An Editorial on Registration of clinical trials
Most clinical trials are undertaken in order to determine that one treatment produces a better result for the patient than another treatment or no treatment at all. Very early in the design of a study, before any patients are involved, the investigators must decide how many participants are needed to provide a reasonable possibility that they will detect a difference between the two interventions. They have to agree why Treatment A should be better than Treatment B or no treatment, how they are going to accurately and reproducibly measure it and how big a difference between the measurements is reasonable to expect if Treatment A was to be considered better than Treatment B