Machine learning algorithm to predict the in-hospital mortality in critically ill patients with chronic kidney disease

Abstract

Background: This study aimed to establish and validate a machine learning (ML) model for predicting in-hospital mortality in critically ill patients with chronic kidney disease (CKD).

Methods: This study collected data on CKD patients from 2008 to 2019 using the Medical Information Mart for Intensive Care IV. Six ML approaches were used to build the model. Accuracy and area under the curve (AUC) were used to choose the best model. In addition, the best model was interpreted using SHapley Additive exPlanations (SHAP) values.

Results: There were 8527 CKD patients eligible for participation; the median age was 75.1 (interquartile range: 65.0-83.5) years, and 61.7% (5259/8527) were male. We developed six ML models with clinical variables as input factors. Among the six models developed, the eXtreme Gradient Boosting (XGBoost) model had the highest AUC, at 0.860. According to the SHAP values, the sequential organ failure assessment score, urine output, respiratory rate, and simplified acute physiology score II were the four most influential variables in the XGBoost model.

Conclusions: In conclusion, we successfully developed and validated ML models for predicting mortality in critically ill patients with CKD. Among all ML models, the XGBoost model is the most effective ML model that can help clinicians accurately manage and implement early interventions, which may reduce mortality in critically ill CKD patients with a high risk of death.

Keywords: Chronic kidney disease; critically care; intensive care unit; machine learning; mortality.

Figure 1.

The flowchart of patient selection. MIMIC IV: Medical Information Mort for Intensive Care IV; ICU: intensive care unit; CKD: chronic kidney disease; LIME: Local Interpretable Model-Agnostic Explanations.

Figure 2.

ROC curves for the ML models and the traditional severity of illness scores to predict in-hospital mortality. (A) ROC curves for the six ML models used to predict in-hospital mortality; (B) ROC curves for the traditional severity of disease scores used to predict in-hospital mortality. ROC: receiver operating characteristic; SVM: support vector machine; KNN; k-nearest neighbors; AUC: area under the curve; SOFA: sequential organ failure assessment; SAPS II: simplified acute physiology score II.

Figure 3.

The top 20 important features derived from the XGBoost model. SHAP indicates the importance of ranking features. Each line represents a feature, and the abscissa is the SHAP value. The matrix plot represents the significance of each covariate in constructing the final predictive model. The higher the SHAP value for each clinical variable, the higher risk of death. SHAP: SHapley Additive exPlanations; SOFA: sequential organ failure assessment; SAPS II: simplified acute physiology score II; BUN: blood urea nitrogen; SpO₂: oxygen saturation; MAP: mean arterial pressure; PTT: partial thromboplastin time.

Figure 4.

SHAP summary plot of the top 20 features of the XGBoost model. The importance matrix plot of clinical variables is derived using the XGBoost model. The matrix plot ranks the importance of the variables, revealing the contribution of each variable to death vs. survive. The greater the SHAP value of a characteristic, the greater the likelihood of death development. The abscissa represents the SHAP value, and each line represents a feature. Red dots indicate greater feature values, whereas blue dots indicate lower feature values. SHAP: SHapley Additive exPlanations; SOFA: sequential organ failure assessment; SAPS II: simplified acute physiology score II; BUN: blood urea nitrogen; SpO₂: oxygen saturation; MAP: mean arterial pressure; PTT: partial thromboplastin time.

Figure 5.

SHAP dependence plot of the XGBoost model. (A) SOFA score; (B) SAPS II; (C) respiratory rate; (D) urine output. SHAP values for specific features exceed zero, representing an increased risk of death. The greater the SHAP value of a characteristic, the greater the likelihood of death development. SHAP: SHapley Additive exPlanations; SOFA: sequential organ failure assessment; SAPS II: simplified acute physiology score II.