Results of the Experimentation of a Platform for Automated Detection of Situations at Risk of Cardiac Decompensation (E-CarePlatform) in Elderly Patients with Multiple Comorbidities

Introduction: Monitoring patients with heart failure by telemedicine systems is a potential means susceptible to optimize the management of these patients and avoid life-threatening emergencies. In this context, we experimented in patients with chronic disorders, especially heat failure, an e-platform E-care dedicated to automated, intelligent detection of situations at risk of heart failure.

Methods: The E-care platform based on medical sensors, communicating, to go up, in real time, to an intelligent physiological information and an analysis of the ontology medical, leading ultimately to the generation of alerts. These are related to a deterioration of the health status of patients in relation to a decompensation of chronic pathologies ( criterion dHS), in particular of heart failure (criterion dCHF), leading to a potential hospitalization. To validate these alerts, an experiment was conducted between February 2014 and April 2015. The E-care platform was deployed to patients followed in internal medicine. During this phase, alerts were collected and analyzed retrospectively in terms of sensitivity (se), specificity (spe), positive predictive values (ppv) and negative (npv) in relation to clinical data.

Results: One hundred and eighty patients were included and 1,500 measurements were obtained. The patient profile included in this experiment was an elderly patient, with comorbidity in 90% of cases –mean Charlson score of 4.1), with chronic heart failure in more than 60% of cases. The analysis of the alerts shows: 1) for criterion dHS: se, spe, ppv and npv values of respectively 100%, 30%, 89% and 100%; and 2) for criterion dCHF: se, spe, ppv and npv values of 100%, 72%, 90% and 100%, respectively.

Conclusion: These results show that the remote monitoring platform E-care can detect 100% of cardiac decompensation and that in ¾ cases, the alerts are related to the latter. Only 10% of alerts are not directly related to HF. In this experiment, the results highlight that in the absence of alert, the patient has no problem, at the cost of many false alarms. In practice, the telemedicine system E-care therefore allows automatically, non-intrusive, generate alerts related to the deterioration of the patient’s health and especially cardiac decompensation.

Keywords: Telemedecine, Heart failure, Detection of signs of cardiac decompensation, Chronic pathology.