Recently, it was argued that clinically important differences should play no role in sample size calculations. Instead, it was proposed that sample size calculations should focus on setting realistic estimates of treatment benefit. We disagree, and argue in this article that considering the importance of a target difference is necessary in the context of randomised controlled trials of effectiveness, particularly definitive phase III trials. Ignoring clinical importance could have serious ethical and practical consequences.
Assumed benefit, Clinical trial, Clinically relevant difference, Effect size, Minimum important difference, Power, RCT, Target difference, Humans, Clinical Relevance, Sample Size, Randomized Controlled Trials as Topic