Improving the reporting of non-inferiority trials by incorporating non-efficacy benefits: not all non-inferiority trials are created equal

Non-inferiority trials are used to test if a novel intervention is not worse than a standard treatment by more than a prespecified amount, the non-inferiority margin (ΔNI). The ΔNI indicates the amount of efficacy loss in the primary outcome that is acceptable in exchange for non-efficacy benefits in other outcomes. However, non-inferiority designs are sometimes used when non-efficacy benefits are absent. Without non-efficacy benefits, loss in efficacy cannot be easily justified. Further, non-efficacy benefits are scarcely defined or considered by trialists when determining the magnitude of and providing justification for the non-inferiority margin. This is problematic as the importance of a treatment's non-efficacy benefits are critical to understanding the results of a non-inferiority study. Here we propose the routine reporting in non-inferiority trial protocols and publications of non-efficacy benefits of the novel intervention along with the reporting of non-inferiority margins and their justification. The justification should include the specific trade-off between the accepted loss in efficacy (ΔNI) and the non-efficacy benefits of the novel treatment and should describe whether patients and other relevant stakeholders were involved in the definition of the ΔNI.