• Soccer Injury Epidemiology

    Epidemiology has been defined as “a study of the distribution and determinants of health related states and the application of this study to the control of health problems” (Cherkasski 1991). Although epidemiological research has become more complex and specialised with the accumulation of medical and scientific knowledge, the essence of a good epidemiological study is defined by a simple approach, that allows the reader to follow the argument and draw their own conclusions (Hennekens and Buring 1994).

    On an individual level the clinician incorporates epidemiology into their practice as part of the clinical reasoning process, without perhaps knowing it. This manifests itself in many aspects of practice, ranging from having knowledge of the incidence of a particular injury, which helps the expert clinician to quickly focus the assessment, to injury prevention, where the clinicians knowledge of attributable risk may lead them to advocate an intervention in an attempt to reduce the risk of injury.

    Soccer has been described by Fevre (1998) as a “collision sport” due to its physical nature, but epidemiological studies suggest that many injuries are not due to direct contact, but other causes such as overuse or overstrain (Ekstrand and Gillquist, 1983). The fact that expert clinicians have, apparently, diverse opinions on the nature and etiology of injury within soccer is due to differing interpretation of the available epidemiological evidence.

    Hennekens and Buring (1994) state that the interpretation of epidemiological data is dependent on two assumptions. The first is that the activity in question – soccer – causes the injury. The second assumption is that other causes of an injury are equally distributed between participants who get injured and those who don’t. The first assumption can be satisfied, as all injuries in the research that is examined within this review indisputably occurred during soccer. However, the second assumption may not apply due to methodological problems in the literature.

    Soccer injury epidemiological research takes place in the clinical environment and there is a lack of experimental control. The main outcome of this type of research is to form the basis of epidemiological hypotheses, that are founded upon observations of associations between two variables. Because of the lack of experimental control, the questions is “are the associations valid?” or “are they due to bias or error?”

    There have been numerous epidemiological studies related to soccer, undertaken in numerous countries and employing various methods of data collection. The diversity of the research poses a problem to the reader, as it makes it difficult to interpret the findings in a comparable manner. Kraus and Burg (1970) state that the inconsistent collection and recording of data is a fundamental problem with sports epidemiological data, which decreases uniformity and subsequent reliability. They argue that the dominant element to this end is the research objective which must be clearly stated. Furthermore, the data collection source must be reliable, otherwise the data will be of poor quality and of insufficient depth. Kraus and Burg (1970) state that probably the biggest problem in sports epidemiology is the definition of an injury, since “a sports injury is a relative event…where is the cut off point between pre-clinical unapparent injury and gross or severe anatomic trauma?”

    Ekstrand and Gillquist (1983) undertook a prospective study of 180 amateur football players to elucidate the incidence of injury during the year. The authors found that 88% of injuries involved the lower extremity and that strains accounted for 48% of all injuries. This lead the authors to advocate a lower limb stretching and strengthening regime as a strategy to reduce the incidence of these injuries – these suggestions being based in convention rather than being backed by strong scientific evidence. Ekstrand and Gillquist (1983) also found that 69% of knee injuries were caused by trauma and that 61% of these were caused by a collision. On face value this may be useful data for the clinician to consider and the strategies advocated by the authors may conceivably be beneficial. However, there is a sense that the data does not inform the clinician more than they subjectively know from experience. This together with methodological issues, reduce the clinical worth of the study and make it of limited use in the clinical reasoning process.

    Ekstrand and Gillquist (1983) provide a confusing definition of injury, where return to practice is the main criteria, but this may be related to other factors such as the player, for some personal reason, not wanting to return to practice, rather than the severity of the injury. The study only incorporated about ten teams in total and each team coach selected 15 players to take part. The authors do not state the selection criteria used by the coaches and this may affect the outcome. Although all of the injuries were examined by the same orthopaedic consultant the data was recorded retrospectively and therefore not all the information may be recalled. In research terms, this can be classified as systematic error. This level of reliability may be no better than the sports clinicians subjective recall, and for this reason it is likely the personal experience of the clinician will have a greater bearing on the clinical reasoning process than research of this standard.

    Keller et al (1987) reviewed six major epidemiological studies in soccer. The authors again highlight the array of definitions of injury in the literature. This aside, based on the available descriptive evidence the authors concluded that, because of the increase in size and intensity of play, there were more injuries due to collision in older more elite players, but that the number of muscle injuries was less in these groups, because of better physical preparation. This hypothesis has face validity but needs to be subjected to scrutiny by further research.

    So too does the theory put forward by Keller et al (1987), that muscle strain and tendonitis is significantly higher in players with reduced flexibility. This epidemiological data can provide clues leading to the formulation of an epidemiological hypothesis that is consistent with existing knowledge, but an association does not prove cause and effect. Analytical epidemiology endeavours to test a hypothesis to judge whether a particular exposure causes or prevents injury. This data is used by clinicians to allocate resources and implement injury prevention strategies. Longitudinal and experimental studies are required to test these hypotheses but, until this research is completed, clinicians must make decisions based on the available evidence and current theories which have biological grounding.

    In keeping with this, it is conventional practice for clinicians to advocate a warm up and stretching program in an effort to prevent muscle and tendon injuries. More research is needed to evaluate the efficacy of these regimes. There is a lack of scientific credibility behind stretching as an injury prevention strategy, although there is biological credibility to this approach.

    Hawkins and Fuller (1999) examined the incidence of injury within professional football players in England, in a study that can be classified as observational epidemiology and is not strictly a correlational study. The authors tend to exceed the limitations of this type of research by drawing conclusions that infer cause and effect. This is not appropriate, as the observations merely show a relationship. The authors found a higher injury frequency in matches compared to training per 1000 hours of exposure. On the face of it, this data may falsely lead the reader to conclude that more injuries occur in matches but, given that training time to match time is approximately 7:1, more injuries actually occur during training.

    Hawkins and Fuller (1999) also demonstrated that significantly more injuries occurred during the last fifteen minutes of each half, and that there were significantly more injuries in the second half compared to the first. The authors felt that this was due to fatigue, and advocated endurance training and carbohydrate intake to counter this problem. This hypothesis has good face validity and would be difficult to ethically research, which means it is difficult to falsify. As endurance training would have no serious ill effects on the player, it seems sensible and practical to introduce this as an aspect of the training program.

    Overall, the value of football injury epidemiological data to the clinician working with football players is limited. The findings of such research should be interpreted with care. Often the causes of injury are multi-factorial, with each of the contributing factors difficult to measure objectively. In epidemiological terms, it is pragmatic to refer to component causes of injury which comprise necessary and sufficient causes of injury. In football injury these causes comprise of collisions, fatigue, overuse, muscle imbalance, poor technique, inadequate strength, inappropriate diet and other such factors which have been described in the literature as etiological.

    In research terms, the difficulty arises when attempts are made to operationalise these concepts. The fact that epidemiogical research takes place in the sporting environment means that there is a lack of research control and a subsequent question mark over the validity of this research. The clinician can draw a general understanding of the etiology and mechanism of injury based on epidemiological research, but the wholesale implementation of strategies designed to reduce injury based on current evidence is not exact, and therefore not completely reliable. Although causation is difficult to pin down, the evidence at present allows the clinician to identify associations. In the absence of experimental evidence, these factors help guide the clinical reasoning process.

    Cherkasski, BL. (1991) The modern interpretation of the basic categories of epidemiology. Zhurnal Mikrobiologii, Epidemiologii i Immunobiologii. Vol 2, February, page 75-8.

    Ekstrand, J. and Gillquist, J. (1983) Soccer injuries and their mechanism: a prospective study. Medicine and Science in Sports and Exercise Volume 15, Number 3, page 267 – 270.

    Ekstrand J. Gillquist J.(1983) Soccer injuries and their mechanisms: a prospective study. Medicine & Science in Sports & Exercise. Vol 15, No 3, page 267-70.

    Wiktorsson-Moller M. Oberg B. Ekstrand J. Gillquist J. (1983) Effects of warming up, massage, and stretching on range of motion and muscle strength in the lower extremity. American Journal of Sports Medicine. Vol 11, No 4, page 249-52.

    Ekstrand J. Gillquist J. The avoidability of soccer injuries. (1983) International Journal of Sports Medicine. Vol 4, No 2, page 124-8.

    Ekstrand, J. Wiktorsson, M. Oberg, B. Gillquist, J. Lower extremity goniometric measurements: a study to determine their reliability. (1982) Archives of Physical Medicine & Rehabilitation. Vol 63, No 4, page 171-175.

    Fevre, D.J. (1998) Collision Sports: Injury and Repair Butterworth Heinemann, Oxford.

    Hawkins, R.D. and Fuller C.W. (1999) A prospective epidemiological study of injuries in four English professional football clubs. British Journal of Sports Medicine Volume 33, page 196 – 203.

    Hennekens C. H. and Buring J. E. (1994) Contributions of Observational Evidence to Prevention Research, Preventive Medicine Volume 23, Issue 5, September, Pages 584-586.

    Keller, C.S., Noyes, F.R., and Bunchner, C.R. (1987) The medical aspects of soccer injury epidemiologogy. The American Journal of Sports Medicine Vol 15, No 3, page 230 – 237.

    Kraus, J.F. and Burg, F.D. (1970) Injury reporting and recording: some essential elements in the collection and retrieval of sports injury information. Journal of the American Medical Association Vol 213, No 3, page 438 – 437.

    Leave a comment