Epidemiology and predictive models of injuries in professional soccer

Abstract

ABSTRACT Soccer (also known as football) requires players to perform many repeated high intensity movements such as sudden acceleration and deceleration, rapid changes of direction, jumping and landing tasks; as well as many situations in which players are involved in tackling to keep possession of or to win the ball. At professional level, the combination of these high physical demands alongside stress and anxiety caused by the congested match calendar may place players at high risk of injury. In fact, soccer is one of the sports with higher injury incidence rates, all of this despite the substantive effort made by the scientific community and physical trainer practitioners to reduce their number and severity. The inefficacy of the preventive measures applied might be caused, in part, by the limitations present in the scientific literature which hinder: a) the accurate estimation of the most frequent soccer-related injuries; b) the identification of professional athletes at high risk of injury; and c) the design of effective neuromuscular training interventions. Therefore, and based on these limitations, the main objectives of the current doctoral thesis were: 1) to carry out a systematic review and a novel meta-analysis of epidemiological data of injuries in professional male soccer; 2) to describe the lower extremity joint ranges of motion profile in professional soccer players; 3) to analyse and compare the behaviour of some machine learning methods in order to select the best performing injury risk factor model to identify professional athletes at risk of lower extremity muscle injuries and hamstring strains; and 4) to analyse the relationships between several parameters of neuromuscular performance with unilateral dynamic balance in such cohort of athletes. To achieve these objectives, a systematic literature review and meta-analysis, a descriptive study, two prospective cohort studies, and a correlational study were conducted. The main findings of the first study report that professional male soccer players are exposed to a substantial risk of sustaining injuries, especially during matches (32.9 injuries per 1000 hours of player exposure). In particular, the lower extremity is the most frequently injured part of the body, being the thigh the anatomical region in which injuries occurs more. Likewise, the most common type of injury is muscle/tendon strains. On the other hand, the results of study two show the necessity of prescribing exercises aimed at improving hip flexion with knee extended and ankle dorsiflexion with knee flexed ranges of motion within soccer training routines. In addition, as some bilateral deficits were observed, unilateral training should be considered where appropriate. Studies three and four present two different injury risk factor models (personal, psychological and neuromuscular risk factors) to identify players at high risk of lower extremity muscle injuries (area under the receiver operating characteristic curve = 0.747) and hamstring strains (area under the receiver operating characteristic curve = 0.867), respectively. Both models are generated by the SmooteBoost technique with a cost-sensitive alternating decision tree as base classifiers. Finally, the findings of study five indicate that, although male and female professional soccer players report similar unilateral dynamic balance scores, but different measures of neuromuscular performance seem to have influenced this fundamental ability. Thus, for males, those variables related to movement patterns in the sagittal plane (hip flexion and ankle dorsiflexion range of motion measures) were important in the overall balance score obtained. However, for females, variables related to the performance of movement patterns in the frontal plane (such as core stability and hip abduction strength and range of motion) were considered predictor variables of this ability. Overall, both the results and methodology used in the present doctoral thesis might be used by coaches, physical trainers and clinicians to improve the decision-making process to reduce the number and impact of injuries in professional soccer. Keywords: football, injury, prevention, hamstring strain, muscle injury, learning algorithm, data mining, dynamic balance, core stability, performance, range of motion.