Every article in the library carries an evidence label near the title. It describes how strong the research behind the article’s main conclusions is, not how important the topic is. This page explains the four labels and the kinds of study you will see cited.
The four labels
| Label | What it means | Typical basis |
|---|---|---|
| Strong | The main conclusions are unlikely to change with new research | Several consistent systematic reviews or large high quality trials |
| Moderate | The direction is clear, but the size of the effect or its limits may change | A systematic review or meta-analysis with some limitations, or several consistent trials |
| Limited | Some evidence exists, but it is small, short, indirect or inconsistent | Small trials, cross-sectional studies, laboratory studies |
| Expert opinion | Little direct research; based on guidelines, principles and clinical experience | Guidelines, consensus, reasoning from related evidence |
A page can have a moderate label and still contain claims with weaker support. In that case the text says so next to the claim.
Kinds of study we cite
- Systematic reviews and meta-analyses pool many studies on one question. They carry the most weight when the studies they pool are good.
- Randomised controlled trials assign people to a treatment or a comparison by chance, which is the best way to test whether something works.
- Cohort studies follow people over time. They can show what predicts an outcome, but not prove cause.
- Cross-sectional studies are snapshots. They show associations, such as more neck pain in heavy phone users, but cannot show which came first.
- Laboratory studies measure short term changes in controlled conditions.
- Guidelines summarise evidence and expert consensus for practice.
How to read a number
- Association is not cause. “Linked to” means the two travel together, not that one causes the other.
- Check who was studied. Results in young healthy adults may not apply to older people or people in pain.
- Look at the size. A statistically significant effect can still be too small to notice.
- One study is one study. Look for whether reviews agree.
Each claim in an article links to its source, and the full list is on sources. How we decide what to say when studies disagree is on how we check claims.
General information, not medical advice. Every claim carries its source; where the evidence is thin, the page says so.Editorial policy · How we grade evidence · All sources · Disclaimer