Saltar al contenido principal

Planetary Health Diet (PHD)

The PHD alignment indicator measures how closely a food portfolio matches the reference diet defined by the EAT-Lancet Commission — category by category, not as a single average.

What this indicator describes

PHD alignment describes the purchased or recorded portfolio, not the food actually eaten by an individual guest. Plate composition, overproduction and plate waste cannot be derived from procurement or sales data. Treat it as a structural measure of a site's purchasing and offering profile.

Reference framework

The Planetary Health Diet comes from the EAT-Lancet Commission report (Willett et al. 2019, Food in the Anthropocene). It describes a diet that could feed roughly 10 billion people healthily by 2050 without exceeding planetary boundaries.

The Commission defines 13 food groups, each with a daily reference intake and a healthy range in grams per person per day.

Eaternity evaluates against the ranges, not the midpoints. The ranges are the flexibility the Commission itself provides, and they permit different but equally valid dietary styles.

The 13 categories

Food groupReference (g/day)Healthy range (g/day)
Grains (whole)1930–193
Starchy vegetables500–100
Vegetables300200–600
Fruits200100–300
Dairy products2500–500
Red meat140–28
Poultry290–58
Eggs130–25
Fish280–100
Legumes750–100
Nuts & seeds500–75
Added fats500–80
Added sugars310–31

These values are stored as code constants and do not change between evaluations.

How the calculation works

Step 1 — Category mapping

Every product is mapped to one of the 13 categories. Composite (convenience) products are broken down into their ingredients first.

Products with no PHD equivalent — beverages, herbs, spices, non-food — are assigned to no category. This follows the EAT-Lancet model, which does not cover these groups. They still contribute to total energy (see Step 3).

Step 2 — Milk equivalents

The 250 g/day dairy reference refers to liquid milk, so dairy products are converted to milk equivalents — how much raw milk the product required:

MilkYogurtCreamCheeseButterMilk powder
×1.1×1.8×3.8×5.0×10.1×8.1

100 g of butter therefore counts as 1,010 g toward the dairy total. Without this conversion, a butter-intensive site would be scored systematically too well.

Step 3 — Normalisation to person-days

Purchase and sales volumes are absolute quantities; PHD reference values are per-capita daily figures. Energy content bridges the two:

person_days       = total portfolio energy (kcal) / 2,808.8 kcal
grams_per_day[c] = category_grams[c] / person_days

The reference of 2,808.8 kcal per person per day is the assumed actually consumed energy in Germany. It is derived from the FAO daily caloric supply for Germany (3,647.81 kcal/person/day) less 23% consumer-level food waste (Beretta et al. 2013):

3,647.81 kcal − 23% = 2,808.8 kcal / person / day

Step 4 — Status per category

Each normalised value is compared against its range and classified as within, above or below the healthy range.

Step 5 — PHD score

deviation[c] = 0                                                    if within range
= min(100, |actual − nearest_boundary| / nearest_boundary × 100) otherwise

PHD score = 100 − mean(deviation over all 13 categories)

100% means every category sits inside its healthy range.

Each category's deviation is capped at 100% before averaging. Without the cap a single outlier can dominate the result — in one real test case a dairy deviation of 417% pulled the score from 75% down to 38% — and the score could go negative. With the cap the score is bounded to 0–100% and no single category can move it by more than 100/13 points.

Reading the chart

Each category shows the actual intake, a band for the healthy range, and a marker for the EAT-Lancet reference value.

Energy coupling — the key interpretation point

Because every category is normalised against the same total energy, the categories are coupled: energy missing from one category necessarily reappears as excess in another. A low vegetable value and a high meat value are typically one finding, not two.

Also note that 11 of the 13 ranges start at 0 g/day. A category can therefore be formally "within range" while not being purchased at all — so always read the distance to the reference marker alongside the status.

Data requirements

  • Portfolio completeness. The indicator normalises over total energy. If entire assortment segments are missing — typically beverages and snacks, which are not part of menu planning — the denominator shifts, and with it every category value, not just the missing ones. Evaluation based on complete goods deliveries is therefore more robust than one based on planned or sold menu lines.
  • Non-food exclusion. Non-food articles belong to no PHD category and must not enter the mass or energy totals. They are excluded by product classification.
  • Level of detail. The more precisely an article is described (name, quantity, unit, ideally GTIN), the more accurate the mapping.

Limitations

  1. The energy reference is a national average. 2,808.8 kcal/person/day is a German average. For sites with a markedly different guest profile (heavy manual labour, clinics, day-care), the person-day conversion is correspondingly imprecise. The relative composition of the portfolio — the actionable signal — is largely unaffected.
  2. Lower bounds of 0 g/day. 11 of 13 ranges have no lower bound, so "within range" is not always equivalent to "nutritionally desirable".
  3. Category mapping is a model. EAT-Lancet defines "starchy vegetables" as true tubers only (potato, cassava, yam). Other root vegetables such as carrots or beetroot count as ordinary vegetables.
  4. PHD is a health indicator, not a climate indicator. It does not replace the CO₂ assessment. The two often correlate, but not necessarily.

Sources

  • Willett W., Rockström J., Loken B., et al. (2019). Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems. The Lancet 393(10170), 447–492. doi:10.1016/S0140-6736(18)31788-4
  • FAO — Daily caloric supply (Germany): 3,647.81 kcal/person/day.
  • Beretta C., Stoessel F., Baier U., Hellweg S. (2013). Quantifying food losses and the potential for reduction in Switzerland. Waste Management 33(3), 764–773.
  • Deutsche Gesellschaft für Ernährung (DGE) — Referenzwerte Energiezufuhr.