Skip to content
Radish — back to home
← All articles Education

ANSES, EFSA, WHO: How the Nutrition Numbers You See Everywhere Are Set

In short

A \u201cnutrition reference value\u201d (e.g. 25g of fibre/day) isn't an average of what people actually eat: it's a statistical calculation that starts from a population's real requirement and adds a safety margin to cover 97.5% of individuals. The method differs slightly between ANSES, EFSA and the WHO, which is why the numbers vary from country to country.

Radish · 9 min

\u201c25g of fibre a day\u201d, \u201c2000 kcal for a woman\u201d, \u201cno more than 5g of salt\u201d — these benchmarks show up everywhere: on labels, in tracking apps, in the press. Few people know where they actually come from, or why a figure published by France's ANSES sometimes differs slightly from what the WHO recommends globally.

This guide explains the mechanics behind these numbers: they're neither a consumption average nor an arbitrary figure, but a statistical calculation with a precise logic.

Download the Radish app

"Nutrition reference value": a term that changed name

Long called ANC (recommended nutritional intakes) in France, these figures are now officially referred to as RNP (population nutritional reference), a terminology shift that better reflects what they represent: a statistical reference built for a group, not an individual prescription tailored to each person.

How a requirement becomes a recommendation

The construction happens in several steps, documented by ANSES in its work updating French national nutrition programme benchmarks:

  1. The Average Requirement (AR): the amount of a nutrient that covers the physiological needs of half of a given sub-population (by age, sex, physiological status such as pregnancy).
  2. The standard deviation: since requirements vary from person to person, statistical spread around the AR is measured, typically estimated at 15% of the AR when precise individual data is missing.
  3. The safety margin: two standard deviations are added to the AR. Since the standard deviation is usually estimated at 15% of the AR, the population reference value most often works out to roughly 1.3 times the AR.

The final figure thus covers the needs of about 97.5% of individuals in the studied sub-population — not 100%, and not an average of actual consumption.

Ready to try Radish?Private calorie counter, Nutri-Score, no ads. Free.Get Radish →

Who does what: ANSES, EFSA, WHO

In practice, EFSA largely supplies the European scientific base, ANSES refines it for France, and the WHO reasons at a global scale with its own expert review processes.

Why the numbers vary from country to country

Several reasons explain different benchmarks for the same nutrient:

This isn't a sign of scientific inconsistency: it's the normal result of independent agencies working from different samples and priorities, using a shared statistical method.

What this means for you day to day

A nutrition reference value is neither a strict survival minimum nor a ceiling never to exceed: it's a population target, built to cover almost all individual needs with a built-in safety margin. Putting it in the context of a real diet remains useful — see our guides on protein, fibre, salt or sugar — but a one-off day below or above doesn't carry the same meaning as a chronic deficiency or excess documented by a health professional.

A reference value is not a personal target

The statistical logic behind these numbers also explains why they shouldn't be read as a personal ceiling or floor to hit exactly. Because the population reference already includes a built-in safety margin (roughly 1.3 times the Average Requirement), most individuals in a sub-population need meaningfully less than the published figure to stay perfectly healthy — the margin exists precisely to cover the people at the higher end of natural variation, not to describe what a "typical" person requires. Falling slightly under a reference value on any given day is therefore rarely a sign of deficiency; a persistent, large gap sustained over months is a different matter, and one worth discussing with a health professional rather than diagnosing from an app.

Ready to try Radish?Private calorie counter, Nutri-Score, no ads. Free.Get Radish →

FAQ

Is a nutrition reference value an average of what people eat?

No. It's a statistical calculation built from a population's average physiological requirement, plus a safety margin to cover about 97.5% of individuals, not an average of what people actually eat.

Why do ANSES and the WHO sometimes publish different numbers?

Because they're independent agencies working from different population samples and public health priorities, with revision dates that don't always line up. That doesn't mean one of the two figures is wrong.

Does ANSES do its own scientific work or just reuse EFSA's?

Both: ANSES relies largely on EFSA's scientific opinions as a base, then adapts them to the French context (dietary habits, the population's nutritional status) to produce the official references used in France.

A nutrition reference value is therefore never a \u201cmagic\u201d number: it's a statistical construct, transparent in its method, but still a population average — not an individual prescription. In Radish, the benchmarks shown (calories, macros, micronutrients) rely on these official reference frameworks, with their source indicated. Try the app, free, your diary stays on your phone — never sent to our servers (except AI or barcode scanning, if you use them).

General information, not medical advice.

🔎 Methodology & sources. Every nutrition figure in this article is cross-checked against official sources (ANSES, EFSA, WHO, CIQUAL/USDA, French Public Health agency), and health topics are reviewed before publication against our internal reference framework. See our full methodology →

Ready to try Radish?Private calorie counter, Nutri-Score, no ads. Free.Get Radish →

Prefer the installed app? Get Radish on iPhone or Android →