Optimal Use of the Non-Inferiority Trial Design

Abstract

Superiority trials are conducted to test the hypothesis that a treatment or strategy A is superior to (i.e., better than) treatment strategy B in reducing the impact of disease. However, A may be considerably safer, more convenient, or cheaper than B. These features may make A more attractive than B even if the burden of disease is reduced comparably by the two treatments, or even a bit worse by A over B. In this context, non-inferiority trials have become increasingly popular to test the hypothesis that a new treatment is not 'unacceptably worse' than an active comparator by more than a predefined non-inferiority margin. Non-inferiority trials have unique design features and methodology and require a different analysis than traditional superiority trials. The main aim of this overview is to analyze the role of non-inferiority trials in the development of new treatments, involving some scientific, statistical, and practical considerations. We elucidate some aspects of non-inferiority trials that contribute to the validity of the results. The unique design features and methodology of non-inferiority trials are summarized and key findings to consider when evaluating a non-inferiority trial are illustrated.