Predicting mortality risk in patients with COVID-19 using machine learning to help medical decision-making

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Epidemiology

Background:

• The COVID-19 pandemic necessitated rapid advancements in patient management and risk stratification.
• Existing healthcare systems faced challenges in efficiently allocating resources and prioritizing care for SARS-CoV-2 infected individuals.

Purpose of the Study:

• To develop and validate an Artificial Intelligence (AI) and Machine Learning (ML) model for predicting health and mortality risks in COVID-19 patients.
• To provide a tool for healthcare facilities to optimize patient triage, resource allocation, and timely intervention.

Main Methods:

• Utilized a large-scale global dataset of over 2.67 million confirmed COVID-19 cases from 146 countries.
• Employed various ML algorithms, including Support Vector Machine (SVM), Artificial Neural Networks, Random Forest, Decision Tree, Logistic Regression, and K-Nearest Neighbor (KNN).
• Identified key symptoms and features indicative of severe outcomes and validated model performance using a separate dataset and confusion matrix analysis.

Main Results:

• The developed AI model achieved an overall accuracy of 89.98% in predicting COVID-19 patient mortality.
• The study successfully identified critical symptoms and patient features associated with increased mortality risk.
• Sensitivity and specificity analyses were conducted for in-depth classifier evaluation.

Conclusions:

• The AI-powered predictive model offers a reliable solution for risk assessment and mortality prediction in COVID-19 patients.
• This technology can significantly enhance hospital decision-making processes, improve patient triage efficiency, and reduce delays in critical care.
• The findings underscore the potential of AI and ML in managing public health crises and optimizing healthcare delivery.