|
Activity Number:
|
441
- Bayesian (and Other) Clinical Trials Designs
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 12, 2021 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #318255
|
|
|
Title:
|
An Updated Pick-the-Winner Design Allowing a Flexible Allocation Ratio at Stage Two
|
|
Author(s):
|
Bing Liu* and Jianghua He
|
|
Companies:
|
University of Kansas, Medical Center and University of Kansas Medical Center
|
|
Keywords:
|
Clinical Trials;
Multiple Treatments;
Two-Stage Design;
Sample Size Calculation;
Binomial Distribution;
Ranking and Selection
|
|
Abstract:
|
Pick-The-Winner (PTW) design is a clinical trial procedure aiming to select the best treatment arm among multiple candidates. A two-stage PTW design combines a screening process (as stage one) and a hypothesis testing (as stage two) to select the best treatment arm based on least favorable configuration. The two-stage PTW design was initially proposed in 1988 for binomial endpoints based on normal approximation. A similar design was developed in 1989 based on the exact binomial distribution. Though these designs have been extended in various ways, some gaps remained in the literature for binary outcomes. In this manuscript, we proved the inaccuracy of the sample size formula for Stage 2 proposed in 1989 and proposed a modified sample size formula. Our formula not only overcomes the issue of inaccuracy, but also enables flexible allocation ratios for Stage 2. Compared to the original designs, our design with balanced allocations shows a reduction in the sample size needed. We will also present some designs with unbalanced allocation ratios.
|
Authors who are presenting talks have a * after their name.
