[How I explore...the secrets of a meta-analysis]

Meta-analysis is becoming a key partner of Evidence-Based Medicine. This statistical tool is currently used in most systematic reviews and allows to quantitatively synthesize the results of several studies by a single weighted average whose precision is indicated by the confidence interval. A standardized graphical representation allows, at first glance, to derive a lot of useful information, especially the number of studies taken into account for the analysis, the mean effect observed in each study with its confidence interval, the relative size (number of subjects) of each study, the homogeneity/heterogeneity of between-study results, finally the overall mean effects with its confidence interval. The meta-analysis plays a key-role in epidemiology, for the evaluation of a risk or a protective factor, and in therapeutics, for summarizing the results of numerous clinical trials with a specific drug or a pharmacological class in a given disease.