Multivariate meta-analysis of multiple outcomes: characteristics and predictors of borrowing of strength from Cochrane reviews

Abstract

\n Objectives\n\nMultivariate meta-analysis allows the joint synthesis of multiple outcomes accounting for their correlation. This enables borrowing of strength (BoS) across outcomes, which may lead to greater efficiency and even different conclusions compared to separate univariate meta-analyses. However, multivariate meta-analysis is complex to apply, so guidance is needed to flag (in advance of analysis) when the approach is most useful.\nStudy design and setting\n\nWe use 43 Cochrane intervention reviews to empirically investigate the characteristics of meta-analysis datasets that are associated with a larger BoS statistic (from 0 to 100%) when applying a bivariate meta-analysis of binary outcomes.\nResults\n\nFour characteristics were identified as strongly associated with BoS: the total number of studies, the number of studies with the outcome of interest, the percentage of studies missing the outcome of interest, and the largest absolute within-study correlation. Using these characteristics, we then develop a model for predicting BoS in a new dataset, which is shown to have good performance (an adjusted R2 of 50%). Applied examples are used to illustrate the use of the BoS prediction model.\nConclusions\n\nCochrane reviewers currently use univariate meta-analysis methods, but our prediction model for BoS helps to flag when a multivariate meta-analysis may also be beneficial in Cochrane reviews with multiple binary outcomes. Extension to non-Cochrane reviews and other outcome types is still required.