Boundaries for Harm in Clinical Trials Designed to Establish Safety

• Activity Number: 503
• Type: Invited
• Date/Time: Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
• Sponsor: ENAR
• Author(s): Alex Fleishman*+ and Barry Davis*+ and Susan Ellenberg*+ and Gordon Lan*+ and Michael Proschan*+ and Janet Wittes*+
• Companies: Amgen, Inc. and The University of Texas School of Public Health and University of Pennsylvania and Johnson & Johnson PRD and National Institute of Allergy and Infectious Diseases and Statistics Collaborative, Inc.
• Keywords: Interim monitoring ; Non-inferiority studies

Abstract

When the goal of a clinical trial is to establish safety, the drug under study should be no worse than an alternative treatment. These trials can involve large numbers of patients when "no worse" is defined by a narrow non-inferiority margin which is often the case when the safety endpoint is mortality. Evidence of the drug being harmful, however, could be detected in much fewer patients. Therefore, another goal of the trial should be to stop promptly if harm has been determined. We present ways of pre specifying stopping boundaries and their operating characteristics. Following this presentation, panelists will comment on the proposal and how they---as experienced statisticians working on or for DSMBs---would view the proposal. We expect that this discussion will provide insight into the best practices for a problem that has yet to find a well established solution.