Towards generating a patient's timeline: Extracting temporal relationships from clinical notes

Azadeh Nikfarjam, Ehsan Emadzadeh, Graciela Gonzalez

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

Clinical records include both coded and free-text fields that interact to reflect complicated patient stories. The information often covers not only the present medical condition and events experienced by the patient, but also refers to relevant events in the past (such as signs, symptoms, tests or treatments). In order to automatically construct a timeline of these events, we first need to extract the temporal relations between pairs of events or time expressions presented in the clinical notes. We designed separate extraction components for different types of temporal relations, utilizing a novel hybrid system that combines machine learning with a graph-based inference mechanism to extract the temporal links. The temporal graph is a directed graph based on parse tree dependencies of the simplified sentences and frequent pattern clues. We generalized the sentences in order to discover patterns that, given the complexities of natural language, might not be directly discoverable in the original sentences. The proposed hybrid system performance reached an F-measure of 0.63, with precision at 0.76 and recall at 0.54 on the 2012 i2b2 Natural Language Processing corpus for the temporal relation (TLink) extraction task, achieving the highest precision and third highest f-measure among participating teams in the TLink track.

Original languageEnglish (US)
JournalJournal of Biomedical Informatics
Volume46
Issue numberSUPPL.
DOIs
StatePublished - 2013

Fingerprint

Hybrid systems
Directed graphs
Learning systems
Processing
Natural Language Processing
Signs and Symptoms
TimeLine
Language

Keywords

  • Automatic patient timeline
  • Clinical text mining
  • Machine learning
  • Natural Language Processing
  • Temporal graph
  • Temporal relation extraction

ASJC Scopus subject areas

  • Computer Science Applications
  • Health Informatics

Cite this

Towards generating a patient's timeline : Extracting temporal relationships from clinical notes. / Nikfarjam, Azadeh; Emadzadeh, Ehsan; Gonzalez, Graciela.

In: Journal of Biomedical Informatics, Vol. 46, No. SUPPL., 2013.

Research output: Contribution to journalArticle

Nikfarjam, Azadeh ; Emadzadeh, Ehsan ; Gonzalez, Graciela. / Towards generating a patient's timeline : Extracting temporal relationships from clinical notes. In: Journal of Biomedical Informatics. 2013 ; Vol. 46, No. SUPPL.
@article{1ff95a2000914719ad18ff8de374cfdf,
title = "Towards generating a patient's timeline: Extracting temporal relationships from clinical notes",
abstract = "Clinical records include both coded and free-text fields that interact to reflect complicated patient stories. The information often covers not only the present medical condition and events experienced by the patient, but also refers to relevant events in the past (such as signs, symptoms, tests or treatments). In order to automatically construct a timeline of these events, we first need to extract the temporal relations between pairs of events or time expressions presented in the clinical notes. We designed separate extraction components for different types of temporal relations, utilizing a novel hybrid system that combines machine learning with a graph-based inference mechanism to extract the temporal links. The temporal graph is a directed graph based on parse tree dependencies of the simplified sentences and frequent pattern clues. We generalized the sentences in order to discover patterns that, given the complexities of natural language, might not be directly discoverable in the original sentences. The proposed hybrid system performance reached an F-measure of 0.63, with precision at 0.76 and recall at 0.54 on the 2012 i2b2 Natural Language Processing corpus for the temporal relation (TLink) extraction task, achieving the highest precision and third highest f-measure among participating teams in the TLink track.",
keywords = "Automatic patient timeline, Clinical text mining, Machine learning, Natural Language Processing, Temporal graph, Temporal relation extraction",
author = "Azadeh Nikfarjam and Ehsan Emadzadeh and Graciela Gonzalez",
year = "2013",
doi = "10.1016/j.jbi.2013.11.001",
language = "English (US)",
volume = "46",
journal = "Journal of Biomedical Informatics",
issn = "1532-0464",
publisher = "Academic Press Inc.",
number = "SUPPL.",

}

TY - JOUR

T1 - Towards generating a patient's timeline

T2 - Extracting temporal relationships from clinical notes

AU - Nikfarjam, Azadeh

AU - Emadzadeh, Ehsan

AU - Gonzalez, Graciela

PY - 2013

Y1 - 2013

N2 - Clinical records include both coded and free-text fields that interact to reflect complicated patient stories. The information often covers not only the present medical condition and events experienced by the patient, but also refers to relevant events in the past (such as signs, symptoms, tests or treatments). In order to automatically construct a timeline of these events, we first need to extract the temporal relations between pairs of events or time expressions presented in the clinical notes. We designed separate extraction components for different types of temporal relations, utilizing a novel hybrid system that combines machine learning with a graph-based inference mechanism to extract the temporal links. The temporal graph is a directed graph based on parse tree dependencies of the simplified sentences and frequent pattern clues. We generalized the sentences in order to discover patterns that, given the complexities of natural language, might not be directly discoverable in the original sentences. The proposed hybrid system performance reached an F-measure of 0.63, with precision at 0.76 and recall at 0.54 on the 2012 i2b2 Natural Language Processing corpus for the temporal relation (TLink) extraction task, achieving the highest precision and third highest f-measure among participating teams in the TLink track.

AB - Clinical records include both coded and free-text fields that interact to reflect complicated patient stories. The information often covers not only the present medical condition and events experienced by the patient, but also refers to relevant events in the past (such as signs, symptoms, tests or treatments). In order to automatically construct a timeline of these events, we first need to extract the temporal relations between pairs of events or time expressions presented in the clinical notes. We designed separate extraction components for different types of temporal relations, utilizing a novel hybrid system that combines machine learning with a graph-based inference mechanism to extract the temporal links. The temporal graph is a directed graph based on parse tree dependencies of the simplified sentences and frequent pattern clues. We generalized the sentences in order to discover patterns that, given the complexities of natural language, might not be directly discoverable in the original sentences. The proposed hybrid system performance reached an F-measure of 0.63, with precision at 0.76 and recall at 0.54 on the 2012 i2b2 Natural Language Processing corpus for the temporal relation (TLink) extraction task, achieving the highest precision and third highest f-measure among participating teams in the TLink track.

KW - Automatic patient timeline

KW - Clinical text mining

KW - Machine learning

KW - Natural Language Processing

KW - Temporal graph

KW - Temporal relation extraction

UR - http://www.scopus.com/inward/record.url?scp=84882737650&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84882737650&partnerID=8YFLogxK

U2 - 10.1016/j.jbi.2013.11.001

DO - 10.1016/j.jbi.2013.11.001

M3 - Article

C2 - 24212118

AN - SCOPUS:84882737650

VL - 46

JO - Journal of Biomedical Informatics

JF - Journal of Biomedical Informatics

SN - 1532-0464

IS - SUPPL.

ER -