Clinical research in rare diseases is difficult for a number of reasons, including limited number of patients, and patient attitudes toward active and placebo therapy. Oftentimes, a number of available drugs are used for a rare disease with limited information about the efficacy of any of the drugs. In such situations, it is logical to first design a trial which would determine if any of the drugs shows promising efficacy. In cooperation with the Vasculitis Clinical Research Consortium under the NIH, we have recently developed a sequential multiple assignment randomized trial (SMART) for rare vasculitis diseases. The goal of the design is to determine the best out of a number of potential drugs and to compare the chosen drug to the best of the remaining drugs. In this talk, I will present the design, analysis, and operating characteristics of the SMART for three drugs when the number of subjects is limited.