ABSTRACT
This case study illustrates how a model may be chosen for a dose response relationship when there is uncertainty in the shape of the dose-response curve. The example is based on a real trial where dose-finding was used, but for confidentiality reasons we have changed some aspects and will use simulated data. However, the main features of the design and the analysis have been retained. The drug under study was for the treatment of COPD (chronic obstructive pulmonary disease) and the primary endpoint in the trial was trough forced expiratory volume over 1 second (TFEV1) following seven days of treatment. The endpoint was recorded in liters just prior to 24 hours post-dose. We will refer to the investigational drug as A. Also included in the trial was an active comparator, which we will refer to as B. The main purpose of the study was to evaluate the efficacy of five doses of A (0, 12.5, 25, 50 and 100 μg once a day) and open-label B (18 μg), so that an optimal dose of A could be chosen for future Phase III studies.
