Design and Validation of an Intelligent System for Decision-making on Telemonitoring Intensity in Heart Failure Patients

Heart failure is a chronic and progressive disease that affects an increasing number of adults worldwide, especially over 65 years of age. It is a leading cause of hospitalization, posing significant challenges for healthcare management. The early detection of (sub)clinical deteriorations presents an opportunity for intervention, potentially averting or delaying the progression to acute clinical events that require hospital admissions. Remote monitoring has shown promise, allowing physicians to have selfmeasured data from the patient to monitor the progress of the disease, and thus intervene timely in case of worsening. Nevertheless, for effective implementation of personalized care programs, physicians have full responsibility for deciding and adapting monitoring plans, which challenge their workload and timely responses. To tackle these challenges, this paper proposes a novel rule-based intelligent system designed to suggest and modify monitoring intensity for heart failure patients undergoing home treatment. It relies on 30 input variables concerning the patient's medical history, vital signs, reported symptoms, and 283 rules for decision-making. The validation results show a promising 0.85 macro F1-score, demonstrating its technological feasibility and applicability in remote patient monitoring contexts.