Understanding the progression of chronic diseases is important to inform early diagnosis, personalize care and ensure effective healthcare system management. Data from clinical and administrative systems have the potential to advance this understanding, but traditional methods for modeling disease progression are ill-suited to many healthcare datasets, where samples may be collected at irregular intervals and cohorts are large and heterogeneous.
The impact of treatment intervention, and the discovery of optimal adaptive treatment strategies (ATSs), is of great importance in primary and clinical care. Most optimal causal discovery approaches have been attempted in relatively small-scale settings. In this project, we will develop digital twins for ATS discovery using Bayesian and machine learning approaches, which scale better to large, complex datasets. These methods promise to have a major impact on clinical decision making, opening the door to data-driven personalized care.