A Preventive Model for Hamstring Injuries in Professional Soccer: Learning Algorithms

Abstract

Hamstring strain injury (HSI) is one of the most prevalent and severe injury in professional soccer. The purpose was to analyze and compare the predictive ability of a range of machine learning techniques to select the best performing injury risk factor model to identify professional soccer players at high risk of HSIs. A total of 96 male professional soccer players underwent a pre-season screening evaluation that included a large number of individual, psychological and neuromuscular measurements. Injury surveillance was prospectively employed to capture all the HSI occurring in the 2013/2014 season. There were 18 HSIs. Injury distribution was 55.6% dominant leg and 44.4% non-dominant leg. The model generated by the SmooteBoostM1 technique with a cost-sensitive ADTree as the base classifier reported the best evaluation criteria (area under the receiver operating characteristic curve score=0.837, true positive rate=77.8%, true negative rate=83.8%) and hence was considered the best for predicting HSI. The prediction model showed moderate to high accuracy for identifying professional soccer players at risk of HSI during pre-season screenings. Therefore, the model developed might help coaches, physical trainers and medical practitioners in the decision-making process for injury prevention.

Prediction models for musculoskeletal injuries in professional sporting activities: A systematic review

“…The median mean age was 24.4 (2.4) years, median mean height was 185.4 (7.9) cm, and median mean body mass was 87.3 (3.4) kg. The distribution of sex was all‐males in 8 studies 17,19‐24,26 . The remaining 2 studies did not report their distribution of sex 18,25 .…”

“…The literature search resulted in 10 studies that met the eligibility criteria and therefore included in the systematic review (Figure 1). 17‐26 …”

“…Item‐specific scores of the Newcastle‐Ottawa Scale are illustrated in Appendix 1. The definition of a musculoskeletal injury was reported to be specifically a “time‐loss” injury in 7 studies 19‐22,24‐26 . The remaining 3 studies did not state whether the injury definition was exclusive to time‐loss injuries 17,18,23 .…”

“…Data components employed for data pre‐processing included body composition data, cardiopulmonary assessments, injury history, neuromuscular assessments, player profile, players statistics, psychological assessments and workload data. The most commonly employed data component was body composition data (7 studies), 17,20‐23,25,26 followed by player profile, that ism age/position (6 studies) 17,18,20‐22,26 then injury history 20‐22,25,26 and workload data (5 studies each) 17‐19,22,24 …”

“…The most commonly constructed machine learning technique was a decision tree followed by logistic regression, then both random forest and supporting vector machine. Sampling techniques were reported in 6 studies with the synthetic minority oversampling technique most commonly employed 17,18,20‐22,26 . Cross‐validation was reported in 9 studies with 10‐fold cross‐validation, the most common method.…”