Evaluating equivalence and noninferiority trials

The conceptual issues related to the design of equivalence and non-inferiority trials and considerations for interpreting the findings of such trials are described.

Comparative effectiveness research (CER) has recently gained importance in the evaluation of different treatment alternatives. Large, prospective, randomized controlled trials (RCTs) conducted with patient populations under routine practice conditions can yield high-quality CER results. A Phase III RCT, usually conducted for establishing superiority of one treatment over another, is called a superiority trial, and the statistical test associated with it is known as a superiority test. In a pragmatic equivalence trial, a researcher aims to test if two treatments are identical (within a specified range) with respect to some predefined clinical criteria. Pragmatic noninferiority trials aim to show if a test therapy is no worse than a standard therapy with respect to achieving the primary treatment outcome. A nonsignificant result obtained from a superiority test does not indicate that the two treatment options are similar. In other words, the lack of evidence of superiority does not guarantee a lack of difference in the performance shown by the therapies. A researcher can only demonstrate identical effects of two treatments in an equivalence trial. In a noninferiority trial, the test therapy is preferred when there is evidence about its benefits over the standard treatment in terms of secondary outcomes such as cost, adherence, and adverse effects.

Equivalence and noninferiority trials are designed differently from superiority trials. The overall quality of equivalence and noninferiority studies depends on study design and the manner in which the results are reported.