A trial claims a new voice-restoration device is non-inferior, but the prespecified margin would allow a clinically important loss of intelligibility. What is the main concern?
A statistically successful result is not clinically reassuring if the margin accepts unacceptable harm. The margin should be justified from prior evidence and patient-important outcomes.
A. Non-inferiority trials never require a comparator: They require an appropriate active comparator and rigorous conduct.
B. The non-inferiority margin is too permissive to preserve meaningful benefit: A statistically successful result is not clinically reassuring if the margin accepts unacceptable harm. The margin should be justified from prior evidence and patient-important outcomes.
C. A larger sample automatically fixes the margin: Precision cannot make an unjustified clinical threshold acceptable.
D. Only p-values determine clinical equivalence: Effect estimates, confidence intervals and the margin matter.
E. Intention-to-treat analysis alone is always conservative: In non-inferiority trials, both intention-to-treat and per-protocol analyses are informative.
If the confidence interval crossed the justified margin, non-inferiority would not be demonstrated.
Independent Clinora educational preparation material; not official JCIE content, an accredited programme, clinical advice, or a substitute for current guidance, local policy and specialist judgement.
Back to the Section 1 bank