Abstract:
|
Methods that allow for unblinded sample size re-estimation while protecting the overall type I error in a clinical trial have been around for close to 20 years, yet use of these methods are not very widespread. Potential reasons may include regulatory challenges, operational concerns, software limitations, or uncertainty over how to implement such methods for more complex situations such as those involving time-to-event endpoints and testing of secondary endpoints. In this roundtable, we will discuss obstacles to implementing unblinded sample size re-estimation and ideas for navigating around such obstacles with a particular focus on issues relating to more complex trials.
|