Risk Factor Analysis and Multiple Predictive Machine Learning Models for Mortality in COVID-19: A Multicenter and...

Area of Science:

• Medical Informatics
• Computational Biology
• Epidemiology

Background:

• The COVID-19 pandemic necessitated rapid development of effective screening and management strategies.
• Accurate and cost-effective methods are crucial for preventing infections and guiding patient care.

Purpose of the Study:

• To develop and validate machine learning models for predicting COVID-19 mortality.
• To identify clinical risk factors associated with severe COVID-19 outcomes.

Main Methods:

• Utilized routine clinical data from 4711 COVID-19 patients.
• Applied three machine learning algorithms: random forest (RF), partial least squares discriminant analysis (PLS-DA), and support vector machine (SVM).
• Compared the predictive performance and identified common risk factors across models.

Main Results:

• Developed three predictive models for COVID-19 mortality.
• The RF model achieved the highest predictive accuracy (ROC 0.859).
• Identified nine consistent risk factors: age, procalcitonin, ferritin, C-reactive protein, troponin, blood urea nitrogen, mean arterial pressure, aspartate transaminase, and alanine transaminase.

Conclusions:

• Machine learning models effectively predict COVID-19 mortality using accessible clinical data.
• The random forest model offers the best performance for mortality prediction.
• The identified nine clinical variables may serve as important prognostic indicators for severe COVID-19.