A sequential procedure for a phase II efficacy trial in hypercholesterolemia

The design of a phase II placebo-controlled efficacy trial to evaluate a new treatment for hypercholesterolemia using a sequential procedure is presented. The sequential procedure used is the triangular test which offers flexibility of monitoring and allows early stopping even when the null hypothesis is "accepted." The primary response is reduction in total serum cholesterol level, which is assumed to be normally distributed with unknown variance. The treatment difference is parameterized in terms of the standardized difference between the means. The accuracy of the error rates of the triangular test using this parameterization is compared with the test based on the difference between the means.