An Adaptive Noninferiority Study

• Activity Number: 221
• Type: Topic Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Pilar Lim*+ and Qing Liu and Rosanne Lane and Barry Schwab and Allan Sampson
• Companies: Johnson & Johnson and Johnson & Johnson and Johnson & Johnson and Johnson & Johnson and University of Pittsburgh
• Keywords: non-inferiority ; adaptive design ; sample size re-estimation ; interim analysis

Abstract

This talk will discuss a recent and successfully completed non-inferiority (NI) study with an adaptive design for sample size re-estimation. Previous placebo-controlled trials with the compound of interest indicated the possibility of varying treatment effects over a baseline ordered categorical variable. Minor modifications to the dosing regimen were implemented to circumvent this difficulty. Nonetheless, for this NI study, an adaptive design was implemented to protect against some degree of residual treatment heterogeneity. At the interim analysis, a sophisticated algorithm was implemented to detect a treatment-by-baseline categorical variable interaction and adjust the remaining sample size to assure that the power was maintained for the prespecified NI margin. The new methodology which was required to develop and implement this design will be presented.