PS15: SMART Designs for Evaluating Treatment Effects, Including Induction and Maintenance of Treatment Effects

Evaluating induction and maintenance effects in clinical trials often relies on traditional designs, such as the randomized withdrawal and treat-through approaches. While widely used, these designs have limitations, including operational complexity and challenges in obtaining robust long-term efficacy data. To address the limitations of traditional clinical trial methodologies, this session will introduce innovative advancements in two-stage designs, specifically focusing on the Sequential Parallel Comparison Design (SPCD) and Sequential, Multiple Assignment, Randomized Trials (SMART). A SPCD trial can reduce the duration patients spend on ineffective treatments, such as placebo or standard care, thereby minimizing the risk of extensive missing data and participant dropout. This design allows for an initial randomization between active treatment and placebo, followed by a second stage where placebo non-responders are re-randomized, enhancing the overall efficacy of the trial. A new version of hybrid SPCD with a binary endpoint is proposed for induction, maintenance and other treatment effect assessments. Furthermore, the implementation of SMART designs enables patients to be randomized or re-randomized across multiple treatment regimens. This flexibility not only improves trial efficiency but also allows for the assessment of stage-specific questions and adaptive interventions tailored to individual patient needs. The session will also cover various enrichment designs, discussing their advantages and disadvantages, as well as the application of analytical strategies and sample size considerations that take into account patient preferences and treatment exclusions. These enhancements aim to improve external validity, streamline trial recruitment, and better reflect real-world clinical practices, ultimately leading to more effective and patient-centered research outcomes. We will have three speakers from academia, industry, and regulatory agencies invited to share their research and experience on innovative SMART designs and advanced statistical methodologies . They will offer insights and practical examples to foster discussion and enhance understanding of how these approaches can be applied in real-world settings.