序贯多分配随机试验设计样本量估算方法及应用OACSTPCD

The Sequential Multi-Assignment Randomized Trial(SMART)design incorporates sequential and dynamic treatment strategies in clinical research.Determining an appropriate sample size poses a significant challenge in SMART studies.Sample size estimation methods for SMART design vary according to four primary research objective.Pilot studies,conducted as preparatory stages,primarily focus on precision and feasibility in sample size estimation.Estimating the main effects of first-line and second-line treatments in SMART design aligns with the commonly used sample size estimation methods in randomized controlled trials.Comparative analyses of embedded adaptive treatment measures encompass optimal adaptive treatment selection,paired dominance testing,paired non-inferiority testing,and multiple comparisons with the best-performing interventions.Sample size estimation for SMART designs aimed at developmental optimization primarily employs methods based on the normal distribution.This study aims to provide practical guidance and examples for sample size estimation in SMART designs,thereby advancing research in SMART design within the field of TCM.