|
Abstract:
|
In clinical trials, endpoints are classified as primary, secondary, etc. To control trial-wise, type-I error, the Hochberg procedure is commonly used for handling multiple testing among primary endpoints. For secondary endpoints which can be used for regulatory claims but are potentially under-powered, the PAAS has been proposed. For secondary endpoints, which play only a supportive role, gate-keeping procedures have been developed. While these procedures are proposed for general use, they do not incorporate a unique characteristic of late-phase clinical trials: The primary endpoints are powered by design. This discussion will focus on how to utilize this important property in analysis of multiple endpoints. Recent developments in statistical methodology and their impacts on clinical trials such as CAPRICORN trial, ATLAS study, DEFINITE trial, and PROactive study, will be included.
|
