Statistics Colloquium, Dr. Jing Zhang

Title: Bayesian hierarchical methods for meta-analysis combining randomized-controlled and single-arm studies

Abstract:   Meta-analysis of interventions usually relies on randomized controlled trials (RCTs). However, when the dominant source of information comes from single-arm studies, or when the results from RCTs lack generalization due to strict inclusion and exclusion criteria, it is vital to synthesize both sources of evidence. One challenge of synthesizing both sources is that single-arm studies are usually less reliable than RCTs due to selection bias and confounding factors. In this paper, we propose a Bayesian hierarchical framework for the purpose of bias reduction and efficiency gain. Under this framework, six models are proposed: two of them treat single-arm studies equally with RCTs, two adjust for design difference and potential biases, and the rest two further downweight single-arm studies adaptively and thus are recommended as primary methods for evidence synthesis. We illustrate our methods by applying all six models to two motivating datasets and evaluate their performance through simulation studies.