Consumer awareness of ultra-processed foods (UPFs), and efforts to avoid them, are growing. Elected leaders across the political spectrum have gotten the message, and begun pushing reforms to reduce harms caused by UPFs. The food industry has sought to discredit these efforts by arguing that no one can agree on how to define UPFs. A report released today should put that talking point to rest.
The report reflects the deliberations of a 14-member expert panel convened by Healthy Eating Research last year (full disclosure: I participated on the panel, which also included experts in nutrition, food science, and epidemiology). The panel had two charges: 1) recommend a UPF definition suitable for guiding policy development, and 2) recommend policies to reduce UPF harms. It offers policymakers at all levels a variety of practical, evidence-based solutions to curb UPF harms now.
The emphasis on defining UPFs may seem unnecessary. After all, hundreds, if not thousands, of studies have examined the health outcomes associated with eating ultra-processed foods, using the definition supplied by the NOVA food classification system. Nutrition researchers have developed a robust methodology for applying the NOVA definition to foods and beverages to dietary recall surveys, and conducted sensitivity analyses to measure how much potential areas of ambiguity—e.g. misclassifying an “all natural” potato chip as a UPF—might skew results. These analyses suggest that such edge cases are largely irrelevant. In general, when applying NOVA to dietary recall surveys, researchers consistently find that people who eat lots of UPFs are at greater risk of adverse health outcomes, from cardiovascular disease to depression. These associations are consistent with randomized controlled trials that show UPFs cause increased calorie intake.
Still, industry critics insist that this evidence cannot support policy reforms due to the ambiguity in the UPF definition. In the words of the Consumer Brands Association, “Because there is not an agreed-upon scientific definition of ‘ultra-processed,’ the NOVA system arbitrarily classifies foods based on perception and therefore cannot be considered fact or science-based.” This critique reflects an understandable perplexity with NOVA, which originated with the goal of testing a “generally acknowledged” observation that “increased production and consumption of industrially processed foods and drinks is an important cause of the current pandemics of obesity and related chronic diseases.” Unlike classifications that turn on macronutrients like added sugar, the food industry cannot easily tweak its products to avoid the NOVA “ultra-processed” definition, because the term evolved to describe what the modern food industry produces.
And in part because the modern food industry has infiltrated our diets so comprehensively, the NOVA UPF definition is broad. It captures mostly “junk food” but also many foods that might contribute to a healthy diet. This breadth, however, does not mean the definition is imprecise, much less “arbitrary.” To the contrary, consumers can rely on well-established criteria to identify UPFs based on packaged foods’ ingredient labeling. To the extent that “ultra-processing” does not appear on a product’s label—e.g. high pressure, high temperature extrusion that renders “corn” or “wheat flour” a “processed refined carbohydrate”—that says more about food labeling laws than NOVA. In any event, many if not most, such products also contain some other characteristic UPF ingredient (e.g. “natural flavor,” “high fructose corn syrup”).
What are these “characteristic UPF ingredients”? Under NOVA, a food product qualifies as a UPF if it contains one or more of two types of ingredients: 1) non-culinary ingredients, and 2) cosmetic additives. The first of these are defined as “food substances never or rarely used in kitchens (such as high-fructose corn syrup, hydrogenated or interesterified oils, and hydrolysed proteins).” The second, cosmetic additives, are “classes of additives designed to make the final product palatable or more appealing (such as flavours, flavour enhancers, colours, emulsifiers, emulsifying salts, sweeteners, thickeners, and anti-foaming, bulking, carbonating, foaming, gelling and glazing agents).”
The report released today provides a comprehensive list of both ingredient types—“non-culinary” and “cosmetic additives”—in two appendices that cross-reference functional ingredient class definitions within FDA rules and the international Codex Alimentarius. The report takes a conservative approach. The list of “non-culinary ingredients” is short. If an herb or spice serves as a “cosmetic additive” (e.g. paprika as a coloring agent), it does not count as a characteristic UPF ingredient.
Notably, the vast majority of UPFs have more than one characteristic UPF ingredient. Defining UPFs solely on the basis of “cosmetic additives,” for example, yields a list that includes 98% of the products defined under the full NOVA definition. In other words, just 2% of products made with a “non-culinary” ingredient like high-fructose corn syrup do not also contain a “cosmetic additive.” At the same time, the report reveals shortcomings of using alternative lists of additives to define UPFs. Some recently proposed state laws, for example, would cover as few as 13% of the products captured under NOVA. These laws, while well-intentioned, pose a genuine risk of creating confusion.
Because we know how to define UPFs. It does not mean that macronutrients no longer matter, or that consumers should avoid UPFs at all costs. But all else equal, choosing the less processed alternative to a UPF is good advice for health.
Of course, the food industry has created UPFs in large part because they cost less to produce than “real food.” Given this reality, the evidence that UPFs are increasingly replacing “real food” should cause alarm. It also points to the need for more targeted approaches. Requiring a school district to simply reduce or eliminate UPFs from its menu could end up compromising nutritional quality, not to mention costing a lot of money. The report therefore recommends UPF restrictions on school meals, but with exemptions for UPFs that meet the FDA definition of “healthy” (and also do not contain artificial sweeteners). It also endorses policies like soda taxes that target the worst UPF offenders, as well as countermarketing campaigns, and front-of-pack labeling to help consumers identify UPFs.
As the report notes, researchers working with the project found that approximately 72% of the U.S. packaged food supply now qualify as UPFs under NOVA. These products vary enormously in their healthfulness, with many contributing critical macronutrients to consumer diets. Many of us would be hard pressed to find replacements for them, at least within the current food system. The ubiquity of UPFs in the food supply, however, is no excuse for policymakers to ignore the evidence that these foods are undermining public health, and to adopt reforms that will reduce their harms. Today’s report should help to move more of these reforms forward.