noninferiority

Noninferiority is a trial design aimed at showing that a new treatment is not unacceptably worse than an active comparator by a prespecified noninferiority margin. It is used when placebo-controlled trials are unethical or when the new therapy offers advantages such as better safety, tolerability, convenience, or lower cost that may offset a small loss of efficacy.

A noninferiority trial requires an active comparator with proven efficacy and a predefined margin that defines the largest clinically acceptable loss of efficacy. The null hypothesis states that the new treatment is worse than the control by more than the margin; the alternative claims it is not.

Analysis relies on confidence intervals. The population choice matters: intention-to-treat data can bias toward noninferiority if there is nonadherence or missing data, while per-protocol analyses are often considered in parallel. Noninferiority is claimed if the confidence interval for the treatment difference remains above the negative margin on the chosen scale.

Margin selection uses historical data, clinical judgment, and the goal of preserving some portion of the active treatment’s effect. Regulators expect prespecified margins and sensitivity analyses, and may require demonstration of assay sensitivity and, in some cases, a route to superiority.

Limitations include the risk of an inappropriately large margin or misinterpretation as showing equivalence. Noninferiority trials require careful design, transparent reporting, and clear labeling. They are common in infectious diseases and oncology, where safety or convenience trade-offs can justify accepting modest efficacy losses.