Development and internal validation of an interpretable machine learning model for predicting vancomycin-induced nephrotoxicity in hospitalized children
5 affiliations
- 1Department of Pharmacy Children’s Hospital of Chongqing Medical University, National Clinical Research Center for Children and Adolescents’ Health and Diseases, Chongqing, China
- 2Ministry of Education Key Laboratory of Child Development and Disorders, Chongqing, China
- 3Chongqing Key Laboratory of Pediatric Metabolism and Inflammatory Diseases, Chongqing, China
- 4Department of Child Healthcare, The First People’s Hospital of Zigong, Zigong, Sichuan China
- 5Department of Pharmacy, The First People’s Hospital of Zigong, No.1 Shangyihao Road, Ziliujing District, Zigong, Sichuan China
Abstract
Introduction: Vancomycin is widely used for severe Gram‑positive infections in children, but vancomycin‑induced nephrotoxicity (VIN) limits its safe application. Existing prediction models rely primarily on traditional logistic regression with limited discriminative performance in pediatric populations, and machine learning (ML) approaches have not been systematically evaluated.
Aim: This study aimed to develop and internally validate an interpretable ML model for predicting VIN in pediatric patients using electronic health record (EHR) data.
Method: We conducted a retrospective EHR-based study at a tertiary children’s hospital. Children aged 1 month to 18 years who received vancomycin for ≥ 3 consecutive days between December 2017 and November 2024 were included. Patients with moderate-to-severe pre-existing renal dysfunction [estimated glomerular filtration rate (eGFR) ≤ 59 mL/min/1.73 m2], preterm birth, blood purification therapy, or major missing data were excluded. VIN was defined as an increase in serum creatinine of ≥ 0.5 mg/dL or > = 50% from baseline on at least two consecutive measurements. Eight ML models were constructed and compared. Model performance was evaluated using the area under the receiver-operating-characteristic curve (AUROC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy (ACC), F1 score, Brier score and the area under the precision recall curve (AUPR). The optimal model was interpreted using the Shapley Additive explanation (SHAP) method. Feature sets of 6, 8, 10, and full features were compared using DeLong's test. Clinical utility was assessed via decision curve analysis (DCA).
Results: A total of 1,452 children were included, of whom 206 (14.2%) developed VIN. EXtreme gradient boosting (XGboost) achieved the best predictive performance (AUROC = 0.90, AUPR = 0.74, Brier score = 0.07). A simplified 6‑feature XGBoost model (eGFR, amphotericin B, early vancomycin trough concentration, albumin, immunosuppressant use, and acyclovir/ganciclovir co-administration) achieved comparable discrimination to the full model (AUROC = 0.87; P > 0.05 by DeLong's test) with acceptable calibration (ECE 3.8%). DCA suggested favorable net benefit across clinically relevant risk thresholds.
Conclusion: Our interpretable 6-feature XGBoost model demonstrated promising internal discrimination for pediatric VIN prediction. Pending external validation and temporal evaluation, this model may serve as a hypothesis-generating decision-support framework for risk stratification.
Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1007/s11096-026-02204-1.
Related
- Development and validation of an interpretable machine learning model for predicting in-hospital mortality among critically ill patients with liver cirrhosis
- Development and Validation of an Interpretable Machine‐Learning Model for Predicting Suboptimal Clinical Response After Sleeve Gastrectomy: A Multicenter Study
- An interpretable machine learning model for predicting intensive care unit mortality risk in elderly patients with concomitant cardiogenic shock and sepsis: Development and external validation using two large critical care databases
- Development and validation of interpretable machine learning models for predicting cumulative live birth rates in assisted reproductive technology: a SHAP-based analysis.
- Development and validation of an interpretable machine learning model for predicting intravenous contrast media extravasation risk in CT imaging
BibTeX
@article{wang2026development,
title = {{Development and internal validation of an interpretable machine learning model for predicting vancomycin-induced nephrotoxicity in hospitalized children}},
author = {Xiuling Wang and Yao Fu and Jinxu Chen and Jie Wan and Wenqiang Kong},
journal = {International Journal of Clinical Pharmacy},
year = {2026},
month = aug,
doi = {10.1007/s11096-026-02204-1},
url = {https://doi.org/10.1007/s11096-026-02204-1},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Development and internal validation of an interpretable machine learning model for predicting vancomycin-induced nephrotoxicity in hospitalized children
AU - Xiuling Wang
AU - Yao Fu
AU - Jinxu Chen
AU - Jie Wan
AU - Wenqiang Kong
PY - 2026
DA - 2026/08/25
JO - International Journal of Clinical Pharmacy
DO - 10.1007/s11096-026-02204-1
UR - https://doi.org/10.1007/s11096-026-02204-1
AB - Introduction: Vancomycin is widely used for severe Gram‑positive infections in children, but vancomycin‑induced nephrotoxicity (VIN) limits its safe application. Existing prediction models rely primarily on traditional logistic regression with limited discriminative performance in pediatric populations, and machine learning (ML) approaches have not been systematically evaluated. Aim: This study aimed to develop and internally validate an interpretable ML model for predicting VIN in pediatric patients using electronic health record (EHR) data. Method: We conducted a retrospective EHR-based study at a tertiary children’s hospital. Children aged 1 month to 18 years who received vancomycin for ≥ 3 consecutive days between December 2017 and November 2024 were included. Patients with moderate-to-severe pre-existing renal dysfunction [estimated glomerular filtration rate (eGFR) ≤ 59 mL/min/1.73 m2], preterm birth, blood purification therapy, or major missing data were excluded. VIN was defined as an increase in serum creatinine of ≥ 0.5 mg/dL or > = 50% from baseline on at least two consecutive measurements. Eight ML models were constructed and compared. Model performance was evaluated using the area under the receiver-operating-characteristic curve (AUROC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy (ACC), F1 score, Brier score and the area under the precision recall curve (AUPR). The optimal model was interpreted using the Shapley Additive explanation (SHAP) method. Feature sets of 6, 8, 10, and full features were compared using DeLong's test. Clinical utility was assessed via decision curve analysis (DCA). Results: A total of 1,452 children were included, of whom 206 (14.2%) developed VIN. EXtreme gradient boosting (XGboost) achieved the best predictive performance (AUROC = 0.90, AUPR = 0.74, Brier score = 0.07). A simplified 6‑feature XGBoost model (eGFR, amphotericin B, early vancomycin trough concentration, albumin, immunosuppressant use, and acyclovir/ganciclovir co-administration) achieved comparable discrimination to the full model (AUROC = 0.87; P > 0.05 by DeLong's test) with acceptable calibration (ECE 3.8%). DCA suggested favorable net benefit across clinically relevant risk thresholds. Conclusion: Our interpretable 6-feature XGBoost model demonstrated promising internal discrimination for pediatric VIN prediction. Pending external validation and temporal evaluation, this model may serve as a hypothesis-generating decision-support framework for risk stratification. Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1007/s11096-026-02204-1.
SN - 2210-7703
SN - 2210-7711
ER -