A0459
Title: Delayed reveal randomization in platform trials: A simulation study
Authors: Rajenki Das - MRC Biostatistics Unit, University of Cambridge (United Kingdom) [presenting]
Sofia Villar - University of Cambridge (United Kingdom)
Abstract: Platform trials are increasingly adopted in intensive care and emergency medicine, comparing multiple interventions simultaneously. This presents a challenge of randomizing critically ill patients to numerous options at a sensitive point in their care. Consequently, "revealed randomization" (where treatment assignment is known immediately) is often preferred to reduce study burden. However, a delayed reveal, while practical for operational reasons, can complicate the rigorous delivery of randomization. A solution is presented through an example trial where participants are randomized to four treatment groups: Three active treatment + standard of care (SOC) arms, and one placebo + SOC arm. A unique aspect is that the fourth treatment group is reserved for a small fraction of the enrolled population meeting a specific eligibility criterion. Stratified randomization is employed, and the properties of this scheme are assessed, with particular emphasis on balance across treatment arms and strata. Findings illustrate a practical solution for a delayed-reveal randomization system that maintains robust performance while addressing the inherent challenges of multi-arm trials in acute care settings.
