Abstract
Noninferiority (NI) clinical trials have been used to assess antibacterial drug efficacy in treating nosocomial and ventilator-associated pneumonia. Previously published trials have employed prespecified NI margins of 15% or 20% based on clinical response or microbiological endpoints. However, as those studies do not describe the statistical and clinical considerations underpinning the margins selected, their scientific plausibility cannot be substantiated. In this report, a fixed NI margin of 7% with respect to all-cause mortality is determined based on the 29% cross-study difference between the two-sided 95%, confidence intervals (CI) for the placebo estimate of 62% (95% CI: 52%, 71%) and the active control estimate of 20% (95% CI: 18%, 23%) obtained from meta-analyses of various published clinical studies. After applying discounting, we estimated the active control treatment benefit to be 14%. Due to clinical concerns, 50% of the active control treatment benefit was preserved, yielding a 7% NI margin.
Get full access to this article
View all access options for this article.
