Fig. 2From: Seamless phase 2/3 design for trials with multiple co-primary endpoints using Bayesian predictive powerThe Overall power difference between BPP and CP when the null hypothesis is falseBack to article page