In this video interview, David Morton, PhD, director of biostatistics at Certara, explores the practical challenges of implementing Bayesian designs, including the need for simulation, cross-functional alignment, and clear communication with regulators.
In a recent video interview with Applied Clinical Trials, David Morton, PhD, director of biostatistics, Certara, discussed how the FDA’s increased
Editor's note: This transcript is a lightly edited rendering of the original audio/video content. It may contain errors, informal language, or omissions as spoken in the original recording.
ACT: What new capabilities or planning will clinical teams need to successfully incorporate prior data and external evidence into trial design?
Morton: To successfully implement these methods, teams need to develop several technical and strategic capabilities. They must provide strong evidence for the relevance of borrowed data and justify how much reliance is placed on it.
The prior distribution is really key. You can have informative priors or non-informative priors, which are more like flat probability distributions. Simulation and sensitivity analyses at the design stage are also critical to explore scenarios where prior data may conflict with new data.
This helps ensure results remain interpretable. Clinical and scientific input is important for assessing external data and defining clinically meaningful thresholds. For example, in biomarker analysis, teams need to understand the probability of exceeding certain thresholds.
From an operational standpoint, infrastructure for adaptive decision-making becomes critical. You need real-time data flow for Bayesian monitoring, so you can trigger early stopping if needed or escalate promising results.
The guidance also emphasizes transparency and documentation. There’s a detailed list of expectations. Bayesian trials require more upfront planning, but in the long run, they can really pay off.
