ABSTRACT
One of the recurring challenges in clinical study design is proper sample size estimation. Often, the information necessary for determining the sample size is not available or poorly reliable. For example, the variance for an outcome endpoint, such as ACR20 reduction, may not be well described for the population of interest. The information may not be reliable because there is an insufcient number of studies, the studies were different in some way from the new study, or the execution of the studies were poor.
