Practiceby Lab 10 Get started ›
Crossfire

Is "ultra-processed food" a category you can actually shop by?

What the experts say

Yvette d'EntremontBS chemistry, MSc forensic science (self-stated, not independently verified) · science writer, former analytical chemist

No. The term has no operational definition, the descriptions in circulation bin incompatible foods together, and a category that large manufactures correlations. Watch added sugar and fat and mind portions instead.

What it rests onHer column of 25 Jan 2024. She separates processed, which has a legal meaning covering pasteurised milk and flash-frozen vegetables, from ultra-processed, which she says has none, and points to the British Heart Foundation's list putting ice cream, sausages, bread, crisps, cereal, fruit yoghurts, canned fish, carbonated drinks and only some alcoholic drinks in one category. She grants that some studies link poor outcomes to foods in that category. What she is for is the 2023 Nordic Nutrition Recommendations, which she quotes at length and approvingly for setting limits on exactly the components at issue without using the term.

Source: web.archive.org ↗
Marion NestlePhD, MPH · Paulette Goddard Professor of Nutrition, Food Studies, and Public Health, Emerita, New York University

Yes, and it is not all that complicated. A food is ultra-processed if it contains at least one cosmetic additive or one non-culinary ingredient, with named exemptions.

What it rests onHer post of 14 May 2026. She is answering the objection that the category is too hard to define: she quotes Secretary Kennedy saying it 'may be too complicated to define these foods for policy purposes' and answers 'Apparently not.' She then backs the Healthy Eating Research expert panel's three-part test: a product is ultra-processed if it contains at least one Cosmetic additive (flavours, emulsifiers, sweeteners both sugar and non-sugar, colours, thickeners, bulking, gelling, glazing, carbonating and anti-foaming agents) and/or one Non-culinary ingredient (casein, dextrose, high-fructose corn syrup, maltodextrin, modified starch, protein isolates, hydrogenated or interesterified oils, mechanically separated meats, lactose, lecithin); vitamins, minerals, herbs, spices and yeast-derived ingredients are exempt; and foods meeting the FDA's Healthy definition are exempt. Her verdict: 'FDA: take note. This ought to work.'

Overview

The gap here is time, not audience type, and it matters. d'Entremont wrote on 25 Jan 2024 and Nestle's post is dated 14 May 2026, so the definition Nestle backs did not exist when d'Entremont argued that none did. Her objection is to the definitions she had, chiefly the British Heart Foundation's list, and she has published nothing about the Healthy Eating Research panel. She has not rejected Nestle's test.

They are also speaking to different people: d'Entremont is telling a shopper what to do in a supermarket, Nestle is telling a regulator what to write into policy, and a definition can be good enough for the second job and useless for the first.

Where they agree

Substantial, and both say so. Neither thinks people should build a diet on crisps. d'Entremont's positive advice, watch added sugar and fat and mind portions, is close to the Nordic recommendations she quotes approvingly, and those overlap heavily with what Nestle has advised for two decades: do not eat too much junk food, no definition required. Her ultra-processed work is the attempt to turn that instinct into a usable test.

Both are arguing against the same opponent, a food industry that benefits from the confusion. They disagree about whether it has now been cleared up.

What would settle it

For d'Entremont: exactly what Nestle published. A two-part ingredient-list test a shopper can apply, with stated exemptions, that does not put gin, canned tuna and breakfast cereal in one bin, answers her objection directly. Her reply to it, whenever it comes, is the thing to watch for. For Nestle: the FDA declining to adopt the panel definition, or evidence that the test gets enough foods wrong to be useless in a shop.

Practice does not pick a winner here

You do, and your call is dated, kept, and yours to change. Follow the people you already trust and what they publish turns into a short daily practice.

Get started

Free to sign up and use. We do not ask for a card.