The federal government is getting closer to answering one of the most consequential questions in U.S. nutrition policy: What exactly is an ultra-processed food (UPF)?
The answer could shape far more than terminology. A definition emerging from the Make America Healthy Again (MAHA) agenda could eventually influence dietary guidance, front-of-package labeling, institutional food procurement and how consumers think about thousands of foods.
That makes getting the description right important. But defining ultra-processed foods is only half the job.
Policymakers also need to understand how Americans actually consume the foods captured by that definition, that is, how much they eat, how often and what role those foods play in their overall diets.
Otherwise, we risk confusing classification with public health impact.
As I wrote recently in Forbes, “Consumers don’t eat classifications. They eat diets.” That distinction should guide the next stage of the UPF debate.
Not All Consumption Is Equal
One challenge with broad UPF classifications is the enormous diversity of foods they can capture.
A food consumed several times a day and contributing substantial calories can sit in the same category as something eaten occasionally and in modest quantities. Yet their potential contributions to obesity and chronic disease may be very different.
My earlier Georgetown University research illustrates the distinction. Using NHANES data, we found that products such as sugar-sweetened beverages and packaged pastries contributed higher levels of calories and sugar to Americans’ diets, while candy accounted for less than 1.8% of total calories and only 6.4% of added sugar. The consumer research also found that the healthiest consumer segment actually purchased candy more frequently than the population overall. In other words, foods that may share an ultra-processed classification can have very different consumption patterns and contributions to the diet.
This doesn’t mean processing is irrelevant. Research has associated diets high in foods commonly classified as ultra-processed with obesity, cardiovascular disease and other adverse health outcomes. But important questions remain about how and why these foods affect health.
If the objective is healthier Americans, classification should therefore be the beginning of the analysis, not the end.
Before translating a UPF definition into warning labels, procurement standards or other interventions, policymakers should ask: How much of these foods are Americans consuming? How frequently? What do they contribute to calories and nutrients of concern? And would reducing their consumption meaningfully improve the overall diet?
Those questions can help distinguish meaningful population-health targets from foods whose UPF classification may matter considerably less.
The Lesson From Chile
Chile provides an important real-world test.
Beginning with a sweetened-beverage tax change in 2014 and followed by its Food Labeling and Advertising Law in 2016, Chile implemented one of the world’s most comprehensive packages of nutrition policies. Products exceeding specified levels of calories, sugar, sodium or saturated fat received prominent black stop sign warnings. In 2018, Chile went further, banning television advertising of products carrying the warnings from 6 a.m. to 10 p.m., regardless of the audience.
These policies changed purchasing behavior. UNC-Chapel Hill research determined net reductions in purchases of 20.2% for sugar, 13.8% for sodium, 9.6% for saturated fat and 8.3% for calories from affected foods and beverages.
Those are meaningful accomplishments. But as I documented in a 2025 Georgetown University report examining front-of-pack labeling systems around the world, changing purchases is not the same as demonstrating better diets or improved health. Chile illustrates that distinction particularly well.
Despite this unusually comprehensive combination of taxes, warning labels and advertising restrictions, overweight and obesity have continued to rise. Pan American Health Organization (PAHO) data show adult overweight and obesity increasing from 72.6% in 2015 (the year before warning labels were introduced) to 78.8% in 2022. PAHO projects the rate could reach 87% by 2030.
That does not prove the policies had no health benefit; obesity is influenced by many factors, and without them rates conceivably could have risen even faster. But it does demonstrate why reductions in purchases of targeted products cannot by themselves be treated as evidence that the underlying public health problem has been solved.
Consumers can purchase less of a targeted food but compensate elsewhere in their diets. Manufacturers can reformulate products without substantially changing total caloric intake. And policies covering packaged foods can miss substantial consumption from restaurants and other foodservice venues.
Chile therefore shouldn’t be viewed simply as evidence that warning labels work or don’t work. Its experience demonstrates why nutrition policy must follow the full chain – from how foods are classified, to how consumers respond, to what they actually eat, and ultimately to whether health outcomes improve.
Stopping halfway tells us only part of the story.
Labels Don’t Eat Food. People Do.
Food labels can provide useful information and influence purchasing decisions. But research also suggests that the same label does not affect every food, or every consumer, the same way.
Research on Chile’s warning label system found substantial purchasing changes in some food categories, including juices and cereals, while finding little or no effect for chocolates, candy and cookies. Other research suggests consumers may pay less attention to warnings over time as the labels become more familiar. The lesson isn’t that labels don’t work. It’s that their impact depends partly on what is being labeled and who is making the choice.
That distinction is especially important because the consumers most attentive to nutrition information may not be those most in need of dietary change. Consumer segmentation research conducted by the Natural Marketing Institute for Georgetown University’s Portion Balance Coalition found that nutrition information influenced purchases for 71% of the healthiest consumer segment, compared with just 23% of the least health-active segment. The least health-active consumers also had substantially higher rates of overweight and obesity.
That has important implications for UPF policy.
Consider two foods that warrant the same warning. One might be candy or cookies consumed occasionally and in modest amounts. Another might be a sugary beverage or packaged pastry consumed more frequently and contributing substantially more calories or added sugars.
The warning may be the same. But the consumption pattern and potential public health impact may be quite different.
Measure What Matters
A consumption lens would make a federal UPF definition more useful, not less. It could help policymakers prioritize foods according to their actual contribution to American diets rather than treating every qualifying product as an equivalent public health concern.
That’s particularly important when the category is so broad. FDA estimates that roughly 70% of packaged products are commonly considered ultra-processed and that children receive more than 60% of their calories from such foods.
When a definition captures that much of the food supply, classification alone tells us too little. Policymakers need to know which foods people consume most, how much and how often, what they consume instead when behavior changes, and whether those changes ultimately improve health.
A definition can tell us what an ultra-processed food is. How Americans consume it tells us how much it matters. And ultimately, the measure of good nutrition policy isn’t what we classify. It’s whether Americans become healthier.