A Bayesian Adaptive Trial for CER: Case Study in Status Epilepticus
View Presentation View Presentation
Kristine Broglio, Berry Consultants  *Jason Connor, Berry Consultants  Jordan Elm, Medical University of South Carolina  Jaideep Kapur, University of Virginia 

Keywords: Bayesian, adaptive trial, comparative effectiveness research

We demonstrate a Bayesian adaptive comparative effectiveness design. The Established Status Epilepticus Treatment Trial (ESETT) is a multicenter, randomized, double-blind,CER study of three commercially available treatments for the treatment of benzodiazepine-refractory Status Epilepticus (SE) in an emergency department setting. We incorporate adaptive randomization, arm dropping, and frequent interim analyses for early stopping in order to achieve multiple goals: identify the best and worst treatment for SE, increase power, improve treatment of patients within the trial, and terminate the trial as soon as the clinical question is answered or it becomes evident that it is unlikely to be answered within this trial. We highlight differences in CER trials and why Bayesian adaptive trials are especially beneficial in these scenarios, even in settings such as ERs.