Modification of the sample size and the schedule of interim analyses in survival trials based on data inspections.

A method is presented which allows us to adapt the sample size as well as the number and time points of interim analyses to the treatment difference observed at an interim look during the course of a clinical trial with censored survival time as the endpoint. The method allows the inclusion of data inspections during the course of the trial and redesign of the trial on the basis of the observed treatment difference without affecting the type I error risk. Formulae for recalculating the required number of events and the number of further patients to be randomized as a function of the observed hazard rates and the detectable hazard ratio are given. Copyright 2001 John Wiley & Sons, Ltd.