R J Elbin1, Shawn R Eagle2, Gregory F Marchetti3, Morgan Anderson4, Philip Schatz5, Melissa N Womble6, Katie Stephenson1, Tracey Covassin4, Michael W Collins2, Anne Mucha2, Anthony P Kontos2. 1. Department of Health, Human Performance and Recreation/Office for Sport Concussion Research, University of Arkansas, Fayetteville, Arkansas, USA. 2. Department of Orthopaedic Surgery, UPMC Sports Medicine Concussion Program, University of Pittsburgh, Pittsburgh, Pennsylvania, USA. 3. Rangos School of Health Sciences, Duquesne University, Pittsburgh, Pennsylvania, USA. 4. Department of Kinesiology, Michigan State University, East Lansing, Michigan, USA. 5. Department of Psychology, Saint Joseph's University, Philadelphia, Pennsylvania, USA. 6. Inova Sports Medicine Concussion Program, Fairfax, VA, USA.
Abstract
OBJECTIVE: To develop clinical cutoffs using change scores for the VOMS individual items and an overall VOMS change score that identified concussion in adolescent athletes. METHODS: Change score clinical cutoffs were calculated from a sample of adolescents (13-18 years) with SRC (n = 147) and a sample of uninjured adolescents CONTROL (n = 147). Receiver operating characteristic (ROC) curves, with area under the curve (AUC), based on Youden's J statistic were used to identify optimal cutoffs for identifying SRC from CONTROLS using VOMS individual item change scores, an overall VOMS change scores, and NPC distance (cm). RESULTS: AUC values for VOMS item change scores ranged from .55 to .71. Optimal change score cutoffs were ≥1 for VOMS items and ≥3 for overall VOMS change score. The optimal cutoff for NPC distance was ≥3 cm. A ROC analysis revealed a three-factor model (AUC = .76) for identifying SRC that included vertical vesibular ocular reflex (VVOR), visual motion sensitivity (VMS), and NPC distance items. The AUC (.73) for the overall VOMS change score was higher than any individual VOMS AUC values. CONCLUSIONS: This study supports an alternate scoring approach and clinical interpretation of VOMS items involving change scores that account for pretest symptoms.
OBJECTIVE: To develop clinical cutoffs using change scores for the VOMS individual items and an overall VOMS change score that identified concussion in adolescent athletes. METHODS: Change score clinical cutoffs were calculated from a sample of adolescents (13-18 years) with SRC (n = 147) and a sample of uninjured adolescents CONTROL (n = 147). Receiver operating characteristic (ROC) curves, with area under the curve (AUC), based on Youden's J statistic were used to identify optimal cutoffs for identifying SRC from CONTROLS using VOMS individual item change scores, an overall VOMS change scores, and NPC distance (cm). RESULTS: AUC values for VOMS item change scores ranged from .55 to .71. Optimal change score cutoffs were ≥1 for VOMS items and ≥3 for overall VOMS change score. The optimal cutoff for NPC distance was ≥3 cm. A ROC analysis revealed a three-factor model (AUC = .76) for identifying SRC that included vertical vesibular ocular reflex (VVOR), visual motion sensitivity (VMS), and NPC distance items. The AUC (.73) for the overall VOMS change score was higher than any individual VOMS AUC values. CONCLUSIONS: This study supports an alternate scoring approach and clinical interpretation of VOMS items involving change scores that account for pretest symptoms.