What We Eat Is Killing Us. The Signal Was There All Along.
Metabolic disease — obesity, type 2 diabetes, cardiovascular disease, fatty liver — now accounts for the majority of US healthcare spending. The dietary signals that predicted this have been visible for decades.
Signal
Approximately 74 percent of American adults are now overweight or obese by BMI criteria — a proportion that has increased from approximately 56 percent in 1990. Type 2 diabetes affects approximately 38 million Americans, roughly 11.6 percent of the population, up from approximately 3 percent in 1960. Non-alcoholic fatty liver disease, which requires ultrasound to diagnose and produces no symptoms until late stages, is estimated to affect 25 to 30 percent of the US adult population — a condition that was essentially unknown before the 1980s.
These numbers do not represent a mystery. They represent a predictable biological response to a specific dietary environment that changed, with remarkable speed, in the second half of the twentieth century.
The dietary environment that changed: the American diet, from roughly the 1960s onward, shifted toward a pattern of foods that are hyper-palatable — engineered to override normal satiety signaling through combinations of fat, sugar, salt, and texture — low in fiber and micronutrients, highly processed in ways that alter their metabolic effects beyond what their macronutrient composition would predict, and consumed at frequencies and portion sizes that the metabolic system was not designed to manage.
Interpretation
The framing of this as an individual behavioral problem — people eating too much because of poor choices — is both empirically inadequate and politically convenient for the industries that produce and market the foods responsible.
The metabolic disease epidemic is a public health crisis with specific industrial causes: the seed oil displacement of traditional fats in processed food manufacturing from the 1960s onward, the addition of high-fructose corn syrup as a cheap sweetener after commodity policies made corn syrup economically favorable in the 1970s, the development of flavor enhancement technologies that produce foods that are difficult to stop eating through normal appetite regulation mechanisms, and the advertising and distribution infrastructure that made ultra-processed foods the default, accessible, affordable option for most American households.
These were not naturally occurring dietary changes. They were the outcome of deliberate product development, marketing, lobbying (of dietary guidelines, school lunch programs, food labeling requirements), and pricing decisions by a food industry that understood the addictive properties of the products it was designing.
The analogy to the tobacco industry is not perfect but it is instructive. The tobacco industry knew that its products caused cancer and heart disease for decades before this was publicly acknowledged; it funded research designed to create doubt, lobbied against regulation, and marketed directly to children. The processed food industry has followed a substantially similar playbook, with the difference that the mechanism of harm — metabolic dysfunction — takes longer to develop and is harder to attribute to specific products than lung cancer is to cigarette smoking.
Scenario
Scenario A — the pharmacological solution — is already underway. GLP-1 receptor agonists (we cover this in our science section) produce weight loss and metabolic improvement that directly addresses the obesity component of metabolic disease. If these drugs become cheap and widely accessible, they may reduce the metabolic disease burden substantially within a decade, regardless of whether the underlying dietary environment changes.
Scenario B — the public health intervention — involves treating ultra-processed food consumption as a public health emergency and deploying the toolkit used for smoking reduction: taxation, marketing restrictions (particularly to children), mandatory nutrition labeling that includes processing level, and school food policy reforms. This scenario is politically contested; the food industry's lobbying power exceeds the tobacco industry's at its peak, and the food industry has more successfully diversified its political alliances.
Scenario C — the default — is the continuation of the current trajectory, in which metabolic disease continues to grow, healthcare costs continue to increase, and the pharmacological solution manages symptoms without addressing causes.
Probability
Metaculus forecasts a 66 percent probability that US prevalence of type 2 diabetes will exceed 15 percent of the adult population before 2030, continuing the existing trend. Kalshi was trading a contract on whether the FDA will implement a mandatory ultra-processed food labeling scheme — comparable to the EU's Nutri-Score system — before 2028 at 17 percent.
The labeling contract probability reflects the political reality: the FDA's history with nutrition labeling reform is one of industry-captured delay, and the current political environment is less favorable to new food industry regulation than any since the 1980s.
Indicators to Watch
— USDA dietary guidelines revision: the 2025 dietary guidelines will either acknowledge or continue to downplay the ultra-processed food evidence; the political process by which they are developed is itself a signal about regulatory capture — GLP-1 prescription rates: whether the drugs are reaching the high-need populations (lower-income, less insured) or primarily the affluent will determine their public health impact — School lunch program funding and standards: federal school meal standards have been weakened and strengthened repeatedly; their level is a proxy for the political seriousness of dietary public health — Soda tax adoption: Philadelphia, Seattle, and other cities have adopted soda taxes; state-level adoption would signal that the political will for dietary public health intervention is growing
Dr. Amara Singh is a staff writer at The Auguro covering medicine, science, and public health.