Incorporating Continuity Corrections When Illustrating Noninferiority with Binomial Proportions

• Activity Number: 522
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Brianna Bright*+ and Julia Soulakova
• Companies: University of Nebraska and University of Nebraska
• Keywords: asymmetric comparisons ; asymptotic tests ; binary outcomes

Abstract

The aim of a non-inferiority clinical trial is to show that an experimental treatment is no less effective than a standard treatment. Non-inferiority is typically evaluated by determining if a 1-sided 97.5% confidence interval around the mean difference due to experimental and standard treatments does not include a pre-specified non-inferiority margin. We consider the problem of illustrating non-inferiority for studies with binary outcomes. While many methods have been proposed in the literature regarding the calculation of these confidence intervals, the issue of determining an optimal continuity correction for asymmetric comparisons, i.e., where one group is treated as an experimental one and the other group is treated as a standard one, has not been resolved. We evaluate the performance of methods described by Anbar (1983) and Falk & Koch (1998) along with different continuity correction terms to a binomial distribution. These methods are compared in terms of power and the coverage probability for multiple parameter settings.