Ultra-Processed Foods: My Verdict

Guilty, as charged, in case you are quite short on time.? If you have a minute, or ten, by all means- please read on.

Not long ago, what is now known as “ultra-processed” food (UPFs) was simply known as “junk food”.? If the overlap between those two categories is not perfect, it is substantial enough to make the designations largely interchangeable.? As a Venn diagram, it would look something like this:

The evolution of “junk” to “ultra-processed” is important for reasons that go well beyond nomenclature.? Junk food long resided - in the company of another infamous exposure - in the realm of things we couldn’t reliably define, but know when we see them.? That is rate limiting for scientific scrutiny, because an operational definition is required in research.? There can be no systematic study of the properties and impacts of “J” if there is no universal, or at least prevailing, agreement on what is, and is not, “J;” if there is no standardized inventory of properties that differentiate “J” from “non-J.”

Enter the work of Prof. Carlos Monteiro and colleagues, and the development of the NOVA classification of food processing.? NOVA specifies criteria for degrees of processing, allowing for a standardized definition of ultra-processed.? That, in turn, has allowed for research into the specific effects of ultra-processing.? That research, and media attention to it, have engendered a lively preoccupation and societal dialogue.? What we long took for granted- junk food- is having a cultural moment.

We have Prof. Monteiro and colleagues to thank for that, and thank them, I do.? Absent the operational definition of UPFs, we would not have randomized trials (for which, among others, I thank, in particular, Kevin Hall); we would likely not have the national dialogue; and we would not have- at long last- a class action lawsuit.

As with all operational definitions and rating schemes, NOVA can- and probably should- evolve.? As edge cases and exceptions are evinced, they can help refine the definitions and distinctions so they are more universally useful in all of the intended ways.? By way of illustration, imagine that the first ornithologist (Prof. Monteiro might be a founding “UPFologist”) defined birds as “vertebrate animals capable of flight.”? That definition would be overwhelmingly right, and would distinguish most birds from most non-birds, but it would also- of course- be wrong.? The definition would include bats and exclude, among others, penguins, ostriches, and emus.? One imagines an emu afficionada might really get her feathers fuffled by the misclassification.? A refined operational definition of birds would be along the lines of: “feathered, egg-laying, beaked, winged vertebrate animals, the vast majority of which are capable of flight.”

In the case of NOVA, I think it would be more right to define “ultra-processed” not merely by the inclusion of select ingredients, but also by the “willful combination of those and other ingredients in the service of producing hyperpalatable, and addictive or nearly addictive food items.”*? (A shout-out of thanks here to Michael Moss.). The ingredients that currently invite the use of the UPF designation might be deemed necessary, but not quite sufficient on their own.? Here, as an example of an edge case, a relatively simple, mass-produced whole wheat bread with some added texturizer and emulsifier would likely qualify under NOVA as ultra-processed, and perhaps it shouldn’t.? On the other hand, such a bread recipe with additions of sugar, salt, and several chemical compounds perhaps should qualify after all.

Either way, there is room for debate at the margins.? That’s healthy, and how science improves itself over time.

For now, those edge cases are proving a source of friction and discord, as recently chronicled in The New Yorker.? As is so often the case where science meets media meets pop culture, I think the dissent is much overblown, and the consensus much overlooked.? For the most part, ultraprocessed foods are bad actors because they are engineered to be bad actors- and the right set of questions would lead to highly confluent answers across the expanse of expertise in nutrition and public health.

Obscuring that consensus are some prevailing fallacies, both in the public understanding of science, and in its application.

Of part and whole: Science tends to look for “active ingredients.”?? We have advanced our understanding in all areas with such reductionism, but it has important limitations.? Imagine studying avalanche fatalities by trying to identify “which snowflake” was responsible for the deaths.? Imagine studying flood control by requiring that one sandbag in a levee be the “active ingredient.”? There are sums with properties beyond their isolated parts, and these can only meaningfully be studied in their entirety.? The methods of science often struggle with this, and that may be the case in the study of UPFs.? The properties of UPFs that both mortgage public health and fatten corporate coffers are, for the most part, more recipe-based than ingredient-based.

The edge case issue, noted above, is another liability.? Science seeks out operational definitions that are universal.? If A causes B, for instance, then A should always cause B.? But if one considers that the ultra-processing of food is a means to an intentional end- making food hyperpalatable and addictive- it stands to reason that there would be both successes and failures.? Sometimes those means would fail to produce the intended ends.? More innocently, at other times, the contributing ingredients of ultra-processing would be applied to different intentions- such as preventing spoilage and extending shelf life. Accordingly, stringent scholars might identify some UPFs that don’t impose the expected harms, and conclude that the classification is unhelpful, or even harmful.? In my view, that makes perfect the enemy of good.? The operational definition of ultra-processing, in my view, is very good.? That’s good enough, even if not perfect.

The edge case limitation leads to a greater one: the baby and bathwater problem.? If we consider the vast inventory of UPFs that fully satisfy the food industry intentions of addictive hyperpalatabilty as their “successes,” their periodic failures will be UPFs that don’t reveal such harms.? More importantly, the same methods of processing could, at times, serve entirely laudable intentions- such as the fabrication of plant-based alternatives to satisfy carnivorous palates.? These, too, would be “UPFs” because of overlapping methods and formulations.? Some would- erroneously, in my opinion- label these products as harmful, simply because they share in the elements of ultra-processing; others would impugn the whole classification system because of the highly occasional visit by such a baby to a vast sea of bathwater.? I find it regrettable that the plant-based meat market suffered the blight of the UPF label.? As alternatives to meat for people who otherwise would be eating hamburgers, these products offered clear benefits to planetary health and our treatment of fellow creatures.? We later learned they appear, despite the processing required to mimic meat, to offer personal health benefits as well.? Ideally, the definition of UPFs would provide for suitable carve-outs where bathwater ends, and baby begins.

Finally, and perhaps most importantly, is the subset of fallacies we might bundle under the adage: “absence of evidence is not evidence of absence.”

Imagine a machine shop using methods “Z” to manufacture both…shock absorbers, and grenade launchers (I am not saying it’s plausible; it’s just a “what if?”).? Imagine lumping those products together as ‘POZs,’ namely: products of manufacturing method Z.? And now imagine the quest to determine if POZs are lethal.? One could look at grenade launchers and conclude YES.? One could look at shock absorbers and conclude NO.? Or one could look at the mix of two divergent items and conclude that the POZ designation is useless, or worse.?

But the remedy is the obvious one: when the methods of Z are combined with the intent to make a lethal product, the product is lethal.? When intentions are otherwise, so, too, are the products.? The means of Z can be applied to disparate ends.

So, too, the ultra-processing of foods.? Intentions matter.

The recently completed work of the 2025 Dietary Guidelines Advisory Committee failed to offer any clear guidance on UPFs for relative want of randomized controlled trials.? But given the transparency of food industry intentions and the apparent harms on full display, I believe the precautionary principle, rather than some arbitrary mass of randomized trials, should have decided the matter.? I would advise minimizing the role of UPFs in your diet and generally steering clear of them in any food category where you have access to less processed alternatives.? As an example of where that might not be the case, plant-based meats again come to mind.? These are processed not to be addictive, but to mimic the properties of meat so that carnivores might be won over.? I fully support that agenda.?

Presumably because the study of UPFs has (predictably) implicated them in overconsumption and weight gain, and by extension the obesity pandemic, reflections on UPFs seem to reanimate reflections on that pandemic as well.? Those, in turn, routinely harbor the contention that pandemic obesity is an insoluble mystery.

I disagree, emphatically.? Obesity is readily explained where body mass meets a critical mass of considerations.

Consider, for instance, if the drowning of Titanic passengers were declared an insoluble mystery as the “iceberg density” theory competed with the “hull strength” theory, and these in turn competed with the “water temperature” theory, and then as all of these competed with theories pertaining to age, sex, social class, berth, and swimming ability.? Imagine further that each of these is studied, as best it can be, independently of the others, leading to the conclusion that no one of them fully explains the drownings.? Hence, the insolubility of the mystery.

But of course, there is no mystery at all; the explanation for all the tragic drownings was the sinking of the ship, in which all of those factors and others played their contributory part.? The parallel in the obesity pandemic to “the sinking of the ship” is “the making of an obesigenic environment.”? Here, the components are -with particular emphasis on the ~eight decades since World War II- ever more constant exposure to an ever expanding inventory of ever larger portions of willfully hyperpalatable, addictive food stuffs; ever more constant exposure to an ever growing diversity of labor-displacing technologies; ever more modern contributions to stress, sleep deprivation, and distraction; and so on.? Study any one of these- or, for that matter, the “carbohydrate” hypothesis in competition with the “calorie” hypothesis - and you might well reach the conclusion that obesity is a daunting enigma.? Sure- just like the drowning of Titanic passengers.

The lesson, it seems to me, is that the reductionism of science is at times the wrong lens, notably when the answers reside in a recipe that is the sum of its parts, and more.? In such situations- as when snowflakes “gang up” to cause a lethal avalanche, or sandbags “collaborate” in a levee to stop a flood- the whole truth…is the only truth.

Three final reflections.? First, science can at times, invoking specific methods and measures, overlook the glaringly obvious.? We know that putting our hands in fires is apt to get us burned, but scholars charged with an evidence review of this topic would conclude they couldn’t conclude anything- because no one has volunteered to participate in randomized trials involving real and placebo fires.?

Second, science can at times, and much to our peril, cause people to overlook the essential value of sense.?? Mere sense tells us that energy-dense, often inexpensive foods engineered to be hyperpalatable are going to make a meaningful contribution to over-consumption.

And third, logic, like sense, may fail to get the consideration it warrants.? UPFs don’t represent all “badness” in the world of food.? Just as there are UPFs that may not be so bad, there are some pretty bad non-UPFs.? Topping that list for me is beef, in all its incarnations.? The global demand for beef, and its satisfaction, result in incalculable abuse of animals; devastation of the environment; and harm to human coronaries (and brains) into the bargain.? We have, for far too long, been fixated on unitary explanations for our dietary misadventures, seeking silver bullets and scapegoats.? There is more than one thing wrong with modern diets, and even UPFs are part of a larger story from which we should not be diverted.? The overall pattern and quality of diet are what matter most to the outcomes that matter most, namely- years in life, life in years, and into the bargain, the fate of our fellow creatures and planet.

UPFs mostly do just what most UPFs were designed to do: make safe the bet that nobody can eat just one.? They sabotage public health for the sake of corporate profits.? In an age of pandemic obesity and diabetes, this rises to the level of crimes against humanity- but not all of humanity.? These are crimes against the many, for the sake of the few.

On that basis, I reiterate: guilty, as charged.? Time for a snack.? Best to take no chances; I am thinking…apple.

?

-fin

*A decade before the provocative disclosures of Pulitzer Prize winner, Michael Moss, a 4-part expose in the Chicago Tribune had revealed the collaborative efforts of food-industry scientists and tobacco-industry scientists to formulate addictive products.? Long accessible on-line, that expose- since a change in ownership at the Tribune- leads to a dead link.

?

Dr. David L. Katz is a board-certified specialist in Preventive Medicine/Public Health, and past president of the American College of Lifestyle Medicine.??He is the founder of Diet ID, Inc, which has developed and validated a fundamentally new and infinitely scalable way to conduct comprehensive dietary intake assessment, including instantaneous generation of overall diet quality by means of the Healthy Eating Index 2020.? He serves as Chief Medical Officer for Tangelo, a leading food-as-medicine company focused preferentially on the needs of underserved and food insecure populations.

?

?

?