FDA Draft Guidance: Use of Bayesian Methodology in Clinical Trials of Drug and Biological Products

Great post by @Stephen . It is up to all of us Bayesians to respect experimental design principles and to allow the use of extra-study data only when justifications are extreme and unavoidable. But even with very rare diseases, patients are ill-served when concurrent controls are not part of the design. Non-study data may possibly be used to augment the concurrent controls, but seldom to replace them.