How to Perform and Interpret a Network Meta-Analysis for Indirect and Mixed Comparisons: Key Methodological Strategies

Abstract

Background: Meta-analysis techniques were traditionally used to assess the effectiveness and safety of a treatment through direct comparison with a single comparator. More recently, network meta-analyses allowed the comparison of multiple interventions and the incorporation of clinical evidence of direct and indirect comparisons. We consider it relevant to address the main methodological considerations on this theme because of its topicality. Objectives: To demonstrate how to perform and interpret a network meta-analysis for indirect and mixed comparisons. Main topics under analysis: We present some key concepts of a network meta-analysis for indirect and mixed comparisons and the assumptions for its application. We point out some of the methodological challenges and opportunities for application of this statistical technique resulting from the theoretical and methodological analysis of articles published in the last 5 years. Conclusions: The use of indirect and mixed comparisons to increase the statistical power is an essential contribution to the current meta-analysis techniques, allowing researchers to improve the process of synthesis and health professionals to apply the best available evidence.