Posterior Predictive Design for Phase I Clinical Trials

Interval-based designs represent cutting-edge adaptive methodologies for phase I clinical trials to identify the maximum tolerated dose (MTD). These designs exhibit robust performance comparable to more intricate, model-based designs, and their pretabulated decision rule enables them to be implemented as simply as the conventional algorithm-based designs. In this paper, we introduce the posterior predictive (PoP) design, a novel interval-based design that leverages advanced Bayesian predictive hypothesis testing techniques for dose escalation and de-escalation. Our work moves beyond the existing model-assisted interval-based designs by achieving global optimality in dose transition. Theoretically, the global optimality ensures that the proposed design can consistently select the true MTD at an impressive convergence rate of <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup><mml:mrow><mml:mi>n</mml:mi></mml:mrow> <mml:mrow><mml:mo>-</mml:mo> <mml:mn>1</mml:mn> <mml:mo>/</mml:mo> <mml:mn>2</mml:mn></mml:mrow> </mml:msup> </mml:math> . Through extensive simulation studies, we demonstrate that the PoP design yields substantial improvement in operating characteristics to identify MTD, thereby presenting a valuable upgrade to the popular interval-based designs in practice. Supplementary materials for this article are available online, including a standardized description of the materials available for reproducing the work.