|
An equivalence test is performed by setting equivalence margins and verifying if a confidence interval for the parameter is completely contained within those margins. The margins are usually symmetric either in the original scale or in the log scale implying that equal importance is given for non-inferiority and non-superiority. However, there are situations where one of non-inferiority and non-superiority is more important than the other. When symmetric margins are used, the sample size required to achieve a target power with a given level of significance may become prohibitively large. Use of asymmetric margins, depending on which side is more important, may alleviate this problem. This presentation will discuss construction of such asymmetric margins which can be applied to situations where a very large sample size is infeasible.
This presentation reflects the views of the author and should not be construed to represent FDA's views or policies.
|
