SC04: AI-Powered Adaptive Designs: Bringing Innovative Clinical Trial Designs into Practice

Clinical trials consume up to 70% of the time and cost in drug development. Traditional designs rely on lengthy, iterative processes, creating inefficiency and ethical concerns. Improved trial designs and covariate-adjusted analyses offer a solution by reducing variability and increasing efficiency. However, these designs and analyses are often under-utilized in practice. This short course introduces the latest advances in adaptive clinical trial designs, including covariate-adaptive randomization (CAR) and covariate-adjusted response-adaptive randomization (CARA), with a special focus on their implementation. It also introduces an associated covariate-adjusted analysis framework. This short course also demonstrates innovative AI tool kits that can streamline the implementation of novel designs. These tools enable covariate selection, evidence synthesis and large-scale simulation (AI agent hub). Emerging methods such as synthetic data generation (SynthIPD) further expand the design toolbox. Together, these approaches lower barriers to modern trial designs, making them both practical and powerful.