TL27: Multiplicity Adjustments for Testing Endpoints in Clinical Trials - A Check-Up on Current Practice
*Davis F Gates, Merck 

Keywords: multiplicity

In a clinical trial with multiple efficacy endpoints of interest to support claims, regulatory guidance has been very clear on the need to adjust for multiple testing of hypotheses. However, there is still room for further discussion of the methods used to provide statistical control. Even though this issue has been recently discussed at several conferences, it would be of interest to hear from industry and FDA statisticians regarding the progression of implementing multiplicity adjustments in clinical trials. An overview of the impact multiplicity has had on successful statistical testing across previous clinical trials is desired, especially in cases where post-hoc examinations indicate successful testing of endpoints are a function of applied methods of control. QUESTIONS: 1) In a typical late phase clinical trial, how many endpoints require some type of statistical control of the overall alpha level? (Discuss the break-down of co-primary and controlled secondary endpoints); 2) what methods of multiplicity controls are being implemented? (Examples include step-wise, splitting the alpha, Hochberg tests); 3) what have been the pros and cons of each multiplicity control approach? (Consider the interaction between sponsors and health authorities regarding approaches)