A video review of multiple concussion signs in National Rugby League match play

Video review has been introduced in many professional sports worldwide to help recognize concussions. However, to date, there has been very little research on the accuracy of using video analysis to identify signs of concussion and the various combinations of observed signs.

The objective of the study is to determine the accuracy of combinations of clinical signs of concussion identified using video analysis to identify concussions in the National Rugby League (NRL). Incidences of players using of the concussion interchange rule (CIR) (n = 156), including those where athletes were diagnosed with a concussion (n = 60), were used to calculate sensitivity and specificity of various combinations of concussion signs (unresponsiveness, slow to get up, clutching/shaking head, gait ataxia, vacant stare, and apparent seizure) and their independent association with an eventual diagnosis of concussion.

Using video analysis, players who were diagnosed with a concussion showed a significantly greater total number of signs at the time of injury (mean = 3.4, SD = 1.3) than those who were removed from play but not diagnosed with a concussion (mean = 3.0, SD = 0.9 signs; p = .046). Players who did not return to play during the same game demonstrated a significantly greater number of total signs than those who did return to play in the same game following CIR activation (mean = 3.4, SD = 1.2 versus mean = 2.9, SD = 0.9; p = 0.002). The most common combination of signs that was observed was clutching/shaking the head and slowness in getting up (17.3%). The sensitivity of the total number of signs observed decreased as the number of signs increased (range = 0.13–0.62), while the specificity increased as more signs were observed (range = 0.29–0.90). Most of the combinations of different observed signs at the time of potential injury were highly specific (> 0.80), but not sensitive to an eventual diagnosis of concussion. When considering all potential predictor variables in a logistic regression model, anticipating the oncoming collision (OR = 3.92, 95% CI = 1.28–12.03), fewer number of defenders involved in the tackle (OR = 0.58, 95% CI = 0.36–0.92), and the presence of a blank or vacant stare (OR = 2.97, 95% CI = 1.26–7.01) were each significantly associated with concussion diagnoses.

The use of video review in the NRL is challenging, but being aware of the combinations of possible concussion signs and the likelihood that various presentations result in a concussion diagnosis can provide a useful addition to sideline concussion identification and removal from play decisions.