Considerations for the Interpretation of Epidemiological Studies of Injuries in Team Sports: Illustrative Examples

Injury surveillance is an important component of the “sequence of prevention”¹ for reducing sports injury incidence, and a number of studies evaluating injury surveillance methodology have aimed to ensure that reliable and comparable data are produced. Arguably, the most important methodological issue surrounding injury surveillance is the definition of injury. This determines the threshold at which an injury is deemed reportable for surveillance purposes and has been addressed by a number of authors.^2-6 The injury definition debate focuses on whether to use a broad definition or whether to place restrictions on a definition that might limit recognition to, for example, an injury that limits game participation. It has been suggested that 70% to 92% of all injuries are excluded when the injury definition is restricted,⁶ but it has been pointed out that no study using a broad definition has demonstrated good reliability and thus a “match time loss only” injury definition should be adopted.⁵

There is an emerging body of evidence that highlights issues of a sociocultural and psychological nature that may affect the validity of injury surveillance findings in team sports. These issues may result in both underreporting and overreporting of injuries. In rugby union, for example, it has been suggested that concussions may be underreported to avoid International Rugby Board rules restricting participation after such injuries.^7,8 These matters, and their impact on injury surveillance, have not been discussed or assessed in detail. We believe that their further discussion is essential for both the development and accurate interpretation of injury surveillance data. Such matters are of particular significance not only for researchers designing and interpreting surveillance studies but also for team sport clinicians who must transfer research findings into evidence-based practice. In addition, they may raise ethical dilemmas more common in a team sports environment.

UNDERREPORTING OF INJURIES

In professional football (soccer), the fear of losing one's place in the team, avoiding the loss of contract bonuses, and wanting to play due to the significance of forthcoming games have been shown to be common incentives to compete with injuries.^9,10 Players have been observed “managing” chronic injuries and delaying surgery to the off-season while playing against medical advice to avoid appearing a burden to the club and being released from a contract.¹¹

Both underreporting of injuries and overestimation of injury recurrence may occur as a result of these practices. Where injury definitions incorporate a greater level of severity and increased degrees of time loss, the possible effect on surveillance findings becomes more pronounced and greater potential appears for underreporting to occur (Table 1). Widely encompassing definitions have a limited effect on the loss of injury data, supporting proposals for the use of broader definitions.⁶ The argument in support of a match time loss only definition is based on reliability; it is suggested that reliable comparable surveillance data have only been produced with match time loss only definitions.⁵ The issue raised by this editorial is one of validity—are the findings of such studies truly representative of all that has occurred within the observed cohort?

TABLE 1: Underreporting: Relationship Between Issues in Surveillance and Varying Injury Definitions

If an injury is detected, overestimation of recurrence rates may occur. Recurrence has been defined in methodology consensus statements as an injury of the same type to the same body part occurring after return to full participation.^12-14 Playing in a match while injured and then continuing injury treatment could result in a new and recurrent injury being recorded, which arguably should be documented as 1 injury with a longer duration of absence. We therefore strongly support the use of “a framework for recording recurrences, reinjuries, and exacerbations in injury surveillance,”¹⁵ which allows the recording of issues such as players participating in matches when only able to partake in limited training activities due to having a nonrecovered injury.

Overuse injuries present a unique problem to researchers in injury epidemiology. The implications of such injuries are dependent on when in the chronic injury continuum they are identified. Subclinical episodes of failed tissue adaptation during a period of overuse are initially asymptomatic,¹⁶ and a time lapse to clinical presentation occurs. This reduces the likelihood of the injury being detected by a time loss definition, particularly if observation periods are short, highlighting a serious issue for researchers investigating sports in which cumulative microtraumatic injuries are common. Standard surveillance models may not be able to detect the magnitude of the overuse injury problem fully; furthermore, disparity in the detection of such injuries may occur between teams with differing resources. A greater level of player monitoring and availability of imaging, leading to a greater detection rate, may be more likely in wealthier organizations.

OVERREPORTING OF INJURIES

Overreporting may occur when players avoid competition and exposure to injury. This has been observed during preseason friendly matches in players who experience mild musculoskeletal discomfort (eg, muscle stiffness) not participating to prevent an injury developing. Similarly, players have been observed feigning or exaggerating an injury to avoid match injury exposure, ensuring that they are able to pass a medical examination when seeking an alternative contract at another club.¹¹

Using an objective, clinical, “tissue damage” injury definition would prevent such occurrences being recorded as surveillance injuries; however, a time loss definition would more likely result in an injury being reported (Table 2). This highlights a “gray area” in injury surveillance methods, where not playing through injury could be differentiated from players being rested or being involved in rotational systems for team selections. Pragmatic criteria to distinguish these phenomena have not been outlined to date. Similarly, whether conditions such as muscle stiffness or muscle soreness should be considered as reportable injuries or conditions is likely to depend on the aims of a particular study, which will therefore determine whether they result in overreporting. Exercising clinical judgment may minimize the likelihood of overreporting of injuries; however, the reliability of applying opinion in these circumstances is unknown.

TABLE 2: Overreporting: Relationship Between Issues in Surveillance and Varying Injury Definitions

OVERCOMING THE PROBLEM

We suggest that factors extraneous to medical issues may influence the data recorded during injury surveillance; some injuries are not detected because the players do not want them to become detected. Because of the nature of the issues discussed, the magnitude of the effect on surveillance study findings is completely unknown and may be difficult to establish. Furthermore, important ethical issues are raised by these observations. For example, placement of team performance before an individual's injury status may increase the risk of longer-term sequelae, and delaying treatment until the off-season may limit recovery time because the off-season becomes encroached by late season games. Conflicts of interest may arise among players, managers, and medical staff during complex return-to-play decisions.

A major decision that must be made when embarking on a surveillance study is the identification of the level of restriction placed on the definition of injury, that is, to consider the trade-off that occurs between the reliability of applying the definition and the validity of the data the definition captures. Broadening some aspects of surveillance may reveal information about some of the issues raised. If players delay effective treatment to the off-season, it would be reasonable to suggest that injury treatments and incidence of surgical intervention would be maintained or increased during the off-season period. Therefore, continuing surveillance throughout the off-season may capture these injuries, which were not identified during the season, giving a truer picture of injury rates. It may be possible to indirectly estimate the frequency that injured players are “patched up” to enable match participation via the auditing of the use of strapping, protective garments, injections, or other similar interventions. Strict unambiguous guidelines on the definition of recovery and adoption of the framework for recording recurrences¹⁵ may also help to improve validity and reliability and improve intrastudy and interstudy consistency.

CONCLUSIONS

Issues affecting the accuracy of injury surveillance have been highlighted, and although the magnitude and effects of the issues are unknown, raising awareness of such occurrences and their associated ethical issues is important to aid interpretation of published data and explain interstudy differences. As a result of the issues discussed, the following questions are raised:

1. How much of an impact do these and other examples have on surveillance findings?
2. Is it enough to acknowledge the flaws in a system, although we may not be able to quantify their size or effect?
3. Is it feasible to incorporate sociocultural and psychological factors into injury definitions and surveillance methods to improve the validity of surveillance findings?

Further research may assist in answering these questions and quantifying the extent and impact of these issues, and others, in injury surveillance.

ACKNOWLEDGMENTS

The authors thank Dr. Lisa Hodgson for her helpful comments.