How We Grade
Three or more double-blind, placebo-controlled trials from at least two independent groups, agreeing on a benefit that matters, at a dose that can actually be taken.
A positive record that is thin on a leg or two — fewer trials, a single group, or an effect too small to be sure it matters.
Limited, small, or mixed evidence, leaning positive.
A weak C — the low end of the grade.
Tested in adequate placebo-controlled trials, and the record as a whole does not show it works — predominantly null results, failed primary outcomes, or a clean null in a trial big enough to settle it. A negative verdict, not a weak record.
Insufficient: Tested, but too thin or too mixed to earn a grade — the record cannot settle the claim either way. Claims that were never tested are not printed at all.
- A trial run by a product's own sponsor caps its grade at C.
- Results published only as an abstract cannot earn a grade.
- Grades measure benefit only — harms are printed as cautions, never netted against the grade.