|
Activity Number:
|
541
- Bayesian Design in Clinical Trial and Some Challenging Issues
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 11, 2022 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
|
Abstract #322058
|
|
|
Title:
|
Practical Steps for Dynamic Borrowing of Historical Control Data in Clinical Trials
|
|
Author(s):
|
Farid Jamshidian* and Ron Xiaolong Yu
|
|
Companies:
|
Gilead Sciences and Gilead Sciences
|
|
Keywords:
|
Borrowing;
Historical controls;
Bayesian;
Clinical Trials
|
|
Abstract:
|
Borrowing historical data in clinical trials can reduce sample size and cost. In practice, investigators designing such trials may shy away from incorporating borrowing into trial design due to practical and regulatory hurdles. At the time of the analysis, patient characteristics or the outcome for the current and historical cohorts may differ. Such lack of similarity between the cohorts may prevent borrowing of historical data and may in turn result in the study having lower number of subjects than required to have sufficient power for testing the primary endpoint. We propose a new practical approach to borrowing historical data. We demonstrate that relying on an independent data monitoring committee (DMC) to evaluate a pre-specified criteria for borrowing may grant the investigators an additional option to continue enrolling prospective subjects if the similarity conditions for borrowing are not met, while preserving type-I error control.
|
Authors who are presenting talks have a * after their name.
