Predictive Analysis Using the Internet of Medical Things to Develop a Smart Patient Monitoring System

Abstract

The current patient monitoring system is primarily designed to cater to emergency and elderly patients, aiming to protect them from critical situations. It relies on the Internet of Things and related medical devices and equipment strategically placed around the patients’ bodies, beds, and the ward where they stay to record their health and related data. However, the existing system has its limitations. For instance, patients' health needs to be monitored, and this may change due to their emergency condition, staying in the intensive care unit, age, and other factors. These factors need to be considered to provide immediate and accurate treatment. However, the patient versus medical practitioner ratio is 1:25, and caring nurses is 1:10, which is not sufficient and efficient in providing smart healthcare. This underscores the need for a more advanced and efficient patient monitoring system. The earlier methods proposed defined computer-aided systems to analyze the patients' records and monitor them indoors and outdoors. IoT devices can provide various kinds of data but cannot be processed simultaneously using a single analytical model. This paper proposed a Multi-Modality Split Learning (MMSL) model for analyzing different medical data obtained from multiple Internet of Medical Things and aggregating the predicted output to the server to allocate the doctors and nurses for immediate treatment to save the patients. The IoMT data is transferred from the devices to the server in parallel and sequential models to increase the transmission speed. The proposed MMSL model is implemented in Python, and the results have been verified. From the output, it is verified that it outperforms the others.