Machine Learning for Predicting Chronic Renal Disease Progression in COVID-19 Patients with Acute Renal Injury: A...

Area of Science:

• Nephrology
• Infectious Diseases
• Artificial Intelligence

Background:

• Coronavirus disease (COVID-19) can lead to acute kidney injury (AKI), increasing the risk of chronic kidney disease (CKD).
• Assessing CKD progression in COVID-19 patients with AKI is crucial for timely intervention.

Purpose of the Study:

• To evaluate the feasibility of machine learning (ML) models for predicting CKD progression in COVID-19 patients with AKI.
• To compare the performance of different ML algorithms in this predictive task.

Main Methods:

• Retrospective study of adult COVID-19 patients with AKI admitted to a hospital in Mérida, Yucatán, México.
• Development and validation of four ML models: Support Vector Machine (SVM), Random Forest, Logistic Regression, and Boosting.
• Utilized variable selection methods to optimize model performance.

Main Results:

• The Boosting model achieved the highest Area Under the Receiver Operating Characteristic Curve (AUC-ROC) of 0.841, followed closely by Logistic Regression (0.840).
• The Support Vector Machine (SVM) model demonstrated excellent classification rates: 99.8% ± 0.1 (training) and 98.43% ± 1.79 (validation).
• All models showed promising performance in predicting CKD progression.

Conclusions:

• Machine learning models are feasible for assessing CKD progression in COVID-19 patients with AKI.
• The Boosting and Logistic Regression models show particular promise for early detection and management of CKD in this population.
• Further research is warranted to validate these findings and refine predictive capabilities.