The argument over ultra-processed foods is partly an argument about words. Now a new analysis suggests it is also an argument about interests: researchers found that nine in ten papers disputing the definition of ultra-processed foods involve authors connected to industry. The study stops short of saying those papers are wrong. Instead, it raises a familiar concern in nutrition science, where commercial funding is common and where definitions can influence everything from warning labels to advertising rules and public guidance.
Inside the analysis
To produce their figure, the team gathered papers that take issue with how ultra-processed foods are defined and then inspected the fine print most readers skip: affiliation lines, acknowledgements and competing-interest declarations. Industry ties, in their counting, include employment, consultancy, research funding and links to trade or advocacy bodies supported by companies. Roughly 90 percent of the disputing papers showed at least one connection. The researchers present the result as descriptive evidence about who participates in the definitional debate, not as proof that any particular conclusion is false.
The definition at the centre of the row
Understanding the row requires a minute with NOVA, the four-group scale at its heart. Group one is foods as they come from nature with minimal alteration, like vegetables, meat and milk. Group two covers extracted ingredients used in cooking. Group three is recognisably processed food, such as canned vegetables or cheese. Group four, the contested one, is ultra-processed: industrial formulations designed for convenience, shelf life and palatability, often containing flavourings, colours and emulsifiers. Whether that fourth category is a useful scientific tool, or too broad to guide policy, is precisely what the disputing papers argue about. Our earlier health reporting, including this background article from our archive, explains how the classification reached dietary guidelines.
Conflicts of interest, explained plainly
Scepticism is healthy in science, including scepticism funded by companies with something at stake; arguments should ultimately stand on data and logic. The new study operates at a different level, asking who is making the arguments and whether readers can tell. Its authors point out that disclosure in nutrition journals remains inconsistent, with ties sometimes buried, abbreviated or omitted. Stronger, standardised declarations would not end disputes over ultra-processed foods, but they would let policymakers see the debate’s composition clearly instead of treating every published paper as an equally disinterested voice.
Labelling decisions on the horizon
The practical stakes sit on supermarket shelves. Several governments already use processing in dietary guidance, and labelling schemes under discussion could flag ultra-processed products or incorporate processing into front-of-pack scores. If the definition is scientifically shaky, such labels risk misleading shoppers; if it is sound, industry-linked criticism could delay useful information. Regulators reading the new study are likely to take from it a procedural lesson rather than a scientific one: weigh the whole evidence base, insist on transparent funding, and be alert to how debate in journals can echo into consultation rooms. General nutrition guidance is published by the World Health Organization and food safety information by the US Food and Drug Administration.
Nutrition research is unusually hard to do well. People misremember what they ate, diets differ in dozens of ways at once, and randomised trials lasting decades are impractical. That is why the field leans on classification systems like NOVA to compare large populations, and why arguments about those systems carry such weight. A definition that misclassifies foods can distort every study built on it, which is exactly why critics say scrutiny matters, and why the funding behind scrutiny is a legitimate subject for research in its own right.
How to read food headlines now
The study’s limits deserve the same attention as its headline. It examines papers about a definition, not trials of health outcomes, and industry ties were counted from disclosures that vary in completeness. It cannot show that any critic was insincere, nor that the definition is correct; serious scholars continue to debate both. What it can do, and what its authors ask of the field, is simpler: clearer declarations, more independent funding, and a debate whose participants are visible. On those terms, the argument over ultra-processed foods may become easier for everyone to judge. Further research context is available from the Harvard T.H. Chan School of Public Health, and related coverage appears here.
