Contents
How do you determine non-inferiority?
In order to demonstrate non-inferiority, the recommended approach is to pre-specify a margin of non- inferiority in the protocol. After study completion, a two-sided 95% confidence interval (or one-sided 97.5% interval) for the true difference between the two agents will be constructed.
What does a non-inferiority trial mean?
in-FEER-ee-OR-ih-tee TRY-ul) A study that tests whether a new treatment is not worse than an active treatment it is being compared to. Non-inferiority trials are sometimes done when a placebo (an inactive treatment) cannot be used.
What does it mean to be non inferior?
A study that tests whether a new treatment is not worse than an active treatment it is being compared to. Non-inferiority trials are sometimes done when a placebo (an inactive treatment) cannot be used.
How to determine non-inferiority margins with Type 1?
An alternative method is to use the point estimate, instead of the lower confidence limit, of the active-control effect. The fixed-margin approach based on the lower limit may be ultra-conservative with unconditional Type 1 error rate much smaller than target α/2level, while the margin based on the point estimate is liberal.
What is the confidence interval around the observed difference?
Instead, we calculate a confidence interval (CI) around the observed difference, which allows a degree of uncertainty associated with the observed value of 10 mmHg, and it is within this CI that the true difference will likely lie.
How is the noninferiority of a trial determined?
The noninferiority margin is determined by halving the control effect, based on a historical placebo-controlled trial. The upper limit of the 95% confidence interval (CI) for the treatment difference by the new treatment compared to the control is less than the margin to conclude noninferiority, although there is an implication of incomparability.
When to use upper confidence limit for Ni?
If the upper confidence limit μˆC−E+z1−α/2σˆC−E<Δλ, the null hypothesis is rejected and the NI conclusion is reached. Because the upper limit of a two-sided (1−α)CI is used, the Type 1 error rate is α/2.