Network Meta-analysis

Abstract

Network meta-analysis is a popular way to combine results from several studies (usually randomized trials) comparing several treatments or interventions. It has usually been performed in a Bayesian setting, but recently it has become possible in a frequentist setting using multivariate meta-analysis and meta-regression, implemented in Stata with mvmeta. I describe a suite of Stata programs for network meta-analysis that perform the necessary data manipulation, fit consistency and inconsistency models using mvmeta, and produce various graphics.