A trial finds a new surgical device has a complication rate 1.5% higher than standard treatment; its prespecified non-inferiority margin was 3%, but the upper confidence-limit is 3.4%. What is the correct conclusion?
The confidence interval must remain within the predeclared margin. A point estimate below the margin does not rescue a result whose plausible harm exceeds it.
A. The device is non-inferior because 1.5% is below 3%: This ignores uncertainty represented by the confidence interval.
B. The device is superior because the point estimate is small: The estimate favours neither superiority nor established non-inferiority.
C. The margin can be widened after results are known: Post-hoc changes undermine the trial’s design and interpretation.
D. Non-inferiority has not been demonstrated: The confidence interval must remain within the predeclared margin. A point estimate below the margin does not rescue a result whose plausible harm exceeds it.
E. Confidence intervals are irrelevant when p is reported: The interval is essential to judge the pre-specified clinical margin.
If the upper confidence-limit were 2.7%, the result would meet the stated non-inferiority criterion.
Draft Clinora educational material only; not official JCIE content or clinical advice.
Back to the Section 1 bank