“If a man does not know to which port he sails, no wind is favorable.”
— Seneca
If you listen long enough to the current conversation about food and health, you can get dietary whiplash.
One week, the focus is on ultraprocessed foods.
Next week, the problem is carbohydrates.
The week after the solution is: “Eat more protein.”
Meanwhile, scientists debate whether ultraprocessed foods should even be considered harmful as a category. Industry groups argue that the classification system used to identify ultraprocessed foods is flawed (see my previous column from last week). Policy advocates push back, pointing to mounting epidemiologic evidence linking ultraprocessed diets to chronic disease. Researchers publish meta-analyses showing consistent associations between ultraprocessed food intake and cardiometabolic risk. Government recommendations attempt to simplify the entire discussion into a single prescription: “Just flip the pyramid.”
What can we conclude from this, if anything?
The uncomfortable answer may be that nutrition science is experiencing a period of conceptual confusion—a kind of growing pain. Not because the science is failing, but because the framework through which we interpret that science is increasingly being exposed as incomplete.
The Study
Rather than focusing on a single paper, this week’s spotlight reflects a broader body of literature examining the role of ultraprocessed foods (UPFs) in modern diets.
Over the past decade, dozens of cohort studies across multiple populations have identified strong associations between higher consumption of ultraprocessed foods and increased risk of obesity, type 2 diabetes, cardiovascular disease, depression, and premature mortality. Much of this work relies on the NOVA classification system, which categorizes foods by the degree of industrial processing rather than by nutrient composition alone.
In other words, there is something about UPFs that appears to influence human health independent of their nutrient panels. These findings have helped elevate ultraprocessed foods to the center of current nutrition debates.
Yet at the same time, criticism of the NOVA classification system has intensified. Critics continue to argue that the definition of ultraprocessing is too broad, that some ultraprocessed foods have acceptable nutrient profiles, and that focusing on the risk attributable to processing ignores the biochemical principles of traditional nutrition science. The result has been a peculiar sort of public discourse.
One group argues that ultraprocessed foods are a central driver of modern chronic disease.
Another argues that ultraprocessing itself is not the real problem.
A third suggests that traditional nutrient recommendations, such as increasing protein intake, are all that is needed to address our metabolic health challenges.
Like the blind men describing an elephant, each perspective captures part of the truth.
Taken together, however, they generate more confusion than clarity.
The Caveat
What is striking about these debates is that they all share the same underlying assumption. They assume the central task of nutrition science is to identify an absolutely correct rule. The result is that we receive directives from experts that contradict one another.
Reduce carbohydrates.
Increase protein.
Avoid ultraprocessed foods.
Optimize macronutrient ratios.
The specific rules offered up differ, but they are children of the same dogmatic logic.
Food is treated primarily as a collection of measurable components, e.g., nutrients, calories, or biochemical attributes, whose manipulation produces predictable health outcomes. This perspective has been enormously successful in certain contexts. For example, nutrient-based approaches helped identify vitamin deficiencies, eliminate industrial trans fats, and address iodine deficiency disorders.
But the dominant health challenges of the twenty-first century are no longer diseases of deficiency. They are diseases of excess, dysregulation, and environmental mismatch. Under those conditions, focusing exclusively on nutrient composition misses something important.
Ultraprocessed foods have brought that “something” to the forefront.
They are not simply “foods with poor nutrients.” They are engineered systems designed for shelf stability, hyper-palatability, convenience, and profit. Their formulations often include combinations of refined ingredients, industrial additives, flavor enhancers, emulsifiers, and textural modifiers. These characteristics influence not only nutrient intake but also eating behavior, satiety signaling, microbiome interactions, food environments, and even the structure and context of our meals.
Attempts to rehabilitate certain ultraprocessed foods by emphasizing favorable nutrient profiles, e.g., high-protein snack bars, fiber-fortified cereals, and vitamin-enriched beverages, risk returning nutrition discourse to the same reductionist thinking that helped produce the problem in the first place. At the same time, simplistic counter-solutions such as “eat more protein” may only add confusion without addressing broader dietary knowledge gaps.
This is why the current debate often feels circular.
One camp argues that nutrients are what matter most.
Another insists that processing is the key variable.
Still others emphasize simple “eat this, not that” answers as a panacea.
Meanwhile, we are left to navigate an ever-expanding dietary fog.
The critical question we need to ask is whether diet should be understood primarily in terms of isolated components at all.
A meal is not simply chemistry. It is also structure, context, timing, culture, environment, and behavior. We experience our food as complex biological systems that involve hormonal regulation, microbial ecosystems, neural reward pathways, and social environments every time we eat.
If we understand that, then we can seize the current debate as a transitional moment.
The evidence linking ultraprocessed diets to chronic disease continues to accumulate. Yet the conceptual framework used to interpret those findings remains rooted in older reductionist models of nutrition. The result is a field attempting to solve twenty-first-century health problems with twentieth-century analytical tools.
This tension helps explain why the public conversation about diet often seems contradictory. Researchers are observing new patterns in population health while interpreting those patterns through frameworks designed for a different era. As a result, we find ourselves simultaneously arguing about ultraprocessed foods, macronutrients, glycemic responses, and protein intake, without a clear, unifying model that explains how these factors interact.
The science itself is not failing. The framework upon which we try to place it may be.
Cognitive scientist Thomas L. Griffiths has observed that scientific confusion often arises not because the evidence is wrong, but because the model used to interpret that evidence is inadequate. When the world becomes more complex than the frameworks we use to explain it, debates multiply, contradictions appear, and researchers begin proposing competing rules that capture fragments of the truth without explaining the whole.
Perhaps the answer lies not in arguing over which data along the road we are on is wrong, but in asking whether we are on the wrong road altogether.
References
Hall KD, Ayuketah A, Brychta R, et al. Ultra-processed diets cause excess calorie intake and weight gain: an inpatient randomized controlled trial. Cell Metab. 2019;30:67–77.
Lane MM, Gamage E, Du S, et al. Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses. BMJ. 2024;384:e077310.
Monteiro CA, Cannon G, Levy RB, et al. Ultra-processed foods: what they are and how to identify them. Public Health Nutr. 2019;22:936–941.
Srour B, Fezeu LK, Kesse-Guyot E, et al. Ultra-processed food intake and risk of cardiovascular disease. BMJ. 2019;365:l1451.
Fardet A. Minimally processed foods are more satiating and less hyperglycemic than ultra-processed foods. Adv Nutr. 2016;7:134–138.
Juul F, Parekh N, Martinez-Steele E, et al. Ultra-processed food consumption and risk of obesity among US adults. Am J Clin Nutr. 2022;115:211–221.
Popkin BM, Barquera S, Corvalan C, et al. Toward unified global action on ultra-processed foods. Lancet. 2025;406.
Be sure to check out Chef Dr. Mike’s new book, Dinner with God: Understanding The Language of Food. Available at Amazon.com in e-book (available for preorder) and print on the release date, 16 March 2026.
