A Novel Group Sequential Design for Sequential Multiple Assignment Randomized Trial

ABSTRACT The sequential multiple assignment randomized trial (SMART) design offers an innovative approach for efficiently selecting optimal adaptive treatment strategies (ATSs) from a set of candidate treatment strategies. By incorporating interim monitoring, the group sequential SMART design reduces both the trial duration and sample size while maintaining overall efficiency. However, existing SMART designs that incorporate interim analysis rely on global tests, which are inadequate for addressing the challenges of selecting the optimal subset of ATSs and performing early termination of ineffective strategies. In this study, we propose a novel group‐sequential SMART design to improve the efficiency of clinical trials in identifying the most optimal subset of ATSs in a way that aligns with the nuances of clinical practice. This design allows for the early termination of less efficacious treatments, reducing the required sample size and increasing the likelihood of accurately identifying the optimal ATSs. The effectiveness of the proposed approach is demonstrated through simulation results.