Network Meta-Analysis

Abstract

This chapter discusses two major approaches, a frequentist as well as a Bayesian hierarchical model, and how they can be implemented in R and Stata. Network meta-analysis can combine both direct and indirect evidence to estimate the effect sizes from several randomized trials. A simple random-effects model can be used in network meta-analysis as well. Now the network meta-analysis is applied to the same observed effect sizes as in the standard pairwise meta-analysis but uses the new observed variances. The chapter shows how to perform a network meta-analysis under the Bayesian framework. Currently, there are two R packages bnma and gemtc available for Bayesian network meta-analysis. A forest plot is a commonly used approach to summarize and present the meta-analysis results, which could also be used in reporting Bayesian network meta-analysis models. The chapter presents how to carry out frequentist network meta-analysis in Stata using network package.