Wednesday, September 24, 2025

10:30 AM - 12:00 PM

Grand Copthorne Waterfront Hotel - Waterfront Ballroom I

---

Details

Model-based dosing requires a suitably predictive model. Often, multiple models are available for a given drug, and so the clinician is faced with a decision of how to best treat their patient. In recent years, two approaches have been proposed: (i) model selection, in which one model of several is selected for a given patient and used to inform dose selection, and (ii) model averaging, in which the predictions from multiple models are combined to inform dose selection. Both model selection and model averaging can be implemented in different ways, ranging from statistical approaches to machine learning algorithms. These approaches come with clinical implementation trade-offs, regulatory differences, and differences in predictive performance. In this symposium, considerations for each approach are explored and compared.