Using mixed treatment comparison meta-analysis to compare interventions in bovine practice

Abstract

Traditional meta-analysis involves a pairwise comparison of two interventions with one outcome. In this review mixed treatment comparisons (MTC) meta-analysis is discussed, which enables the comparison of multiple treatments using direct and indirect information from a network of trials. Mixed treatment comparisons meta-analysis is also known as network analysis. Although a potentially powerful tool for decision-making, several key assumptions are necessary for validity, and frequently these will not be met. As with any meta-analysis, MTC meta-analysis has the potential to propagate biases that occur at the trial level. Further, because of the assumptions necessary for MTC, bias can be introduced at the meta-analysis level, especially when trial populations compared are not comparable. In this review, we discuss how MTC meta-analysis may be used in bovine practice and raise concerns about its use.