Meta-analysis
Also known as: pooled analysis, systematic review with meta-analysis, evidence synthesis
Definition
A meta-analysis is a quantitative research method that pools data from many individual studies (usually randomised controlled trials) to generate a combined estimate of the treatment effect. Rather than relying on a single study, which may have limited sample size or be affected by chance findings, a meta-analysis aggregates information across studies to provide a more robust and precise estimate. Researchers conduct meta-analyses by retrieving published (and sometimes unpublished) studies, assessing their quality, extracting relevant data, and using statistical models to combine results while accounting for differences in study design, population, and outcome measurement. Meta-analyses produce summary statistics (such as pooled effect sizes, confidence intervals, and number-needed-to-treat) that summarize the overall evidence base. A systematic review is the rigorous process underlying a meta-analysis, involving comprehensive literature searching, explicit inclusion/exclusion criteria, and quality assessment.
Meta-analyses are valuable for identifying consistency of effects across studies, for detecting publication bias (the tendency for studies with positive results to be published more often), and for examining effect modification (whether some subgroups benefit more than others). Researchers use graphical displays called forest plots to visualize individual study results and the pooled estimate. Heterogeneity (differences in study results) is assessed statistically and explored through subgroup analyses and meta-regression to understand what factors explain variations between studies.
The quality of a meta-analysis depends critically on the quality of included studies and the appropriateness of combining them. Poor-quality studies with high bias can skew pooled estimates, and combining studies with very different designs or populations may produce misleading results. Transparent reporting of meta-analyses is essential; guidelines such as PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) specify how results should be documented to allow assessment of validity.