Non-inferiority clinical trials: What they are and when they should be done.

Abstract

Randomized controlled trials (RCTs) are the gold standard design of studies evaluating novel interventions, pharmaceutical and otherwise. The standard RCT design is a 'superiority' study to assess if the tested intervention is better than existing treatments. However, a new intervention that has additional benefits, such as better adverse effect profile, may be useful even if it is not more effective than existing treatments. In such cases, a non-inferiority (NI) design may be preferable. In NI designs, once NI is demonstrated, sequential analysis for superiority may be done, provided it is pre-specified. This article discusses the utility and principles of an NI design. Examples to illustrate the design are taken from published trials of anticoagulation in atrial fibrillation.