Fig. 1
From: Bayesian adaptive designs for multi-arm trials: an orthopaedic case study
Flow diagram showing the process for the virtual re-execution of designs 3–6. Response adaptive randomisation or arm dropping was performed every 50 patients until the final analysis (at N = 584). Early stopping for efficacy or futility was assessed every 200 patients. The process depicted in this figure was repeated 1000 times. CAST Collaborative Ankle Support Trial