DHA supplementation for early preterm birth prevention: An application of Bayesian finite mixture models to adaptive clinical trial design optimization.

Simulation reveals several advantages of the RAR designs, such as higher allocation to the more promising dose and a trial duration reduction. The proposed Hybrid RAR Designs addresses the statistical power drop observed in Adaptive RAR. The new design model shows robustness to hyperprior choices. We recommend Hybrid RAR Design 1 for ADORE Precision, anticipating that it will yield precise determinations, which is crucial for advancing our understanding in this field.