Concussion is a common injury in rugby union (‘rugby’) and yet its diagnosis is reliant on clinical judgment. Oculomotor testing could provide an objective measure to assist with concussion diagnosis. NeuroFlex® evaluates oculomotor function using a virtual-reality headset. This study examined differences in NeuroFlex® performance in clinician-diagnosed concussed and not concussed elite male rugby players over three seasons. NeuroFlex® testing was completed alongside 140 head injury assessments (HIAs) in 122 players. The HIA is used for suspected concussion events. Of these 140 HIAs, 100 were eventually diagnosed as concussed, 38 were not concussed (2 were unclear) Eight of the 61 NeuroFlex® metrics were analysed as they were comparable at all time points. These eight metrics, from three oculomotor domains (vestibulo-ocular reflex, smooth pursuit and saccades), were tested for their ability to distinguish between concussed and not concussed players using mean difference / odds ratios and corresponding 95% confidence intervals (CI’s). General and generalised linear mixed models, accounting for baseline test performance, were used to determine any meaningful differences in concussed and not concussed players. The diagnostic accuracy of these differences was provided by the area under the receiver operating curve (AUC). Only one of the eight metrics (number of saccades, smooth pursuit domain) had clear differences in performance between concussed and not concussed players at the HIA during the match (odds ratio: 0.76, 95%CI: 0.54–0.98) and after 48 hours (0.74, 95%CI: 0.52–0.96). However, the direction of this difference was contrary to clinical expectations (concussed performed better than not concussed) and the AUC for this outcome was also poor (0.52). NeuroFlex® was unable to distinguish between concussed and not concussed players in this elite male cohort. Future research could study other cohorts, later time points before return to play, and the tool’s role in rehabilitation.

Concussion is a common injury in rugby union (‘rugby’) and yet its diagnosis is reliant on clinical judgment. Oculomotor testing could provide an objective measure to assist with concussion diagnosis. NeuroFlex® evaluates oculomotor function using a virtual-reality headset. This study examined differences in NeuroFlex® performance in clinician-diagnosed concussed and not concussed elite male rugby players over three seasons. NeuroFlex® testing was completed alongside 140 head injury assessments (HIAs) in 122 players. The HIA is used for suspected concussion events. Of these 140 HIAs, 100 were eventually diagnosed as concussed, 38 were not concussed (2 were unclear) Eight of the 61 NeuroFlex® metrics were analysed as they were comparable at all time points. These eight metrics, from three oculomotor domains (vestibulo-ocular reflex, smooth pursuit and saccades), were tested for their ability to distinguish between concussed and not concussed players using mean difference / odds ratios and corresponding 95% confidence intervals (CI’s). General and generalised linear mixed models, accounting for baseline test performance, were used to determine any meaningful differences in concussed and not concussed players. The diagnostic accuracy of these differences was provided by the area under the receiver operating curve (AUC). Only one of the eight metrics (number of saccades, smooth pursuit domain) had clear differences in performance between concussed and not concussed players at the HIA during the match (odds ratio: 0.76, 95%CI: 0.54–0.98) and after 48 hours (0.74, 95%CI: 0.52–0.96). However, the direction of this difference was contrary to clinical expectations (concussed performed better than not concussed) and the AUC for this outcome was also poor (0.52). NeuroFlex® was unable to distinguish between concussed and not concussed players in this elite male cohort. Future research could study other cohorts, later time points before return to play, and the tool’s role in rehabilitation.