income, smoking and many other ways — but the consistency, the dose-response pattern and the feeding-trial result together make a serious case that something about these diets is harmful.
Where the Category Strains
The difficulty is breadth. The ultra-processed label covers cola and confectionery, but also wholemeal sliced bread, many plant-based meat alternatives, flavoured yoghurts and some fortified cereals that dietitians have recommended for years. Subgroup analyses increasingly suggest the risk is concentrated in particular categories — processed meats and sugary drinks especially — while a few ultra-processed subgroups show neutral or even favourable associations, possibly because they displace worse options or carry fortification.
Critics also point out that processing itself is not a mechanism. The plausible culprits bundled inside the label include high energy density, soft textures that encourage fast eating, hyper-palatable combinations of fat, salt and sugar, low fibre, and possibly certain additives or packaging contaminants — each with different evidence and different policy implications. Treating the category as one thing may misdirect both shoppers and regulators.
What This Means at the Supermarket
For individuals, the pragmatic reading is not very different from older dietary advice, which is itself informative. Cook more from basic ingredients where you can; be sceptical of products engineered to be eaten quickly and in quantity; read long ingredient lists of unrecognisable additives as a useful warning sign. But perfection is neither possible nor necessary, and several convenient processed foods can sit comfortably in an otherwise sound diet.
For policymakers, the nuance argues for targeted tools — taxes on sugary drinks, restrictions on marketing to children, clearer front-of-pack labelling — alongside the broader goal of making minimally processed food cheaper and more available. Several governments are watching whether processing-based labels add anything beyond existing nutrient labels before committing to them in law.
The Bottom Line
The ultra-processed concept has usefully shifted attention from single nutrients to the industrial food environment as a whole. Its weakness is treating a very mixed category as uniformly dangerous. Holding both truths at once is less satisfying than a headline, but it is how the evidence actually reads — much as home diagnostics reward careful interpretation, explored in our piece on sleep apnoea home testing.
