Food patterns and the food matrix

Updated 11 Oct 2026

The nutrient table is one view of the meal

I want meals I enjoy, can finish and can buy with the storage available to me. The earlier rice, chicken, beans, olives and fish combination failed that test: I disliked parts of it, became too full and discarded food. A later day was more enjoyable. Those observations matter even before asking whether a table reaches a particular nutrient reference.

The food table answers questions about estimated amounts. A dietary pattern asks what foods replace other foods, how a meal is prepared, how often I eat it and whether I can sustain it. A food’s physical structure also affects chewing and digestion. Neither view replaces the other. Reaching a vitamin reference does not prove that a diet improves training or long-term health; a named diet does not tell me whether today’s protein or energy intake fits my needs.

What the corrected PREDIMED trial supports

The 2018 corrected PREDIMED report studied 7,447 adults in Spain at high cardiovascular risk: men aged 55–80 and women 60–80. Median follow-up was 4.8 years. Major cardiovascular events occurred in 3.8% with Mediterranean-diet/olive-oil advice, 3.4% with Mediterranean-diet/nuts advice and 4.4% with reduced-fat advice; adjusted hazard ratios were 0.69 and 0.72. Estruch et al., 2018.

The original report was withdrawn after randomization departures, including household and clinic assignments. Corrected analyses adjusted for these problems; results were similar after excluding 1,588 participants with known or suspected departures.

Participants received advice and supplied foods; control counseling was initially less intensive. Public funding and food donations were disclosed; sponsors reported no role in design, analysis or reporting. This supports a dietary strategy in that population without isolating a nut or herb, predicting my training outcomes or prescribing the trial’s oil quantities.

Processing can affect intake beyond the offered nutrient totals

Hall’s inpatient crossover trial gave 20 adults an ultra-processed diet and an unprocessed diet for two weeks each, in random order. Participants could eat freely. The offered diets were designed to match several nutrient totals, but actual intake was about 508 kcal/day higher on the ultra-processed diet. Participants gained about 0.9 kg during that phase and lost about 0.9 kg during the other. Hall et al., 2019.

This is experimental evidence that those two menus changed eating behavior. It does not identify a universal toxic ingredient or prove that equal calories consumed would have different effects on weight. The diets also differed in eating rate and the energy density of their non-beverage foods. Matching a menu’s offered fiber and macronutrients did not match all its physical properties or the amounts ultimately eaten.

For my current goal, the implication is broader than avoiding excess intake. A very bulky menu can also make it difficult to eat enough comfortably. Enjoyment, portion size, texture and meal timing belong in the log alongside calories.

A healthier ultra-processed menu is a useful counterexample

A 2025 community crossover trial provided two eight-week diets that both followed the UK Eatwell Guide. It randomized 55 adults with overweight or obesity; 50 contributed primary-outcome data for at least one diet. Average weight changed by −2.06% on the minimally processed diet and −1.05% on the ultra-processed diet, a 1.01 percentage-point difference. Thus both menus produced weight loss, with more on the minimally processed one. Dicken et al., 2025.

The result concerns supplied menus in a predominantly female, higher-BMI population. It does not establish weight loss as my goal or show that every ultra-processed food is equivalent. There was attrition, and the diets differed in more than a category label. Published 2026 correspondence challenges how confidently the weight difference can be attributed to processing itself; the authors published a reply. Robinson and Forde, 2026, Dicken et al., reply.

Texture can change intake within the same category

In another crossover experiment, 41 adults ate two diets supplying roughly 95% of energy from ultra-processed foods for two weeks each. The menus were matched for several properties, including non-beverage energy density, while their textures encouraged slower or faster eating. Intake was 369 kcal/day lower on the slower-eating menu. Body weight did not significantly differ between diets. Forde et al., 2026.

That is evidence for a specific mechanism, not a finding that processing never matters. It shows why the category alone is incomplete: foods can vary in texture, energy density, nutrient content and convenience. The study excluded people doing more than four hours of moderate-to-vigorous activity weekly, so its precise intake difference is not an athlete prediction.

Structure also changes how a nutrient total is used

An 18-person crossover study measured 4.42 kcal/g for whole natural almonds and 6.53 kcal/g for pure almond butter. These are experimental diet averages, not personal calorie corrections. Gebauer et al., 2016. The nut-energy companion examines processing comparisons and balance measurements.

What changes in my shopping routine

The useful default is a repeatable set of foods I enjoy, with enough variety to cover ordinary nutrient needs and enough flexibility to adjust energy and meal size. Packaging, freezing, canning or cooking can make that feasible; the presence of a package is not itself a reason to reject rice, yogurt or fish.

I would keep the daily record focused on what I actually ate, the amounts I can verify, enjoyment, appetite, digestion and discarded food. A proposed change should specify its purpose: a more enjoyable meal, easier preparation, a nutrient source or a training-related outcome. That is more informative than maximizing a processing score or collecting foods because each has one promising paper.

Back to investigations