Multiplicity in Clinical Trials involving Multiple Treatment Groups

• Activity Number: 405
• Type: Contributed
• Date/Time: Thursday, August 12, 2004 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Kim Hung Lo*+ and Mani Lakshminarayanan
• Companies: Centocor, Inc. and Centocor, Inc.
• Keywords: multiple comparison ; Type I error rate ; dose response ; power ; Dunnett's Test ; pairwise comparison

Abstract

Clinical trials involving multiple treatment groups are very common in most of the late-stage drug development programs. For example, multiple treatment groups may be included in a Phase 2 trial designed to investigate the dose response of an experimental drug or in a Phase 3 trial with a primary objective of confirming one or more effective doses to be marketed. When there are multiple treatment groups, the question of multiplicity is of serious concern from the point of view of controlling the Type I error rate. Techniques that are typically used to control the Type I error rate includes an overall test (say, F-test in the case of linear models) followed by pairwise comparisons, Bonferroni adjustment, and closed test procedures. In Chapter 5 of "Multiple Comparisons: Theory and Methods," Jason Hsu provides arguments and other details of choosing one or more methods under certain conditions. One such method that is commonly used in practice involves performing an overall test comparing a combined group of dose levels of the test drug versus the control, followed by pairwise comparisons.