Objectives
The purpose of this study was to investigate head kinematic variables in elite men’s and women’s rugby union and their ability to predict player removal for an off-field (HIA1) head injury assessment.
Methods
Instrumented mouthguard (iMG) data were collected for 250 men and 132 women from 1865 and 807 player-matches, respectively, and synchronised to video-coded match footage. Head peak resultant linear acceleration (PLA), peak resultant angular acceleration (PAA) and peak change in angular velocity (dPAV) were extracted from each head acceleration event (HAE). HAEs were linked to documented HIA1 events, with ten logistical regression models for men and women, using a random subset of non-case HAEs, calculated to identify kinematic variables associated with HIA1 events. Receiver operating characteristic curves (ROC) were used to describe thresholds for HIA1 removal.
Results
Increases in PLA and dPAV were significantly associated with an increasing likelihood of HIA1 removal in the men’s game, with an OR ranging from 1.05–1.12 and 1.13–1.18, respectively. The optimal values to maximise for both sensitivity and specificity for detecting an HIA1 were 1.96 krad⋅s−2, 24.29 g and 14.75 rad⋅s−1 for PAA, PLA and dPAV, respectively. Only one model had any significant variable associated with increasing the likelihood of a HIA1 removal in the women’s game—PAA with an OR of 8.51 (1.23–58.66). The optimal values for sensitivity and specificity for women were 2.01 krad⋅s−2, 25.98 g and 15.38 rad⋅s−1 for PAA, PLA and dPAV, respectively.
Conclusion
PLA and dPAV were predictive of men’s HIA1 events. Further HIA1 data are needed to understand the role of head kinematic variables in the women’s game. The calculated spectrum of sensitivity and specificity of iMG alerts for HIA1 removals in men and women present a starting point for further discussion about using iMGs as an additional trigger in the existing HIA process.
ORCID: 
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)