Bayesian approach to evidence synthesis

Abstract

We briefly present the advantages and opportunities available to umbrella reviews from the use of Bayesian techniques while taking into account that the concerns commonly arising in Bayesian meta-analysis procedures are also present in umbrella reviews. This is the case, for example, of sparse data, for which the hierarchical logit-normal model can give very poor results. An additional concern in this context is that of the choice of noninformative priors, which can lead to a significant variation in the final conclusions drawn. Accordingly, this chapter highlights the potential for Bayesian approaches in umbrella reviews, overviews of reviews, and meta-epidemiologic studies while acknowledging their limitations and complexities.