Multi-Agent Healthcare Monitoring System

Abstract

The failure to detect changes in a patient's vitals could have significant consequences on the patient's welfare. This makes us launch Neulife, an Artificial Intelligence-based system that could continuously monitor physiological parameters and provide on-time risk evaluation. Neulife has a 3 -level full-stack approach with patient, healthcare, and administrative-specific dashboards. Neulife is based on a hybrid machine learning architecture, as opposed to fixed thresholds. An Isolation Forest model identifies unusual physiological patterns, and a Random Forest classifier estimates the risk levels of the patients under the National Early Warning Score (NEWS2) standard. Key measurements measured include oxygen saturation, blood pressure and heart rate. Neulife is able to distinguish between benign changes and true critical changes and drastically reduce false positives, reducing alarm fatigue. High accuracy and response times while detecting risks are achieved using experimental validation through simulation of clinical scenarios. In the end, Neulife offers a scalable and efficient platform for data-driven, proactive patient care in both patient support areas, hospitals, and remote.