TY - CONF

T1 - Non-contact lower limb injuries in Rugby Union: a two-year pattern recognition analysis of injury risk factors

AU - Evans, Seren

AU - Owen, Robin

AU - Whittaker, Gareth

AU - Davis, Oran Elphinstone

AU - Jones, Eleri

AU - Hardy, James

AU - Owen, Julian

PY - 2024/7/8

Y1 - 2024/7/8

N2 - Lower limb musculoskeletal injuries, notably lateral ankle sprains, consistently emerge as predominant injuries in injury surveillance studies within Rugby Union. The cause of sport injuries is multifactorial and requires sophisticated statistical approaches to accurately identify risk factors predisposing athletes to injury. Pattern recognition analyses may be useful in injury risk prediction due to their ability to account for repeated measures, non-linear interactions, and imbalanced datasets, however there are limited examples of their use in injury risk prediction. Senior Regional Academy Rugby Union players were monitored over two consecutive seasons which included 1740 individual weekly data points includingtraining load, performance testing, musculoskeletal screening, and injury history parameters. Predictive models (injured vs. non-injured) were generated for non-contact ankle and severe lower limb non-contact injuries using Bayesian pattern recognition from a pool of 36 Senior Academy Rugby Union athletes. Compared to non-injured players, lower dorsiflexion angle measures were predictive of non-contact ankle injuries(32.5±11°), along with slower sprint times over 10m (1.80±0.2s) and 40m (5.43±0.4s), greater body mass (105.7±15.2kg), previous concussion, and previous ankle injury (Area Under ROC = 0.76). Predictors of severe non-contact lower limb injuries included greater changes in mean weekly training load (2022.4±397.4AU), slower 10m (1.76±0.1s) and 40m sprint time (5.44±0.34s), reduction in hamstring (-4.83±15.5mmHg) and adductor strength (-8.4±15.7mmHg), reduction in dorsiflexion angle(-1.13±5.4°), greater increases in perceived muscle soreness (+0.76±1.24), and playing as a forward (Area Under ROC = 0.72). These findings provide evidence supporting targeted prospective injury risk analysis from routine monitoring of athletes, enabling coaches and medical practitioners to have actionable thresholds to tailor training regimes and injury prevention protocols for non-contact ankle and lower limb non-contact injuries.

AB - Lower limb musculoskeletal injuries, notably lateral ankle sprains, consistently emerge as predominant injuries in injury surveillance studies within Rugby Union. The cause of sport injuries is multifactorial and requires sophisticated statistical approaches to accurately identify risk factors predisposing athletes to injury. Pattern recognition analyses may be useful in injury risk prediction due to their ability to account for repeated measures, non-linear interactions, and imbalanced datasets, however there are limited examples of their use in injury risk prediction. Senior Regional Academy Rugby Union players were monitored over two consecutive seasons which included 1740 individual weekly data points includingtraining load, performance testing, musculoskeletal screening, and injury history parameters. Predictive models (injured vs. non-injured) were generated for non-contact ankle and severe lower limb non-contact injuries using Bayesian pattern recognition from a pool of 36 Senior Academy Rugby Union athletes. Compared to non-injured players, lower dorsiflexion angle measures were predictive of non-contact ankle injuries(32.5±11°), along with slower sprint times over 10m (1.80±0.2s) and 40m (5.43±0.4s), greater body mass (105.7±15.2kg), previous concussion, and previous ankle injury (Area Under ROC = 0.76). Predictors of severe non-contact lower limb injuries included greater changes in mean weekly training load (2022.4±397.4AU), slower 10m (1.76±0.1s) and 40m sprint time (5.44±0.34s), reduction in hamstring (-4.83±15.5mmHg) and adductor strength (-8.4±15.7mmHg), reduction in dorsiflexion angle(-1.13±5.4°), greater increases in perceived muscle soreness (+0.76±1.24), and playing as a forward (Area Under ROC = 0.72). These findings provide evidence supporting targeted prospective injury risk analysis from routine monitoring of athletes, enabling coaches and medical practitioners to have actionable thresholds to tailor training regimes and injury prevention protocols for non-contact ankle and lower limb non-contact injuries.

KW - Machine learning

KW - Sports Medicine

KW - Rugby

KW - Sport injury

M3 - Poster

T2 - UK Collaborating Centre on Injury and Illness Prevention in Sport (UKCCIIS) conference

Y2 - 8 July 2024 through 9 July 2024

ER -