Non-inferiority trials

Non-inferiority trials In many diseases, effective treatments are already available. Thus, expected benefits with new treatments in development are often small. In this context, the statistical demonstration of superiority of these new treatments is often difficult even impossible to be proven. Consequently, many current trials aim to show that a new treatment is safer and/or easy to use without loss of efficacy, rather than to demonstrate superiority. Such trials, called "non-inferiority" trials, need be distinguished from superiority trials regarding both methodological and statistical aspects. In this note, we examine the statistical rationale and methodological features of non-inferiority trials. We present analysis methods using confidence intervals approach and key aspects of their design. Finally, we consider the possibility to switch from non-inferiority to superiority testing.