We appreciate Drs. Fisher's and Shafer's interest in our work.1We agree with their view that our model did not optimally fit all data points. There is a large variation in the placebo and 2.0 mg/kg group, and it was difficult to define a model that optimally fits these data points. For higher doses of sugammadex, the model fits the data very well.
We have conducted a Phase 2 clinical trial. Those studies attempt to learn what is a good (if not optimal) drug regimen to achieve useful clinical value (acceptable benefit/risk). In contrast to the confirming phases of drug development, the learning phases entail so-called explanatory analyses; i.e. , analyses that estimate the quantitative relationship between inputs and outcomes according to some mechanistic view of the relationship.2In a Phase 2 study, a nominal design, including all ostensibly controllable factors affecting the conduct of the trial, is an abstract ideal. In fact, in any real study, deviations from nominal design are inevitable.2We decided to apply the model to our data which has been defined a priori , and has been used for several data sets on sugammadex which already have been published.3–5We did not want to retrospectively change the predefined approach of our statistical efficacy analysis. In future confirmatory studies on sugammadex it will be possible to develop and apply a more sophisticated model. The suggestions of Drs. Fisher and Shafer will be very useful in that context.
*Medical University Innsbruck, Innsbruck, Austria. email@example.com