When Patients Recover From COVID-19: Data-Driven Insights From Wearable Technologies

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Computational Biology

Background:

• Coronavirus disease 2019 (COVID-19) presents a significant global health challenge, overwhelming healthcare systems.
• Effective patient management requires accurate and timely classification of disease severity to optimize resource allocation.

Purpose of the Study:

• To develop a cost-effective model for early and accurate classification of COVID-19 disease stages.
• To enhance patient care and resource management through precise disease staging.

Main Methods:

• A hybrid model integrating Long-short Term Memory (LSTM) for temporal feature extraction and Deep Neural Network (DNN) for attribute feature extraction.
• Utilizing vital indicators and temporal features for disease stage classification.

Main Results:

• The proposed LSTM-DNN model significantly outperforms existing state-of-the-art methods in prediction accuracy.
• Identified critical vital indicators for different COVID-19 stages, aiding in early detection of severity.
• Case studies demonstrated the model's ability to differentiate between severe and mild cases and identify recovery patterns.

Conclusions:

• The developed model provides a valuable tool for patients to understand their disease progression.
• Enables healthcare providers to offer timely, remote treatment plans tailored to specific disease stages, optimizing medical resource utilization.