Citation

Ferris LM, Kontos AP, Eagle SR, Elbin RJ, Clugston JR, Ortega J, Port NL. Vision Res. 2022; 200: e108081.

Copyright

(Copyright © 2022, Elsevier Publishing)

DOI

10.1016/j.visres.2022.108081

PMID

35926346

Abstract

The Vestibular/Ocular-Motor Screening (VOMS), an important component in acute (<72 h) sport-related concussion (SRC) assessment, is increasingly used alongside the Sport Concussion Assessment Tool (SCAT) and as part of the Military Acute Concussion Evaluation 2 (MACE2). VOMS demonstrates clinically useful diagnostic accuracy for acute SRC and improves the overall utility when added to the SCAT3. However, potential overlap among VOMS's vestibular and oculomotor items suggests the possibility of a more efficient version. VOMS and SCAT3 scores were analyzed for 3,958 preseason (47.8% female) and 496 acute-SRC (37.5% female) NCAA-DoD Concussion Assessment, Research, and Education (CARE) consortium collegiate athlete evaluations. Analyses revealed very large effect sizes (d = 2.39-2.45) and high correlations (rho = 0.95-0.99) among all VOMS items except near point of convergence distance (d = 0.79, rho ≤ 0.341). Receiver operating characteristic (ROC) curve analyses showed clinically useful discriminative utility for VOMS Total (AUC = 0.85) and the VOMS Total change score, where pretest symptoms were incorporated (AUC = 0.81). A modified VOMS (mVOMS) consisting of four items (smooth pursuits, horizontal saccades, horizontal vestibulo-ocular reflex, visual motion sensitivity) yielded identical AUCs to VOMS Total. Integer cutoff analyses suggest a score of ≥4 for VOMS Total and ≥4 for mVOMS Total optimizes concussion identification. Incorporating VOMS or mVOMS into SCAT3 (AUC = 0.79) significantly improved the combined tool's acute utility for acute concussion identification by a maximum of 4% (SCAT3+VOMS AUC = 0.84, SCAT3+mVOMS AUC = 0.83). Future versions of SCAT or MACE may want to consider incorporating a more parsimonious VOMS for the purpose of identifying acute concussion.

Language: en

Keywords

Machine learning; Sport-related concussion; Saccades; Vestibular-oculomotor reflex; Vestibular/ocular-motor screening