Brain Behav. 2025 Oct;15(10):e70958. doi: 10.1002/brb3.70958.
ABSTRACT
OBJECTIVE: To develop and validate a machine learning (ML)-based predictive model for intensive care unit (ICU) readmission in patients with cerebral infarction using neurophysiological and clinical data from the MIMIC-IV database.
METHODS: A retrospective cohort of 3,348 patients diagnosed with cerebral infarction was identified from the MIMIC-IV database. Feature selection was conducted using the least absolute shrinkage and selection operator (LASSO) regression, followed by multivariable logistic regression analysis. Various ML models, including Decision Tree, K-Nearest Neighbors, LightGBM, Naïve Bayes, Random Forest, Support Vector Machine, and XGBoost, were developed and evaluated based on model performance metrics.
RESULTS: The logistic regression model achieved the highest area under the receiver operating characteristic curve (AUC) of 0.682 (95% CI: 0.630-0.733). Significant predictors of ICU readmission included peptic ulcer disease, glucocorticoid use, potassium levels, and red blood cell count.
CONCLUSIONS: This study demonstrates that ML models can effectively predict ICU readmission in CI patients. Logistic regression provides a clinically interpretable approach for risk stratification.
PMID:41076549 | DOI:10.1002/brb3.70958
