Title: Bayesian Methods for Sample Size Determinations for Hypotheses Testing
Speaker: Dr. Sujit K. Ghosh, NC State University
Under the classical frequentist statistical framework, sample size calculations for a hypothesis test of interest maintain pre-specified Type-I and Type-II error rates whereas majority of the Bayesian methods are not designed to control for such errors despite the need for such controls from regulatory perspectives. However, the success of classical methods crucially depend on finding a pivotal quantity which becomes increasingly difficult for general composite null hypothesis (e.g., as in bio-equivalence and non-inferiority tests).
Thus, the need for a methodology that theoretically controls for desired level of errors and is broadly applicable for composite null hypothesis is desirable.
This webinar presents (i) a general Bayesian framework for hypothesis testing and sample size determination using recently developed concepts of "Bayesian average errors," (ii) provides theoretical and numerical illustrations of controlling two types of errors; and (iii) applications of the methodology for a few popular clinical trials. Some challenging aspects that remains to be explored in this area are also presented which calls for further collaborations within the Bayesian DIA community.
You can join the Skype meeting at: https://meet.lilly.com/natanegara_fanni/6T31HJQ8
or join by phone: 1 (317) 277-1498, 1 (855) 545-5910; with conference ID: 1407450