4099 — Development & Evaluation of mTBI Case Ascertainment Methods Using the VA Comprehensive TBI Evaluation Database

Lead/Presenter: Alexandra Schneider, 
All Authors: Schneider AL (Rocky Mountain MIRECC) Hostetter TA (Rocky Mountain MIRECC) Forster JE (Rocky Mountain MIRECC; University of Colorado Anschutz Medical Campus, Department of Physical Medicine & Rehabilitation) Brenner LA (Rocky Mountain MIRECC; University of Colorado Anschutz Medical Campus, Departments of Psychiatry, Physical Medicine & Rehabilitation, and Neurology) Bahraini NB (Rocky Mountain MIRECC, University of Colorado Anschutz Medical Campus, Departments of Psychiatry and Physical Medicine & Rehabilitation)

Objectives:
One challenge of using national Veteran's Administration (VA) data for research is confirming its validity. Transparent methods to determine how data from the VA Patient Care Services (PCS) centralized database on the Comprehensive TBI Evaluation (CTBIE) should be used for mild traumatic brain injury (mTBI) case ascertainment are limited. Moreover, previous studies of CTBIE data examining the concordance of clinical judgement of mTBI with diagnostic based criteria have revealed inconsistencies. Methods for identifying a cohort of Veterans with mTBI from the CTBIE database based on diagnostic criteria were developed and evaluated as part of a study examining the psychometric properties of the Neurobehavioral Symptom Inventory among OEF/OIF Veterans.

Methods:
Records of individuals given the CTBIE between October 1, 2008 and September 30, 2011 were evaluated for mTBI. Records where TBI status could not be determined due to missing and/or inconsistent data were removed (n = 8,309), as were those with any indication of moderate or severe TBI (n = 4,854) or no history of TBI (n = 13,441). Only individuals who met criteria for mTBI based on VA/DoD clinical classification guidelines were included in the final cohort. As such, records needed to indicate at least one of the following: loss of consciousness (LOC) < = 30 minutes, alteration of consciousness (AOC) < = 24 hours, or post-traumatic amnesia (PTA) < = 1 day.

Results:
39,490 records from the complete CTBIE database (n = 66,094) were identified as having probable history of mTBI. Differences between diagnostic based criteria and clinician judgement will be examined.

Implications:
Case ascertainment of mTBI using CTE data should be informed by rigorous, standardized methods based on diagnostic criteria. Strict reliance on clinician judgment without consideration of how it corresponds to standard diagnostic criteria may contribute to misclassification of individuals with mTBI. Efforts to reduce missing data and automated tools that ensure diagnostic indicators in the CTBIE agree with clinician diagnosis may improve the reliability and validity of mTBI classification, strengthening the utility of the CTBIE for TBI research.

Impacts:
Applying these methods in research studies that aim to identify Veterans with mTBI can help decrease the likelihood of misclassification and lead to more accurate results.