Interoperable and explainable machine learning models to predict morbidity and mortality in acute neurological injury in the pediatric intensive care unit: secondary analysis of the TOPICC study

被引:1
|
作者
Munjal, Neil K. [1 ,2 ]
Clark, Robert S. B. [2 ]
Simon, Dennis W. [2 ]
Kochanek, Patrick M. [3 ]
Horvat, Christopher M. [2 ]
机构
[1] Univ Wisconsin, Dept Pediat, Madison, WI USA
[2] UPMC Childrens Hosp Pittsburgh, Dept Crit Care Med, Pittsburgh, PA 15224 USA
[3] Univ Pittsburgh, Safar Ctr Resuscitat Res, Pittsburgh, PA USA
来源
FRONTIERS IN PEDIATRICS | 2023年 / 11卷
关键词
machine learning; clinical decision support; pediatric intensive care unit; acute neurological injury; predictive modeling; LOGISTIC-REGRESSION; SERUM BIOMARKERS; BRAIN-INJURIES; DISEASE; RISK; CHILDREN; OUTCOMES; SEPSIS; SYSTEM; SCALE;
D O I
10.3389/fped.2023.1177470
中图分类号
R72 [儿科学];
学科分类号
100202 ;
摘要
BackgroundAcute neurological injury is a leading cause of permanent disability and death in the pediatric intensive care unit (PICU). No predictive model has been validated for critically ill children with acute neurological injury. ObjectivesWe hypothesized that PICU patients with concern for acute neurological injury are at higher risk for morbidity and mortality, and advanced analytics would derive robust, explainable subgroup models. MethodsWe performed a secondary subgroup analysis of the Trichotomous Outcomes in Pediatric Critical Care (TOPICC) study (2011-2013), predicting mortality and morbidity from admission physiology (lab values and vital signs in 6 h surrounding admission). We analyzed patients with suspected acute neurological injury using standard machine learning algorithms. Feature importance was analyzed using SHapley Additive exPlanations (SHAP). We created a Fast Healthcare Interoperability Resources (FHIR) application to demonstrate potential for interoperability using pragmatic data. Results1,860 patients had suspected acute neurological injury at PICU admission, with higher morbidity (8.2 vs. 3.4%) and mortality (6.2 vs. 1.9%) than those without similar concern. The ensemble regressor (containing Random Forest, Gradient Boosting, and Support Vector Machine learners) produced the best model, with Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.91 [95% CI (0.88, 0.94)] and Average Precision (AP) of 0.59 [0.51, 0.69] for mortality, and decreased performance predicting simultaneous mortality and morbidity (0.83 [0.80, 0.86] and 0.59 [0.51, 0.64]); at a set specificity of 0.995, positive predictive value (PPV) was 0.79 for mortality, and 0.88 for mortality and morbidity. By comparison, for mortality, the TOPICC logistic regression had AUROC of 0.90 [0.84, 0.93], but substantially inferior AP of 0.49 [0.35, 0.56] and PPV of 0.60 at specificity 0.995. Feature importance analysis showed that pupillary non-reactivity, Glasgow Coma Scale, and temperature were the most contributory vital signs, and acidosis and coagulopathy the most important laboratory values. The FHIR application provided a simulated demonstration of real-time health record query and model deployment. ConclusionsPICU patients with suspected acute neurological injury have higher mortality and morbidity. Our machine learning approach independently identified previously-known causes of secondary brain injury. Advanced modeling achieves improved positive predictive value in this important population compared to published models, providing a stepping stone in the path to deploying explainable models as interoperable bedside decision-support tools.
引用
收藏
页数:11
相关论文
共 50 条
  • [41] A Validation Study of Administrative Health Care Data to Detect Acute Kidney Injury in the Pediatric Intensive Care Unit
    D'Arienzo, David
    Hessey, Erin
    Ali, Rami
    Perreault, Sylvie
    Samuel, Susan
    Roy, Louise
    Lacroix, Jacques
    Jouvet, Philippe
    Morissette, Genevieve
    Dorais, Marc
    Lafrance, Jean-Philippe
    Phan, Veronique
    Pizzi, Michael
    Chanchlani, Rahul
    Zappitelli, Michael
    CANADIAN JOURNAL OF KIDNEY HEALTH AND DISEASE, 2019, 6
  • [42] Machine Learning-Based Mortality Prediction for Acute Gastrointestinal Bleeding Patients Admitted to Intensive Care Unit
    Liu, Zhou
    Zhang, Liang
    Jiang, Gui-jun
    Chen, Qian-qian
    Hou, Yan-guang
    Wu, Wei
    Malik, Muskaan
    Li, Guang
    Zhan, Li-ying
    CURRENT MEDICAL SCIENCE, 2025, 45 (01) : 70 - 81
  • [43] Machine Learning-Based Mortality Prediction for Acute Gastrointestinal Bleeding Patients Admitted to Intensive Care Unit
    Liu, Zhou
    Zhang, Liang
    Jiang, Gui-jun
    Chen, Qian-qian
    Hou, Yan-guang
    Wu, Wei
    Malik, Muskaan
    Li, Guang
    Zhan, Li-ying
    CURRENT MEDICAL SCIENCE, 2025,
  • [44] Machine Learning-Based Mortality Prediction for Acute Gastrointestinal Bleeding Patients Admitted to Intensive Care Unit
    Zhou Liu
    Liang Zhang
    Gui-jun Jiang
    Qian-qian Chen
    Yan-guang Hou
    Wei Wu
    Muskaan Malik
    Guang Li
    Li-ying Zhan
    Current Medical Science, 2025, 45 (1) : 70 - 81
  • [45] Critical analysis of acute kidney injury in pediatric COVID-19 patients in the intensive care unit
    Rupesh Raina
    Ronith Chakraborty
    Isabelle Mawby
    Nirav Agarwal
    Sidharth Sethi
    Michael Forbes
    Pediatric Nephrology, 2021, 36 : 2627 - 2638
  • [46] Critical analysis of acute kidney injury in pediatric COVID-19 patients in the intensive care unit
    Raina, Rupesh
    Chakraborty, Ronith
    Mawby, Isabelle
    Agarwal, Nirav
    Sethi, Sidharth
    Forbes, Michael
    PEDIATRIC NEPHROLOGY, 2021, 36 (09) : 2627 - 2638
  • [47] Transfusion Related Acute Lung Injury (TRALI) in Pediatric Intensive Care Unit: A Multicenter Cohort Study
    Emeriaud, G.
    Thibodeau, G. Du Pont
    Menon, K.
    Fontela, P.
    Gilfoyle, E.
    Hutchison, J.
    Robitaille, N.
    Gauvin, F.
    Lacroix, J.
    AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE, 2018, 197
  • [48] Prediction of persistent acute kidney injury in postoperative intensive care unit patients using integrated machine learning: a retrospective cohort study
    Xuandong Jiang
    Yongxia Hu
    Shan Guo
    Chaojian Du
    Xuping Cheng
    Scientific Reports, 12
  • [49] Prediction of persistent acute kidney injury in postoperative intensive care unit patients using integrated machine learning: a retrospective cohort study
    Jiang, Xuandong
    Hu, Yongxia
    Guo, Shan
    Du, Chaojian
    Cheng, Xuping
    SCIENTIFIC REPORTS, 2022, 12 (01)
  • [50] ShockSurv: A machine learning model to accurately predict 28-day mortality for septic shock patients in the intensive care unit
    Zheng, Fudan
    Wang, Luhao
    Pang, Yuxian
    Chen, Zhiguang
    Lu, Yutong
    Yang, Yuedong
    Wu, Jianfeng
    BIOMEDICAL SIGNAL PROCESSING AND CONTROL, 2023, 86