TY - GEN
T1 - An Efficient and Secure Framework for Smart Healthcare Using IoT and Machine Learning
AU - Pandey, Vivek Kumar
AU - Prakash, Shiv
AU - Gupta, Tarun Kumar
AU - Rathore, Vandana
AU - Singh, Abhishek
AU - Yang, Tiansheng
AU - Rathore, Rajkumar Singh
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2025/1/17
Y1 - 2025/1/17
N2 - The healthcare systems worldwide are challenged with the increased demand for rapid and efficient patient care. In this context, an advanced solution is required to meet the growing needs. This paper presents an intelligent healthcare framework integrating IoT devices with machine learning algorithms to enable real-time predictive patient monitoring. It comprises a four-layer architecture: data acquisition from IoT sensors, data preprocessing, analytics for predictive modeling, and an application layer for user interactions. Validated with data from the MIMIC-III and PhysioNet datasets along with 12 months of IoT sensor data from a university medical center, the framework achieved 96% prediction accuracy with precision, recall, and ROC AUC scores of 0.95, 0.96, and 0.98, respectively. The system had a great response time of 200 milliseconds. It has greatly improved early detection by 36.9% and reduced false alarms by 68%. False alarms decreased from 25% to 8%, and response time was reduced from 45 to 12 minutes, or 73.3 %. Resource utilization analysis reflects optimal performance since the analytics layer hit a spike at 45% CPU utilization, thereby pointing out optimal load management. Patient satisfaction also rose to 22.7%, increasing from 75% to 92%, reflecting the framework's impact on care quality. All these results render the framework an effective and scalable solution that optimizes healthcare resources while improving care for the patient.
AB - The healthcare systems worldwide are challenged with the increased demand for rapid and efficient patient care. In this context, an advanced solution is required to meet the growing needs. This paper presents an intelligent healthcare framework integrating IoT devices with machine learning algorithms to enable real-time predictive patient monitoring. It comprises a four-layer architecture: data acquisition from IoT sensors, data preprocessing, analytics for predictive modeling, and an application layer for user interactions. Validated with data from the MIMIC-III and PhysioNet datasets along with 12 months of IoT sensor data from a university medical center, the framework achieved 96% prediction accuracy with precision, recall, and ROC AUC scores of 0.95, 0.96, and 0.98, respectively. The system had a great response time of 200 milliseconds. It has greatly improved early detection by 36.9% and reduced false alarms by 68%. False alarms decreased from 25% to 8%, and response time was reduced from 45 to 12 minutes, or 73.3 %. Resource utilization analysis reflects optimal performance since the analytics layer hit a spike at 45% CPU utilization, thereby pointing out optimal load management. Patient satisfaction also rose to 22.7%, increasing from 75% to 92%, reflecting the framework's impact on care quality. All these results render the framework an effective and scalable solution that optimizes healthcare resources while improving care for the patient.
KW - Healthcare Analytics
KW - Internet of Things
KW - Machine Learning
KW - Real-time Monitoring
KW - Smart Healthcare
UR - http://www.scopus.com/inward/record.url?scp=85217224180&partnerID=8YFLogxK
U2 - 10.1109/dasa63652.2024.10836569
DO - 10.1109/dasa63652.2024.10836569
M3 - Conference contribution
SN - 979-8-3503-6911-3
T3 - 2024 International Conference on Decision Aid Sciences and Applications (DASA)
SP - 1
EP - 5
BT - 2024 International Conference on Decision Aid Sciences and Applications (DASA)
PB - Institute of Electrical and Electronics Engineers (IEEE)
T2 - 2024 International Conference on Decision Aid Sciences and Applications (DASA)
Y2 - 11 December 2024 through 12 December 2024
ER -