Investigating the Dynamics of Cardio-Metabolic Comorbidities and Their Interactions in Ageing Adults Through Dynamic Bayesian Networks

With increased longevity, the likelihood of developing multiple chronic diseases also increases. Among these, cardio-metabolic comorbidities represent a burden both in terms of individual quality of life and public health. Understanding the impact of risk factors and unravelling possible cross-effects between comorbidities themselves can facilitate care management and prevention strategies. In this work, we present a model of ageing progression and the onset of three cardio-metabolic diseases, namely type 2 diabetes, hypertension, and heart problems, together with survival, based on socio-demographic, clinical, and biomarkers data of more than 11,000 subjects available in the English Longitudinal Study of Ageing. Leveraging dynamic Bayesian networks, our model effectively captures the probabilistic relationships between risk factors and morbidities over time, with many biological interactions known from the literature correctly encoded, such as the effect of body mass index and physical activity on the onset of cardio-metabolic diseases. Noticeably, some cross-relationships between outcomes’ occurrence also emerge, with an increased risk of heart problems in the presence of hypertension. In addition to the graphical description of the ageing process, we propose a simulation strategy that allows us to predict in silico the progression of the clinical state of a new patient population (iAUC between 0.62–0.83 for all outcomes), as well as a stratification analysis that allows investigating the effect of selected variables on the risk of developing morbidity. This approach provides valuable support for the acquisition of knowledge aimed at designing prevention strategies and targeted interventions to improve the health status of the ageing population.