Predicting Prolonged Length of ICU Stay through Machine Learning

被引:20
作者
Wu, Jingyi [1 ,2 ]
Lin, Yu [3 ]
Li, Pengfei [2 ]
Hu, Yonghua [4 ,5 ]
Zhang, Luxia [1 ,2 ,6 ]
Kong, Guilan [1 ,2 ]
机构
[1] Peking Univ, Natl Inst Hlth Data Sci, Beijing 100191, Peoples R China
[2] Peking Univ, Adv Inst Informat Technol, Hangzhou 311215, Peoples R China
[3] Chinese Univ Hong Kong, LKS Inst Hlth Sci, Dept Med & Therapeut, Hong Kong, Peoples R China
[4] Peking Univ, Sch Publ Hlth, Dept Epidemiol & Biostat, Beijing 100191, Peoples R China
[5] Peking Univ, Med Informat Ctr, Beijing 100191, Peoples R China
[6] Peking Univ First Hosp, Peking Univ Inst Nephrol, Div Renal, Dept Med, Beijing 100034, Peoples R China
基金
中国国家自然科学基金;
关键词
prolonged length of ICU stay; machine learning; clinical decision rules; medical informatics; INTENSIVE-CARE-UNIT; OF-STAY; MORTALITY PREDICTION; HOSPITAL MORTALITY; MODELS; SCORE; VALIDATION; ADULTS;
D O I
10.3390/diagnostics11122242
中图分类号
R5 [内科学];
学科分类号
1002 ; 100201 ;
摘要
This study aimed to construct machine learning (ML) models for predicting prolonged length of stay (pLOS) in intensive care units (ICU) among general ICU patients. A multicenter database called eICU (Collaborative Research Database) was used for model derivation and internal validation, and the Medical Information Mart for Intensive Care (MIMIC) III database was used for external validation. We used four different ML methods (random forest, support vector machine, deep learning, and gradient boosting decision tree (GBDT)) to develop prediction models. The prediction performance of the four models were compared with the customized simplified acute physiology score (SAPS) II. The area under the receiver operation characteristic curve (AUROC), area under the precision-recall curve (AUPRC), estimated calibration index (ECI), and Brier score were used to measure performance. In internal validation, the GBDT model achieved the best overall performance (Brier score, 0.164), discrimination (AUROC, 0.742; AUPRC, 0.537), and calibration (ECI, 8.224). In external validation, the GBDT model also achieved the best overall performance (Brier score, 0.166), discrimination (AUROC, 0.747; AUPRC, 0.536), and calibration (ECI, 8.294). External validation showed that the calibration curve of the GBDT model was an optimal fit, and four ML models outperformed the customized SAPS II model. The GBDT-based pLOS-ICU prediction model had the best prediction performance among the five models on both internal and external datasets. Furthermore, it has the potential to assist ICU physicians to identify patients with pLOS-ICU risk and provide appropriate clinical interventions to improve patient outcomes.
引用
收藏
页数:18
相关论文
共 76 条
[1]   Low Glasgow Coma Score in Traumatic Intracranial Hemorrhage Predicts Development of Cerebral Vasospasm [J].
Al-Mufti, Fawaz ;
Amuluru, Krishna ;
Lander, Megan ;
Mathew, Melvin ;
El-Ghanem, Mohammad ;
Nuoman, Rolla ;
Park, Seami ;
Patel, Vikas ;
Singh, Inder Paul ;
Gupta, Gaurav ;
Gandhi, Chirag D. .
WORLD NEUROSURGERY, 2018, 120 :E68-E71
[2]   LIBSVM: A Library for Support Vector Machines [J].
Chang, Chih-Chung ;
Lin, Chih-Jen .
ACM TRANSACTIONS ON INTELLIGENT SYSTEMS AND TECHNOLOGY, 2011, 2 (03)
[3]  
Dzaharudin F., 2020, MORTALITY PREDICTION
[4]   Severity scoring for prognostication in patients with severe acute pancreatitis - Comparative analysis of the ranson score and the APACHE III score [J].
Eachempati, SR ;
Hydo, LJ ;
Barie, PS .
ARCHIVES OF SURGERY, 2002, 137 (06) :730-736
[5]   Evaluation of the PaO2/FiO2 ratio after cardiac surgery as a predictor of outcome during hospital stay [J].
Esteve, Francisco ;
Lopez-Delgado, Juan C. ;
Javierre, Casimiro ;
Skaltsa, Konstantina ;
Carrio, Maria L. L. ;
Rodriguez-Castro, David ;
Torrado, Herminia ;
Farrero, Elisabet ;
Diaz-Prieto, Antonio ;
Ventura, Josep L. L. ;
Manez, Rafael .
BMC ANESTHESIOLOGY, 2014, 14
[6]   Prediction Models for Prolonged Intensive Care Unit Stay After Cardiac Surgery Systematic Review and Validation Study [J].
Ettema, Roelof G. A. ;
Peelen, Linda M. ;
Schuurmans, Marieke J. ;
Nierich, Arno P. ;
Kalkman, Cor J. ;
Moons, Karel G. M. .
CIRCULATION, 2010, 122 (07) :682-U16
[7]  
Evans J, 2018, CAN J ANESTH, V65, P627, DOI 10.1007/s12630-018-1087-1
[8]   Elevated blood urea nitrogen is an independent risk factor of prolonged intensive care unit stay due to acute necrotizing pancreatitis [J].
Faisst, Maik ;
Wellner, Ulrich F. ;
Utzolino, Stefan ;
Hopt, Ulrich T. ;
Keck, Tobias .
JOURNAL OF CRITICAL CARE, 2010, 25 (01) :105-111
[9]   Common intensive care scoring systems do not outperform age and glasgow coma scale score in predicting mid-term mortality in patients with spontaneous intracerebral hemorrhage treated in the intensive care unit [J].
Fallenius, Marika ;
Skrifvars, Markus B. ;
Reinikainen, Matti ;
Bendel, Stepani ;
Raj, Rahul .
SCANDINAVIAN JOURNAL OF TRAUMA RESUSCITATION & EMERGENCY MEDICINE, 2017, 24
[10]   Speed up grid-search for parameter selection of support vector machines [J].
Fayed, Hatem A. ;
Atiya, Amir F. .
APPLIED SOFT COMPUTING, 2019, 80 :202-210