Statistical Considerations in Meta-Analysis

Abstract

The increase in medical research has led to a large body of related studies. The huge volume of research brings about a problem of how to organize and summarize the findings of studies. Meta-analysis is a statistical technique for combining the results from two or more studies, which addresses a similar hypothesis in a similar way. Meta-analysis includes the complete coverage of all relevant studies, and describes the results of each study via a quantitative index of effect size. Meta-analysis presents the precise estimate of treatment effect via combining these estimates across studies. Further, meta-analysis looks for the presence, degree and cause of heterogeneity, and explores the robustness of the main findings using statistical techniques. The author dealt with the some statistical issues and considerations which should be considered in conducting and presenting meta-analysis with explanation (ie. Effect size, Fixed and Random effect model, Heterogeneity, Reporting bias, and Meta-analysis Packages). This article may remind readers to conduct and evaluate the meta-analysis systematically and comprehensively.