Fig. 2From: Designs for adding a treatment arm to an ongoing clinical trialComparing a single trial to separate trials when a treatment arm is added at different time points during the trial, for varying power (70%, 80% and 90%) and FWER (2.5%, 5% and 10%). It is assumed here that randomisation is 1:1:1 and the treatment effect to be detected is the same for all treatments.Back to article page