Inference and sample size calculation for clinical trials with incomplete observations of paired binary outcomes

摘要: We investigate the estimation of intervention effect and sample size determination for experiments where subjects are supposed to contribute paired binary outcomes with some incomplete observations. propose a hybrid estimator appropriately account mixed nature observed data: from those who complete pairs observations unpaired either pre-intervention or post-intervention outcomes. theoretically prove that if data evenly distributed between periods, proposed will always be more efficient than traditional estimator. A numerical research shows when distribution is unbalanced, superior there moderate-to-strong positive within-subject correlation. further derive closed-form formula help researchers determine how many need enrolled in such studies. Simulation results suggest calculated maintains empirical power type I error under various design configurations. demonstrate method using real application example. Copyright © 2016 John Wiley & Sons, Ltd.