In reply
We thank Dr Faasse for his thoughtful comments. We agree that it would have been clearer to use the term “inconclusive” instead of the one we used.
As far as we are aware, there are no validated rules to determine noninferiority margins. Nor can there be, since almost every clinical situation has its own peculiarities. What constitutes a clinically acceptable difference is ultimately a matter of judgment and may vary across patients, physicians, investigators, regulators, or payers.1 It is striking but perhaps not surprising that in trial registrations of noninferiority trials, the noninferiority margins are very often absent.2 A rational approach to planning noninferiority trials would involve performing a cost-utility analysis to derive the noninferiority margins before embarking on the trial. Such analyses could incorporate the different opinions of the various stakeholders.