Exploring heterogeneity in meta-analyses: needs, resources and challenges


Dr William D. Fraser, Department of Obstetrics and Gynecology, Université de Montréal, 3175 Chemin de la Côte Sainte-Catherine, Montréal (Quebéc), Canada H3T 1C5.
E-mail: william.fraser@umontreal.ca


The investigation of heterogeneity remains an essential but difficult issue in the conduct of meta-analysis. We reviewed standard and graphical methods used to explore heterogeneity in meta-analysis and publications from January 2005 to April 2007 regarding meta-analyses that focused on perinatal health topics. We assessed their approaches to the investigation of heterogeneity, including: (1) whether statistical testing for heterogeneity was performed and, if so, which test was used, (2) how a finding of statistically significant heterogeneity was handled, and (3) how the analyses were conducted in the presence of heterogeneity.