Data-Driven and Expert-Informed Causal Discovery for Type 2 Diabetes Risk in Primary Care

The aim of this study is to extract causal relationships from a large set of routinely collected primary care data (biomarkers, medical conditions, risk factors, and medications) from across Canada, with a focus on Type 2 Diabetes risk. The causal discovery process combined data-driven insights with prior expert knowledge, using iterative refinement to construct a causal Directed Acyclic Graph (DAG). The retrieved DAG, which aligns with medical knowledge and performs satisfactorily in predicting future Type 2 Diabetes onset, could serve as a foundation for developing interpretable tools to support medical decision-making.