Multiscale modeling of comorbidity relations in hypertensive outpatients

Abstract Investigation of comorbidity relations plays crucial role in patients’ stratification and risk analysis. Having a huge dataset of electronic health records of hypertensive outpatients from Federal Almazov North-West Medical Centre, we propose a hierarchical modeling scheme and present our initial results. At each level of disease hierarchy (including time domain), we consider different problems; from causality links to disease progression and optimized treatment. Bayesian networks are used as a main tool for modeling as they are ideal for uncertain and noisy medical high-dimensional datasets.