Literature DB >> 34916699

[Predicting prolonged length of intensive care unit stay via machine learning].

J Y Wu1,2, Y Lin1, K Lin1, Y H Hu1,3, G L Kong2,4.   

Abstract

OBJECTIVE: To construct length of intensive care unit (ICU) stay (LOS-ICU) prediction models for ICU patients, based on three machine learning models support vector machine (SVM), classification and regression tree (CART), and random forest (RF), and to compare the prediction perfor-mance of the three machine learning models with the customized simplified acute physiology score Ⅱ(SAPS-Ⅱ) model.
METHODS: We used medical information mart for intensive care (MIMIC)-Ⅲ database for model development and validation. The primary outcome was prolonged LOS-ICU(pLOS-ICU), defined as longer than the third quartile of patients' LOS-ICU in the studied dataset. The recursive feature elimination method was used to do feature selection for three machine learning models. We utilized 5-fold cross validation to evaluate model prediction performance. The Brier value, area under the receiver operation characteristic curve (AUROC), and estimated calibration index (ECI) were used as perfor-mance measures. Performances of the four models were compared, and performance differences between the models were assessed using two-sided t test. The model with the best prediction performance was employed to generate variable importance ranking, and the identified top five important predictors were pre-sented.
RESULTS: The final cohort in our study consisted of 40 200 eligible ICU patients, of whom 23.7% were with pLOS-ICU. The proportion of the male patients was 57.6%, and the age of all the ICU patients was (61.9±16.5) years.Results showed that the three machine learning models outperformed the customized SAPS-Ⅱ model in terms of all the performance measures with statistical significance (P < 0.01). Among the three machine learning models, the RF model achieved the best overall performance (Brier value, 0.145), discrimination (AUROC, 0.770) and calibration (ECI, 7.259). The calibration curve showed that the RF model slightly overestimated the risk of pLOS-ICU in high-risk ICU patients, but underestimated the risk of pLOS-ICU in low-risk ICU patients. Top five important predictors for pLOS-ICU identified by the RF model included age, heart rate, systolic blood pressure, body tempe-rature, and ratio of arterial oxygen tension to the fraction of inspired oxygen(PaO2/FiO2).
CONCLUSION: The RF algorithm-based pLOS-ICU prediction model had a best prediction performance in this study. It lays a foundation for future application of the RF-based pLOS-ICU prediction model in ICU clinical practice.

Entities:  

Keywords:  Intensive care units; Length of stay; Machine learning; Random forest; Simplified acute physiology score

Mesh:

Year:  2021        PMID: 34916699      PMCID: PMC8695140     

Source DB:  PubMed          Journal:  Beijing Da Xue Xue Bao Yi Xue Ban        ISSN: 1671-167X


  30 in total

1.  Predicting in-hospital mortality of patients with acute kidney injury in the ICU using random forest model.

Authors:  Ke Lin; Yonghua Hu; Guilan Kong
Journal:  Int J Med Inform       Date:  2019-02-12       Impact factor: 4.046

2.  Device-associated infection rates, device use, length of stay, and mortality in intensive care units of 4 Chinese hospitals: International Nosocomial Control Consortium findings.

Authors:  Bijie Hu; Lili Tao; Victor D Rosenthal; Kun Liu; Yang Yun; Yao Suo; Xiandong Gao; Ruisheng Li; Danxia Su; Hungmei Wang; Chunxia Hao; Wei Pan; Catherine L Saunders
Journal:  Am J Infect Control       Date:  2012-10-05       Impact factor: 2.918

3.  Heart rate before ICU discharge: a simple and readily available predictor of short- and long-term mortality from critical illness.

Authors:  Wilhelm Grander; Kathrin Müllauer; Bernhard Koller; Herbert Tilg; Martin Dünser
Journal:  Clin Res Cardiol       Date:  2013-04-27       Impact factor: 5.460

4.  [A model based on random forests in prediction of 28-day prognosis in patients with severe sepsis/septic shock].

Authors:  Yang Wang; Shangzhong Chen; Caibao Hu; Changqin Chen; Jing Yan; Guolong Cai
Journal:  Zhonghua Wei Zhong Bing Ji Jiu Yi Xue       Date:  2017-12

5.  Elevated body temperature independently contributes to increased length of stay in neurologic intensive care unit patients.

Authors:  Michael N Diringer; Nancy L Reaven; Susan E Funk; Gwen C Uman
Journal:  Crit Care Med       Date:  2004-07       Impact factor: 7.598

6.  What is the importance of age on treatment of the elderly in the intensive care unit?

Authors:  C Brandberg; H Blomqvist; M Jirwe
Journal:  Acta Anaesthesiol Scand       Date:  2013-02-04       Impact factor: 2.105

7.  ICU physicians are unable to accurately predict length of stay at admission: a prospective study.

Authors:  Antonio Paulo Nassar; Pedro Caruso
Journal:  Int J Qual Health Care       Date:  2015-12-13       Impact factor: 2.038

8.  Data mining in the Life Sciences with Random Forest: a walk in the park or lost in the jungle?

Authors:  Wouter G Touw; Jumamurat R Bayjanov; Lex Overmars; Lennart Backus; Jos Boekhorst; Michiel Wels; Sacha A F T van Hijum
Journal:  Brief Bioinform       Date:  2012-07-10       Impact factor: 11.622

9.  Heterogeneous ensemble approach with discriminative features and modified-SMOTEbagging for pre-miRNA classification.

Authors:  Supatcha Lertampaiporn; Chinae Thammarongtham; Chakarida Nukoolkit; Boonserm Kaewkamnerdpong; Marasri Ruengjitchatchawalya
Journal:  Nucleic Acids Res       Date:  2012-09-24       Impact factor: 16.971

10.  Length of Hospital Stay Prediction at the Admission Stage for Cardiology Patients Using Artificial Neural Network.

Authors:  Pei-Fang Jennifer Tsai; Po-Chia Chen; Yen-You Chen; Hao-Yuan Song; Hsiu-Mei Lin; Fu-Man Lin; Qiou-Pieng Huang
Journal:  J Healthc Eng       Date:  2016       Impact factor: 2.682

View more

北京卡尤迪生物科技股份有限公司 © 2022-2023.