ABSTRACT
Dose-finding is a trade-off between efficacy and safety. The ultimate goal is to determine the best dose for each patient. However, this goal will never be fully reached. A series of experiments are needed to, hopefully, find a dose that is beneficial to a population of patients. Dose-finding should therefore be seen as an extended scientific process, not as a single clinical trial. Despite this, we will focus on dose-finding in late clinical development, mainly the dose-finding trial in phase IIB but also on the possibility of including two (say) candidate doses in phase III. The design of this late stage clinical program should build firmly on the dose-response information from earlier stages, clinical and preclinical. A key to good dose decisions and to a good design of the development program is to have well thought-through and clearly defined objectives. The optimal design theory can readily be applied to suggest study designs. The answer to what the best design is, however, depends on the study objective, the previously available information, and the study constraints.
