Title: Methodological advances in evidence synthesis

Network meta-analysis (NMA) is an extension of the usual (pairwise) meta-analysis. It is a statistical tool for synthesizing evidence obtained from studies comparing multiple competing interventions for the same disease. In this lecture, we will go through some recent advances in the field. First, we will discuss a new model for the NMA of binary outcomes. This model generalizes the well-known Mantel-Haenszel method, and can be especially valuable for the case of rare events, e.g. when synthesising data on mortality or serious adverse events. The method has been implemented in R in freely available, easy-to-use routines. Second, we will discuss models for including non-randomized studies in NMA. Non-randomized studies can reveal whether or not interventions are effective in real-life clinical practice and there is a growing interest in including such evidence in the decision-making process. Here we present and compare an array of alternative methods, and we apply some of the methods in previously published clinical examples. Finally, we will discuss methods for individual participant data network meta-analysis (IPD-NMA). IPD are considered the gold standard in evidence synthesis, and inclusion of IPD in NMA offers unique advantages, such as increase in precision, decrease in heterogeneity, as well as the capacity to individualize the treatment according to a patient’s characteristics. We showcase our methods using an example from depression.

Note. This is a joint RSS South West Local Group and Exeter Health Statistics Event