Skip to content
Menu

Editorial policy

How we choose sources, check accuracy, grade evidence and correct mistakes.

Sources

We write from peer-reviewed research, clinical guidelines and systematic reviews. Every factual claim is numbered and linked to its source in the references list, using the DOI or PubMed entry where one exists.

Medical review

Research summaries are always checked by a medical reviewer before they are published. Other articles may be reviewed too. A reviewer never reviews their own writing. If an article changes after review, it is reviewed again, and the review date shown is only ever the date of the latest review.

How we grade evidence

Each post shows the strongest kind of evidence it relies on. Stronger evidence means we can be more confident; weaker evidence is not wrong, just less certain.

  1. Meta-analysis

    Pools the results of several studies, usually randomized trials.

  2. Randomized controlled trial

    People are randomly assigned to a treatment or a comparison, which limits bias.

  3. Clinical study

    Tests a treatment in patients, but without full randomization or control.

  4. Observational study

    Looks for links in existing data. Shows association, not cause.

  5. Expert opinion

    Based on clinical experience and guidelines rather than new data.

  6. Anecdotal

    Individual reports or common practice, not formal research.

Updates and corrections

We revisit articles as new research appears. Meaningful changes are shown with an "Updated" date. If you spot an error, tell us and we will correct it.

Independence

Research content is never sponsored. Nobody can pay for coverage, for a score, or for a place in a roundup, and no brand sees a write-up before it is published.

Reviews and roundups carry affiliate links, and every one of those pages says so above the first link. We may earn a commission if you buy through one, at no extra cost to you. What we earn does not depend on what we score a product.

Commercial pages are kept visually and structurally separate from research. A testing score grades a product we handled; it is not a statement about the evidence for what that product claims to do. Those are two different scales and they never share a slot.

We only score products we have used. Every score records who tested it, when, and how we obtained the unit — and a unit the brand supplied says so next to the score. See how we test.