|
Abstract:
|
Outcome reporting bias (ORB) and publication bias (PB) are recognized as a threat to the validity of both pairwise and network meta-analysis (NMA). In recent years, multivariate meta-analysis (MMA) approaches have been proposed to handle the impact of potential ORB in the pairwise setting. These methods have shown that MMA can reduce bias and increases efficiency of pooled effect sizes. However, it is unknown whether multivariate NMA (MNMA) applied to multiple treatments (network) systematic reviews can similarly reduce ORB and PB. To determine whether MNMA can reduce the impacts of ORB and/or PB on pooled treatment effect sizes, we present an extensive simulation study adopting a Bayesian MNMA model similar to Efthimiou et al. Via extensive simulation studies, we show MNMA reduces the bias of pooled effect sizes under a variety of missingness scenarios, including missing at random and missing not at random. Further, MNMA also improves the precision of estimates producing narrower credible intervals. We demonstrate the applicability of the approach via application of MNMA to a large multi-treatment systemic review of second-generation antidepressants.
|
