Incorporating Historical Controls in Analyses of Clinical Trial Data: An Empirical Power Prior Approach

• Activity Number: 29 - Statistical Issues Specific to Therapeutic Areas, Power and Sample Size Calculations, and Trial Monitoring
• Type: Contributed
• Date/Time: Sunday, August 8, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Duncan Rotich* and Yun Shen and Kay Tatsuoka
• Companies: Bristol Myers Squibb and Bristol Myers Squibb and Bristol Myers Squibb
• Keywords: Historical Controls; Power Prior; Dynamic Borrowing; Bayesian Analysis; Clinical Trials

Abstract

Recently, there has been an increasing interest to integrate information from historical controls for informed decisions in new intervention assessments. Integration of historical data with randomized controls creates an evidence-based synthesis that may potentially increase precision in treatment effect evaluation. Besides, during the design phase of a trial, borrowing from historical controls has the potential to reduce the number of subjects randomized to the control arm due to availability of complementary data, hence, allowing more participants to receive a potentially life-saving experimental drug especially in trials with enrolment challenges. One challenge with borrowing lies in specifying the amount of information to borrow from identified studies. Currently, several approaches have been developed to borrow from historical controls e.g., power prior. In this presentation, we extend the power prior approach by creating a dynamic borrowing framework, which utilizes the level of similarity between the controls to address the amount of information borrowed. Using simulations, our approach demonstrates better performance regarding power and control of type I error.