Activity Number:
|
245
- Bayesian Inference in the Life Sciences and Medicine
|
Type:
|
Contributed
|
Date/Time:
|
Monday, July 29, 2019 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Bayesian Statistical Science
|
Abstract #302909
|
|
Title:
|
Assessing Go/No-Go Decisions in Drug Development Under a Bayesian Paradigm Using Stan
|
Author(s):
|
Xiangyi Zhao* and Alan Hartford
|
Companies:
|
AbbVie Inc. and Takeda Pharmaceutical Company
|
Keywords:
|
Bayesian decision making;
Stan;
Drug development;
Go/No-Go decision
|
Abstract:
|
It can take over 10 years and 1 billion dollars to develop a drug. One major challenge is that an underlying drug effect is unknown. Phase II trials, which are designed to obtain an estimate of treatment effect for developing Phase III trials for regulatory approvals, are usually with a restricted and relatively small sample size resulted in a very high failure rate (around 50%) in Phase III programs. Thus how to effectively quantify uncertainty in the estimated treatment effect for a good Go/No-Go decision for Phase III trials has always been a central topic in drug development. Bayesian methodology has been shown to be very helpful for addressing this. Stan is a state-of-the-art probability computation language for statistical modeling with full Bayesian statistical inference abilities through the latest computation methods, such as Hamiltonian Monte Carlo for MCMC sampling. This talk discusses how to incorporate probability distributions that characterize current evidence about a treatment effect at a specific clinical milestone by using Stan under a Bayesian decision making framework in aiming to systematically increase the quality of decisions during life cycle of a drug.
|
Authors who are presenting talks have a * after their name.