Data-driven Rules for Individualized Lifestyle Recommendations



Journal Title

Journal ISSN

Volume Title



Heart failure (HF) is a global pandemic affecting more than 26 million patients worldwide. Effective management of risk factors is extremely important for reducing heart failure. Lifestyle modification can effectively reduce the risk of heart failure but clinical guidelines are generalized and not tailored for individuals. This project developed a rule-based framework that automatically generates personalized lifestyle modification recommendations for heart failure risk reduction. The proposed framework integrates an ensemble learning-based rule discovery model (RuleFit) that translates the patient-level profiles into actionable patterns (rules), and a rule-based optimization algorithm that searches for the optimal lifestyle modification recommendations based on the patient’s unchangeable profile. The proposed framework was applied to a large population in the Atherosclerosis Risk in Communities (ARIC) study to manage patient risk of fatal coronary heart disease events.



Individualized lifestyle recommendations, Expert systems, Patient-centered healthcare, Supervised machine learning, RuleFit ensemble, Network-assisted optimization, Operations research