Permutation and Bootstrap Testing for the Sequential Parallel Comparison Design

Sequential parallel comparison design (SPCD) was proposed to reduce placebo response in randomized trials with placebo comparators. This design is achieved by randomizing subjects between placebo and drug in stage 1 and after the parallel first stage, participants who do not respond to placebo in stage 1 are randomly re-assigned to treatment or placebo in stage 2. The efficacy analysis of the study data includes all data from stage 1 and all placebo nonresponding participants from stage 2. This article examines the validity of permutation and bootstrap testing for SPCD. We examine two permutation tests and three bootstrap tests for SPCD and give recommendations on which test to use.