POWER ANALYSES OF CUTOFF-BASED RANDOMIZED CLINICAL-TRIALS

A cutoff-based randomized clinical trial couples cutoff-based assignme nt on an appropriate covariate with random assignment to help balance ethical and scientific concerns in certain situations. A statistical p ower algorithm based on the Fisher Z method is developed that is parti cular to and inclusive of cutoff-based random clinical trials and the single cutoff-point (regression-discontinuity) design, which has no ra ndomization. This article quantifies power and sample size estimates f or varying levels of randomization and cutoff-based assignment. Althou gh more randomization engenders greater statistical power less randomi zation requires a much larger increase in sample size for small treatm ent effects.