ABSTRACT
Because of ethical and economical reasons, clinical trials are often designed as sequential experiments, as discussed in Chapter 2. A multistage sequential design can be described as follows: At each stage of the trial, a predetermined number of patients are treated. Then the accumulated number of treatment responses is compared to the stopping boundaries. If the accumulated number of responders is smaller than or equal to the lower boundary, the trial is terminated for lack of treatment efficacy. If the accumulated number of responders is larger than or equal to the upper boundary, the trial may be terminated for high treatment efficacy. Obviously, early termination for high efficacy in this setting is not as ethically imperative. Otherwise, an additional fixed number of patients will be treated at the next stage. This will continue until early termination or until the predetermined number of stages. Multistage designs have been described and investigated by Schultz, Nichol, Elfring, and Weed (1973), Herson (1979), Fleming (1982), and Chang, Therneau, Wieand, and Cha (1987), among others.
