What is a meta-analysis?
A meta-analysis is a study of studies: it pools the data of several comparable investigations in one statistical analysis. This allows an overall effect to be estimated more precisely than from a single small trial. Its quality depends on the studies included.
Meta-analysis explained
Individual studies are often small and give mixed results. A meta-analysis therefore collects several studies on the same question, combines their results into one overall effect and gives larger, more precise studies more weight. The result is often shown in a “forest plot”: each study is a dot with error bars, and the pooled effect sits at the bottom.
It usually starts with a systematic review, where studies are searched for and selected by fixed rules. The meta-analysis is then the statistical part that merges the numbers. Not every systematic review contains a meta-analysis, for example when the studies are too different.
Where does a meta-analysis sit in the evidence?
Simplified ranking of study types. A rule of thumb, not a law of nature.
| Study type | Strength for effects in humans |
|---|---|
| Study type Meta-analysis / systematic review of RCTs | Strength for effects in humans highest, if the individual studies are good |
| Study type Randomised controlled trial (RCT) | Strength for effects in humans high, especially with placebo and blinding |
| Study type Observational study | Strength for effects in humans shows associations, not cause |
| Study type Animal and cell experiments | Strength for effects in humans give hints, do not automatically transfer to humans |
What a meta-analysis can and cannot do
Strengths: more participants, more precise estimates, and it shows whether several studies point in the same direction.
Limits: a result is only as good as the individual studies (“garbage in, garbage out”). If studies differ a lot (dose, preparation, duration), results are heterogeneous and a single pooled value says little. There is also publication bias: studies with positive results are more likely to be published, which can skew the overall picture.
Meta-analyses on curcumin
There are several meta-analyses on turmeric and curcumin, for example on joint complaints in osteoarthritis. Our articles describe them this way: on average they show a moderate, statistically measurable improvement in pain compared with placebo; the quality of the individual studies varies, however, and some authors regard the trials as too small for a firm recommendation. That makes curcumin no miracle remedy. At the same time the articles point out that results depend strongly on the formulation and thus on absorption, and that the individual studies vary in duration, dose and quality.
To judge studies yourself, look at sample size, placebo group and blinding, as described in the entry on the randomised controlled trial. Understanding what “evidence” means also helps.
Check: how many studies and participants are included? How different are the studies (heterogeneity)? Were quality and publication bias assessed? A single “significant” value is no guarantee of a relevant benefit in daily life.
Frequently asked questions about meta-analysis
What is the difference between a meta-analysis and a review? A classic review article summarises studies in narrative form. A systematic review searches by fixed rules; a meta-analysis additionally calculates the numbers together.
Is a meta-analysis always better than a single study? Not automatically. A meta-analysis of many weak or very different studies can be less informative than one large, well-run study.
What does “moderate effect” mean? It is a classification of effect size. Moderate means measurable but not dramatic. Whether it is noticeable in daily life depends on the person and the question.
How do I recognise a good meta-analysis? It defines in advance which studies are included, states its search strategy, assesses study quality and openly discusses differences between results. Reliable sources link the original paper so you can check the claims yourself.
