Statistical and Clinical Characteristics of Composite Endpoint in Clinical Studies

• Activity Number: 584
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Jeng Mah*+
• Companies: Regulatory and Clinical Research Institute, Inc.
• Keywords: composite endpoint ; parameterization ; estimation ; hypothesis testing ; prediction

Abstract

A composite endpoint is a group of measurements which together provide a single-dimensioned evaluation of a specific clinical outcome. Because of their flexibility and convenience, composite endpoints have become widely used in clinical studies, especially for medical devices. A composite endpoint typically includes multiple standard measurements and a decision rule to combine these into a final binary result; however, the implications and consequences of this simplification are sometimes not carefully considered before directly applying standard statistical analyses and hypothesis testing.

This paper examines the statistical properties of a typical composite endpoint, such as distribution, parameterization, and estimation, and their impacts on clinical interrelation of a composite endpoint. Simulated data are used to verify and illustrate derived results.