Predicting mortality in the surgical intensive care unit using artificial intelligence and natural language processing of physician documentation

Joshua Parreco, Antonio Hidalgo, Robert Kozol, Nicholas Namias, Rishi Rattan

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

The purpose of this study was to use natural language processing of physician documentation to predict mortality in patients admitted to the surgical intensive care unit (SICU). The Multiparameter Intelligent Monitoring in Intensive Care III database was used to obtain SICU stays with six different severity of illness scores. Natural language processing was performed on the physician notes. Classifiers for predicting mortality were created. One classifier used only the physician notes, one used only the severity of illness scores, and one used the physician notes with severity of injury scores. There were 3838 SICU stays identified during the study period and 5.4 per cent ended with mortality. The classifier trained with physician notes with severity of injury scores performed with the highest area under the curve (0.88 6 0.05) and accuracy (94.6 6 1.1%). The most important variable was the Oxford Acute Severity of Illness Score (16.0%). The most important terms were “dilated” (4.3%) and “hemorrhage” (3.7%). This study demonstrates the novel use of artificial intelligence to process physician documentation to predict mortality in the SICU. The classifiers were able to detect the subtle nuances in physician vernacular that predict mortality. These nuances provided improved performance in predicting mortality over physiologic parameters alone.

Original languageEnglish (US)
Pages (from-to)1190-11994
Number of pages10805
JournalAmerican Surgeon
Volume84
Issue number7
StatePublished - Jul 1 2018

Fingerprint

Natural Language Processing
Artificial Intelligence
Critical Care
Documentation
Intensive Care Units
Physicians
Mortality
Injury Severity Score
Area Under Curve
Databases
Hemorrhage

ASJC Scopus subject areas

  • Surgery

Cite this

Predicting mortality in the surgical intensive care unit using artificial intelligence and natural language processing of physician documentation. / Parreco, Joshua; Hidalgo, Antonio; Kozol, Robert; Namias, Nicholas; Rattan, Rishi.

In: American Surgeon, Vol. 84, No. 7, 01.07.2018, p. 1190-11994.

Research output: Contribution to journalArticle

@article{df6269379b6e44118e5b0ffad2b54111,
title = "Predicting mortality in the surgical intensive care unit using artificial intelligence and natural language processing of physician documentation",
abstract = "The purpose of this study was to use natural language processing of physician documentation to predict mortality in patients admitted to the surgical intensive care unit (SICU). The Multiparameter Intelligent Monitoring in Intensive Care III database was used to obtain SICU stays with six different severity of illness scores. Natural language processing was performed on the physician notes. Classifiers for predicting mortality were created. One classifier used only the physician notes, one used only the severity of illness scores, and one used the physician notes with severity of injury scores. There were 3838 SICU stays identified during the study period and 5.4 per cent ended with mortality. The classifier trained with physician notes with severity of injury scores performed with the highest area under the curve (0.88 6 0.05) and accuracy (94.6 6 1.1{\%}). The most important variable was the Oxford Acute Severity of Illness Score (16.0{\%}). The most important terms were “dilated” (4.3{\%}) and “hemorrhage” (3.7{\%}). This study demonstrates the novel use of artificial intelligence to process physician documentation to predict mortality in the SICU. The classifiers were able to detect the subtle nuances in physician vernacular that predict mortality. These nuances provided improved performance in predicting mortality over physiologic parameters alone.",
author = "Joshua Parreco and Antonio Hidalgo and Robert Kozol and Nicholas Namias and Rishi Rattan",
year = "2018",
month = "7",
day = "1",
language = "English (US)",
volume = "84",
pages = "1190--11994",
journal = "The American surgeon",
issn = "0003-1348",
publisher = "Southeastern Surgical Congress",
number = "7",

}

TY - JOUR

T1 - Predicting mortality in the surgical intensive care unit using artificial intelligence and natural language processing of physician documentation

AU - Parreco, Joshua

AU - Hidalgo, Antonio

AU - Kozol, Robert

AU - Namias, Nicholas

AU - Rattan, Rishi

PY - 2018/7/1

Y1 - 2018/7/1

N2 - The purpose of this study was to use natural language processing of physician documentation to predict mortality in patients admitted to the surgical intensive care unit (SICU). The Multiparameter Intelligent Monitoring in Intensive Care III database was used to obtain SICU stays with six different severity of illness scores. Natural language processing was performed on the physician notes. Classifiers for predicting mortality were created. One classifier used only the physician notes, one used only the severity of illness scores, and one used the physician notes with severity of injury scores. There were 3838 SICU stays identified during the study period and 5.4 per cent ended with mortality. The classifier trained with physician notes with severity of injury scores performed with the highest area under the curve (0.88 6 0.05) and accuracy (94.6 6 1.1%). The most important variable was the Oxford Acute Severity of Illness Score (16.0%). The most important terms were “dilated” (4.3%) and “hemorrhage” (3.7%). This study demonstrates the novel use of artificial intelligence to process physician documentation to predict mortality in the SICU. The classifiers were able to detect the subtle nuances in physician vernacular that predict mortality. These nuances provided improved performance in predicting mortality over physiologic parameters alone.

AB - The purpose of this study was to use natural language processing of physician documentation to predict mortality in patients admitted to the surgical intensive care unit (SICU). The Multiparameter Intelligent Monitoring in Intensive Care III database was used to obtain SICU stays with six different severity of illness scores. Natural language processing was performed on the physician notes. Classifiers for predicting mortality were created. One classifier used only the physician notes, one used only the severity of illness scores, and one used the physician notes with severity of injury scores. There were 3838 SICU stays identified during the study period and 5.4 per cent ended with mortality. The classifier trained with physician notes with severity of injury scores performed with the highest area under the curve (0.88 6 0.05) and accuracy (94.6 6 1.1%). The most important variable was the Oxford Acute Severity of Illness Score (16.0%). The most important terms were “dilated” (4.3%) and “hemorrhage” (3.7%). This study demonstrates the novel use of artificial intelligence to process physician documentation to predict mortality in the SICU. The classifiers were able to detect the subtle nuances in physician vernacular that predict mortality. These nuances provided improved performance in predicting mortality over physiologic parameters alone.

UR - http://www.scopus.com/inward/record.url?scp=85050889095&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050889095&partnerID=8YFLogxK

M3 - Article

C2 - 30064586

AN - SCOPUS:85050889095

VL - 84

SP - 1190

EP - 11994

JO - The American surgeon

JF - The American surgeon

SN - 0003-1348

IS - 7

ER -