Noninferiority and Superiority Tests for Multiple Immunogenicity Endpoints in Vaccine Clinical Studies

• Activity Number: 183
• Type: Contributed
• Date/Time: Monday, August 1, 2011 : 10:30 AM to 12:20 PM
• Sponsor: International Chinese Statistical Association
• Author(s): Lihan Yan*+
• Companies: U.S. Food and Drug Administration
• Keywords: non-inferiority ; superiority ; multiple endpoints ; vaccine ; type I error

Abstract

Non-inferiority and superiority of a new vaccine to an existing one in terms of immunogenicity in a vaccine study often involves multiple endpoints corresponding to multiple strains or serotypes. The statistical hypothesis is often constructed as an intersection-union test among multiple endpoints, which allows the hypothesis test to be done for each endpoint at the desired overall Type I error rate with the overall rate being still conservatively preserved. Following the closed testing principle, the Type I error for the superiority test performed subsequently to the non-inferiority test is also controlled without further adjustment. It is not uncommon that the investigators weigh in on success of a trial by individual hypothesis test results and make a declaration even though the success criterion constrained by the hypothesis is not met. This presentation explores the impacts on the Type I error rate control from a variety of scenarios where the claims are made deviating from the decision rule indicated by the hypothesis testing. These results may serve as a potential reference for the investigators as to cautions as well as levels of comfort during decision making.