Abstract:
|
In clinical studies together with uncertainties associated with observed endpoints we face uncertainties caused by an enrollment process that often can be viewed as stochastic process. If one observes time to event, then the subject exposure intervals and censoring times are random. Thus, unlike the traditional optimal design setting, the amount of information that can be gained during the execution of planned study is random and becomes known only after completion of this study. We show that for moderately large sample size the maximization of the average information is a sound strategy both theoretically and computationally. For the sake of simplicity we stayed with the Poisson enrollment model. Our approach is based on the concept of the elemental Fisher information matrix that allows the derivation of the very general results that work for a large family of survival models and convex optimality criteria.
|