Ultra-processed food is a phrase you now see on packaging, in headlines, and in arguments online, usually as a synonym for bad. But it is a specific technical term with a real definition, and it does not simply mean processed, or unhealthy, or full of chemicals. Tinned beans are processed and good for you. A home-baked cake is not ultra-processed but is still a treat. Getting the definition straight matters, because once you can actually spot ultra-processed food, the advice about it stops being vague and starts being useful.
The short version
- Ultra-processed food is defined by the NOVA system, which sorts foods by how and why they are processed, not by how healthy they are.
- The giveaway is the ingredient list: ultra-processed foods contain substances and additives you would never cook with at home, like flavourings, emulsifiers, and modified starches.
- It is not the same as processed. Group 1 to 3 are unprocessed and lightly processed foods and kitchen ingredients. Ultra-processed is group 4, the industrial formulations.
- It matters because in a tightly controlled trial, people on an ultra-processed diet ate about 500 more calories a day than on a matched unprocessed diet.
- A large 2024 review linked higher ultra-processed intake to higher risk of several conditions, but not all ultra-processed foods are equal, so use judgement, not fear.
The NOVA system in plain terms
The term comes from a classification called NOVA, developed by a team of researchers led by Carlos Monteiro. Instead of grouping foods by nutrients, NOVA groups them by the extent and purpose of processing they have been through. That is a deliberately different lens. It is not asking how much fat or sugar is in something, it is asking what was done to it and why.
There are four groups. The first three cover food as most people have always understood it. The fourth is the new category that the concern is really about.
| NOVA group | What it is | Examples |
|---|---|---|
| Group 1 | Unprocessed or minimally processed | Fresh fruit, veg, eggs, milk, plain nuts, dried beans, plain frozen veg |
| Group 2 | Culinary ingredients | Oil, butter, salt, sugar, flour, honey |
| Group 3 | Processed foods | Tinned beans, cheese, fresh bread, tinned fish, salted nuts |
| Group 4 | Ultra-processed formulations | Soft drinks, packaged snacks, most breakfast cereals, mass-produced bread, ready meals, vegan and meat nuggets |
NOVA is not a health ranking. A group 3 processed food like tinned sardines is excellent for you. Ultra-processed is a description of how a food was made, not a verdict on how good it is.
The real test: read the ingredients
The most reliable way to spot ultra-processed food is the one the researchers themselves suggest: look at the ingredient list. Ultra-processed foods are industrial formulations made mostly from substances extracted or refined from foods, plus additives whose job is to make the product cheap, shelf-stable, and hyperpalatable. The tell is the presence of ingredients you would simply never have in your own kitchen.
- Ingredients you cannot buy for home cooking: high-fructose corn syrup, hydrogenated oils, protein isolates, invert sugar, maltodextrin.
- Additives for texture and shelf life: emulsifiers, thickeners, modified starches, anti-caking agents, humectants.
- Cosmetic additives: artificial flavours, flavour enhancers, colours, and non-sugar sweeteners that make the product taste and look better than its ingredients would suggest.
- A long list, often with names most people do not recognise, on a product designed to be eaten straight from the pack.
Contrast that with a processed group 3 food. Fresh bread from a bakery might be flour, water, yeast, and salt, all things you could use at home. Mass-produced sliced bread with emulsifiers, preservatives, and added gluten and dextrose crosses into ultra-processed. Same basic food, different category, and the ingredient list is what tells them apart.
If the ingredient list is short and reads like a recipe you could shop for, it is probably not ultra-processed. If it contains things you would never keep in your kitchen cupboard, it probably is.
Why it matters
The strongest single piece of evidence is a tightly controlled trial run by Kevin Hall and colleagues. Twenty adults lived in a research unit and ate either an ultra-processed or an unprocessed diet, each matched for calories, sugar, fat, fibre, and salt on offer, and were told to eat as much or as little as they liked. On the ultra-processed diet, people ate about 500 more calories a day and gained weight. On the unprocessed diet, the same people ate less and lost weight. The food, not the person, drove the difference.
On top of that, a large 2024 umbrella review pulled together dozens of studies covering nearly 10 million people and found that higher ultra-processed intake was associated with higher risks of several outcomes, including cardiovascular-disease-related death, type 2 diabetes, and some mental-health outcomes. The evidence is mostly observational, so it shows a link rather than proof of cause, but the trial and the reviews point the same way often enough to take seriously.
Where the definition gets fuzzy
Honesty matters here, because the category is broad and some of it is misleading. NOVA lumps a can of cola in with wholegrain packaged bread, a supermarket hummus, and a fortified plant milk, all as ultra-processed. Common sense says those are not equally worth worrying about. A soft drink and a wholemeal loaf with a couple of additives do not belong in the same mental bucket, even if the classification puts them there.
Some ultra-processed foods are genuinely convenient and reasonable choices: baked beans, wholegrain cereals with real fibre, plain fortified soy or oat milk, and many plant-based meat alternatives that help people eat less red and processed meat. The label ultra-processed tells you how a food was made, not whether it earns a place in your week. Use it as a prompt to check the food, not as an automatic red card.
Fear of the label can push people toward all-or-nothing eating, or toward guilt over ordinary foods. Ultra-processed is a useful flag, not a moral category. A fortified soy milk and a fizzy drink are both group 4 and are not remotely the same choice.
What to actually do about it
You do not need to eliminate ultra-processed food, which is nearly impossible and not the point. The useful move is to shift the balance of your week toward whole and lightly processed foods, and to be choosy about which ultra-processed items you keep.
- Build meals around groups 1 to 3. Vegetables, fruit, beans, lentils, eggs, dairy, fish, meat, tofu, plain grains, and tinned staples should make up most of your plate.
- Keep the useful ultra-processed foods. Baked beans, wholegrain cereal, fortified plant milks, and wholemeal bread are convenient and fine. Do not throw out the good with the bad.
- Cut the ones that add least. Soft drinks, sweets, crisps, and sugary snacks are the group 4 foods easiest to overeat and hardest to justify. Start there.
- Swap where it is easy. Whole fruit for fruit snacks, plain yoghurt with fruit for the flavoured pot, nuts for crisps, home-cooked batches for ready meals. Small swaps, repeated, do the work.
That is the whole approach. Not fear, not purity, just tilting most of your eating toward food that looks like food, and being deliberate about the packaged items you let in.
Questions people ask
Does ultra-processed just mean unhealthy?
What is the difference between processed and ultra-processed?
Is all ultra-processed food bad for you?
How can I tell if a food is ultra-processed?
References
- Monteiro CA, et al. Public Health Nutrition. Ultra-processed foods: what they are and how to identify them. 2019. doi.org/10.1017/S1368980018003762
- Hall KD, et al. Cell Metabolism. Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake. 2019. pubmed.ncbi.nlm.nih.gov/31105044
- Lane MM, et al. BMJ. Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses. 2024. pubmed.ncbi.nlm.nih.gov/38418082
Last reviewed 18 July 2026. We check health-condition articles against current guidelines and update the date above when we do.
KF.Social Editorial