Sample size

Clinical studies, whether involving a retrospective single-center study or a prospective multicenter randomized trial, require significant resources—including time, energy, and financial support—to perform. Therefore, it is essential that a determination of feasibility is made before investing these resources in the study. A critical component of this feasibility ascertainment is determining the sample size that will be required to observe a difference between the control and experimental group. For prospective studies, this calculation approximates the size of the study, allowing an estimation of costs and time commitments. For retrospective studies, it can help determine whether or not there will be sufficient patients or observations to make the process of data collection and statistical analysis worthwhile. When interpreting studies designed to identify a difference between two groups, in which none is ultimately found, it is pertinent to examine whether the study was adequately powered to justify the conclusions. In this chapter, we examine the key components of developing an appropriate sample size calculation; these include the type of study, variance in the outcome, the significance level, the power of the statistical testing, and the minimal clinically relevant difference.