➢ With any study, readers should be cautious and critical when the conclusion is that “these treatments are the same.” ➢ If only superiority testing was performed, failing to find a difference does not mean that the treatments are the same, even when the study was adequately powered. ➢ Noninferiority analysis is the correct method to compare treatments that researchers and clinicians think may be “the same” for the primary outcome. ➢ The most important aspect of a noninferiority analysis is the selection of the noninferiority margin, which is the minimum difference between groups that would be considered meaningful. ➢ To perform noninferiority testing, the difference in an outcome measure of interest between experimental and control groups must be examined with respect to the noninferiority margin of the same outcome measure. Assuming that a greater value indicates improvement in an outcome measure, if the lower bound of a 95% confidence interval of a difference in means based on a 1-sided test is greater than the noninferiority margin, then the experimental treatment can be considered noninferior to the control.
Burks et al. (Tue,) studied this question.