The determination of practically useful doses of new drugs: some methodological considerations.

An important aim of Phase 2 clinical trials is the firm determination of the doses and the dose regimes to be used both in later clinical trials and in clinical practice. Comparisons of findings made during clinical development and recommendations contained in the package leaflets of marketed preparations indicate that this objective is rarely achieved. The possible reasons for this difficulty are investigated by reviewing the methodology used in the so-called dose-finding trials performed during drug development. Since the terminology used in this context is frequently misleading, definitions are given and alternative approaches, referring to the direct and indirect methods of bioassay, are suggested. The practical use of these models depends on a number of medical and ethical considerations, on the indications for treatment and on the characteristics of the drug under investigation. Some examples are given, indicating that dose-response relationships can be determined more easily for efficacy, where endpoints more reliably fit commoner distributions, than for tolerability, where the responses frequently follow irregular patterns. This area deserves further investigation to improve the criteria for assessment and to develop adequate statistical methods.