Comparing multiple interventions workshop: Estimating treatment effects and evaluating the evidence from network meta-analysis

Nikolakopoulou A1, Chaimani A2, Higgins J3
1Institute of Social and Preventive Medicine, University of Bern, 2Research Center of Epidemiology and Statistics Sorbonne Paris Cité (CRESS-UMR1153), Paris Descartes University, 3School of Social and Community Medicine, Bristol

Objectives:To understand the statistical methodology of network meta-analysis and the assumption of consistency. To present a methodology that can be used to evaluate the credibility of evidence.

Description:This is the second of two workshops offered by the Cochrane Comparing Multiple Interventions Methods Group. The workshop will provide insight into network meta-analysis models that can be used to derive estimates for the relative effects of all treatments of interest. We will present approaches to check for and account for inconsistency in the results.

It is important to consider the confidence with which produced treatment effects and treatment ranking can enable clinicians and decision makers to make informed decisions. We will present a framework based on the GRADE (Grading of Recommendations, Assessment, Development and Evaluations) system that can be used to evaluate the credibility of the results from network meta-analysis. Core aspects of the approach include considerations about the plausibility of the transitivity assumption underlying network meta-analysis and understanding the relative contributions of direct and indirect evidence. We will illustrate the process using networks of different size and complexity and we will show how a web application, CINeMA (Confidence In Network Meta-Analysis), simplifies the evaluation of the quality of NMA results with semi-automation of methods and via a guided on-line process.