Nine in ten papers that contest the definition of ultra-processed foods have industry affiliations, a new study concludes, reviving long-running concerns about commercial influence in nutrition science. Because the definition feeds directly into labelling proposals and dietary guidelines in several countries, the authors say transparency about funding should be treated as a basic requirement rather than an optional courtesy.
Guidelines raise the stakes
The classification behind the controversy was designed to track how modern food systems change what people eat. NOVA treats extensive industrial processing as a marker in itself, separate from fat, sugar or salt content. That idea has influenced guidelines in several countries and the design of warning labels, which explains why papers disputing the definition attract attention well beyond nutrition journals.
Beneath the institutional questions sits a human one: most researchers, including those with industry ties, believe their own conclusions. Motivated reasoning is not corruption, and it operates on publicly funded academics with reputations invested in a framework just as it can on company scientists. That is precisely why the field builds structural safeguards, disclosure, replication, adversarial collaboration, rather than relying on anyone’s purity. The study is best understood as an argument for those safeguards, applied evenly to a debate with real consequences.
Direction of travel in the debate
Another subtlety is direction of travel. Companies fund research on questions they care about, so ties will naturally cluster where commercial stakes exist; that alone does not prove influence over results. The study’s contribution is to quantify the cluster in this particular debate and to argue that labelling policy, which affects every shopper, should be built with that landscape clearly in view.
Reformulation and moving goalposts
Whatever governments decide, definitions will keep evolving as products change. Reformulation, where companies adjust recipes to escape a category, can be a genuine public-health win or a cosmetic shuffle, depending on the definition’s quality. Monitoring frameworks discussed by the World Health Organization increasingly try to capture those shifts, which again raises the premium on definitions built from independent evidence.
Foundations that remain solid
Ultimately, consumers benefit when disagreements happen in the open with interests declared. The alternative, quiet influence over definitions that reach every supermarket shelf, serves no shopper. Until labelling rules settle, the familiar foundations, variety, moderation and mostly minimally processed staples, remain the least controversial guidance available, as this article in our archive also concluded.
Finally, perspective on scale is warranted. The papers disputing the definition are few, and most nutrition research, including most work critical of industry products, is publicly financed. The study does not depict a captured field; it depicts a contested corner where one side’s megaphone is disproportionately commercial. Keeping that proportion in view prevents both complacency and cynicism, and models the measured reading the authors ask of everyone else approaching the ultra-processed debate.
Disclosure technology may soon make analyses like this routine. Machine-readable funding registries, standardised affiliation records and journal-level dashboards could let readers see a paper’s interests at a glance. Until such systems exist, periodic studies auditing the literature remain the main instrument. Their findings, including this one’s striking ratio, function as both evidence and reminder that the instrument is needed at all.
In the end, the controversy is a sign of a maturing field. Nutrition science is acquiring the self-scrutiny, audits of funding, methods and reproducibility, that older disciplines built after their own crises of confidence. That process is uncomfortable for everyone involved and valuable for everyone else. A shopper reading a label in five years’ time will benefit from it without ever knowing this study’s name, which is rather the point.
Slow self-correction in nutrition
Science corrects itself slowly, and nutrition science slower than most, because people cannot be randomised to decades of diet easily. Studies of the literature itself, like this one, are part of that correction mechanism. They do not tell us what to eat. They tell us how much confidence to place in the people arguing about what we eat, which may matter just as much.
