Ultra-Processed Food: The Gap Between the American and European Plate

Here’s a number that does a lot of work on this site: somewhere around 55 to 60% of the calories an average American eats come from ultra-processed food. In parts of Europe, the same figure drops as low as 14%. That’s not a rounding difference. That’s a different way of eating.

If you’ve ever come back from a trip to Italy feeling lighter and clearer and wondered what was in the water, this gap is a big part of the honest answer. Not the only part. But a big one.

The trouble is that “ultra-processed” has become one of those phrases that means everything and therefore risks meaning nothing. So let me do what we do here: define the term properly, show you the real numbers, explain what’s actually been proven, and then tell you where the science is still arguing with itself. Because it is.

What “ultra-processed” actually means

The phrase isn’t vibes. It comes from a specific classification system called NOVA, developed by a Brazilian research team led by Carlos Monteiro at the University of São Paulo. NOVA sorts food not by nutrients but by degree and purpose of processing, into four groups.

Group 1 is unprocessed or minimally processed: an apple, a bag of dried lentils, plain milk, a chicken breast. Group 2 is culinary ingredients: oil, butter, salt, sugar, the things you cook with. Group 3 is processed food: bread, cheese, canned beans, foods made by combining groups 1 and 2 with traditional methods. Group 4, the controversial one, is ultra-processed: industrial formulations made largely from substances you’d never find in a home kitchen. Think protein isolates, modified starches, emulsifiers, and a roster of additives engineered for shelf life, texture, and crave-ability. Soda, packaged snack cakes, chicken nuggets, most breakfast cereals, a lot of “health” bars.

The key idea is that ultra-processed isn’t a synonym for unhealthy on a nutrition label. It’s a description of how industrially a food was built. That distinction is the whole ballgame, and we’ll come back to why it’s both useful and slippery.

The actual gap, in numbers

The American figure is striking. Ultra-processed foods account for roughly 55 to 60% of total daily calories in the United States, depending on the survey year and method. It’s the highest rate in the world.

Europe is not one number, it’s a spread, and the spread is the interesting part. Across 22 European countries, the share of energy from ultra-processed food ranged from about 14% to 44%, lowest in Italy and Romania, highest in the United Kingdom and Sweden (Mendonça and colleagues, European UPF consumption analysis, PubMed). Look at household food availability and you see the same pattern: from around 10% in Portugal and 13% in Italy up to roughly 46% in Germany and 50% in the UK (household availability of UPFs across nineteen European countries, PMC).

So the clean “Europe eats less processed food” line needs an asterisk. Mediterranean and Eastern European countries eat dramatically less. The UK eats almost as much as America. When you “feel better in Europe,” where in Europe matters a great deal. A week in Bologna and a week in Birmingham are not the same experiment.

There’s also a cost dimension that rarely makes the conversation. A 2024 analysis of supermarket staples found that mainstream US stores are so dominated by ultra-processed options that Americans, more than Europeans, are effectively forced to trade off health against price (medRxiv, US vs European supermarket staples). The American food environment doesn’t just permit UPF. In a lot of places it’s the cheap default, and the unprocessed option is the splurge.

The one experiment that changed the debate

For years, the case against ultra-processed food was observational. Big population studies kept finding that people who ate more UPF had worse health outcomes. But observational data has a famous weakness: people who live on soda and snack cakes differ from people who cook lentils in a hundred ways besides the food. Income, exercise, smoking, sleep. Correlation, not proof.

Then in 2019 a team led by Kevin Hall at the National Institutes of Health ran the experiment that observational studies couldn’t. Twenty adults lived inside the NIH Clinical Center and were fed, in random order, two weeks of an ultra-processed diet and two weeks of an unprocessed one. Here’s the clever part: the two diets were matched for presented calories, sugar, sodium, fat, fiber, and macronutrients. On paper, identical. People could eat as much or as little as they wanted.

On the ultra-processed diet, they ate about 500 calories more per day, and they gained weight. On the unprocessed diet, they lost it (NIH news release on the Hall trial; Hall et al., Cell Metabolism, 2019). Same nutrients on the label. Different bodies at the end. The processing itself, the form of the food, seemed to be doing something: people ate faster and consumed more before feeling full.

This is the strongest single piece of evidence that ultra-processing is more than a proxy for junk. But hold the celebration for one second, because rigor cuts both ways.

Where the science is still honestly uncertain

The Hall trial was a landmark, and it was small. Twenty people, four weeks. A tightly controlled metabolic-ward study can prove a mechanism is real without telling you how large it is across millions of free-living people eating thousands of different products. It needs replication at scale, and that work is ongoing.

Then there’s NOVA itself, which has real critics. The category “ultra-processed” lumps a diet soda in with a packaged whole-grain bread, a fruit-and-nut bar, and a tub of plain Greek yogurt that happens to use a stabilizer. Some of those almost certainly aren’t doing the same thing to your body. The classification can also be inconsistent: studies have shown that trained coders sometimes sort the same product into different NOVA groups. When your independent variable is fuzzy at the edges, you should be a little humble about precise claims built on top of it.

And the observational mortality studies, while they keep pointing the same direction, still can’t fully escape the confounding problem. A large multi-country European cohort did find associations between more highly processed diets and higher mortality (Lancet Regional Health Europe, 9-country cohort), which is consistent with the Hall mechanism. Consistent is not the same as airtight.

So the calibrated position, the one I’d actually stake out: the ultra-processed framework is pointing at something real. The Hall trial shows the form of food, not just its nutrients, can drive overeating. The America-versus-Europe calorie gap is genuine and large. But the boundary of the category is blurry, the causal size is not yet nailed down, and “ultra-processed” should be read as a useful rule of thumb rather than a verdict on every single product that falls inside it.

What to do with all this

You don’t need a NOVA classifier app to act on this. The practical signal is simpler than the science. The foods driving the American calorie surplus are the obvious ones: sugary drinks, packaged snacks, fast food, the engineered-to-keep-eating stuff. You already know what they are.

The reason a Mediterranean trip resets something isn’t that Italy banned a chemical. It’s that the default plate there is built mostly from groups 1 through 3: real bread, olive oil, vegetables, cheese, smaller portions, fewer hyper-palatable industrial products within arm’s reach at every hour. The environment does a lot of the work that willpower gets blamed for.

That’s the honest and slightly liberating conclusion. The gap between the American and European plate is real, it’s measurable, and you don’t have to move to Bologna to narrow it. You mostly have to shift the default, more food that looks like food, less that looks like a formulation, and let the form of what you eat do some of the work for you.