|
Abstract:
|
In an era of personalized medicine and biological discovery, multi-arm adaptive platform trials offer a novel approach to clinical trial design to evaluate therapies for complex diseases. We discuss a large Australian platform trial design for the treatment of patients with Cystic Fibrosis (CF). The purpose of the trial is to find the best treatment(s) for management of pulmonary exacerbations in a heterogeneous population of CF patients. The trial has multiple treatment options in primary and adjunctive antibiotics, and evaluates benefit in multiple subgroups of patients. A Bayesian repeated measures model is used to evaluate the primary endpoint of change in percent predicted FEV1, in which patients can have multiple exacerbations and re-randomization at subsequent episodes. A master protocol with frequent interim analyses is used to implement response adaptive randomization in the various treatment regimens and patient subgroups, and to allow dropping and adding new therapies to the ongoing trial. Simulations studies are used to explore the performance of the design, and to illustrate the benefits compared to traditional strategies.
|
