March 9, 2023 Webinar
On the treatment effect estimation of count or recurrent event data
Tu Xu, PhD
Abstract
In clinical research of pulmonary diseases, the occurrence of acute pulmonary exacerbation events (PEx) is often selected as a key efficacy endpoint. As a type of recurrent event data, the PEx rate is often analyzed by generalized linear models (e.g. negative binomial regression, Poisson regression). Recent FDA draft guidance (2021) for randomized control trials specifically discusses the covariate adjustment on non-linear may yield conditional treatment effect that differs from the marginal treatment effect. In this presentation, we show that the PEx rate estimated with covariate adjustment is always less than the marginal event rate and further discuss the collapsibility of between-group rate ratio.
Short Bio
Dr. Tu Xu is an Associate Director of Biostatistics at Vertex Pharmaceutical. He is a Program Biostatistics Lead for the clinical development of cystic fibrosis therapies. Dr. Xu is motivated to advance innovative statistical approaches to accelerate clinical development process and his current research focuses on clinical trial design and real-world data analysis.