Keywords: Sample Size Re-estimation; Multiple Dose
When designing confirmatory phase 3 studies, due to uncertainty of treatment effect, a study with two doses of investigational drug and placebo with preplanned unblinded sample size re-estimation is often suggested. Although unblinded sample size re-estimation with one dose and placebo trial has been discussed extensively in the literature, the two dose and placebo trial adds much more complexities. For example, the amount of sample size increase has to be determined based on one or two doses that entered into the promising zone. In addition, if we are allowing dropping of one dose and subsequently change of randomization ratio after the interim analysis, it adds more complexity regarding the type 1 error control for the final analysis. In this round table discussion, the existing statistical approach and practical implementation consideration will be shared and discussed.