Meta-analysis in hematology and oncology

Meta-analysis increasingly is being used to synthesize evidence from randomized, controlled trials in order to help make treatment decisions. This article discusses important prestatistical and quantitative issues in summarizing evidence. For meta-anaylsis to be reliable, a thorough literature search and rigorous assessment of the methodological quality of studies included in the meta-anaylsis must be performed. Most meta-analyses combine treatment effects across studies and report weighted averages. Heterogeneous data, however, are more appropriately analyzed using regression analysis or individual patient data modeling.