Knowledge Papers


The classical approach to decision-making in clinical research is the randomized trial. Patients are assigned to treatment conditions according to sample sizes that are determined a priori by power calculations. The data are not analyzed until the end of the study, at which time assignment codes are typically broken. Most clinical trials employ the Intent to Treat strategy which involves keeping all patients enrolled in the trial to its end. Positive decisions as to drug efficacy generally cannot be made until the study has been completed. Due to the cost and time involved in recruiting and processing subjects, size determination is extremely important. Unfortunately, this relies on two parameters that are rarely available during the study of new drugs – within-subject variability and the estimated magnitude of experimental effect.