If Two ‘Healthy’ Diets Differ, What Are We Missing?
Last updated: March 2026
Nutrition guidance is built on a stable assumption: if nutrient composition is aligned with recommendations, outcomes should follow predictably. That assumption is becoming less reliable.
What the evidence shows
Recent controlled feeding studies challenge this model. A 2025 randomized crossover trial published in Nature Medicine compared diets differing in processing level despite similar nutrient profiles.The metabolic outcomes differed.
Earlier work by Kevin Hall and colleagues, published in Cell Metabolism (2019), showed similar results: ultra-processed and minimally processed diets led to different energy intake and physiological responses, despite comparable macronutrient composition.
What this challenges
When diets aligned with the same nutritional targets produce different physiological effects, nutrient composition alone cannot explain the outcome.
This does not invalidate current recommendations, but it limits what they capture.
Where the difference may come from
Research on gut microbiota and metabolic variability suggests that food responses are partly mediated by biological processes not reflected in nutrient profiles.
A study published in Cell in 2015, showed markedly different glycaemic responses to identical meals.
Processing level, food structure and microbial activity interact with diet in ways that are not captured by composition-based metrics.
If part of the diet–disease relationship is mediated through these pathways, processing becomes more than a classification variable — it becomes biologically relevant.
The structural mismatch
Food systems are designed around: standardisation, comparability, and scale.
Dietary guidance follows the same logic. Biological responses do not.
They vary between individuals, across contexts, and over time. When uniform dietary frameworks interact with non-uniform biology, variation is not noise — it is expected.
What this means
We are not at a point where: personalised nutrition can guide decisions at scale, or where microbiome-based recommendations are operational.
But it is increasingly difficult to argue that nutrient composition alone is sufficient to explain how food affects health.
Why this matters
Nutrient-based frameworks underpin dietary guidelines, product formulation, health claims, and regulatory approaches. If those frameworks capture only part of the system, their outputs are necessarily partial.
The real shift
This is not a replacement of nutrition science. It is a shift from composition-based to interaction-based understanding.
That shift introduces complexity that current systems are not designed to absorb.
Bottom line
Food is not only defined by what it contains. It is defined by how it behaves in biological systems.
Current models capture composition. They do not fully capture response.
Until that gap is addressed, differences in outcomes will remain — even when inputs appear equivalent.