Noninferiority Trials: Is a New Treatment Almost as Effective as Another?
This JAMA Guide to Statistics and Methods describes the reasons for conducting a noninferiority trial and how to analyze and interpret the results from a trial that did.
Sometimes the goal of comparing a new treatment with a standard treatment is not to find an approach that is more effective but to find a therapy that has other advantages, such as lower cost, fewer adverse effects, or greater convenience with at least similar efficacy to the standard treatment. With other advantages, a treatment that is almost as effective as a standard treatment might be preferred in practice or for some patients. The purpose of a noninferiority trial is to rigorously evaluate a new treatment against an accepted and effective treatment with the goal of demonstrating that it is at least almost as good (ie, not inferior).
In an example published in JAMA, Salminen et al described the results of a multicenter noninferiority trial of 530 adults with computed tomography–confirmed acute appendicitis who were randomized either to early appendectomy (the standard treatment) or to antibiotic therapy alone (a potentially less burdensome experimental treatment).1
Why Are Noninferiority Trials Conducted?
In a traditional clinical trial, a new treatment is compared with a standard treatment or placebo with the goal of demonstrating that the new treatment has greater efficacy. The null hypothesis for such a trial is that the 2 treatments have the same effect. Rejection of this hypothesis, implying that the effects are different, is signaled by a statistically significant P value or, alternatively, by a 2-tailed confidence interval that excludes no effect. While the new treatment could be either superior or inferior, the typical trial aims to demonstrate superiority of the new treatment and is known as a superiority trial. Since a superiority trial is capable of identifying both harmful and beneficial effects of a new therapy vs a control (ie, a current therapy), a 2-tailed 95% CI can be used to indicate the upper and lower limits of the difference in treatment effect that are consistent with the observed data. The null hypothesis is rejected, indicating that the new therapy differs from the control, if the confidence interval does not include the result that indicates absence of effect (eg, a risk ratio of 1 or a risk difference of 0).2 This is equivalent to a statistically significant P value.
Although superiority or inferiority of a new treatment can be demonstrated by a superiority trial, it would generally be incorrect to conclude that the absence of a significant difference in a superiority trial demonstrates that the therapies have similar effects; absence of evidence of a difference is not reliable evidence that there is no difference. An active-controlled noninferiority trial is ...