How we rate evidence
The six strength ratings Appetune uses on research briefs, what each one means, what changes a rating, and how nutrition figures are calculated.
Nutrition evidence is uneven. Some questions have decades of consistent work behind them; others have three small studies and a lot of confident writing. Presenting both in the same voice is misleading, so every Appetune research brief carries a rating.
The six ratings
Strong: consistent, replicated in people, and unlikely to reverse. Good studies point the same way and the effect has held up.
Moderate: a real effect, with something still unsettled: how large it is, who it applies to, or how long it lasts. Where the only endpoint is a biomarker rather than an outcome you would notice, a rating stops here and the brief names the biomarker.
Limited: thin. Small, short, single-centre, or only one trial however good it was. Worth knowing about; not worth reorganising your eating around.
Mixed: good studies disagree, and the disagreement has not been resolved. A brief with this rating has to say who disagrees and why, not average them into a shrug.
Insufficient: there is not enough evidence to say either way. Not the same as evidence of no effect, and we will not report it as though it were.
Not studied adequately: nobody has properly looked. A mechanism that only works in a dish or in animals sits here for the human question, however good the mechanism looks.
Those last three are separate ratings on purpose. “Experts disagree”, “the data is too thin to say”, and “nobody has run the study” are three different things to tell a reader, and collapsing them into one hedge is how health writing stops being useful.
This is Appetune’s own scale, shared with our sister publications. It is not GRADE, and we do not describe it as GRADE, which names a specific published methodology with requirements we do not claim to meet.
The rating is a promise about the wording
A rating constrains the verbs a brief is allowed to use. Confident prose under a weak rating, or hedging under a strong one, is a defect rather than a matter of style.
| Rating | How the brief may write |
|---|---|
| Strong | “reduces”, “improves”, with the size of the effect |
| Moderate | “appears to”, “is associated with” |
| Limited | “early data suggest”, “one small trial found” |
| Mixed | “trials disagree”, then the explanation |
| Insufficient | “there is not enough evidence to say” |
| Not studied adequately | “this has not been studied in people” |
What changes a rating
A rating is reviewed when its underlying evidence moves, not on a calendar: a new trial, a changed guideline, a label change, a correction, or a retraction. Anything that would change the verb in the table above changes the rating.
When a rating goes down, that is a normal outcome and not an embarrassment. The change is dated on the brief and, where the substance changed, recorded in corrections and on safety updates.
What every brief must state
- The question, in the form a reader would ask it.
- Who it was studied in. A finding in supervised trial participants is not automatically a finding in someone eating a third of their usual volume.
- What is uncertain. Not a disclaimer at the bottom. A labelled part of the summary, in berry, at the top.
If we cannot state who a finding applies to, we do not publish the finding.
Extrapolation is labelled
A great deal of what is written about nutrition during GLP-1 treatment is extrapolated from research that predates it. Extrapolation is sometimes the best available reasoning, but it is not the same as evidence, and Appetune labels it as extrapolation wherever it appears.
How nutrition figures are calculated
Food values come from structured public food-composition records rather than estimates, calculated against the specific ingredients and weights a recipe lists. A recipe stays in draft until its figures have been recalculated and checked; while it is in draft, its figures are labelled illustrative and should not be relied on.
Which sources count
Ratings are built from peer-reviewed human research, systematic reviews, and current guidance from public health bodies. Official prescribing information is the authority for anything about a medication, and it is linked rather than paraphrased alone. A press release, a conference abstract, or a manufacturer’s own summary is not a source we will rate a brief on.
Who funded a study is part of reading it. Where funding or a declared interest bears on a finding, the brief says so in the same breath as the rating rather than in a footnote. Appetune’s own position on money is on editorial standards.
Last updated
Uncertainty stated, not buried.
See how the ratings read on an actual brief.