Event Projection and Adaptive Adjustment for Event-Driven Trials

• Activity Number: 515
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Xiaoqiang Xue*+
• Companies: Quintiles, Inc.
• Keywords: Event-driven trial ; Poisson-gamma ; Time to events

Abstract

For event driven trials, such as cardiovascular and oncology studies, it is always desirable to know the projected number of events at the trial design stage, in order to properly power the study, with time and financial constraints, and better schedule interim analysis as well as prepare the regulatory submission. Furthermore, it can provide reliable information for sample re-assessment during the study if necessary. The presentation focuses on combining stochastic modeling of multicenter enrolment and the dynamics of the events occurring to predict the time to events. We utilize the Poisson-gamma models for both patient enrolment and event accrual. A Bayesian simulation scheme is used to generate the probability distribution of predicted timelines of specified events, which can help the investigator to make adjustment to accomplish a study within the timeline constraint bounds.