Statistical Considerations in Assessing Assay Sensitivity in a Noninferiority Trial

• Activity Number: 131
• Type: Contributed
• Date/Time: Monday, August 1, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Isaac Nuamah*+
• Companies: Johnson & Johnson Pharmaceutical R&D, LLC
• Keywords: non-inferiority ; assay sensitivity ; gold standard design

Abstract

In clinical trials where non-inferiority of a new experimental drug to an active control has to be shown, it has been suggested to include a placebo arm (if ethically justifiable) in order to have an internal evidence to demonstrate assay sensitivity. In such a three-arm trial (sometimes referred to as 'gold standard design'), testing for non-inferiority follows a two-step hierarchical process. In the first step, superiority of either the experimental or active control to placebo is performed, and in the second step, non-inferiority is assessed. For an active control non-inferiority study (without a concurrent placebo arm), assay sensitivity cannot be directly assessed in that trial and has to depend on appropriate trial conduct and historical evidence of sensitivity of active drug effect. In this talk, we discuss the statistical issues needed to demonstrate assay sensitivity when (1) there is a placebo arm in that trial, the so-called 'gold standard design' and (2) when the only comparator is the active control arm. Data from a psychiatry clinical trial are used to illustrate these issues.