299 – Meta-Analysis and Other Innovative Comparison
Meta-Analysis of Rare Adverse Events in Randomized Controlled Clinical Trials: A Case Study
Suna Barlas
Pfizer Inc.
Eunhee Hwang
Pfizer Inc.
Lingfeng Yang
Pfizer Inc.
Meta-analysis of randomized controlled clinical trials are often used to evaluate the effect of a new drug entity on rare adverse events as safety characterization plays an important role in the overall benefit risk assessment. In this case study, we apply commonly used meta-analysis techniques to our clinical trial data: combining summary statistics across trials via variance inflation or Newcombe method, Bayesian fixed / random effect model, generalized linear model, and generalized linear mixed model. We examine the impact of explicit and implicit assumptions in each technique and provide intuitive interpretation for some surprising results. Our choice of analysis method and metrics for the data are explained considering the issue of continuity correction, large sample approximation, bias, coverage probability and the maximum use of all available data. In this presentation, we limit our focus on the binomial outcomes without adjusting for the exposure.
