Assessing Heterogeneity Among studies’ Results in Meta-analysis

Abstract

Abstract:

Assessing heterogeneity among studies’ results is the most important step in meta-analysis for the purpose of selecting an appropriate model for combining the results to produce an overall statistic from the included studies.  The popular statistics methods are the visual investigation of the heterogeneity such as Forest plot, Galbraith plot and L’Abbé plot. Additionally, the Q statistics and I² can be used to assess the heterogeneity. In practice it is suggested that Forest plot, Q statistics and I² should be used together with the possibility of content heterogeneity in order to assess the heterogeneity between studies’ results.

Key words: heterogeneity, meta-analysis, Forest plot, Galbraith plot, L’Abbé plot, Q statistics, I²