Bayesian adaptive dose optimization with repeated patient outcomes

Patients often contribute multiple outcomes during treatment, while traditional dose-finding designs typically use one outcome per patient. This work develops a hierarchical Bayesian framework that uses repeated outcomes while accounting for within-patient dependence and changes over time. It connects dose-finding methods to a broader class of adaptive clinical designs, motivated by pediatric oncology applications.

This work is in progress. A public manuscript link will be added when available.