|
Activity Number:
|
263
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 1, 2016 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #319648
|
|
|
Title:
|
An Extension of Clinical Trial Assurance to a Setting of Multiple Unknown Parameters in a Single-Arm Binomial Trial
|
|
Author(s):
|
Yizhou Jiang and Lynn Navale* and Allen Xue
|
|
Companies:
|
Kite Pharma and Kite Pharma and Kite Pharma
|
|
Keywords:
|
assurance ;
clinical trial design ;
single arm ;
power ;
historical control ;
Bayesian
|
|
Abstract:
|
O'Hagan introduced the concept of clinical trial assurance to assess the unconditional probability that a trial will yield a "positive outcome," rather than power conditional on a fixed treatment effect. Assurance may be extended to a single arm binomial study designed to test an observed rate against a historical control where the historical control is highly variable. The motivation for this application is a clinical trial setting in which the patient population has not been extensively studied and there is uncertainty in the historical control response rate. Given the importance of the historical control in a single arm study, exploration of the uncertainty in this parameter is critical to the design. The application of assurance presented here extends the methodology to a single sample binomial test where the study power is averaged over the prior distribution for the unknown historical control, and a single sample binomial test where the study power is averaged over the prior distributions for both the unknown historical control and an unknown treatment effect. Bayesian simulation is used to assess the study design under a range of scenarios.
|
Authors who are presenting talks have a * after their name.
