Ultra-processed foods: the evidence against them may be weaker than you think

Show summary Hide summary

• Why “ultra-processed” captured public attention
• What the NOVA system claims and how it works
• Why classification often breaks down
• The evidence so far: observational studies and their limits
• Randomized trials tried to isolate processing — and raised questions
  • What those trials did
  • What critics found when they re-examined the trials
• Debate among researchers: nuance, not nullification
• Real-world stakes: policy, markets, and consumer choices
• Practical guidance for shoppers and policy makers

The phrase “ultra-processed foods” has jumped from obscure research papers into mainstream fear. It now shapes grocery aisles, headlines, and policy debates. Yet the science behind the label is messier than the message suggests.

Why “ultra-processed” captured public attention

In a few years, UPF became shorthand for the industrial foods people should avoid. Media guides teach shoppers how to spot them. Activists use the term to push for change. Companies and regulators have responded.

That popularity rests partly on a powerful idea: modern, industrially made foods may be linked to obesity and disease. But a catchy label does not equal clear science.

What the NOVA system claims and how it works

Most UPF research uses the NOVA classification. It divides foods into four groups to signal processing level and help analyze diets.

• Group 1: Unprocessed or minimally processed items. Think whole fruits, vegetables, beans, milk, and raw cuts of meat.
• Group 2: Culinary ingredients used to cook at home, like oils, butter, sugar, and salt.
• Group 3: Foods made by combining group 1 and 2 ingredients, such as homemade bread, stews, or cured meats.
• Group 4 (UPFs): Industrial formulations that include additives, dyes, emulsifiers, or ingredients rarely found in home kitchens.

Why classification often breaks down

The boundaries in NOVA can be vague. Ingredients or processes are sometimes labeled “ultra-processed” even when they are long-standing culinary tools.

• Wheat gluten can be added to packaged bread and flagged as a UPF marker.
• Yet gluten has centuries of use in many cuisines and can be used at home.
• An industrially extruded edamame pasta can be classed as minimally processed, despite major mechanical changes to the food.

In short, NOVA sometimes groups products by ingredient lists rather than by consistent mechanisms.

The evidence so far: observational studies and their limits

Many claims about UPFs rest on large observational studies. They find links between higher UPF intake and worse health outcomes.

But these studies cannot prove cause and effect. Diet surveys often lack the detail needed to tell whether a reported food was truly ultra-processed.

That makes it easy for correlation to be mistaken for causation.

Randomized trials tried to isolate processing — and raised questions

To test causation, some scientists ran randomized controlled trials. These studies assign participants to different diets and measure outcomes.

What those trials did

Several high-profile trials compared diets rich in UPFs with diets based largely on unprocessed foods. They reported that ultra-processed menus led to higher calorie intake and weight gain.

What critics found when they re-examined the trials

A recent analysis of five landmark trials argued those studies did not isolate processing alone. The diets differed in many important ways.

• UPF menus were often more calorie-dense.
• They usually contained less fiber.
• They tended to be softer in texture, which speeds eating.
• Saturated fat and sodium levels sometimes differed.

Any of these factors can cause overeating or adverse health effects on their own. That means the trials may have captured the impact of calorie density, texture, or fiber — not processing per se.

Debate among researchers: nuance, not nullification

Some scientists defend the trials as valuable steps. They say the reanalysis raises plausible alternative explanations but does not fully overturn the findings.

Others call for more refined experiments that match diets on texture, energy density, fiber, and other factors.

Experts agree on one thing: better-designed trials are needed to separate the effects of processing from the effects of composition.

Real-world stakes: policy, markets, and consumer choices

The UPF label now shapes policy discussions and product reformulation. Governments and schools are debating rules that use the concept to define allowable foods.

Certification schemes market “non-ultra-processed” products. Those labels can give unhealthy packaged items a health aura.

There are political consequences too. Demonizing UPFs has been used to attack plant-based meat and milk alternatives. That may push consumers away from options that lower greenhouse gas emissions and, in some cases, improve health.

When a broad category is used to craft policy, the risk is simple: reformulation or stigmatizing alternatives without improving nutrition.

Practical guidance for shoppers and policy makers

Rather than relying solely on the UPF tag, focus on established product attributes linked to health.

• Check calorie density. Lower calories per gram often help with weight control.
• Choose higher fiber options. Fiber supports satiety and gut health.
• Watch added sugars, refined flours, and excess sodium.
• Prefer whole-food-based meals when possible.
• Don’t automatically assume that a single ingredient signals harm.

Good decisions combine ingredient knowledge with attention to calories, fiber, and food texture. Labels can be a starting point, not the whole story.