ABSTRACT
ABSTRACT Network meta-analysis (NMA) is a statistical technique to assess various treatment effects and compare their benefits or harms simultaneously in a systematic review. In comparative effectiveness research, NMA offers useful information to help patients and their caregivers make better decisions on their health care. Recently, many systematic reviews collecting various endpoints and methods for integrating such multivariate evidence jointly in NMA have been developed. In this chapter, we introduce Bayesian NMA methods under a missing data framework incorporating multiple outcomes by accounting for their inherent correlations. We utilize two different parameterizations which can be applied separately based on the scientific question of interest. In addition, we provide two decision-making tools, best and acceptability probabilities. We illustrate our methods using a real diabetes data example including two outcomes, and a simulation study validates the performance of our methods in terms of model selection and coverage probability. We close this chapter with a brief summary and discussion of potential future work.
