A minimal set of physiomarkers in continuous high frequency data streams predict adult sepsis onset earlier

Franco van Wyk, Anahita Khojandi, Akram Mohammed, Edmon Begoli, Robert Davis, Rishikesan Kamaleswaran

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Purpose: Sepsis is a life-threatening condition with high mortality rates and expensive treatment costs. To improve short- and long-term outcomes, it is critical to detect at-risk sepsis patients at an early stage. Methods: A data-set consisting of high-frequency physiological data from 1161 critically ill patients was analyzed. 377 patients had developed sepsis, and had data at least 3 h prior to the onset of sepsis. A random forest classifier was trained to discriminate between sepsis and non-sepsis patients in real-time using a total of 132 features extracted from a moving time-window. The model was trained on 80% of the patients and was tested on the remaining 20% of the patients, for two observational periods of lengths 3 and 6 h prior to onset. Results: The model that used continuous physiological data alone resulted in sensitivity and F1 score of up to 80% and 67% one hour before sepsis onset. On average, these models were able to predict sepsis 294.19 ± 6.50 min (5 h) before the onset. Conclusions: The use of machine learning algorithms on continuous streams of physiological data can allow for early identification of at-risk patients in real-time with high accuracy.

Original languageEnglish (US)
Pages (from-to)55-62
Number of pages8
JournalInternational Journal of Medical Informatics
Volume122
DOIs
StatePublished - Feb 1 2019

Fingerprint

Sepsis
Critical Illness
Health Care Costs
Mortality

All Science Journal Classification (ASJC) codes

  • Health Informatics

Cite this

A minimal set of physiomarkers in continuous high frequency data streams predict adult sepsis onset earlier. / van Wyk, Franco; Khojandi, Anahita; Mohammed, Akram; Begoli, Edmon; Davis, Robert; Kamaleswaran, Rishikesan.

In: International Journal of Medical Informatics, Vol. 122, 01.02.2019, p. 55-62.

Research output: Contribution to journalArticle

@article{827857de12c54f5b92f202e88c7231e8,
title = "A minimal set of physiomarkers in continuous high frequency data streams predict adult sepsis onset earlier",
abstract = "Purpose: Sepsis is a life-threatening condition with high mortality rates and expensive treatment costs. To improve short- and long-term outcomes, it is critical to detect at-risk sepsis patients at an early stage. Methods: A data-set consisting of high-frequency physiological data from 1161 critically ill patients was analyzed. 377 patients had developed sepsis, and had data at least 3 h prior to the onset of sepsis. A random forest classifier was trained to discriminate between sepsis and non-sepsis patients in real-time using a total of 132 features extracted from a moving time-window. The model was trained on 80{\%} of the patients and was tested on the remaining 20{\%} of the patients, for two observational periods of lengths 3 and 6 h prior to onset. Results: The model that used continuous physiological data alone resulted in sensitivity and F1 score of up to 80{\%} and 67{\%} one hour before sepsis onset. On average, these models were able to predict sepsis 294.19 ± 6.50 min (5 h) before the onset. Conclusions: The use of machine learning algorithms on continuous streams of physiological data can allow for early identification of at-risk patients in real-time with high accuracy.",
author = "{van Wyk}, Franco and Anahita Khojandi and Akram Mohammed and Edmon Begoli and Robert Davis and Rishikesan Kamaleswaran",
year = "2019",
month = "2",
day = "1",
doi = "10.1016/j.ijmedinf.2018.12.002",
language = "English (US)",
volume = "122",
pages = "55--62",
journal = "International Journal of Medical Informatics",
issn = "1386-5056",
publisher = "Elsevier Ireland Ltd",

}

TY - JOUR

T1 - A minimal set of physiomarkers in continuous high frequency data streams predict adult sepsis onset earlier

AU - van Wyk, Franco

AU - Khojandi, Anahita

AU - Mohammed, Akram

AU - Begoli, Edmon

AU - Davis, Robert

AU - Kamaleswaran, Rishikesan

PY - 2019/2/1

Y1 - 2019/2/1

N2 - Purpose: Sepsis is a life-threatening condition with high mortality rates and expensive treatment costs. To improve short- and long-term outcomes, it is critical to detect at-risk sepsis patients at an early stage. Methods: A data-set consisting of high-frequency physiological data from 1161 critically ill patients was analyzed. 377 patients had developed sepsis, and had data at least 3 h prior to the onset of sepsis. A random forest classifier was trained to discriminate between sepsis and non-sepsis patients in real-time using a total of 132 features extracted from a moving time-window. The model was trained on 80% of the patients and was tested on the remaining 20% of the patients, for two observational periods of lengths 3 and 6 h prior to onset. Results: The model that used continuous physiological data alone resulted in sensitivity and F1 score of up to 80% and 67% one hour before sepsis onset. On average, these models were able to predict sepsis 294.19 ± 6.50 min (5 h) before the onset. Conclusions: The use of machine learning algorithms on continuous streams of physiological data can allow for early identification of at-risk patients in real-time with high accuracy.

AB - Purpose: Sepsis is a life-threatening condition with high mortality rates and expensive treatment costs. To improve short- and long-term outcomes, it is critical to detect at-risk sepsis patients at an early stage. Methods: A data-set consisting of high-frequency physiological data from 1161 critically ill patients was analyzed. 377 patients had developed sepsis, and had data at least 3 h prior to the onset of sepsis. A random forest classifier was trained to discriminate between sepsis and non-sepsis patients in real-time using a total of 132 features extracted from a moving time-window. The model was trained on 80% of the patients and was tested on the remaining 20% of the patients, for two observational periods of lengths 3 and 6 h prior to onset. Results: The model that used continuous physiological data alone resulted in sensitivity and F1 score of up to 80% and 67% one hour before sepsis onset. On average, these models were able to predict sepsis 294.19 ± 6.50 min (5 h) before the onset. Conclusions: The use of machine learning algorithms on continuous streams of physiological data can allow for early identification of at-risk patients in real-time with high accuracy.

UR - http://www.scopus.com/inward/record.url?scp=85058448201&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058448201&partnerID=8YFLogxK

U2 - 10.1016/j.ijmedinf.2018.12.002

DO - 10.1016/j.ijmedinf.2018.12.002

M3 - Article

C2 - 30623784

AN - SCOPUS:85058448201

VL - 122

SP - 55

EP - 62

JO - International Journal of Medical Informatics

JF - International Journal of Medical Informatics

SN - 1386-5056

ER -