Covariate Adaptive Randomization in Seamless Phase II/III Clinical Trials

• Activity Number: 342 - Topics in Adaptive-Seamless and Group Sequential Designs
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Hongjian Zhu* and Wei Ma and Mengxi Wang
• Companies: University of Texas Health Science Center at Houston and Renmin University and University of Texas Health Science Center at Houston
• Keywords: Covariate adaptive randomization; Seamless clinical trials; Type I error

Abstract

There has been an urgent need to evaluate new therapies in a time-sensitive and cost-effective manner without compromising the integrity and validity of the development process. We studied the adaptive seamless phase II/III design (ASD) combined with covariate adaptive randomization (CAR) to meet this need. FDA (2018) drafted guidance on seamless clinical trials, representing its effort to broaden acceptance of the design. CAR is the most popular design in randomized controlled trials to ensure valid treatment comparisons. However, the type I error rate can be inflated under the dual influence of multiplicity and selection of ASD; CAR poses tremendous challenges due to its complicated allocation mechanism. Here we establish a theoretical foundation and perform comprehensive numerical studies for the adaptive seamless phase II/III trial with CAR. Three major challenges including avoidance of model misspecification, theoretical properties of the complicated design and analysis procedure, and control of the type I error rate, are addressed in the current study, which will be a real step forward to promote the application of this procedure.