Missing Data in NonInferiority Trials

• Activity Number: 100
• Type: Invited
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Brian Wiens*+ and Gerd K. Rosenkranz
• Companies: Alcon Laboratories and Novartis Pharma AG
• Keywords: clinical trial ; equivalence ; per protocol ; intent to treat ; MMRM

Abstract

Issues pertaining to missing data in clinical trials have been widely discussed in the literature, but issues specific to missing data in non-inferiority trials have rarely been addressed. The goal of this paper is to present discuss analyses of non-inferiority trials when missing data are present, specifically with a mixed model repeated measures analysis. Simulation results show that with a missing at random mechanism and a moderate amount of missing data, MMRM analyses are robust to missing data. Some efficiency is lost due to a loss of information, but the type I error rate is maintained. The same is not true of analyses using observed case data or last observation carried forward. Collecting full data on all subjects, even those who discontinue treatment; using all data collected; and analyzing as randomized are more difficult to understand in non-inferiority trials than in superiority trials. We advocate collecting data to the extent possible for sensitivity analyses.