TL28: Super-superiority trial with the consideration of clinical significance
*Shiling Ruan, FDA/CDRH 


In the planning of a superiority trial, it is very important to calculate the sample size of the study to ensure adequate power with appropriate control of the Type I error to detect a statistical significant difference that is also clinically significant. However, clinical significance is often not clearly defined. Additionally, sometimes the observed treatment effect is statistically significant but the difference (and its confidence interval estimate) might not be clinically meaningful. In an effort to avoid different interpretations of the study results later in the review process, some clinical trials (i.e., some medical device trials) require super-superiority, that is, the lower bound of the confidence interval for the treatment effect has to be greater than a minimal clinically meaningful difference. To meet this requirement, the sample size of the clinical trial might have to increase substantially. Clinicians consider super-superiority trials as problem solvers. Interestingly, statisticians might have different view of its use. This roundtable will discuss the super-superiority trial, its implication and alternative approaches.