Online Program Home
My Program

Abstract Details

Activity Number: 477 - SPEED: Bayesian Methods and Applications in the Life and Social Sciences
Type: Contributed
Date/Time: Wednesday, August 1, 2018 : 8:30 AM to 10:20 AM
Sponsor: Section on Bayesian Statistical Science
Abstract #330077 Presentation
Title: Bayesian Adaptive Design of Phase 2 Dose-Finding Study
Author(s): Tanya Granston* and Huafeng Zhou and Lixia Wang
Companies: CTI BioPharma Corp. and CTI BioPharma Corp. and CTI BioPharma Corp.
Keywords: Adaptive Designs; Interim Analysis; Predictive Probability

Often in dose-finding studies, though the objective is to determine a dose for future development, there's concern about prolonged exposure of patients to doses that do not provide benefit. Futility rules can be built into the design to prevent such a scenario. Here we present a Bayesian approach to evaluate the study outcome during the conduct of the trial, i.e., multiple interim analyses to inform dropping futile doses in a dose-finding study. This Bayesian approach relied on the calculation of Bayesian predictive probability (BPP) defined for the primary efficacy endpoint for each dose arm at repeated interim looks. With normal distribution assumption for the observed data for the primary efficacy endpoint and a conjugated, non-informative normal prior distribution for parameters, a closed-form equation for BPP is available (Dmitrienko and Wang, 2006). This allows the calculation of BPP to be easily performed using commonly available software such as SAS or R. A futility rule based on the BPP was defined and the operating characteristics of the Bayesian approach were evaluated in the context of our Phase 2 dose-finding study.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program