How we grade evidence
Health writing usually presents everything with the same confidence, whether it rests on forty trials or on one podcast. Every recommendation here carries a grade instead.
Well established
We are confident about this.
Supported by multiple randomised trials or meta-analyses that broadly agree.
Supported
The evidence leans this way.
Limited trials, or strong observational data plus a plausible mechanism.
Uncertain
Interesting, but we do not know yet.
Early, small or conflicting studies. Treat as a hypothesis, not a fact.
Practice-based
No good research. Cheap to try.
Widely used and low risk, but not meaningfully studied. We say so rather than dress it up.
The rules we hold ourselves to
- The first thing we ever recommend to someone new is graded A or B. Never C, never D.
- Sources are primary research, not other blogs.
- Every page carries the date it was last checked. Anything older than two years gets reviewed again.
- When we change a recommendation, we write down what changed and why.
- Nothing on this site is paid for by a brand, and no recommendation is ever chosen because it earns money.
- Where a topic touches disordered eating, depression, self-harm, sleep apnoea or pregnancy, the page says so at the top and points to a clinician.
What would change our mind
On anything graded C, the page states what result would move it up or down. If that result arrives, the page changes and the change is logged.
This is not medical advice
It is general information written carefully. It cannot know your history, your medication or your diagnosis. Anything persistent or frightening belongs with a doctor.