The Ultra-Processed Food Debate: What the Science Actually Says

Ultra-processed food has become the nutrition topic of the decade. A three-part Lancet Series published in November 2025 declared that UPFs are globally displacing traditional diets and driving chronic disease. The UK Parliament's POST office published a briefing on UPF health impacts in February 2026. Tim Spector has built an entire media platform partly around the message that UPFs are making us ill.

And yet the UK's own Scientific Advisory Committee on Nutrition has stopped short of recommending blanket UPF avoidance, noting that more direct evidence is needed. Food scientists argue that the NOVA classification system, which defines what counts as ultra-processed, groups together products with wildly different nutritional profiles.

Both sides have legitimate points. The problem is that the public conversation has become binary, and binary thinking about food is almost always wrong.

Here is what the evidence actually shows, where the genuine debate sits, and how we built a scoring system that tries to get this right.

What "ultra-processed" actually means

The NOVA classification, developed by researchers at the University of Sao Paulo, groups all foods into four categories:

NOVA 1: Unprocessed or minimally processed. Fresh fruit, vegetables, eggs, plain meat, milk, nuts, grains.

NOVA 2: Processed culinary ingredients. Oils, butter, sugar, salt, flour. Things you use to cook with but rarely eat on their own.

NOVA 3: Processed foods. Tinned vegetables, cheese, fresh bread from a bakery, cured meats. These are NOVA 1 foods with NOVA 2 ingredients added.

NOVA 4: Ultra-processed foods. Products made mostly from substances derived from foods and additives, with little or no intact NOVA 1 food. They typically contain ingredients you would not find in a home kitchen: emulsifiers, stabilisers, flavour enhancers, hydrogenated oils, high-fructose syrups.

The issue is that NOVA 4 is an enormous category. It contains both a can of Coca-Cola and a pot of Alpro soya yoghurt. It includes a Greggs sausage roll and a Quorn mince. A packet of Haribo and a loaf of Hovis wholemeal bread. The classification was designed to study dietary patterns at a population level. It was never intended as a product-by-product shopping guide.

What the research shows

The evidence linking high UPF consumption to poor health outcomes is now substantial and growing. The Lancet Series aggregated data from multiple large cohort studies and meta-analyses. The associations are real: populations that consume more ultra-processed foods have higher rates of obesity, type 2 diabetes, cardiovascular disease, depression, and several cancers.

The UK specifically has one of the highest UPF consumption rates in Europe. Estimates suggest that roughly 50-60% of calories consumed in the UK come from ultra-processed sources. In lower-income households, that figure is higher.

But association is not the same as mechanism, and this is where the debate gets genuinely interesting.

Why UPFs might cause harm

Several mechanisms have been proposed, and they are not mutually exclusive.

Energy density and hyper-palatability. UPFs tend to be energy-dense, soft in texture, and engineered to be extremely easy to consume quickly. A landmark 2019 randomised controlled trial by Kevin Hall at the NIH found that when people were given unlimited access to ultra-processed meals versus unprocessed meals (matched for calories, fat, sugar, fibre, and salt), they consumed roughly 500 more calories per day on the UPF diet. They ate faster. The food was simply easier to overconsume.

Disrupted food matrices. When you eat an apple, the fibre slows the absorption of the sugar. When you drink apple juice, the fibre is gone and the sugar hits your bloodstream rapidly. Processing breaks down the physical structure of food in ways that change how your body processes the nutrients. This is true even when the nutrient profile on the label looks identical.

Additives. Certain emulsifiers (polysorbate 80, carboxymethylcellulose) have been shown in animal studies and some human research to alter the gut microbiome and increase intestinal permeability. Artificial sweeteners may affect gut bacteria and glucose metabolism. These are not settled questions, but the research direction is concerning.

Displacement. Every UPF calorie replaces a calorie that could have come from a whole food. At a population level, the displacement of traditional diets by industrial food products is associated with worse overall diet quality, reduced intake of protective phytochemicals, and lower dietary diversity.

Where the counter-arguments have merit

Food scientists and the more cautious scientific advisory bodies raise points that deserve honest consideration.

NOVA is blunt. Grouping fortified wholemeal bread with confectionery under the same classification does not produce useful consumer guidance. The SACN in the UK and similar bodies in Scandinavia have acknowledged the associations but questioned whether processing itself is the causal factor, or whether it is the nutrient profile, the additives, or the eating behaviour that UPFs encourage.

Not all processing is harmful. Pasteurisation makes milk safe. Fermentation produces kefir, yoghurt, and sourdough. Fortification adds iron and folic acid to flour. Canning preserves vegetables at peak nutrition. Freezing preserves nutrient content better than keeping fresh vegetables in the fridge for a week. Processing per se is not the enemy. Specific types of industrial processing applied to specific foods, producing specific nutrient profiles, is the concern.

Reformulation potential. If the problem is the nutrient profile rather than the processing itself, then reformulating UPFs to be lower in sugar, salt, and saturated fat could address the health concerns without requiring wholesale dietary change. This is the position many food manufacturers prefer, and it is not entirely without scientific basis.

How we handle this in MyFoodFit

When we built the scoring engine, we had to make practical decisions about UPF that went beyond the academic debate. People scan products in supermarkets and need a score that reflects reality.

Here is what we did and why.

We use NOVA as one input, not the only input. A product's NOVA group affects its score, but it does not determine it. A NOVA 4 product with excellent nutritional credentials (high protein, high fibre, low sugar) will score better than a NOVA 4 product with poor nutritional credentials. This matters because telling someone that their Quorn mince is equivalent to a bag of Haribo is not just unhelpful. It is wrong.

We built a Health-Positive UPF system. We identified eight categories of technically ultra-processed foods with demonstrated health benefits: fortified cereals with adequate fibre, plant milks, fermented beverages like kefir and kombucha, plant sterol products, wholemeal bread, high-protein dairy, tofu and tempeh, and plant-based protein alternatives. These products receive a reduced NOVA penalty, provided they pass sugar and saturated fat safety gates.

The maximum penalty reduction is 60%. We never completely remove the NOVA signal, because even health-positive UPFs are still industrially processed and the long-term implications of specific additives remain uncertain. But we also refuse to penalise a pot of Alpro kefir the same way we penalise a can of Monster Energy. That is not science. That is laziness.

We exclude NOVA 4 from all positive guardrails. This is important. Our scoring system has guardrails that protect genuinely healthy whole foods from being unfairly penalised (olive oil does not deserve a low score just because it is high in fat). NOVA 4 products never benefit from these guardrails. A fish finger does not receive the oily fish guardrail floor. A fruit-flavoured sweet does not receive the fruit guardrail floor. The guardrails exist to protect whole foods, full stop.

We are building towards additive-level scoring. NOVA is a blunt instrument. The next evolution is evaluating specific additives in specific products based on emerging research. Not all emulsifiers carry the same risk profile. Not all sweeteners affect the gut microbiome equally. As ingredient text coverage in our database improves, we will move towards scoring that reflects what is actually in the product, not just how processed it is in the abstract.

The practical takeaway

If you eat mostly whole foods, cook from basic ingredients when you can, and treat obviously industrial products (soft drinks, confectionery, crisps, reconstituted meat) as occasional rather than daily, you are doing what the evidence supports.

But if you also eat wholemeal bread from a supermarket, use plant milk in your coffee, eat Quorn bolognese on a Tuesday evening, and give your kids fortified cereal for breakfast, you are not poisoning yourself or your family. You are making pragmatic food choices in a world where most people work full time, cannot afford a purely whole-food diet, and do not have three hours a day to cook from scratch.

The UPF conversation is important. The research is real. But turning it into a purity test helps nobody, and it disproportionately burdens people on lower incomes who rely on affordable processed foods to feed their families.

What helps is knowing which processed foods are genuinely problematic (high sugar, high saturated fat, nutritionally empty) and which are fine (fortified, high in protein or fibre, serving a genuine nutritional purpose). That distinction is exactly what a personalised scoring system should provide.

---

Coming soon to the App Store. Register your interest for early access.

This content is for information only and does not replace medical advice.