Examples
Worked examples
- Is an instance
A random-effects meta-analysis of 23 RCTs of a drug, with I² = 45% and pooled OR 0.78 (95% CI 0.68-0.89)
Counter-examples
Looks similar, but isn't
- Not an instance
A narrative summary of effect sizes without statistical pooling is not a meta-analysis
Editorial commentary
Meta-analysis is the statistical combination of effect estimates from two or more separate studies addressing a common question, typically using a random-effects model to compute a single pooled estimate with a confidence interval, alongside heterogeneity diagnostics (such as I²) that quantify how much the individual studies actually agree with each other. It is a specific quantitative technique, not a review format in its own right.
How this differs from a systematic review
Meta-analysis is typically embedded within a systematic review, which supplies the registered protocol and the process of identifying and appraising the eligible studies; the meta-analysis is the specific pooling step applied once those studies are in hand, and only where the included studies are judged sufficiently similar in participants, interventions, comparisons, and outcomes to make pooling meaningful. A systematic review can conclude without ever reaching a meta-analysis; conversely, a meta-analysis without a preceding systematic search and appraisal process is a much weaker, and increasingly discouraged, exercise.
Two named variants extend the basic method: network meta-analysis generalises pairwise pooling to three or more interventions compared simultaneously within one statistical model, combining direct and indirect evidence; individual patient data (IPD) meta-analysis uses raw participant-level data supplied by the original study investigators rather than published aggregate results, enabling consistent outcome definitions and more careful handling of missing data across studies — generally regarded as a gold-standard variant where it is feasible.
Example
A random-effects meta-analysis pooling 23 randomised controlled trials of a drug, reporting a pooled odds ratio with a 95% confidence interval and an I² statistic describing between-study heterogeneity.
Reporting standards
The MOOSE guideline (Stroup et al., JAMA 2000) is the reporting standard specific to meta-analyses of observational studies, distinct from PRISMA 2020’s broader systematic-review coverage. GRADE is commonly used alongside meta-analysis to rate the certainty of the pooled estimate across four levels (high, moderate, low, very low).
Also known as
Meta analysis · Quantitative synthesis
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="Meta-analysis"
vocab-term-identifier="https://casrai.org/dictionary/term/meta-analysis" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/meta-analysis",
"name": "Meta-analysis",
"identifier": "https://casrai.org/dictionary/term/meta-analysis",
"description": "A quantitative synthesis of effect estimates from multiple studies on a common question, using statistical methods (typically random-effects models) to compute a pooled estimate with confidence interval and heterogeneity diagnostics.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/research-outputs#set",
"url": "https://casrai.org/dictionary/term/meta-analysis",
"sameAs": [
"Meta analysis",
"Quantitative synthesis"
],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"author": {
"@id": "https://casrai.org/#editorial-team"
},
"datePublished": "2026-05-21T02:05:58",
"dateModified": "2026-08-22T15:29:14",
"inLanguage": "en-GB",
"isAccessibleForFree": true
}
