A trial compares a shorter antibiotic course with standard treatment after contaminated head-and-neck surgery. The risk-difference non-inferiority margin is 5%. The observed difference in infection risk is 2% higher with the short course, with a 95% confidence interval from −1% to +6%. What is the correct conclusion?
Even though the point estimate lies within the margin, the upper confidence limit permits excess harm greater than the prespecified threshold. Failure to prove non-inferiority is not proof that treatments are different.
A. The short course is non-inferior because 2% is below 5%: The full confidence interval, not only the point estimate, must lie within the margin.
B. The short course is superior because the interval includes −1%: The interval includes no difference and clinically important harm; superiority is not established.
C. The treatments are equivalent: Non-inferiority design and this interval do not establish two-sided equivalence.
D. Non-inferiority is not demonstrated because the confidence interval crosses the +5% margin: Even though the point estimate lies within the margin, the upper confidence limit permits excess harm greater than the prespecified threshold. Failure to prove non-inferiority is not proof that treatments are different.
E. The standard course is proven superior: Failure of non-inferiority does not automatically prove superiority of the control.
If the same point estimate had a 95% confidence interval from 0% to +4%, non-inferiority would be demonstrated, although superiority would not.
Independent Clinora educational material; not official JCIE content, an accredited programme, clinical advice, or a substitute for local policy and specialist judgement.
Back to the Section 1 bank