prospective study of charts and how they predict phase III outcomes Probably have fairly similar results as doing an analysis of using Phase 2 retrospective data to predict Phase 3 success. BTW, do you think doctors in the trial, etc., don't ever talk?