What it is
How closely your eating pattern matches a scored diet-quality index, like a Mediterranean-diet adherence score. This is a pattern built from many foods scored together, not one nutrient or one food group on its own.
How it is modeled (not scored)
Technical notes
The curve is Sofi 2013: updated meta-analysis built explicitly to derive a usable adherence score, over 4.1 million subjects pooled. Per 2-point increase in adherence score, all-cause mortality RR 0.92 (0.91–0.93). Main source: Sofi 2013. Supporting: Eleftheriou 2018 (30 studies, 225,600 deaths, highest-vs-lowest RR 0.79 (0.77–0.81); above-median meat intake was the one score component that ran the wrong direction, RR 1.07) / Onvani 2017 (competing instrument family, not a competing direction — 12 studies using the Healthy Eating Index or Alternative Healthy Eating Index instead of a Mediterranean score, highest-vs-lowest RR 0.77 (0.76–0.78), same direction and rough size, different food-group weighting, no head-to-head comparison exists) / Dinu 2018 (umbrella review of 13 observational meta-analyses and 16 RCT meta-analyses; finds robust evidence for lower mortality and several disease outcomes, but explicitly rates evidence for most site-specific cancers and for LDL cholesterol as only suggestive or weak — this is not a uniformly strong predictor across everything it gets marketed against). Sofi 2008 corroborates the per-2-point slope; the 2026 Italian-guideline pooling (Nucci) reports a different unit, per 1 point; the PREDIMED trial (Estruch 2018) is the field's central RCT evidence but reports a composite cardiovascular endpoint, not an all-cause mortality figure.
How to improve it
- Across cohorts, the highest Mediterranean-diet scores went with 21% lower all-cause mortality than the lowest (Eleftheriou 2018)
- Each 2-point higher adherence score went with 8% lower all-cause mortality (Sofi 2013)
- The highest fiber intakes went with 15% lower all-cause mortality than the lowest; the greatest benefit across critical outcomes was at 25–29 g/day (Reynolds 2019)
- In US cohorts, a modeled swap of 3% of daily energy from processed-red-meat protein to plant protein went with 34% lower all-cause mortality (Song 2016)