Online Program

Return to main conference page
Wednesday, September 25
Wed, Sep 25, 2:45 PM - 4:00 PM
Thurgood Marshall South
Learn from History or You're Doomed to Repeat It

Adaptation of Historical Data-Borrowing Methods in Clinical Trials with Event Outcomes (301026)

*Bingzhi Zhang, Sanofi 

In order to improve the efficiency of clinical trials, several Bayesian analysis methods were proposed to borrow existing external or historical control data to supplement the concurrent control data in a new study. In this presentation, application of different borrowing methods ( including but not limited to power prior, commensurate prior) in trials with event outcome will be demonstrated. Their performance in frequentist operation characteristics will be compared by simulation studies. Some practical considerations in implementing these methods will also be discussed.