Active Control Trials with Adaptive Modification of Margin to Address Nonconstancy

• Activity Number: 452
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 8:30 AM to 10:50 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Scott Emerson*+ and Siobhan P. Brown
• Companies: University of Washington and University of Washington

Abstract

When a therapy reduces the rate of mortality or major morbidity for a given condition, it may be unethical to withhold that treatment from subjects. Investigational treatments then need to be tested against an active comparator using a "noninferiority margin" that defines the maximal difference between the new therapy and the active control that would be judged acceptable. Much non-inferiority methodology is based on the "constancy assumption" that presumes that the historical effect of the active comparator will be unchanged. If this assumption is violated, the burden of proof may shift to one of superiority for the new treatment. In this talk we present a method that adaptively modifies the non-inferiority margin based on the strength of evidence for homogeneity of the active comparator's effect across trials. We present results from simulation studies that examine settings in which a flexible margin approach can provide some protection against violations of constancy, though we find such improved behavior comes at the price of less precision when the constancy assumption is valid.