Article Text

Download PDFPDF

Improving the accuracy of sports medicine surveillance: when is a subsequent event a new injury?
  1. Ian Shrier1,
  2. Ben Clarsen2,
  3. Evert Verhagen3,
  4. Kerry Gordon4,
  5. Jay Mellette4
  1. 1Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, McGill University, Montreal, Quebec, Canada
  2. 2Oslo Sports Trauma Research Center, Oslo, Norway
  3. 3Department of Public and Occupational Health, EMGO+ Institute at the VU University Medical Center in Amsterdam, Amsterdam, The Netherlands
  4. 4Department of Performance Medicine, Cirque du Soleil, Montreal, Quebec, Canada
  1. Correspondence to Dr Ian Shrier, Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, 3755 Cote Ste-Catherine Road, Montreal, Quebec, Canada H3T 1E2; ian.shrier{at}


The recent increased use of injury and illness surveillance programmes has the potential to greatly advance our knowledge about risk factors and treatment effectiveness. Maximising this potential requires that data be entered in a format that can be interpreted and analysed. One remaining challenge concerns whether and when an increase in symptoms should be documented within an existing injury record (eg, exacerbation) versus a new injury record. In this review, we address this challenge using the principles of the multistate framework for the analysis of subsequent injury in sport (M-FASIS). In brief, we argue that a new injury record should be documented whenever there is an increase in symptoms due to activity-related exposures that is beyond the normal day-to-day symptom fluctuations, regardless of whether the athlete was in a ‘healthy state’ immediately before the event. We illustrate the concepts with concrete examples of shoulder osteoarthritis, ankle sprains and ACL tears.

  • Chronic
  • Injury
  • Overuse injury
  • Recurrent
  • Surveillance

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The development of effective athlete prevention and rehabilitation programmes can be enhanced through appropriate risk factor analyses based on valid and reliable injury and illness surveillance data. Surveillance methods and software should allow clinicians to easily document injuries and illnesses.

When people have no previous injuries prior to study entry, the calculation of incidence is usually considered straightforward, although the ‘start’ time for overuse injuries can sometimes present a challenge. However, in a study by Clarsen et al,1 only 9% of injuries were clearly identified as new injuries, and most athletes had previous injuries or symptoms at baseline.

When there are no clearly defined surveillance methods to guide clinicians entering data in real-world situations, some will include an event as a new injury record while others may include it only in the notes of follow-up therapy sessions. If different people enter data differently, it is difficult to answer questions that require differentiation between re-injuries, exacerbations and normal fluctuation in symptoms that occur as part of the natural history of chronic conditions. Some authors have proposed surveillance methods to capture the normal fluctuation in symptoms over time.2 In this brief review, we address one of the remaining challenges: when should an event or an increase in symptoms be documented as a new injury (or illness) record instead of being included within the notes of a follow-up therapy session of the initial injury (or illness)?

As examples, we consider shoulder osteoarthritis, ankle sprains and ACL tears. Although our focus is on sport injuries, the same challenges exist for other medical conditions with fluctuating symptoms (eg, asthma). Our approach is based on the multistate framework for the analysis of subsequent injury in sport (M-FASIS).3 In brief, multistate approaches focus on the change in injury state for any analysis.

With respect to defining injury, recent definitions imply that an athlete moves from a healthy state to a non-healthy state. In other words, if an athlete was not in a healthy state following a previous injury, the new event would be considered an exacerbation rather than a new injury. This creates a challenge when trying to document events such as exacerbations (worsening of an unhealed injury) and changes in chronic conditions (conditions that never heal, such as osteoarthritis). Typically, these types of events are documented within follow-up notes of injury files. M-FASIS simply expands the scope and defines an injury as occurring when one moves from any state to a less-healthy state. This small shift leads to a more consistent and logical approach with respect to documenting subsequent injuries and exacerbations of injuries or chronic conditions.

General principles for when a new injury record is required

M-FASIS considers the athlete to be in a particular ‘state’ of health at any one time. In general epidemiology, incidence rate is calculated as the number of events (ie, injuries or illnesses) among participants divided by the total exposure time of the participants at risk. We argue that this means a new injury record should be documented whenever a participant changes from one state to a less-healthy state, and the change was directly linked to exposure, regardless of whether the athlete was ‘healthy’ or not at the time of the injury. For example, if athletes were participating in a sporting competition, an appropriate analysis would consider them to be at risk of injury each time they participate. However, if creating a new injury record requires that the athlete is in only one particular state of health (ie, 100% healed), an athlete who is not 100% healthy counts only in the denominator because they cannot be re-injured. This approach would underestimate the true risk of the activity. Excluding injured participating athletes is a poor option because an athlete with a previous ankle sprain may injure their shoulder. Therefore, one would need to count the participation of this athlete for some injuries, but not for other injuries. Similarly, an incidence calculation should never include an event if it occurred at a time that was not linked to an exposure or cumulative exposures. Otherwise, the analysis would be counting injuries that occurred outside of the ‘risk set’ and overestimating the true risk.

The important perspective of M-FASIS is that the data gathering stage should be independent of the status of the athlete's original injury (ie, healed or not healed). In other words, an exacerbation (worsening of an injury that has not healed) should be documented as a new injury record (as opposed to a visit note related to the old injury record) within the surveillance software because the only difference from a traditional new injury record is the health state immediately before the event. For the reasons explained above (and elsewhere4), this differs from the recommendations of Finch and Cook5 who suggested that investigators only document incidents as new injury records (new index injuries) if they believe the recent incident was not related to the previous injury. When adopted, these methods will provide investigators the flexibility to answer traditional questions such as total number of days missed from the first injury, and also to answer questions such as the incidence of ‘exacerbations’ during rehabilitation or for chronic conditions. These concepts are illustrated in the concrete examples below.

How to record injuries when the initial injury is an acute injury

Subsequent injury is an acute injury (sudden onset, usually traumatic)

Consider an example: an athlete with an initial acute ankle sprain who experiences a new inversion injury because he or she slipped on a wet field. Failing to record the second acute injury would lead to underestimation of risk due to wet field conditions. Yet historically, surveillance methods would usually link the information about this new event to an existing injury and create a new injury record only if the previous ankle sprain was ‘healed’. Therefore, one must define a ‘healed state’, which has sometimes been defined as return to activity6 or end of treatment.3 ,7 However, most clinicians would not consider an injury healed if symptoms persist, and this is often the case even after athletes stop treatments.2 The M-FASIS approach suggests that one should always create a new injury record for this new event because the state of health immediately prior to the new injury should not influence the categorisation.

Subsequent injury is a gradual-onset (overuse) injury

Keeping in mind a history of an acute ankle injury, consider a different example of an athlete who never previously injured his or her ankle and gradually develops ankle pain. This painful injury would be documented as a new injury record in both the traditional and M-FASIS approaches. Now, what if this gradual-onset injury had been preceded by a previous acute ankle sprain? In the traditional approach, the subsequent gradual-onset injury would only be documented as a new injury record if the athlete was pain-free or treatment-free or pain-free and treatment-free. However, there is no fundamental difference in the gradual-onset injury between the examples of the athlete with no previous ankle injury and the athlete with the previous ankle injury. The M-FASIS approach represents a more internally consistent approach where the subsequent gradual-onset injury is treated the same regardless of the state of health immediately prior to the subsequent injury.

Defining the healed state?

In the above examples, both the traditional and M-FASIS approaches require the clinician to decide if the change in health status (state) is enough to warrant the documentation of a new injury, and to decide if the state change occurred because the athlete was exposed to activity. However, the traditional approach requires the additional challenge of defining a healed state. Since the natural history for injuries is that symptoms often fluctuate over time, and athletes can return to unrestricted activity while still having treatment, this can be especially difficult or arbitrary. Further, imaging often suggests persistent structural damage long after symptoms and signs subside,8–10 so the definition of a healed state as ‘pain and treatment free’ still includes some misclassification. Finally, a healed state is conceptually a return to baseline state. However, preinjury and postinjury values for tissue function (eg, proprioception) or general well-being (including psychological state, readiness to play) are rarely available.

Even if these additional challenges were solved, we remain unsure of the advantage of even considering the state of healing prior to the event as a criterion. Rather, documenting all activity-related increases in tissue damage that are outside the usual fluctuation of symptoms and signs as new injury records provides distinct advantages. This will allow researchers to address many different research questions using the surveillance data, whereas current practices of documenting injury events limit the possible analyses. For example, using the date of healing (however that is defined), one would be able to categorise the second ankle sprain6 ,11 as a re-injury (injury date ≥healing date of 1st ankle sprain), an exacerbation (injury date <healing date of 1st ankle sprain), or combine the two injuries into one (because the healing dates of the first and second ankle sprains are the same). The current practice of only documenting injuries as ‘new injuries’ if an old injury has ‘healed’ prevents the analysis of exacerbations, and provides no advantages.

Chronic conditions and gradual-onset initial injuries

Chronic conditions (expected to cause intermittent or constant symptoms for years)

We now apply the same principles to an athlete with osteoarthritis of the shoulder that causes symptoms of stiffness and swelling. Since chronic conditions do not heal, investigators often prefer to use prevalence measures such as ‘how many participants have osteoarthritis?’. Although this information is valuable, one might also want to know how many times a person with osteoarthritis is unable to participate because of his or her condition, or how often they require a series of treatments to manage flare-ups. Further, when one considers recurrent injuries, the distinction becomes even less clear. Does someone with recurrent ankle sprains have several acute injuries, or are they suffering from one condition called chronic ankle instability? Similarly, are investigators and clinicians only interested in knowing how many participants have asthma, or are they also interested in developing treatment and prevention programmes that decrease asthma attacks? Injury and illness surveillance tools need to be able to answer all of these questions.

When the athlete's ‘state’ is stable over time with only minor variations in symptoms, these minor variations should be documented within the notes of follow-up therapy sessions for the ongoing injury rather than in a new injury record. This is because the fluctuations are expected to occur unrelated to sport activity, and exposure time in sports is as such not relevant. However, when there is a reasonable possibility that the change in state occurred because the athlete was exposed to activity (eg, trauma, sudden increase in activity), the incident should be described as a new injury record.

Although diagnosing a second ankle sprain following an initial ankle sprain seems logical, diagnosing a second ‘knee osteoarthritis’ (or a 2nd Achilles tendinopathy) may appear counterintuitive. That said, a second diagnosis of osteoarthritis (or Achilles tendinopathy) is consistent with billing procedures. When, for instance, injury surveillance data are analysed to address research questions, investigators would simply interpret a second visit with a diagnosis of osteoarthritis that required subsequent treatments as a worsening of the condition. Although it might be possible to include ‘exacerbation of OA’ as a diagnostic choice in surveillance software (even though there is no International Classification of Diseases, 10th Revision (ICD-10) coding), consistency would then suggest one would have to document the second ankle injury as ‘exacerbation of ankle sprain’, even though this would not be correct. Further, an ACL tear might experience recurrent episodes of giving way followed by swelling after returning to activity. In this context, the precise diagnosis for the second injury is not a second ACL tear, but rather a ‘tibiofemoral subluxation’ (with or without additional ligament/capsule damage).

Gradual-onset injuries

The same principles apply when the initial injury is a gradual-onset injury, which generally includes some aspects common to acute injuries (they may eventually heal) and some aspects common to chronic conditions (the duration is long and symptom fluctuation is common). Note that the first injury and subsequent injury can occur due to external stress (eg, wrestling or judo throw) or internal stress (eg, eccentric muscle contraction). We recognise that the M-FASIS approach means there might be several new ‘injuries’ due to these activity-related incidents, and we consider this a strength of the approach as described below.

Remaining challenges

The main challenge with our proposal is that different clinicians will most likely interpret ‘reasonable possibility that the change in state occurred because the athlete was exposed to activity’ differently. First, what is a ‘change in state’? Second, was the state change due to being exposed to activity? As noted above, these challenges are not created by the M-FASIS approach but in fact also exist for traditional approaches. In addition, they are increased with overuse injuries, where fluctuations in symptoms may have a wider range and are more common. We believe these discrepancies should be minimised over time as researchers and clinicians begin to use multistate systems such as M-FASIS and consensus develops about the operational definition for ‘stable state’. Third, treatments associated with damaged tissue could be allocated to the first or second injury, but should not be allocated to both because it would double the amount of work and suggest double the amount of treatments in any analysis. We suggest that all treatments after the second event be allocated to the second event. This is consistent with the current approach of allocating treatments to the subsequent injury if the athlete had already fully returned to play, and with the common practice of documenting treatments with chronic conditions. When the data are analysed, the investigators will be able to easily include these treatments with the first injury if the study question justifies this approach. Finally, we also want to emphasise that our proposed methods only address the challenge of when to document a new injury record instead of a note in an old injury record. Having changes in state documented as new injury records enables accurate calculation of incidence for different types of subsequent injuries and exacerbations. The analysis of suffering and other consequences of injury requires detailed data collection on numerous variables, which would all be part of the new injury or old injury record.


In summary, the energy and time put into developing and implementing injury and illness surveillance software based on a multistate framework, and the clinicians’ time required for data gathering and entry, will allow researchers to properly track injury and illness trends, determine causes, and develop and implement effective prevention and rehabilitation programmes. We recommend that a new injury record be created whenever symptoms or signs increase due to activity even if the initial injury had not healed or there is a chronic condition.

What are the findings?

  • We provide an internally consistent framework to document new injuries (illnesses), exacerbations, subsequent injuries and chronic conditions.

  • An event should be documented as a new injury record whenever the athlete moves towards a less-healthy state as a result of activity, where a less-healthy state is defined as an increase in symptoms that is beyond the expected normal day-to-day fluctuation (eg, ‘exacerbations’). This means a new injury record may be created even though the first injury has not ‘healed’.

  • Similarly, an increase in symptoms due to activities in long-standing injuries or conditions should also be documented as a new injury (illness) record.

How might this affect clinical practice?

  • Improved documentation in injury and illness surveillance programs should enable clinicians to better:

  • Understand risk for exacerbations and reinjuries.

  • Evaluate effectiveness of rehabilitation programs.

  • Improve safe return to play decision making.



  • Twitter Follow Benjamin Clarsen at @benclarsen and Evert Verhagen at @Evertverhagen

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.