You’re caught up.

Home/Issues/Issue 09/Study

Cross sectional construct validity analysis of pooled national dietary surveys Online publication date confirmed against the PubMed publication date field.

A simple count of food groups predicted nutrient adequacy better than the Planetary Health Diet Index

AI narration, generated on first listen
Journal
The American Journal of Clinical Nutrition, article 101495
Authors
Hanley-Cook GT, Van Der Meulen E, Beal T
Institution
Ghent University
Published
26 August 2026
Source
PMID 42648644 · DOI 10.1016/j.ajcnut.2026.101495
Design
Cross sectional construct validity analysis. Multilevel linear and modified Poisson regression models quantified associations between each diet score and reference metrics of nutrient adequacy and of moderation of foods linked to diet related non-communicable disease.
Sample
152,004 non-pregnant females and males aged 15 and over, from 45 surveys in 26 countries, accessed through the FAO and WHO Global Individual Food Consumption Data Tool.

What construct validity is

Construct validity asks whether a score actually measures the thing it claims to. Here the claim being tested is that the Planetary Health Diet Index captures a healthy diet, and the test is whether it tracks independent yardsticks of a good diet more closely than rival scores do.

Drawn from standard methodology, not from this paper.

Why they ran it

The Planetary Health Diet Index was developed to measure adherence to the diet proposed by the EAT-Lancet Commissions. The authors note that its construct validity for accurately reflecting a healthy diet, as compared against other global dietary measures, remains to be evaluated.

Drawn from the stated aims of the paper.

Quantitative 24 hour recall and food record data from 152,004 non-pregnant people aged 15 and over, drawn from 45 surveys in 26 countries, were scored four ways: the Planetary Health Diet Index, a food group diversity score, and the positive and negative sub-metrics of the Global Diet Quality Score. Each was then tested against independent reference measures of a healthy diet, chiefly the mean adequacy ratio for nutrients and the proportion of energy coming from ultra-processed food.

The Planetary Health Diet Index performed as designed. A one standard deviation increase was associated with a higher mean adequacy ratio, by 5.49 percentage points (95 percent CI 3.65 to 7.33) in low income countries falling to 1.41 (1.23 to 1.58) in upper middle income countries, with less unprocessed red meat and a lower share of energy from ultra-processed food.

The simpler instruments beat it on nutrient adequacy. The same one standard deviation increase in the food group diversity score was associated with a mean adequacy ratio 11.2 percentage points higher (95 percent CI 9.13 to 13.3) in low income countries and 8.57 (8.42 to 8.72) in upper middle income countries, roughly double the index it was compared against. The pattern repeated for fruit and vegetable intake.

Adherence to the Planetary Health Diet Index was highest in low and lower middle income countries and lowest in upper middle and high income countries, which is the reverse of how diet quality usually sorts by national income.

The numbers

Participants152,004 across 45 surveys in 26 countries
Mean adequacy ratio per 1 SD of the Planetary Health Diet Index5.49 pp (95% CI 3.65 to 7.33) low income, to 1.41 (1.23 to 1.58) upper middle income
Mean adequacy ratio per 1 SD of the food group diversity score11.2 pp (95% CI 9.13 to 13.3) low income, to 8.57 (8.42 to 8.72) upper middle income
Appropriate fruit and vegetable intake, RR per 1 SD of food group diversity score2.78 (2.26 to 3.41) lower middle income, to 1.63 (1.22 to 2.19) high income
Unprocessed red meat per 1 SD of the Planetary Health Diet Indexminus 10.0 g/day (minus 11.4 to minus 8.73) in low income countries
Energy from ultra-processed food per 1 SD of the Planetary Health Diet Indexminus 2.47 pp (minus 3.77 to minus 1.18) in lower middle income countries

Why this might happen

Not available The full discussion is not openly available and the abstract offers no mechanism, so we do not know what the authors think caused this.

Drawn from no open full text; the paper is an index validation rather than a mechanistic study.

What this does not show

  • This does not show that the Planetary Health Diet is unhealthy or that its environmental case is weak. What was tested is how well one scoring instrument tracks nutrient adequacy against rival instruments, not whether the underlying diet is good for people or for the planet.
  • This does not show that anyone's health improved or worsened. Every measurement is a single day or two of recalled intake compared against other dietary measures, with no health outcome, no follow up and no intervention.
  • This does not tell you about individuals. These are population level associations from pooled national surveys, and a score that performs well across a country says nothing about what any one person should eat.
  • This does not settle which index to use everywhere. The authors find the ranking varies by country income level, and they frame the newer index as potentially suitable for a specific context rather than as generally inferior.

Where this leaves us

A sustainability oriented diet index does track nutrient adequacy in the right direction, but a much simpler count of how many food groups someone eats tracks it roughly twice as strongly, across dozens of countries.

People aged 15 and over in 45 surveys across 26 countries, excluding pregnant women.

For anyone choosing a diet quality metric for population monitoring, it argues that the added complexity of the newer index buys less on the health side than the simpler alternatives, particularly where undernutrition is still a concern.

Caveats worth holding

  • Single day 24 hour recall or food record data, which captures usual intake poorly at the individual level.
  • Surveys were collected across different years and by different national methods before pooling.
  • No open full text was reachable, so the analysis here rests on the abstract and the stated aims.

Newsletter

Each issue by email, when the newsletter launches. Leaving your address puts you on the list, nothing is sent yet.