Skip to content

Towards a decision support tool for intensive care discharge: Machine learning algorithm development using electronic healthcare data from MIMIC-III and Bristol, UK

Research output: Contribution to journalArticle

Standard

Harvard

APA

Vancouver

Author

Bibtex

@article{3cf81cbaf28b48e6bf3842869bf58c2c,
title = "Towards a decision support tool for intensive care discharge: Machine learning algorithm development using electronic healthcare data from MIMIC-III and Bristol, UK",
abstract = "ObjectiveThe primary objective is to develop an automated method for detecting patients that are ready for discharge from intensive care.DesignWe used two datasets of routinely collected patient data to test and improve upon a set of previously proposed discharge criteria.SettingBristol Royal Infirmary general intensive care unit (GICU).PatientsTwo cohorts derived from historical datasets: 1870 intensive care patients from GICU in Bristol, and 7592 from MIMIC-III (a publicly available intensive care dataset).ResultsIn both cohorts few successfully discharged patients met all of the discharge criteria. Both a random forest and a logistic classifier, trained using multiple-source cross-validation, demonstrated improved performance over the original criteria and generalised well between the cohorts. The classifiers showed good agreement on which features were most predictive of readiness-fordischarge, and these were generally consistent with clinical experience. By weighting the discharge criteria according to feature importance from the logistic model we showed improved performance over the original criteria, while retaining good interpretability.ConclusionsOur findings indicate the feasibility of the proposed approach to ready-for-discharge classification, which could complement other risk models of specific adverse outcomes in a future decision support system. Avenues for improvement to produce a clinically useful tool are identified.",
keywords = "patient discharge, machine learning, clinical decision support, critical care, patient flow",
author = "Chris McWilliams and Daniel Lawson and Raul Santos-Rodriguez and Iain Gilchrist and Alan Champneys and Timothy Gould and Matthew Thomas and Chris Bourdeaux",
year = "2019",
month = "3",
day = "7",
doi = "10.1136/bmjopen-2018-025925",
language = "English",
volume = "9",
journal = "BMJ Open",
issn = "2044-6055",
publisher = "BMJ Publishing Group",
number = "3",

}

RIS - suitable for import to EndNote

TY - JOUR

T1 - Towards a decision support tool for intensive care discharge

T2 - Machine learning algorithm development using electronic healthcare data from MIMIC-III and Bristol, UK

AU - McWilliams, Chris

AU - Lawson, Daniel

AU - Santos-Rodriguez, Raul

AU - Gilchrist, Iain

AU - Champneys, Alan

AU - Gould, Timothy

AU - Thomas, Matthew

AU - Bourdeaux, Chris

PY - 2019/3/7

Y1 - 2019/3/7

N2 - ObjectiveThe primary objective is to develop an automated method for detecting patients that are ready for discharge from intensive care.DesignWe used two datasets of routinely collected patient data to test and improve upon a set of previously proposed discharge criteria.SettingBristol Royal Infirmary general intensive care unit (GICU).PatientsTwo cohorts derived from historical datasets: 1870 intensive care patients from GICU in Bristol, and 7592 from MIMIC-III (a publicly available intensive care dataset).ResultsIn both cohorts few successfully discharged patients met all of the discharge criteria. Both a random forest and a logistic classifier, trained using multiple-source cross-validation, demonstrated improved performance over the original criteria and generalised well between the cohorts. The classifiers showed good agreement on which features were most predictive of readiness-fordischarge, and these were generally consistent with clinical experience. By weighting the discharge criteria according to feature importance from the logistic model we showed improved performance over the original criteria, while retaining good interpretability.ConclusionsOur findings indicate the feasibility of the proposed approach to ready-for-discharge classification, which could complement other risk models of specific adverse outcomes in a future decision support system. Avenues for improvement to produce a clinically useful tool are identified.

AB - ObjectiveThe primary objective is to develop an automated method for detecting patients that are ready for discharge from intensive care.DesignWe used two datasets of routinely collected patient data to test and improve upon a set of previously proposed discharge criteria.SettingBristol Royal Infirmary general intensive care unit (GICU).PatientsTwo cohorts derived from historical datasets: 1870 intensive care patients from GICU in Bristol, and 7592 from MIMIC-III (a publicly available intensive care dataset).ResultsIn both cohorts few successfully discharged patients met all of the discharge criteria. Both a random forest and a logistic classifier, trained using multiple-source cross-validation, demonstrated improved performance over the original criteria and generalised well between the cohorts. The classifiers showed good agreement on which features were most predictive of readiness-fordischarge, and these were generally consistent with clinical experience. By weighting the discharge criteria according to feature importance from the logistic model we showed improved performance over the original criteria, while retaining good interpretability.ConclusionsOur findings indicate the feasibility of the proposed approach to ready-for-discharge classification, which could complement other risk models of specific adverse outcomes in a future decision support system. Avenues for improvement to produce a clinically useful tool are identified.

KW - patient discharge

KW - machine learning

KW - clinical decision support

KW - critical care

KW - patient flow

UR - http://www.scopus.com/inward/record.url?scp=85062609948&partnerID=8YFLogxK

U2 - 10.1136/bmjopen-2018-025925

DO - 10.1136/bmjopen-2018-025925

M3 - Article

VL - 9

JO - BMJ Open

JF - BMJ Open

SN - 2044-6055

IS - 3

M1 - e025925

ER -