Abstract:
|
Non-inferiority trials are a type of active-control trial conducted for the goal of showing that a new or experimental treatment is not worse (or non-inferior) than the control, as opposed to a superiority trial which intends to show that one drug is better than another. The trial is performed to show that there may exist some small clinically acceptable difference on the effect of the outcome between the new or experimental therapy and the active control (also denoted as standard of care) (Snapinn, 2000; US Food and Drug Administration, 2010). As with any other clinical trial, the measure of the effect on a dichotomous outcome can be presented in various ways (e.g. odds ratio, relative risk, or difference in proportions). In this presentation, we evaluate the relative efficiency of different non-inferiority designs by comparing and contrasting required sample sizes to achieve pre-specified power for a common Type I error and non-inferiority margin while varying the measure of effect chosen to analyze the results of this type of trial.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.