How grading works
The four evidence tiers used across the compound library, and the rule that decides between them.
Last reviewed 24 Jul 2026
The four tiers
| Tier | What it means |
|---|---|
| Approved medicine | An approved, prescribable medicine somewhere in the world, for at least one indication. |
| Human trial data | Completed human trials exist, but the compound is not an approved medicine. |
| Limited human data | Some human use is documented, but it is thin, self-reported, not replicated, or otherwise hedged rather than settled. |
| Preclinical only | The evidence is animal or mechanistic. No credible human trial data exists yet. |
Why grading checks for denials first
An evidence summary that says “no human trials” or “not yet approved” contains the exact words — “human trials”, “approved” — that would otherwise mark it as strong evidence if read carelessly. Grading here checks for that kind of denial before it checks for a positive claim, so a sentence stating what a compound lacks cannot be misread as evidence that it has it.
What would change a grade
A grade moves only when the underlying evidence position changes: a compound gains regulatory approval, a completed human trial is published, or an existing claim turns out to be wrong. It does not move because a compound becomes more or less popular, and it does not move to make a listing more attractive.
Review cadence
The full library is checked against this methodology at least twice a year, and individually whenever new trial data or a correction changes a specific compound's position. The date at the top of this page and of the library index reflects the last full pass.