Issues in Assessing Noninferiority: Proper Discount and Margin Selection

• Activity Number: 100
• Type: Invited
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Ping Gao*+
• Companies: The Medicines Company
• Keywords: Non-inferiority ; discounting ; discount synthesis ; non-inferiority margin ; regulatory margin

Abstract

Frequently used methods in assessing non-inferiority make the assumption of assay constancy (i.e. the concurrent efficacy is the same as the historical efficacy). However, the validity of this assumption is often in question, due to the possible impact of the changes in medical practices (newer drugs, change in patient care). Consequently, the historical efficacy is almost always discounted. Often, the "discounts" are arbitrary, which may not provide adequate type I error control. We propose that it is necessary to investigate the impact of the changes in medical practice, and make "proper" discounts. Such an approach can provide more reliable control of type I error. Another issue is the interpretation of the non-inferiority margin. Historically, "non-inferiority" margins have been calculated from historical data, often with confusions in its interpretation. We propose that two margins, a "non-inferiority" margin and a "regulatory" margin,that will provide clarity in the interpretation of the results of "non-inferiority" trials.