Epidemiology of angina pectoris: Role of natural language processing of the medical record

Serguei S V Pakhomov, Harry Hemingway, Susan A. Weston, Steven J. Jacobsen, Richard Rodeheffer, Veronique Lee Roger

Research output: Contribution to journalArticle

34 Citations (Scopus)

Abstract

Background: The diagnosis of angina is challenging because it relies on symptom descriptions. Natural language processing (NLP) of the electronic medical record (EMR) can provide access to such information contained in free text that may not be fully captured by conventional diagnostic coding. Objective: To test the hypothesis that NLP of the EMR improves angina pectoris ascertainment over diagnostic codes. Methods: Billing records of inpatients and outpatients were searched for International Classification of Diseases, Ninth Revision (ICD-9) codes for angina pectoris, chronic ischemic heart disease, and chest pain. EMR clinical reports were searched electronically for 50 specific nonnegated natural language synonyms to these ICD-9 codes. The 2 methods were compared to a standardized assessment of angina by Rose questionnaire for 3 diagnostic levels: unspecified chest pain, exertional chest pain, and Rose angina. Results: Compared with the Rose questionnaire, the true-positive rate of EMR-NLP for unspecified chest pain was 62% (95% CI 55-67) versus 51% (95% CI 44-58) for diagnostic codes (P < .001). For exertional chest pain, the EMR-NLP true-positive rate was 71% (95% CI 61-80) versus 62% (95% CI 52-73) for diagnostic codes (P = .10). Both approaches had 88% (95% CI 65-100) true-positive rate for Rose angina. The EMR-NLP method consistently identified more patients with exertional chest pain over a 28-month follow-up. Conclusion: EMR-NLP method improves the detection of unspecified and exertional chest pain cases compared to diagnostic codes. These findings have implications for epidemiological and clinical studies of angina pectoris.

Original languageEnglish (US)
Pages (from-to)666-673
Number of pages8
JournalAmerican Heart Journal
Volume153
Issue number4
DOIs
StatePublished - Apr 2007

Fingerprint

Natural Language Processing
Electronic Health Records
Angina Pectoris
Chest Pain
Medical Records
Epidemiology
International Classification of Diseases
Access to Information
Myocardial Ischemia
Inpatients
Epidemiologic Studies
Outpatients
Language

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine

Cite this

Epidemiology of angina pectoris : Role of natural language processing of the medical record. / Pakhomov, Serguei S V; Hemingway, Harry; Weston, Susan A.; Jacobsen, Steven J.; Rodeheffer, Richard; Roger, Veronique Lee.

In: American Heart Journal, Vol. 153, No. 4, 04.2007, p. 666-673.

Research output: Contribution to journalArticle

Pakhomov, SSV, Hemingway, H, Weston, SA, Jacobsen, SJ, Rodeheffer, R & Roger, VL 2007, 'Epidemiology of angina pectoris: Role of natural language processing of the medical record', American Heart Journal, vol. 153, no. 4, pp. 666-673. https://doi.org/10.1016/j.ahj.2006.12.022
Pakhomov, Serguei S V ; Hemingway, Harry ; Weston, Susan A. ; Jacobsen, Steven J. ; Rodeheffer, Richard ; Roger, Veronique Lee. / Epidemiology of angina pectoris : Role of natural language processing of the medical record. In: American Heart Journal. 2007 ; Vol. 153, No. 4. pp. 666-673.
@article{1c53917f163e42f6a3b518565fdfe522,
title = "Epidemiology of angina pectoris: Role of natural language processing of the medical record",
abstract = "Background: The diagnosis of angina is challenging because it relies on symptom descriptions. Natural language processing (NLP) of the electronic medical record (EMR) can provide access to such information contained in free text that may not be fully captured by conventional diagnostic coding. Objective: To test the hypothesis that NLP of the EMR improves angina pectoris ascertainment over diagnostic codes. Methods: Billing records of inpatients and outpatients were searched for International Classification of Diseases, Ninth Revision (ICD-9) codes for angina pectoris, chronic ischemic heart disease, and chest pain. EMR clinical reports were searched electronically for 50 specific nonnegated natural language synonyms to these ICD-9 codes. The 2 methods were compared to a standardized assessment of angina by Rose questionnaire for 3 diagnostic levels: unspecified chest pain, exertional chest pain, and Rose angina. Results: Compared with the Rose questionnaire, the true-positive rate of EMR-NLP for unspecified chest pain was 62{\%} (95{\%} CI 55-67) versus 51{\%} (95{\%} CI 44-58) for diagnostic codes (P < .001). For exertional chest pain, the EMR-NLP true-positive rate was 71{\%} (95{\%} CI 61-80) versus 62{\%} (95{\%} CI 52-73) for diagnostic codes (P = .10). Both approaches had 88{\%} (95{\%} CI 65-100) true-positive rate for Rose angina. The EMR-NLP method consistently identified more patients with exertional chest pain over a 28-month follow-up. Conclusion: EMR-NLP method improves the detection of unspecified and exertional chest pain cases compared to diagnostic codes. These findings have implications for epidemiological and clinical studies of angina pectoris.",
author = "Pakhomov, {Serguei S V} and Harry Hemingway and Weston, {Susan A.} and Jacobsen, {Steven J.} and Richard Rodeheffer and Roger, {Veronique Lee}",
year = "2007",
month = "4",
doi = "10.1016/j.ahj.2006.12.022",
language = "English (US)",
volume = "153",
pages = "666--673",
journal = "American Heart Journal",
issn = "0002-8703",
publisher = "Mosby Inc.",
number = "4",

}

TY - JOUR

T1 - Epidemiology of angina pectoris

T2 - Role of natural language processing of the medical record

AU - Pakhomov, Serguei S V

AU - Hemingway, Harry

AU - Weston, Susan A.

AU - Jacobsen, Steven J.

AU - Rodeheffer, Richard

AU - Roger, Veronique Lee

PY - 2007/4

Y1 - 2007/4

N2 - Background: The diagnosis of angina is challenging because it relies on symptom descriptions. Natural language processing (NLP) of the electronic medical record (EMR) can provide access to such information contained in free text that may not be fully captured by conventional diagnostic coding. Objective: To test the hypothesis that NLP of the EMR improves angina pectoris ascertainment over diagnostic codes. Methods: Billing records of inpatients and outpatients were searched for International Classification of Diseases, Ninth Revision (ICD-9) codes for angina pectoris, chronic ischemic heart disease, and chest pain. EMR clinical reports were searched electronically for 50 specific nonnegated natural language synonyms to these ICD-9 codes. The 2 methods were compared to a standardized assessment of angina by Rose questionnaire for 3 diagnostic levels: unspecified chest pain, exertional chest pain, and Rose angina. Results: Compared with the Rose questionnaire, the true-positive rate of EMR-NLP for unspecified chest pain was 62% (95% CI 55-67) versus 51% (95% CI 44-58) for diagnostic codes (P < .001). For exertional chest pain, the EMR-NLP true-positive rate was 71% (95% CI 61-80) versus 62% (95% CI 52-73) for diagnostic codes (P = .10). Both approaches had 88% (95% CI 65-100) true-positive rate for Rose angina. The EMR-NLP method consistently identified more patients with exertional chest pain over a 28-month follow-up. Conclusion: EMR-NLP method improves the detection of unspecified and exertional chest pain cases compared to diagnostic codes. These findings have implications for epidemiological and clinical studies of angina pectoris.

AB - Background: The diagnosis of angina is challenging because it relies on symptom descriptions. Natural language processing (NLP) of the electronic medical record (EMR) can provide access to such information contained in free text that may not be fully captured by conventional diagnostic coding. Objective: To test the hypothesis that NLP of the EMR improves angina pectoris ascertainment over diagnostic codes. Methods: Billing records of inpatients and outpatients were searched for International Classification of Diseases, Ninth Revision (ICD-9) codes for angina pectoris, chronic ischemic heart disease, and chest pain. EMR clinical reports were searched electronically for 50 specific nonnegated natural language synonyms to these ICD-9 codes. The 2 methods were compared to a standardized assessment of angina by Rose questionnaire for 3 diagnostic levels: unspecified chest pain, exertional chest pain, and Rose angina. Results: Compared with the Rose questionnaire, the true-positive rate of EMR-NLP for unspecified chest pain was 62% (95% CI 55-67) versus 51% (95% CI 44-58) for diagnostic codes (P < .001). For exertional chest pain, the EMR-NLP true-positive rate was 71% (95% CI 61-80) versus 62% (95% CI 52-73) for diagnostic codes (P = .10). Both approaches had 88% (95% CI 65-100) true-positive rate for Rose angina. The EMR-NLP method consistently identified more patients with exertional chest pain over a 28-month follow-up. Conclusion: EMR-NLP method improves the detection of unspecified and exertional chest pain cases compared to diagnostic codes. These findings have implications for epidemiological and clinical studies of angina pectoris.

UR - http://www.scopus.com/inward/record.url?scp=33947306795&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33947306795&partnerID=8YFLogxK

U2 - 10.1016/j.ahj.2006.12.022

DO - 10.1016/j.ahj.2006.12.022

M3 - Article

C2 - 17383310

AN - SCOPUS:33947306795

VL - 153

SP - 666

EP - 673

JO - American Heart Journal

JF - American Heart Journal

SN - 0002-8703

IS - 4

ER -