How we research
Without a medical degree behind it, this site's sourcing discipline has to be stricter than a site with doctors on staff, not looser. Here is exactly how it works.
The short version
Every factual claim traces to a numbered source at the bottom of the article. Every source links to its DOI or PubMed record so you can open it yourself. Every claim about whether something works carries a grade telling you how good that evidence actually is. Where the research is thin or contradictory, the article says so rather than picking a side and sounding confident.
The four grades
Grades apply to a specific outcome, never to a compound as a whole. Most supplements have decent evidence for one thing and very little for the four other things they are sold for — separating those is most of the point.
Consistent findings across several randomised controlled trials, or a meta-analysis of them, in a population close enough to ours to be relevant.
Randomised trials exist, but they are few, small, short, or were run in a different population. Or there is consistent observational data with a plausible mechanism behind it.
Evidence from mechanisms, animals or observation only — or human trials that contradict each other.
A claim being marketed with no human trial behind it in this population at all.
What we cite, in order
We cite as high up this list as the question allows, and only move down it when nothing above exists.
- Systematic reviews and meta-analyses of randomised trialsThe strongest thing we can cite. Used for general claims about whether something works.
- Individual randomised controlled trialsUseful, but always reported with the population, the dose and the limits attached.
- Large prospective cohort studiesGood for association and timing — never for cause. A cohort can tell you two things happen together, not that one causes the other.
- Guidelines from named bodiesThe Menopause Society, NICE, the Endocrine Society, ACOG. Cited with the year, and attributed to them rather than stated in our own voice.
- Mechanistic and animal studiesOnly ever used to explain how something might work. Never used to recommend anything.
What we never cite as evidence
Press releases and university media offices. Supplement brand pages, including their "research" sections. Other content sites — if a big health site cites a study, we go and read the study. And no podcast, video or book is ever a source for a factual claim.
Researchers and clinicians are a source of framing, not a source of fact. They are how you find the study and understand the argument. Then you cite the study. Where one of them sells a product related to what they are saying, we say so.
On the limits of reading studies
Reading the literature is not the same as clinical expertise, and nothing here pretends otherwise. Plenty of things were observed in practice for years before a trial confirmed them."No good trial has shown this" is not the same sentence as "this doesn't work" — published evidence lags reality, and in women's health a great deal simply has never been studied. Where that gap matters to a claim, the article says so.
When we get it wrong
Articles carry a "last reviewed" date, and that date only moves when the sources have actually been read again — not when a word was changed. When new evidence changes a conclusion, the article is updated and the change is described rather than quietly swapped in. If you find something wrong, tell us and it gets fixed.