Online Program

Return to main conference page

All Times EDT

Friday, September 25
Fri, Sep 25, 11:45 AM - 12:45 PM
Virtual
Poster Session

PS01-The Optimal Randomization Ratio in Time-to-Event Trials Is Not 1:1 (1st Place Best Poster Winner) (301060)

View Presentation

Yi Liu, Nektar Therapeutics 
*Godwin Yung, Takeda Pharmaceutical 

Keywords: randomization, survival, time-to-event, hazard ratio

Most late-phase clinical trials randomize subjects 1:1, that is, in equal proportions to the various treatment options. Although it has been suggested that unbalanced randomization may have ethical advantages over balanced designs (since more subjects would be assigned to what is hypothesized as the superior therapy), this suggestion has been generally ignored in practice because 1:1 randomization is thought to optimize statistical power.

While this is true for continuous endpoints, recent research by Yung and Liu (2019) found that, for time-to-event endpoints such as progression free survival (PFS) and overall survival (OS) in oncology, randomizing patients in a 3:2 or 2:1 fashion can in fact increase statistical power. This finding re-raises the question, “Why don’t we randomize more patients to investigational therapy?”

In this presentation, we will discuss the rationale behind the current misconception. We will then explain what is in fact the optimal randomization ratio in a trial with time-to-event endpoint. We will also explore the potential benefits (eg, shortened timeline, less observed deaths) and pitfalls (eg, prolonged timeline) of unbalanced randomization.