SESSION G

Clinical trials are intricate endeavors with multiple objectives, including controlling the type I error rate, enhancing the power to detect treatment differences, allocating more patients to superior treatments, achieving balance in covariates across treatments, and more. To fulfill these objectives, the literature proposes two distinct families of techniques: (i) the analysis approach—where observed data is analyzed sequentially, and (ii) the design approach—where the allocation probability is altered sequentially. Sequential monitoring has administrative, ethical, and economic advantages. Response adaptive randomization (RAR) can also achieve ethical and efficient objectives by skewing the allocation proportion. In addition, it is well known that an imbalance of the confounding covariates across treatments may bias the study results. This imbalance can be mitigated by covariate adaptive randomization (CAR). CAR can also reduce the selection bias, minimize the accidental bias, and improve statistical efficiency. This tutorial delves into the fundamentals of sequential monitoring, responses adaptive randomization, and covariate adaptive randomization. Additionally, we present theoretical and numerical findings from sequential monitoring in various adaptive randomized clinical trials.