Sensitivity and specificity of traumatic brain injury diagnosis codes in United States Department of Veterans Affairs administrative data

Kathleen Carlson, Joan E. Barnes, Emily M. Hagel, Brent C. Taylor, David X. Cifu, Nina A. Sayer

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Objective: To examine the validity of using International Classification of Diseases (ICD) diagnosis codes from United States Department of Veterans Affairs (VA) data to describe prevalence of traumatic brain injury (TBI) among military veterans. Methods: VA clinicians complete a standardized TBI evaluation to determine whether veterans' deployment exposures resulted in TBI. Clinician-confirmed cases and non-cases of TBI were used as recorded on the evaluation as the criterion standard against which to evaluate three series of TBI-related ICD diagnosis codes in national VA datasets. Focusing on codes used within VA, measures of validity were calculated and correlates of discordance examined, including patient characteristics, region and time. Secondarily, it was examined whether TBI codes can differentiate mild from more severe TBI cases. Results: Of 49 962 veterans with completed TBI evaluations, 29 534 (59%) received clinician-confirmed TBI diagnoses. Sensitivity of the VA series of codes was 70%, specificity was 82% and concordance was 75%. Concordance varied by region, but not by patient characteristics or time. Codes were not useful for distinguishing mild TBI. Conclusion: Estimates of TBI prevalence in military veterans are important for national programme development and resource distribution. Estimates derived from ICD diagnosis codes in administrative data should take potential misclassification into account.

Original languageEnglish (US)
Pages (from-to)640-650
Number of pages11
JournalBrain Injury
Volume27
Issue number6
DOIs
StatePublished - Jun 2013

Fingerprint

United States Department of Veterans Affairs
Veterans
Sensitivity and Specificity
International Classification of Diseases
Brain Concussion
Traumatic Brain Injury
Specificity
Program Development

Keywords

  • Clinical coding
  • Epidemiology
  • ICD-9-CM
  • Public health surveillance
  • Traumatic brain injury
  • Validity

ASJC Scopus subject areas

  • Clinical Neurology
  • Arts and Humanities (miscellaneous)
  • Developmental and Educational Psychology

Cite this

Sensitivity and specificity of traumatic brain injury diagnosis codes in United States Department of Veterans Affairs administrative data. / Carlson, Kathleen; Barnes, Joan E.; Hagel, Emily M.; Taylor, Brent C.; Cifu, David X.; Sayer, Nina A.

In: Brain Injury, Vol. 27, No. 6, 06.2013, p. 640-650.

Research output: Contribution to journalArticle

Carlson, Kathleen ; Barnes, Joan E. ; Hagel, Emily M. ; Taylor, Brent C. ; Cifu, David X. ; Sayer, Nina A. / Sensitivity and specificity of traumatic brain injury diagnosis codes in United States Department of Veterans Affairs administrative data. In: Brain Injury. 2013 ; Vol. 27, No. 6. pp. 640-650.
@article{ebdcaffb2e7046d58076c74c603377af,
title = "Sensitivity and specificity of traumatic brain injury diagnosis codes in United States Department of Veterans Affairs administrative data",
abstract = "Objective: To examine the validity of using International Classification of Diseases (ICD) diagnosis codes from United States Department of Veterans Affairs (VA) data to describe prevalence of traumatic brain injury (TBI) among military veterans. Methods: VA clinicians complete a standardized TBI evaluation to determine whether veterans' deployment exposures resulted in TBI. Clinician-confirmed cases and non-cases of TBI were used as recorded on the evaluation as the criterion standard against which to evaluate three series of TBI-related ICD diagnosis codes in national VA datasets. Focusing on codes used within VA, measures of validity were calculated and correlates of discordance examined, including patient characteristics, region and time. Secondarily, it was examined whether TBI codes can differentiate mild from more severe TBI cases. Results: Of 49 962 veterans with completed TBI evaluations, 29 534 (59{\%}) received clinician-confirmed TBI diagnoses. Sensitivity of the VA series of codes was 70{\%}, specificity was 82{\%} and concordance was 75{\%}. Concordance varied by region, but not by patient characteristics or time. Codes were not useful for distinguishing mild TBI. Conclusion: Estimates of TBI prevalence in military veterans are important for national programme development and resource distribution. Estimates derived from ICD diagnosis codes in administrative data should take potential misclassification into account.",
keywords = "Clinical coding, Epidemiology, ICD-9-CM, Public health surveillance, Traumatic brain injury, Validity",
author = "Kathleen Carlson and Barnes, {Joan E.} and Hagel, {Emily M.} and Taylor, {Brent C.} and Cifu, {David X.} and Sayer, {Nina A.}",
year = "2013",
month = "6",
doi = "10.3109/02699052.2013.771795",
language = "English (US)",
volume = "27",
pages = "640--650",
journal = "Brain Injury",
issn = "0269-9052",
publisher = "Informa Healthcare",
number = "6",

}

TY - JOUR

T1 - Sensitivity and specificity of traumatic brain injury diagnosis codes in United States Department of Veterans Affairs administrative data

AU - Carlson, Kathleen

AU - Barnes, Joan E.

AU - Hagel, Emily M.

AU - Taylor, Brent C.

AU - Cifu, David X.

AU - Sayer, Nina A.

PY - 2013/6

Y1 - 2013/6

N2 - Objective: To examine the validity of using International Classification of Diseases (ICD) diagnosis codes from United States Department of Veterans Affairs (VA) data to describe prevalence of traumatic brain injury (TBI) among military veterans. Methods: VA clinicians complete a standardized TBI evaluation to determine whether veterans' deployment exposures resulted in TBI. Clinician-confirmed cases and non-cases of TBI were used as recorded on the evaluation as the criterion standard against which to evaluate three series of TBI-related ICD diagnosis codes in national VA datasets. Focusing on codes used within VA, measures of validity were calculated and correlates of discordance examined, including patient characteristics, region and time. Secondarily, it was examined whether TBI codes can differentiate mild from more severe TBI cases. Results: Of 49 962 veterans with completed TBI evaluations, 29 534 (59%) received clinician-confirmed TBI diagnoses. Sensitivity of the VA series of codes was 70%, specificity was 82% and concordance was 75%. Concordance varied by region, but not by patient characteristics or time. Codes were not useful for distinguishing mild TBI. Conclusion: Estimates of TBI prevalence in military veterans are important for national programme development and resource distribution. Estimates derived from ICD diagnosis codes in administrative data should take potential misclassification into account.

AB - Objective: To examine the validity of using International Classification of Diseases (ICD) diagnosis codes from United States Department of Veterans Affairs (VA) data to describe prevalence of traumatic brain injury (TBI) among military veterans. Methods: VA clinicians complete a standardized TBI evaluation to determine whether veterans' deployment exposures resulted in TBI. Clinician-confirmed cases and non-cases of TBI were used as recorded on the evaluation as the criterion standard against which to evaluate three series of TBI-related ICD diagnosis codes in national VA datasets. Focusing on codes used within VA, measures of validity were calculated and correlates of discordance examined, including patient characteristics, region and time. Secondarily, it was examined whether TBI codes can differentiate mild from more severe TBI cases. Results: Of 49 962 veterans with completed TBI evaluations, 29 534 (59%) received clinician-confirmed TBI diagnoses. Sensitivity of the VA series of codes was 70%, specificity was 82% and concordance was 75%. Concordance varied by region, but not by patient characteristics or time. Codes were not useful for distinguishing mild TBI. Conclusion: Estimates of TBI prevalence in military veterans are important for national programme development and resource distribution. Estimates derived from ICD diagnosis codes in administrative data should take potential misclassification into account.

KW - Clinical coding

KW - Epidemiology

KW - ICD-9-CM

KW - Public health surveillance

KW - Traumatic brain injury

KW - Validity

UR - http://www.scopus.com/inward/record.url?scp=84877852947&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84877852947&partnerID=8YFLogxK

U2 - 10.3109/02699052.2013.771795

DO - 10.3109/02699052.2013.771795

M3 - Article

C2 - 23514276

AN - SCOPUS:84877852947

VL - 27

SP - 640

EP - 650

JO - Brain Injury

JF - Brain Injury

SN - 0269-9052

IS - 6

ER -