Design of an Adaptive Machine Learning Model for Early Prediction of Critical Patient Deterioration in IoT-Enabled Healthcare Systems

Read full text Summarise with VIRAT
Authors: Nwobodo-Nzeribe, Chikezie C. D., Ndulue, Theophilus Okwudili; Chikezie, Chukwuma David, Obonyano, Kingdom Nelson; Chikezie, Chukwuma David, Utibe V.; Asoronye G. O.
Cited by: 0
References: 11

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The dynamic development of the IoT-based technologies in the healthcare sector has allowed to monitor the physiological indicators of the patients continuously, but the methods of the early warning that are used traditionally based on the threshold values frequently cannot provide an opportunity to notice the critical worsening in time. This paper introduces a flexible hybrid deep learning network consisting of Multilayer Perceptron (MLP), One-Dimensional Convolutional Neural Network (1D-CNN), and Gated Recurrent Unit (GRU) networks as a predictor of early patient deterioration in the Internet of Things (IoT)-based healthcare systems. IoT-monitored physiological time-series data (heart rate, blood pressure, respiratory rate, temperature, and SpO 2 ) were gathered using both IoT sensors and retrospective clinical data. Preprocessing of the data included filtering of noise, normalisation, filling in of missing values and breaking down into fixed-length sequences. The hybrid model combines the static, short-term and long-term time data to make real-time predictions to overcome the shortcomings of traditional early-warning models. Experimental results demonstrate that the hybrid model outperforms individual sub-models, achieving 96.8% accuracy, 95.7% precision, 97.2% recall, 96.4% F1-score, and 0.983 ROC-AUC, with significantly reduced false positives and false negatives. The results suggest that the suggested framework will be able to issue reliable and timely alerts about vital patient deterioration, contribute to proactive clinical interventions, and enhance patient safety in IoT healthcare settings.

Fields of Study & Subjects

Machine Learning in Healthcare Non-Invasive Vital Sign Monitoring Artificial Intelligence in Healthcare

Publication details

Journal / source
Iconic Research and Engineering Journals
Volume
9
Issue
9
Publication date
30 March 2026
Language
EN