2021/11/17 by Maxwell Cairns, Cairns, Maxwell, Luke A. Prendergast +1
Decision Sciences · #Meta-analysis and systematic reviews
paper · pdf · doi:10.48550/arxiv.2111.09518
Meta-regression is often used to form hypotheses about what is associated\nwith heterogeneity in a meta-analysis and to estimate the extent to which\neffects can vary between cohorts and other distinguishing factors. However,\nstudy-level variables, called moderators, that are available and used in the\nmeta-regression analysis will rarely explain all of the heterogeneity.\nTherefore, measuring and trying to understand residual heterogeneity is still\nimportant in a meta-regression, although it is not clear how some heterogeneity\nmeasures should be used in the meta-regression context. The coefficient of\nvariation, and its variants, are useful measures of relative heterogeneity. We\nconsider these measures in the context of meta-regression which allows\nresearchers to investigate heterogeneity at different levels of the moderator\nand also average relative heterogeneity overall. We also provide CIs for the\nmeasures and our simulation studies show that these intervals have good\ncoverage properties. We recommend that these measures and corresponding\nintervals could provide useful insights into moderators that may be\ncontributing to the presence of heterogeneity in a meta-analysis and lead to a\nbetter understanding of estimated mean effects.\n