A machine learning approach for predicting urine output after fluid administration

Pei Chen Lin, Hsu Cheng Huang, Matthieu Komorowski, Wei Kai Lin, Chun Min Chang, Kuan Ta Chen, Yu Chuan Li, Ming Chin Lin

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Background and objective: To develop a machine learning model to predict urine output (UO)in sepsis patients after fluid resuscitation. Methods: We identified sepsis patients in the Multiparameter Intelligent Monitoring in Intensive Care-III v1.4 database according to the Sepsis-3 criteria. We focused on two outcomes: whether the UO decreased after fluid administration and whether oliguria (defined as UO less than the threshold of 0.5 mL/kg/h)developed. A gradient tree-based machine learning model implemented with an eXtreme Gradient Boosting algorithm was used to integrate relevant physiological parameters for predicting the aforementioned outcomes. A confusion matrix was computed. Results: A total of 232,929 events in 19,275 patients were included. Using decreased UO as the outcome measure, the optimal model achieved an area under the curve (AUC)of 0.86; for predicting oliguria, most models achieved an AUC greater than 0.86, and the highest sensitivity was 92.2% when the model was applied to patients with baseline oliguria. Conclusions: Machine learning could help clinicians evaluate fluid status in sepsis patients after fluid administration, thus preventing fluid overload-related complications.

Original languageEnglish
Pages (from-to)155-159
Number of pages5
JournalComputer Methods and Programs in Biomedicine
Volume177
DOIs
Publication statusPublished - Aug 1 2019

Fingerprint

Learning systems
Oliguria
Urine
Sepsis
Fluids
Area Under Curve
Resuscitation
Critical Care
Outcome Assessment (Health Care)
Machine Learning
Databases
Monitoring

Keywords

  • Clinical decision support
  • Electronic health records
  • Fluid resuscitation
  • Machine learning
  • Prediction
  • Sepsis

ASJC Scopus subject areas

  • Software
  • Computer Science Applications
  • Health Informatics

Cite this

A machine learning approach for predicting urine output after fluid administration. / Lin, Pei Chen; Huang, Hsu Cheng; Komorowski, Matthieu; Lin, Wei Kai; Chang, Chun Min; Chen, Kuan Ta; Li, Yu Chuan; Lin, Ming Chin.

In: Computer Methods and Programs in Biomedicine, Vol. 177, 01.08.2019, p. 155-159.

Research output: Contribution to journalArticle

Lin, Pei Chen ; Huang, Hsu Cheng ; Komorowski, Matthieu ; Lin, Wei Kai ; Chang, Chun Min ; Chen, Kuan Ta ; Li, Yu Chuan ; Lin, Ming Chin. / A machine learning approach for predicting urine output after fluid administration. In: Computer Methods and Programs in Biomedicine. 2019 ; Vol. 177. pp. 155-159.
@article{9d60be7f688e4f37b6e51fb581986af9,
title = "A machine learning approach for predicting urine output after fluid administration",
abstract = "Background and objective: To develop a machine learning model to predict urine output (UO)in sepsis patients after fluid resuscitation. Methods: We identified sepsis patients in the Multiparameter Intelligent Monitoring in Intensive Care-III v1.4 database according to the Sepsis-3 criteria. We focused on two outcomes: whether the UO decreased after fluid administration and whether oliguria (defined as UO less than the threshold of 0.5 mL/kg/h)developed. A gradient tree-based machine learning model implemented with an eXtreme Gradient Boosting algorithm was used to integrate relevant physiological parameters for predicting the aforementioned outcomes. A confusion matrix was computed. Results: A total of 232,929 events in 19,275 patients were included. Using decreased UO as the outcome measure, the optimal model achieved an area under the curve (AUC)of 0.86; for predicting oliguria, most models achieved an AUC greater than 0.86, and the highest sensitivity was 92.2{\%} when the model was applied to patients with baseline oliguria. Conclusions: Machine learning could help clinicians evaluate fluid status in sepsis patients after fluid administration, thus preventing fluid overload-related complications.",
keywords = "Clinical decision support, Electronic health records, Fluid resuscitation, Machine learning, Prediction, Sepsis",
author = "Lin, {Pei Chen} and Huang, {Hsu Cheng} and Matthieu Komorowski and Lin, {Wei Kai} and Chang, {Chun Min} and Chen, {Kuan Ta} and Li, {Yu Chuan} and Lin, {Ming Chin}",
year = "2019",
month = "8",
day = "1",
doi = "10.1016/j.cmpb.2019.05.009",
language = "English",
volume = "177",
pages = "155--159",
journal = "Computer Methods and Programs in Biomedicine",
issn = "0169-2607",
publisher = "Elsevier Ireland Ltd",

}

TY - JOUR

T1 - A machine learning approach for predicting urine output after fluid administration

AU - Lin, Pei Chen

AU - Huang, Hsu Cheng

AU - Komorowski, Matthieu

AU - Lin, Wei Kai

AU - Chang, Chun Min

AU - Chen, Kuan Ta

AU - Li, Yu Chuan

AU - Lin, Ming Chin

PY - 2019/8/1

Y1 - 2019/8/1

N2 - Background and objective: To develop a machine learning model to predict urine output (UO)in sepsis patients after fluid resuscitation. Methods: We identified sepsis patients in the Multiparameter Intelligent Monitoring in Intensive Care-III v1.4 database according to the Sepsis-3 criteria. We focused on two outcomes: whether the UO decreased after fluid administration and whether oliguria (defined as UO less than the threshold of 0.5 mL/kg/h)developed. A gradient tree-based machine learning model implemented with an eXtreme Gradient Boosting algorithm was used to integrate relevant physiological parameters for predicting the aforementioned outcomes. A confusion matrix was computed. Results: A total of 232,929 events in 19,275 patients were included. Using decreased UO as the outcome measure, the optimal model achieved an area under the curve (AUC)of 0.86; for predicting oliguria, most models achieved an AUC greater than 0.86, and the highest sensitivity was 92.2% when the model was applied to patients with baseline oliguria. Conclusions: Machine learning could help clinicians evaluate fluid status in sepsis patients after fluid administration, thus preventing fluid overload-related complications.

AB - Background and objective: To develop a machine learning model to predict urine output (UO)in sepsis patients after fluid resuscitation. Methods: We identified sepsis patients in the Multiparameter Intelligent Monitoring in Intensive Care-III v1.4 database according to the Sepsis-3 criteria. We focused on two outcomes: whether the UO decreased after fluid administration and whether oliguria (defined as UO less than the threshold of 0.5 mL/kg/h)developed. A gradient tree-based machine learning model implemented with an eXtreme Gradient Boosting algorithm was used to integrate relevant physiological parameters for predicting the aforementioned outcomes. A confusion matrix was computed. Results: A total of 232,929 events in 19,275 patients were included. Using decreased UO as the outcome measure, the optimal model achieved an area under the curve (AUC)of 0.86; for predicting oliguria, most models achieved an AUC greater than 0.86, and the highest sensitivity was 92.2% when the model was applied to patients with baseline oliguria. Conclusions: Machine learning could help clinicians evaluate fluid status in sepsis patients after fluid administration, thus preventing fluid overload-related complications.

KW - Clinical decision support

KW - Electronic health records

KW - Fluid resuscitation

KW - Machine learning

KW - Prediction

KW - Sepsis

UR - http://www.scopus.com/inward/record.url?scp=85067919319&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85067919319&partnerID=8YFLogxK

U2 - 10.1016/j.cmpb.2019.05.009

DO - 10.1016/j.cmpb.2019.05.009

M3 - Article

VL - 177

SP - 155

EP - 159

JO - Computer Methods and Programs in Biomedicine

JF - Computer Methods and Programs in Biomedicine

SN - 0169-2607

ER -