Ultra-processed food and the NOVA groups
Last updated September 21, 2026 · By the Foodmetria editorial team · how we work
“Ultra-processed” is not a synonym for unhealthy, and it is not about whether a food came in a packet. It is a specific category — group 4 of the NOVA system — for industrial formulations built from extracted substances and additives. The distinction matters because two controlled trials now suggest processing affects how much people eat, independently of nutrition content.
The four NOVA groups
NOVA was developed by researchers at the University of São Paulo and sorts foods by what was done to them rather than by their nutrients.
| Group | What it covers | Examples |
|---|---|---|
| 1 | Unprocessed or minimally processed | Fruit, vegetables, dried beans, rice, plain yogurt, eggs, fresh and frozen meat, pasteurised milk |
| 2 | Processed culinary ingredients | Oil, butter, sugar, salt, vinegar — things used to cook with rather than eat alone |
| 3 | Processed foods | Group 1 plus group 2: canned vegetables, tinned fish, cheese, cured meat, freshly baked bread |
| 4 | Ultra-processed | Industrial formulations: soft drinks, packaged snacks, most breakfast cereals, reconstituted meat products, many ready meals |
Group 3 and group 4 are the pair people confuse. A loaf of bread from flour, water, yeast and salt is group 3. The same loaf with emulsifiers, dough conditioners and preservatives is group 4. Nothing about the second one is necessarily harmful — but it is a different kind of object, and that is what the classification records.
How to spot group 4 on a label
- Ingredients that are fractions of foods rather than foods: protein isolate or concentrate, modified starch, maltodextrin, invert sugar, hydrogenated or interesterified oil.
- Cosmetic additives — colours, flavourings, flavour enhancers, non-sugar sweeteners, emulsifiers, thickeners, firming and bulking agents. Additives with a purely technical role, such as a preservative in tinned food, do not by themselves make something ultra-processed.
- A long list of things you could not buy to cook with.
Our additive guides cover what regulators have concluded about the common ones, individually.
What the research shows
The bulk of the evidence is observational: across large cohorts, people who eat more ultra-processed food have higher rates of obesity, type 2 diabetes and cardiovascular disease. That is an association. Diets differ in dozens of ways, and people who eat more packaged food differ from those who don't in income, time and much else. Two randomised trials are more informative.
- NIH, 2019
- Twenty adults lived on a research ward and ate whatever they liked from an ultra-processed or an unprocessed menu for two weeks each, in random order. The menus were matched for calories offered, sugar, fat, fibre and salt. On the ultra-processed menu people ate about 500 calories a day more and gained 0.9 kg; on the unprocessed one they lost about the same. The food was eaten faster, which is the leading explanation.
- University College London, 2025
- Fifty-five adults did two eight-week periods, in random order, on diets that both followed UK healthy-eating guidance — one cooked from minimally processed ingredients, one assembled from supermarket products. Weight fell on both: −2.06% on the minimally processed diet against −1.05% on the ultra-processed one, a difference of about one percentage point. Fat mass and triglycerides improved more on the minimally processed diet; LDL cholesterol happened to fall more on the other.
Read together, these say something more useful than either headline. Processing does appear to matter on its own — but when an ultra-processed diet is built to meet nutrition guidance, most of the gap closes. The strongest case for cutting back is for the sweetened drinks, confectionery and snacks at the energy-dense end of group 4, not for every product with a barcode.
A US federal definition is coming
On 10 August 2026 the Department of Health and Human Services, the FDA and USDA announced that a proposed federal definition of ultra-processed foods had been submitted for final review — the first in the United States — citing the absence of a standard definition as an obstacle to consistent research, and putting ultra-processed food at nearly 60% of the American diet. The FDA separately proposed making GRAS notifications mandatory, so that manufacturers must tell the agency when they conclude an ingredient is generally recognised as safe.
Until that definition is published, NOVA is what research and this site use. We will update this page when the final rule appears.
What NOVA does not tell you
- Nutrition. A group 4 product can be low in sugar and salt, and a group 1 food can be fatty. Read the Nutri-Score and the numbers alongside the group.
- Severity. Group 4 makes no distinction between a fortified wholegrain cereal and a can of cola.
- Specific additives. The group is a summary, not a safety assessment of anything in the list.
- Data quality. Classification needs a complete ingredient list. Where ours is missing, we show no group rather than inventing one.
Frequently asked questions
What counts as ultra-processed food?
Under the NOVA system, group 4 covers industrial formulations made largely from substances extracted from foods — protein isolates, modified starches, hydrogenated oils, invert sugars — usually with additives whose job is to make the result look, taste or last better: colours, flavours, emulsifiers, non-sugar sweeteners, thickeners. A practical test: if the ingredient list contains things you would never find in a home kitchen, it is probably group 4.
Is processing the same as being unhealthy?
No. Group 4 is a very mixed bag: wholemeal packaged bread and a can of energy drink both land in it. Pasteurising milk, freezing vegetables and canning beans are processing too, and none of them make a food worse. The useful question is not whether a food was processed, but what was done to it and what was added.
How much of the American diet is ultra-processed?
The US Department of Health and Human Services put it at nearly 60% of the diet in August 2026, when it announced work on a federal definition. Estimates vary with the data source and the year, and shares are typically higher for children than for adults.
Does the United States have an official definition?
Not yet, but it is close. On 10 August 2026 HHS, the FDA and USDA announced that a proposed federal definition of ultra-processed foods had been submitted for final review, built on feedback from thousands of stakeholders. Until it is published, NOVA remains the working standard in research.
What does the research actually show?
Most of the evidence is observational: people who eat more ultra-processed food have higher rates of obesity, type 2 diabetes and cardiovascular disease, which shows association rather than cause. Two controlled trials go further. A 2019 NIH inpatient trial of 20 adults found people ate about 500 calories a day more on an ultra-processed diet matched for calories, sugar, fat, fibre and salt, and gained 0.9 kg in two weeks. A 2025 UCL trial put 8-week diets that both followed UK healthy-eating guidance head to head, and found a real but modest advantage for the minimally processed one: 2.06% weight loss against 1.05%.
So should I avoid ultra-processed food entirely?
That is neither realistic nor what the evidence supports. The two trials point in the same direction — food made from recognisable ingredients makes it easier to eat less without trying — but the second one also shows the gap narrows a lot when the processed diet still follows nutrition guidance. Reducing the sweetened drinks and snacks is where the evidence is strongest; packaged wholegrain bread is not the problem.
How does Foodmetria decide a product's NOVA group?
We display the group derived from the ingredient list in the Open Food Facts database, not our own classification. Where the ingredient list is missing or too short to classify, we show no group rather than guessing.
Explore the data
Sources
- HHS — food policy reforms announcement, 10 August 2026 (federal ultra-processed food definition submitted for review; GRAS notification rule).
- Hall KD et al., “Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain”, Cell Metabolism, 2019 — and the NIH Clinical Center summary.
- Dicken SJ et al., “Ultraprocessed or minimally processed diets following healthy dietary guidelines on weight and cardiometabolic health”, Nature Medicine, 4 August 2025.
NOVA groups shown on Foodmetria come from Open Food Facts and are refreshed daily. This page is general information, not dietary or medical advice. How we work.