Clinical trials designed to evaluate therapeutic preferences.

We present an experimental design for the evaluation of therapeutic preferences among well established similar modes of therapy. The measure of the degree of preference is based on clinical decisions to continue or discontinue the current treatment at each patient visit. Such a trial simulates decision-making in ordinary clinical practice, while adhering to the scientific principles of experimental design, and it alleviates some of the ethical problems inherent in randomized allocation of a treatment for a period of fixed length in conventional control trials. We discuss the design and methods of analysis and illustrate their application with data from a trial comparing three alternative drug treatments for chronic asthma. Such clinical trials provide a mechanism for examining concordance between the relative efficacy of treatments predicted from conventional controlled clinical trials and their ultimate performance in ordinary clinical practice, expressed in terms of therapeutic preferences.