Treatment effect estimate in adaptive clinical trial
*Gordon Lan, Janssen R&D, Johnson & Johnson 


In this presentation, we will start with treatment effect estimation for fixed designed studies and group sequential trials. In general, only a Bayesian approach could provide a perfect solution. Unfortunately, it is almost impossible to obtain a consensus agreement on the prior distribution. An example will be given to illustrate the difficulty of picking the right prior distribution. Time permitting, we will also discuss the meaning of treatment effect under modeling and simulation.