91% of the papers criticizing the ultra-processed-food category had an industry-tied author

On Tuesday we said to watch who makes the argument against the ultra-processed-food evidence, and who paid for the making. A systematic review in the American Journal of Public Health, described this week by the Oregon Research Institute, has done the counting. The researchers found 69 academic articles published between 2009 and 2025 that criticized the Nova classification — the system that defines ultra-processed food — or described potential benefits of such foods, and traced the affiliations of their 232 authors. Ninety-one percent of the articles had at least one author with an identified food-industry relationship; 68.5% of authors had one. For 71% of those authors, the tie was found only through public records, not disclosed in the article. Two-thirds of the articles appeared in journals linked to nutrition societies with their own industry partnerships.

What this does and does not show. It does not show that the criticism is wrong. The lead author said so directly: the study is not saying that criticism of Nova is illegitimate. Some of the critics' points — the category is broad, intake is self-reported, processing and nutrients are hard to separate — are made by independent scientists too, and we have made them here. What the review shows is the shape of the debate: the doubt has a sponsor, the sponsor is mostly undisclosed, and the undisclosed part is the part that should change how a reader weighs a paper. The pattern is familiar from tobacco and sugar. The method for catching it is the same one we gave you for fake citations last year: check who wrote it, not just what it says.

Ben’s take

This is follow-the-money in its purest form. A food company does not need to disprove the evidence; it needs a literature that says the evidence is contested, and 69 papers is a literature. The cost is small — a consultancy, a conference, a society partnership — and the return is a decade of "the science isn't settled" in every news story. The review puts a number on the sponsorship. Keep the number next to every op-ed defending the category.

Loree’s take

My patients do not read the Nova debates. They read the headline that says ultra-processed food might be fine after all, and they feel relieved, because the box is easier than the kitchen. I now have a figure to put beside that headline. It does not settle the science. It tells you who wanted you relieved.

Why it matters to a recovering optimizer. The longevity trap is built on manufactured complexity, and manufactured doubt is its cousin: one tells you health is too hard to manage without products, the other tells you the evidence against the products is unsettled. Both are funded. The habit that defeats them is the same cheap one — look at the affiliations before the abstract.