Power

To detect a statistical difference in a clinical study, it is imperative to design the trial with the appropriate sample size which depends on properly powering the study. With too few subjects, one risks being unable to delineate if there is a true difference between two study populations. Conversely, if the sample is too large, then one risks inappropriately utilizing resources toward detecting minuscule and likely clinically irrelevant differences. Ensuring that one has the appropriate number of subjects to detect a difference or effect, should one exist, is the definition of power. Power determines if a study hypothesis suggesting a true difference exists can be appropriately accepted, and whether the null hypothesis suggesting no difference exists between study groups in the underlying population can be rejected. It determines the sample size that is needed for prospective studies inclusive of nonrandomized and randomized control trials.