Fig. 1From: Bayesian group sequential designs for phase III emergency medicine trials: a case study using the PARAMEDIC2 trialKey operating characteristics for Bayesian designs across several treatment effects and survival rates. The mean sample size is shown in the left column and the probability of declaring a difference between the trial arms is shown in the right column. The treatment effect is the raw difference between adrenaline and placebo survival rates. A positive treatment effect corresponds to adrenaline being superior; a negative treatment effect corresponds to adrenaline being harmful. The horizontal line in the right column figures represents a type I error of 5%Back to article page