TITLE: Navigating the Complexities of Platform Trials: Design and Simulation
INSTRUCTOR: Kyle Wathen, Cytel Inc
MODERATOR: Jingjing Ye
Abstract:
Adaptive platform trials offer substantial advantages over traditional clinical trial designs, including greater flexibility, improved efficiency, and the ability to evaluate multiple interventions within a single, evolving framework. Their operational and statistical complexity, however, makes careful planning and comprehensive simulation essential. This 3-hour tutorial provides a practical introduction to the design and simulation of adaptive platform trials using OCTOPUS (https://kwathen.github.io/OCTOPUS/), an R package developed for simulating platform trials.
The tutorial begins with a concise overview of platform trial design, including core design features, common sources of complexity, and the situations in which platform approaches may offer meaningful advantages. It then focuses on why it is important to simulate the actual trial intended for implementation, including key real-world features such as the addition or removal of treatment arms over time and other sources of variability that can materially affect operating characteristics and trial performance.
A central component of the tutorial is a hands-on worked example in which participants will develop a simulation plan for a hypothetical platform trial and implement it in OCTOPUS. Using this example, the tutorial will demonstrate how to specify design assumptions, run simulations, and interpret outputs relevant to trial decision-making. Participants will also examine how different design choices influence performance and how simulation can be used to compare alternatives and support more robust design decisions.
The tutorial will additionally introduce the concept of patient-level simulation, including examples of custom patient simulators and approaches that extend beyond the standard analyses available in OCTOPUS. By the end of the session, participants will have a practical understanding of how to structure a platform trial simulation in R, generate informative visualizations, and use simulation results to communicate and justify design choices.
Take-Home Learnings
After completing the tutorial, participants will leave with:
- A practical frameworkfor thinking about the design and simulation of adaptive platform trials.
- Hands-on experienceusing OCTOPUS to build and run a platform trial simulation.
- A clearer understandingof how perpetual features, such as adding or dropping treatment arms, can affect trial behavior.
- Experience interpreting resultsfrom simulated trial scenarios to inform design decisions.
- Reusable ideas and code patternsfor setting up simulations and producing visual summaries in R.
Instructors’ Biography:
Kyle brings experience from a diverse background in academia, consulting, and the life sciences industry to his tutorial. Working on the development and application of novel Bayesian methodology for adaptive clinical trial designs, he is involved in each step of developing new adaptive clinical trial designs, starting from initial concept development through software development/trial simulation and completing with trial conduct and data collection.
Kyle has over 25 years of experience in the design of innovative clinical trials such as Bayesian approaches, platform trials and other adaptive approaches. He has been involved in many innovative clinical trials, especially platform trials, in various disease areas including oncology, neuroscience, infectious diseases, cardiovascular and inflammation. He has released several software packages besides OCTOPUS.
Kyle received his M.S. in statistics from Texas A&M University and M.S. and Ph.D. in Biostatistics from the University of Texas: Graduate School of Biomedical Sciences.