Comparing multiple interventions with network meta-analysis

Article type
Authors
Salanti G1, Chaimani A2, Li T3, Caldwell D4, Higgins J4
1University of Bern
2Paris Descartes University and Cochrane France
3Johns Hopkins Bloomberg School of Public Health
4University of Bristol
Abstract
Background:
Standard meta-analysis methods for clinical trials focus on comparisons of two interventions, such as a drug versus placebo, or a new intervention versus standard practice. In clinical practice, there are rarely only two interventions under consideration. Extensions of meta-analysis to address three or more treatments have been the subject of much methodological research in recent years, and are increasingly being applied. Most simply, indirect comparisons can be performed in ways that respect the randomization within each clinical trial. More complex forms are the so-called network meta-analyses, also known as multiple treatments meta-analyses or mixed treatment comparison meta-analyses. These allow the simultaneous analysis of clinical trials involving different treatments.

Objectives:
To introduce the concepts and methods of indirect comparison and network meta-analysis in the context of a Cochrane systematic review, following the new Handbook Chapter drafted by the Cochrane Comparing Multiple Interventions Methods Group (CMIMG).

Description:
This workshop is aimed at statisticians, epidemiologists and other quantitatively-minded researchers who want to understand state-of-the- art statistical syntheses of clinical trials involving multiple interventions. The workshop will provide insights into network meta-analysis models that can be used to derive estimates for the relative effects of all treatments of interest. We will present practical examples using the netmeta package in R. By the end of this workshop participants will have an understanding of the role and potential of indirect comparisons and network meta- analysis in the evaluation of healthcare interventions; the principles, steps and statistical methods involved and the biases that can distort indirect comparisons and network meta-analysis.