Search

In this paper we expose the design of a system for the remote monitoring of Heart Failure (HF) patients, complemented by an Artificial Intelligence (AI) engine to perform a classification of patients severity on a three levels scale: mild, moderate and severe. The system allows multiple care regimes: a scheme called IHC (Integrated Home Care) and a scheme called CIHC (Continuous Integrated Home Care). The first needs that a health care worker is traveling periodically to the patient's home to perform various measurements of physiological parameters, the second is fully automatic but requires that a kit for the automatic acquisition of the parameters is provided to the patient. In results section we show performances of AI, trained using our clinical partner database, in assessing HF severity and HF type that are respectively 89% and 86% hold out accuracy. This system would facilitate the application of the principles of the Chronic Care Model, in our case regarding the assistance for Heart Failure, but the system is scalable to many other chronic diseases. Due to the amount of input parameters and the fact that HF involves the whole body, we believe that it can be the right disease for the prototype of a disease-specialized system that allows structured communications between hospital and territory. In this paper we expose the design of a system for the remote monitoring of Heart Failure (HF) patients, complemented by an Artificial Intelligence (AI) engine to perform a classification of patients severity on a three levels scale: mild, moderate and severe. The system allows multiple care regimes: a scheme called IHC (Integrated Home Care) and a scheme called CIHC (Continuous Integrated Home Care). The first needs that a health care worker is traveling periodically to the patient's home to perform various measurements of physiological parameters, the second is fully automatic but requires that a kit for the automatic acquisition of the parameters is provided to the patient. In results section we show performances of AI, trained using our clinical partner database, in assessing HF severity and HF type that are respectively 89% and 86% hold out accuracy. This system would facilitate the application of the principles of the Chronic Care Model, in our case regarding the assistance for Heart Failure, but the system is scalable to many other chronic diseases. Due to the amount of input parameters and the fact that HF involves the whole body, we believe that it can be the right disease for the prototype of a disease-specialized system that allows structured communications between hospital and territory.