A method of interim analysis in a clinical trial in which very small p values are required for the early stopping of a trial, whereas later values for stopping are closer to conventional levels of significance.