How we rate evidence
Every rating on this site follows the same four steps. They are written down here so you can check our work, and so an editor cannot quietly move an ingredient up a level.
The four levels
Two or more well-run randomised controlled trials in humans, or a systematic review, pointing the same way with a meaningful effect size.
Example: Psyllium husk for constipation
At least one good randomised trial, or several smaller ones, mostly positive but with limits in size, duration or design.
Example: Melatonin for jet lag
Small, short or uncontrolled human studies, or mixed results across trials. Promising but unproven.
Example: Valerian for sleep quality
A long record of use in one or more traditions, with little or no modern clinical research.
Example: Slippery elm for a sore throat
How a rating is decided
- 1
Start from the best evidence available
We look first for Cochrane and other systematic reviews, then regulators’ assessments (the European Medicines Agency’s herbal monographs, NCCIH, the NHS), then individual randomised trials in humans, searched through PubMed. Every source is one an editor opened and read; nothing is cited from a search snippet.
- 2
Weigh what the studies can show
How many people, how long, whether the trial was blinded and placebo-controlled, who paid for it, and whether the form and dose are stated. Manufacturer-funded or very small trials are named as such on the page.
- 3
Rate each use and form separately
An ingredient gets a separate rating for every use it is claimed for, and the rating applies to the form that was tested: an extract’s trial does not carry over to a tea. A null or negative systematic review caps a use at Limited, however many small positive trials exist.
- 4
Check the citations, then clinical review
Every citation is checked automatically against PubMed or the source’s own page before publishing. A pharmacist or clinician then signs the page off, and their name and the date appear on it. Until then the page says it is awaiting clinical review, and editing a signed-off page voids the sign-off until it is checked again.
What a rating does not mean
Not a safety rating
Strong evidence of effect says nothing about interactions. Safety is assessed separately and shown separately.Not a recommendation
A Strong rating means the research is good, not that it is right for you. Traditional means under-studied, not useless.Not fixed
Ratings move when the evidence moves. Signed-off pages show their review date, and every rating change is logged.Who does the reviewing
Each page is researched and written by the Healthy and Herbal editorial team, then read by a pharmacist or clinician who signs it off. The reviewer is named on the page with the date; who reviews our pages lists them. The site sells nothing and carries no affiliate links.
If you think we have got something wrong, use the form on the corrections log. Any change to a rating, dose range or safety flag is recorded there with the date and the reason.