Comparing multiple treatments 2: Statistical methods for network meta-analysis

Article type
Authors
Salanti G1, Higgins J2, Schmid C3
1University of Ioannina School of Medicine, Greece
2MRC Biostatistics Unit, United Kingdom
3Brown University, USA
Abstract
Objectives:

To understand the statistical methodology of networkmeta-analyses and the assumption of consistency.

Description:

Network meta-analysis is the statistical methodology used to combine evidence in a network of trials that compares more than two interventions. The workshop will provide insight to network metaanalysis models that can be used to derive estimates for the relative effects of all treatments of interest. We will explore the different implementation alternatives (Bayesian versus frequentist) through worked examples and we will discuss extensively the underlying assumption of consistency. We will present models to check for and incorporate inconsistency in the results and present applications of multiple-treatments meta-regression models. Finally, we will discuss concerns regarding the role of bias in network meta-analysis.