ADEPt, a semantically-enriched pipeline for extracting adverse drug events from free-text electronic health records

Ehtesham Iqbal; Robbie Mallah; Daniel Rhodes; Honghan Wu; Alvin Romero; Nynn Chang; Olubanke Dzahini; Chandra Pandey; Matthew Broadbent; Robert Stewart; Richard J. B. Dobson; Zina M. Ibrahim; Tudor Groza

doi:10.1371/journal.pone.0187121

ADEPt, a semantically-enriched pipeline for extracting adverse drug events from free-text electronic health records

Ehtesham Iqbal, Robbie Mallah, Daniel Rhodes, Honghan Wu, Alvin Romero, Nynn Chang, Olubanke Dzahini, Chandra Pandey, Matthew Broadbent, Robert Stewart, Richard J. B. Dobson, Zina M. Ibrahim, Tudor Groza (Editor)

Research output: Contribution to journal › Article › peer-review

25 Citations (Scopus)

244 Downloads (Pure)

Abstract

Adverse drug events (ADEs) are unintended responses to medical treatment. They can greatly affect a patient’s quality of life and present a substantial burden on healthcare. Although Electronic health records (EHRs) document a wealth of information relating to ADEs, they are frequently stored in the unstructured or semi-structured free-text narrative requiring Natural Language Processing (NLP) techniques to mine the relevant information. Here we present a rule-based ADE detection and classification pipeline built and tested on a large Psychiatric corpus comprising 264k patients using the de-identified EHRs of four UK-based psychiatric hospitals. The pipeline uses characteristics specific to Psychiatric EHRs to guide the annotation process, and distinguishes: a) the temporal value associated with the ADE mention (whether it is historical or present), b) the categorical value of the ADE (whether it is assertive, hypothetical, retrospective or a general discussion) and c) the implicit contextual value where the status of the ADE is deduced from surrounding indicators, rather than explicitly stated. We manually created the rulebase in collaboration with clinicians and pharmacists by studying ADE mentions in various types of clinical notes. We evaluated the open-source Adverse Drug Event annotation Pipeline (ADEPt) using 19 ADEs specific to antipsychotics and antidepressants medication. The ADEs chosen vary in severity, regularity and persistence. The average F-measure and accuracy achieved by our tool across all tested ADEs were 0.83 and 0.83 respectively. In addition to annotation power, the ADEPT pipeline presents an improvement to the state of the art context-discerning algorithm, ConText.

Original language	English
Pages (from-to)	e0187121
Journal	PL o S One
Volume	12
Issue number	11
Early online date	9 Nov 2017
DOIs	https://doi.org/10.1371/journal.pone.0187121
Publication status	E-pub ahead of print - 9 Nov 2017

Access to Document

10.1371/journal.pone.0187121Licence: CC BY

ADEPt, a semantically-enriched_IQBAL_Publishedonline9November2017_GOLD VoR (CC BY)Final published version, 2.34 MBLicence: CC BY

Cite this

Iqbal, E., Mallah, R., Rhodes, D., Wu, H., Romero, A., Chang, N., Dzahini, O., Pandey, C., Broadbent, M., Stewart, R., Dobson, R. J. B., Ibrahim, Z. M., & Groza, T. (Ed.) (2017). ADEPt, a semantically-enriched pipeline for extracting adverse drug events from free-text electronic health records. PL o S One , 12(11), e0187121. Advance online publication. https://doi.org/10.1371/journal.pone.0187121

@article{4ca85e61114a4e2fb10cd53ef796fe64,

title = "ADEPt, a semantically-enriched pipeline for extracting adverse drug events from free-text electronic health records",

abstract = "Adverse drug events (ADEs) are unintended responses to medical treatment. They can greatly affect a patient{\textquoteright}s quality of life and present a substantial burden on healthcare. Although Electronic health records (EHRs) document a wealth of information relating to ADEs, they are frequently stored in the unstructured or semi-structured free-text narrative requiring Natural Language Processing (NLP) techniques to mine the relevant information. Here we present a rule-based ADE detection and classification pipeline built and tested on a large Psychiatric corpus comprising 264k patients using the de-identified EHRs of four UK-based psychiatric hospitals. The pipeline uses characteristics specific to Psychiatric EHRs to guide the annotation process, and distinguishes: a) the temporal value associated with the ADE mention (whether it is historical or present), b) the categorical value of the ADE (whether it is assertive, hypothetical, retrospective or a general discussion) and c) the implicit contextual value where the status of the ADE is deduced from surrounding indicators, rather than explicitly stated. We manually created the rulebase in collaboration with clinicians and pharmacists by studying ADE mentions in various types of clinical notes. We evaluated the open-source Adverse Drug Event annotation Pipeline (ADEPt) using 19 ADEs specific to antipsychotics and antidepressants medication. The ADEs chosen vary in severity, regularity and persistence. The average F-measure and accuracy achieved by our tool across all tested ADEs were 0.83 and 0.83 respectively. In addition to annotation power, the ADEPT pipeline presents an improvement to the state of the art context-discerning algorithm, ConText. ",

author = "Ehtesham Iqbal and Robbie Mallah and Daniel Rhodes and Honghan Wu and Alvin Romero and Nynn Chang and Olubanke Dzahini and Chandra Pandey and Matthew Broadbent and Robert Stewart and Dobson, {Richard J. B.} and Ibrahim, {Zina M.} and Tudor Groza",

year = "2017",

month = nov,

day = "9",

doi = "10.1371/journal.pone.0187121",

language = "English",

volume = "12",

pages = "e0187121",

journal = "PL o S One ",

issn = "1932-6203",

publisher = "Public Library of Science",

number = "11",

}

TY - JOUR

T1 - ADEPt, a semantically-enriched pipeline for extracting adverse drug events from free-text electronic health records

AU - Iqbal, Ehtesham

AU - Mallah, Robbie

AU - Rhodes, Daniel

AU - Wu, Honghan

AU - Romero, Alvin

AU - Chang, Nynn

AU - Dzahini, Olubanke

AU - Pandey, Chandra

AU - Broadbent, Matthew

AU - Stewart, Robert

AU - Dobson, Richard J. B.

AU - Ibrahim, Zina M.

A2 - Groza, Tudor

PY - 2017/11/9

Y1 - 2017/11/9

N2 - Adverse drug events (ADEs) are unintended responses to medical treatment. They can greatly affect a patient’s quality of life and present a substantial burden on healthcare. Although Electronic health records (EHRs) document a wealth of information relating to ADEs, they are frequently stored in the unstructured or semi-structured free-text narrative requiring Natural Language Processing (NLP) techniques to mine the relevant information. Here we present a rule-based ADE detection and classification pipeline built and tested on a large Psychiatric corpus comprising 264k patients using the de-identified EHRs of four UK-based psychiatric hospitals. The pipeline uses characteristics specific to Psychiatric EHRs to guide the annotation process, and distinguishes: a) the temporal value associated with the ADE mention (whether it is historical or present), b) the categorical value of the ADE (whether it is assertive, hypothetical, retrospective or a general discussion) and c) the implicit contextual value where the status of the ADE is deduced from surrounding indicators, rather than explicitly stated. We manually created the rulebase in collaboration with clinicians and pharmacists by studying ADE mentions in various types of clinical notes. We evaluated the open-source Adverse Drug Event annotation Pipeline (ADEPt) using 19 ADEs specific to antipsychotics and antidepressants medication. The ADEs chosen vary in severity, regularity and persistence. The average F-measure and accuracy achieved by our tool across all tested ADEs were 0.83 and 0.83 respectively. In addition to annotation power, the ADEPT pipeline presents an improvement to the state of the art context-discerning algorithm, ConText.

AB - Adverse drug events (ADEs) are unintended responses to medical treatment. They can greatly affect a patient’s quality of life and present a substantial burden on healthcare. Although Electronic health records (EHRs) document a wealth of information relating to ADEs, they are frequently stored in the unstructured or semi-structured free-text narrative requiring Natural Language Processing (NLP) techniques to mine the relevant information. Here we present a rule-based ADE detection and classification pipeline built and tested on a large Psychiatric corpus comprising 264k patients using the de-identified EHRs of four UK-based psychiatric hospitals. The pipeline uses characteristics specific to Psychiatric EHRs to guide the annotation process, and distinguishes: a) the temporal value associated with the ADE mention (whether it is historical or present), b) the categorical value of the ADE (whether it is assertive, hypothetical, retrospective or a general discussion) and c) the implicit contextual value where the status of the ADE is deduced from surrounding indicators, rather than explicitly stated. We manually created the rulebase in collaboration with clinicians and pharmacists by studying ADE mentions in various types of clinical notes. We evaluated the open-source Adverse Drug Event annotation Pipeline (ADEPt) using 19 ADEs specific to antipsychotics and antidepressants medication. The ADEs chosen vary in severity, regularity and persistence. The average F-measure and accuracy achieved by our tool across all tested ADEs were 0.83 and 0.83 respectively. In addition to annotation power, the ADEPT pipeline presents an improvement to the state of the art context-discerning algorithm, ConText.

UR - http://www.scopus.com/inward/record.url?scp=85033454279&partnerID=8YFLogxK

U2 - 10.1371/journal.pone.0187121

DO - 10.1371/journal.pone.0187121

M3 - Article

SN - 1932-6203

VL - 12

SP - e0187121

JO - PL o S One

JF - PL o S One

IS - 11

ER -

ADEPt, a semantically-enriched pipeline for extracting adverse drug events from free-text electronic health records

Abstract

Access to Document

Other files and links

Fingerprint

Cite this