Kinney AR, Schneider AL, King SE, Yan XD, Forster JE, Bahraini NH, Brenner LA. Identifying and Predicting Subgroups of Veterans With Mild Traumatic Brain Injury Based on Distinct Configurations of Postconcussive Symptom Endorsement: A Latent Class Analysis.
J Head Trauma Rehabil 2024;
39:247-257. [PMID:
38259092 DOI:
10.1097/htr.0000000000000890]
[Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/24/2024]
Abstract
OBJECTIVE
To identify distinct subgroups of veterans with mild traumatic brain injury (mTBI) based on configurations of postconcussive symptom (PCS) endorsement, and to examine predictors of subgroup membership.
SETTING
Outpatient Veterans Health Administration (VHA).
PARTICIPANTS
Veterans with clinician-confirmed mTBI who completed the Neurobehavioral Symptom Inventory (NSI), determined using the Comprehensive Traumatic Brain Injury Evaluation database. Individuals who tended to overreport symptoms were excluded via an embedded symptom validity scale.
DESIGN
Retrospective cohort study leveraging national VHA clinical data from 2012 to 2020. Latent class analysis (LCA) with a split-sample cross-validation procedure was used to identify subgroups of veterans. Multinomial logistic regression was used to examine predictors of subgroup membership.
MAIN MEASURES
Latent classes identified using NSI items.
RESULTS
The study included 72 252 eligible veterans, who were primarily White (73%) and male (94%). The LCA supported 7 distinct subgroups of veterans with mTBI, characterized by diverging patterns of risk for specific PCS across vestibular (eg, dizziness), somatosensory (eg, headache), cognitive (eg, forgetfulness), and mood domains (eg, anxiety). The most prevalent subgroup was Global (20.7%), followed by Cognitive-Mood (16.3%), Headache-Cognitive-Mood (H-C-M; 16.3%), Headache-Mood (14.2%), Anxiety (13.8%), Headache-Sleep (10.3%), and Minimal (8.5%). The Global class was used as the reference class for multinomial logistic regression because it was distinguished from others based on elevated risk for PCS across all domains. Female (vs male), Black (vs White), and Hispanic veterans (vs non-Hispanic) were less likely to be members of most subgroups characterized by lesser PCS endorsement relative to the Global class (excluding Headache-Mood).
CONCLUSION
The 7 distinct groups identified in this study distill heterogenous patterns of PCS endorsement into clinically actionable phenotypes that can be used to tailor clinical management of veterans with mTBI. Findings reveal empirical support for potential racial, ethnic, and sex-based disparities in PCS among veterans, informing efforts aimed at promoting equitable recovery from mTBI in this population.
Collapse