Implementation of an Ad Hoc Communication-Based Nurse Call System Integrated with Machine Learning for Patient Risk Classification

Traditional nurse call systems often lack an intelligent mechanism for prioritizing urgent cases, resulting in delayed emergency responses and inefficient care delivery. This work presents a novel nurse call system that integrates ad hoc wireless communication, RFID-based nurse authentication, and machine learning-based patient risk classification. The system utilizes a CatBoost classifier trained on a large dataset of patient vital signs to accurately distinguish between high-risk and low-risk patients, achieving a classification accuracy of 99.89 %. When patients initiate a call, their physiological data is analyzed, and emergency alerts are prioritized based on the risk level. The system adopts a 2.4 GHz RF communication network for easy deployment and ensures secure nurse response tracking via RFID authentication. Experimental evaluation in the prototyping stage confirms that the proposed methodology could improve emergency response speed, enhance workflow efficiency, and enable best resource utilization. This integrated solution is a promising implementation in the IoT in healthcare to improve patient safety and hospital operations.