A Note on Hypothesis Test in Binary Data under the Single-Consent Randomized Design

Abstract

When we compare an experimental treatment with the best standard therapy, Zelen first proposed use of the single-consent randomized design to increase the accrual rate of patients into trials. Under the single-consent randomized design, I developed two other simple test procedures to improve the power of the commonly used procedure for the intention-to-treat analysis in testing whether there is a difference in treatment effects in this article. Based on Monte Carlo simulation, I evaluated and compared these three test procedures with respect to type 1 error and power. I demonstrated that the two test procedures developed here outperformed the commonly employed test procedure in a variety of situations. Finally, I employed an example by studying the effect of vitamin A supplementation on reducing mortality in preschool children to illustrate the use of these test procedures.