What the Research Says
What the Nutrition Research Actually Shows
Nutrition research is plagued by confounds, but some findings are robust. Here's what holds up under rigorous scrutiny — and what's mostly noise.
Why Nutrition Research Is Hard
Dietary studies face a methodological mountain: you can't blind people to what they eat, recall data is unreliable, diet and lifestyle co-vary, and most important outcomes (cardiovascular disease, cognitive decline) take decades. The majority of nutrition headlines come from observational studies with uncontrolled confounders. When the RCT evidence is examined separately, it's considerably thinner.
What Replicates Strongly
Caloric balance determines body weight with high consistency. Despite complexity in metabolic adaptation, total energy intake versus expenditure predicts body weight change across populations and intervention types. Metabolic ward studies show this with high precision. Where diets differ dramatically — low-carb vs low-fat — the long-term weight differences largely disappear when calories and protein are matched.
Protein is the most satiating macronutrient and has the highest thermic effect. High-protein diets (1.6–2.2g/kg) consistently produce greater fat loss than isocaloric lower-protein diets in RCTs, primarily through satiety and diet-induced thermogenesis. This effect holds across dietary patterns — it's macronutrient-level, not food-source-level.
Dietary fibre is causally linked to improved gut microbiome diversity and lower CVD risk. The PREDIMED trial and others demonstrate fibre's protective role via multiple mechanisms (microbiome, glucose regulation, satiety). Most people consume less than half the recommended 30g/day. This is one of the highest-confidence actionable findings in nutrition epidemiology.
Ultra-processed food consumption predicts poorer health outcomes independent of macronutrient content. Kevin Hall's 2019 randomised crossover trial — the first to directly compare ultra-processed vs unprocessed diets with matched macros — found spontaneous caloric overconsumption of ~500 kcal/day on ultra-processed diets. Mechanism: food reward engineering, faster eating rate, reduced satiety signalling.
Omega-3 fatty acids (EPA+DHA) reduce systemic inflammation. Marine omega-3 RCTs consistently show reductions in inflammatory markers (CRP, IL-6). Cardiovascular benefit is clearer in high-risk populations; cognitive benefit requires longer study durations. Dose threshold for anti-inflammatory effects: approximately 1g EPA+DHA/day.
What the Research Can't Tell You
Individual glycaemic response to identical foods varies ~2–3× between people (Weizmann Institute CGM studies). The "best diet" in population averages may be poor for you individually. Tracking your own glucose, energy, and satiety response to specific meals is more informative than applying dietary patterns designed for groups. The meal quality signals worth monitoring personally are energy stability, satiety duration, and sleep — not just weight.