Improving Mortality Models in the ICU with High-Frequency Data

James Todd, Adrian Gepp, Brent Richards, Bruce J Vanstone

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Background: Assessment of the performance of Intensive Care Units (ICU) is of vital importance for an effective
healthcare system. Such assessment ensures that the limited resources of the healthcare system are allocated
where they are most needed. Severity scoring systems are employed for this purpose and improving these systems
is a continuing area of research which has focused on the use of more complex techniques and new
variables.
Objectives: This paper investigates whether scoring systems could be improved through use of metrics which
better summarise the high frequency data collected by automated systems for patients in the ICU.
Methods and Data: 3128 admissions to the Gold Coast University Hospital ICU are used to construct three logistic
regressions based on the most widely used scoring system (APACHE III) to compare performance with and
without predictors leveraging available high frequency information. Performance is assessed based on model
accuracy, calibration, and discrimination. High frequency information was considered for existing pulse and
mean arterial pressure physiology fields and resulting models compared against a baseline logistic regression
using only APACHE III physiology variables.
Results: Model discrimination and accuracy were better for models which included high frequency predictors,
with calibration remaining good in all cases. The most influential high frequency summaries were the number of
turning points in a patient’s mean arterial pressure or pulse in the first 24 h of ICU admission.
Conclusions: The findings indicate that scoring systems can be improved by better accounting for high frequency
data.
Original languageEnglish
Pages (from-to)318-323
Number of pages6
JournalInternational Journal of Medical Informatics
Volume129
Early online date13 Jul 2019
DOIs
Publication statusPublished - Sep 2019

Fingerprint

Intensive Care Units
APACHE
Mortality
Calibration
Arterial Pressure
Ghana
Delivery of Health Care
Research

Cite this

@article{3f19656654ea4d07bc5b96f33f4f0b91,
title = "Improving Mortality Models in the ICU with High-Frequency Data",
abstract = "Background: Assessment of the performance of Intensive Care Units (ICU) is of vital importance for an effectivehealthcare system. Such assessment ensures that the limited resources of the healthcare system are allocatedwhere they are most needed. Severity scoring systems are employed for this purpose and improving these systemsis a continuing area of research which has focused on the use of more complex techniques and newvariables.Objectives: This paper investigates whether scoring systems could be improved through use of metrics whichbetter summarise the high frequency data collected by automated systems for patients in the ICU.Methods and Data: 3128 admissions to the Gold Coast University Hospital ICU are used to construct three logisticregressions based on the most widely used scoring system (APACHE III) to compare performance with andwithout predictors leveraging available high frequency information. Performance is assessed based on modelaccuracy, calibration, and discrimination. High frequency information was considered for existing pulse andmean arterial pressure physiology fields and resulting models compared against a baseline logistic regressionusing only APACHE III physiology variables.Results: Model discrimination and accuracy were better for models which included high frequency predictors,with calibration remaining good in all cases. The most influential high frequency summaries were the number ofturning points in a patient’s mean arterial pressure or pulse in the first 24 h of ICU admission.Conclusions: The findings indicate that scoring systems can be improved by better accounting for high frequencydata.",
author = "James Todd and Adrian Gepp and Brent Richards and Vanstone, {Bruce J}",
year = "2019",
month = "9",
doi = "10.1016/j.ijmedinf.2019.07.002",
language = "English",
volume = "129",
pages = "318--323",
journal = "International Journal of Bio-Medical Computing",
issn = "1386-5056",
publisher = "Elsevier",

}

Improving Mortality Models in the ICU with High-Frequency Data. / Todd, James; Gepp, Adrian; Richards, Brent; Vanstone, Bruce J.

In: International Journal of Medical Informatics, Vol. 129, 09.2019, p. 318-323.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Improving Mortality Models in the ICU with High-Frequency Data

AU - Todd, James

AU - Gepp, Adrian

AU - Richards, Brent

AU - Vanstone, Bruce J

PY - 2019/9

Y1 - 2019/9

N2 - Background: Assessment of the performance of Intensive Care Units (ICU) is of vital importance for an effectivehealthcare system. Such assessment ensures that the limited resources of the healthcare system are allocatedwhere they are most needed. Severity scoring systems are employed for this purpose and improving these systemsis a continuing area of research which has focused on the use of more complex techniques and newvariables.Objectives: This paper investigates whether scoring systems could be improved through use of metrics whichbetter summarise the high frequency data collected by automated systems for patients in the ICU.Methods and Data: 3128 admissions to the Gold Coast University Hospital ICU are used to construct three logisticregressions based on the most widely used scoring system (APACHE III) to compare performance with andwithout predictors leveraging available high frequency information. Performance is assessed based on modelaccuracy, calibration, and discrimination. High frequency information was considered for existing pulse andmean arterial pressure physiology fields and resulting models compared against a baseline logistic regressionusing only APACHE III physiology variables.Results: Model discrimination and accuracy were better for models which included high frequency predictors,with calibration remaining good in all cases. The most influential high frequency summaries were the number ofturning points in a patient’s mean arterial pressure or pulse in the first 24 h of ICU admission.Conclusions: The findings indicate that scoring systems can be improved by better accounting for high frequencydata.

AB - Background: Assessment of the performance of Intensive Care Units (ICU) is of vital importance for an effectivehealthcare system. Such assessment ensures that the limited resources of the healthcare system are allocatedwhere they are most needed. Severity scoring systems are employed for this purpose and improving these systemsis a continuing area of research which has focused on the use of more complex techniques and newvariables.Objectives: This paper investigates whether scoring systems could be improved through use of metrics whichbetter summarise the high frequency data collected by automated systems for patients in the ICU.Methods and Data: 3128 admissions to the Gold Coast University Hospital ICU are used to construct three logisticregressions based on the most widely used scoring system (APACHE III) to compare performance with andwithout predictors leveraging available high frequency information. Performance is assessed based on modelaccuracy, calibration, and discrimination. High frequency information was considered for existing pulse andmean arterial pressure physiology fields and resulting models compared against a baseline logistic regressionusing only APACHE III physiology variables.Results: Model discrimination and accuracy were better for models which included high frequency predictors,with calibration remaining good in all cases. The most influential high frequency summaries were the number ofturning points in a patient’s mean arterial pressure or pulse in the first 24 h of ICU admission.Conclusions: The findings indicate that scoring systems can be improved by better accounting for high frequencydata.

UR - http://www.scopus.com/inward/record.url?scp=85068924127&partnerID=8YFLogxK

U2 - 10.1016/j.ijmedinf.2019.07.002

DO - 10.1016/j.ijmedinf.2019.07.002

M3 - Article

VL - 129

SP - 318

EP - 323

JO - International Journal of Bio-Medical Computing

JF - International Journal of Bio-Medical Computing

SN - 1386-5056

ER -