How we read the evidence

The four levels of strength

From a practice carried by tradition alone, up to a finding more than one research group has reproduced. A label changes when the studies behind it change.

04evidence-based

Supported by controlled trials or meta-analysis, replicated by more than one research group.

03evidence-informed

Real signal in the research, with small studies, small effects, or few controls holding it back.

02emerging

Early findings that look promising and need better studies before anyone can be sure.

01traditional

Long practised in a living tradition, with limited clinical study behind it.

Three labels outside the scale

Some ideas are the wrong shape for a trial. A philosophy is a different kind of claim, so putting it on a strength scale would misrepresent it. These labels say what kind of thing you are reading.

··experiential

Reported consistently by people who practise it, without objective measurement.

··philosophical

A way of understanding experience whose value sits outside experimental trials.

··metaphorical

A useful image for something real whose meaning sits outside a proposed mechanism.

Which sources we use

Primary research and public health bodies first. Every link was opened and checked on the review date. Where a record remained inaccessible, the text names the study without linking it.

• PubMed Central and PubMed
• The World Health Organization
• The National Institutes of Health and the National Institute on Aging
• Harvard Health
• The Sleep Foundation
• The Centers for Disease Control and Prevention
• University research groups publishing in peer-reviewed journals

Where a study was funded or authored by a company selling the thing being tested, the brief says so next to the finding. Company involvement calls for measured confidence while the study still belongs in the evidence record.

How often we revisit a page

Every brief shows the date it was last reviewed. Pages are revisited as new work appears, and a brief whose verdict changes says what changed and why.

The boundaries of these pages

These briefs are educational summaries of findings across groups of people. Clinical diagnosis and treatment belong with a clinician who knows your history.

• Outcomes remain uncertain; cure, treatment, and healing claims require separate clinical evidence.
• Studies, statistics, quotes, and links must be verifiable.
• Product recommendations stay separate from research findings.

Where the evidence is weak, the page says so in the first paragraph. A verdict only means something when it could have gone the other way.