Meta-analysis: pooling studies into one estimate
A meta-analysis combines (pools) the numbers from two or more similar studies into one weighted estimate, so larger, more precise studies count for more.
Remember
A systematic review is the structured search for the relevant studies and the appraisal of each one (a structured check of its quality), and a meta-analysis is the sum on top of it: the statistical pooling of their results, which a review may include but does not have to.
Memory aid
the systematic review is the method and the meta-analysis is the sum
Use it to keep the two terms apart: the review finds the studies and the meta-analysis adds up their numbers.
A meta-analysis combines the numbers from two or more similar original studies (the primary studies). It gives one pooled (combined) estimate of the effect, such as the average change in blood pressure with a drug, together with a 95% confidence interval (the range of values compatible with the data).
Each study is weighted by its size and precision (how tightly its result is pinned down). A larger trial with a narrower confidence interval therefore counts for more than a small, imprecise one. Pooling also makes the combined estimate less uncertain than the result of any one small trial, and it increases statistical power (the ability to detect a real effect).
Pooling only makes sense when the studies are similar enough. If they differ too much in patients, treatments or outcomes (clinical heterogeneity), pooling would be misleading, and the results are described in words instead.
Keep the two linked terms apart. A systematic review is the structured search for, and appraisal of, the relevant studies, where appraisal means a structured check of each study's quality. The meta-analysis is the statistical step that may be added to it, so a review may include one but does not have to. The usual picture is the forest plot: each study is shown as a horizontal line, with a diamond near the bottom for the pooled result.
Key facts
- Definition
- the statistical pooling of two or more similar studies into one weighted estimate of the effect.
- Output
- one pooled estimate of the effect with a 95% confidence interval (the range of values compatible with the data).
- Weighting
- each study is weighted by its size and precision, so larger trials with narrower confidence intervals count for more.
- When to pool
- the studies need to be similar enough; if they are too clinically different, pooling can be misleading and the results are reported in words instead.
- Link to a systematic review
- a systematic review may include a meta-analysis but does not have to.
- Standard picture
- the forest plot, in which each study is a horizontal line and a diamond shows the pooled estimate.
Exam traps
- A heterogeneity analysis (a check of how much study results differ from one another) is not a meta-analysis, because it assesses consistency rather than producing a single overall estimate.
- A sensitivity analysis repeats an analysis under different decisions, such as leaving out the smaller studies, to see whether the conclusion changes. It tests how robust a pooled result is rather than creating it.
- A simple average of study results is not a meta-analysis, because each study is weighted by its size and precision, so a large precise trial counts for more than a small imprecise one.
Read the full notes
Tested in 6 SimpleMed+ questions
Practice it in Pharmacology, with the answer and a full explanation after every question.
Practice these questions