Article Text

Download PDFPDF

Do not neglect injury severity and burden when assessing the effect of sports injury prevention interventions: time to paint the whole picture
  1. Evert Verhagen1,
  2. Ben Clarsen2,3,
  3. Larisa van der Graaff1,
  4. Roald Bahr2,4
  1. 1Amsterdam Collaboration for Health and Safety in Sports, Department of Public and Occupational Health, Amsterdam Movement Sciences, Amsterdam UMC Location VUmc, Amsterdam, The Netherlands
  2. 2Oslo Sports Trauma Research Center, Department of Sports Medicine, Norwegian School of Sports Sciences, Oslo, Norway
  3. 3Federation Internationale de Football Association, Zurich, Switzerland
  4. 4Aspetar Orthopaedic and Sports Medicine Hospital, Doha, Qatar
  1. Correspondence to Professor Evert Verhagen, Department of Public and Occupational Health, Amsterdam UMC Location VUmc, Amsterdam, Noord-Holland, Netherlands; e.verhagen{at}

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 outcomes typically reported from injury prevention trials are incidence rates. In this editorial, we show that this approach only paints part of a bigger picture and could even lead to misinterpretations. We argue that injury severity and injury burden should be presented in conjunction with incidence to provide a more complete assessment of the effectiveness of sports injury prevention interventions.

Principles of risk management

Sports participation carries an inherent risk of injury, and anticipating every athlete may sustain several injuries during their career is reasonable. No preventive measure or programme has ever been, or will ever be, 100% effective. As a result, rather than eradicating injuries, the focus should be on reducing risks to an acceptable level. To prioritise our efforts, we must first identify which risks are unacceptable.1 2

Risk management principles have previously been discussed in the context of sports.1 3 A risk management approach adopts a broad set of mitigation measures that can reduce the number and severity of injuries.1 The International Olympic Committee, among others, has encouraged the use of risk management principles in the consensus on the recording and reporting of injury and illness in sports.4 These recommendations emphasise that choices for preventive foci should not be based solely on the incidence of injury but should also consider their severity in terms of days lost from sports.

This is not to say that the severity of an injury can be defined by time loss alone. Where injury incidence is universally regarded as the number of events observed for some time spent in the population, no equivalent characterisation is available for severity. Severity is a latent measure involving many factors, such as the nature and diagnosis of an injury and limitations to sport-specific functional capacity. Among these, time loss resulting from an injury is an easy-to-observe parameter of interest.

Risk and severity

Figure 1A presents a risk matrix as initially proposed by Fuller and Ward.2 For illustration purposes, this plots original data on ankle and shoulder injuries in a cohort of recreational volleyball players.5 Risk matrices such as these help understand the impact of injuries by illustrating their frequency and severity, as shown in figure 1B.1 An added benefit of this quantitative presentation is that both axes combine to form a composite measure—‘injury burden’—which describes the impact injuries have on a specific population by providing the cumulative number of time-loss days per 1000 hours of participation.1 6 Figure 1B highlights that ankle injuries represent a much greater injury burden than other injury types. From this knowledge, the authors decided to focus on ankle injuries by developing a targeted injury prevention programme.7

Figure 1

Risk matrices for injuries in recreational volleyball players, recalculated from the original data of Verhagen et al.7 (A, B) The injury frequency is considered on the x-axis, and on the y-axis, the injury’s severity. On the left (A) presents a simplified risk matrix based on a qualitative categorisation,2 whereas (B) presents the risk matrix based on quantitative data.1 In (B) the product of both axes is the injury burden, that is, the number of days lost per 1000 hours of exposure. The dotted lines present the ‘Risk Contour’ along which the injury incidence and severity combinations have an injury burden equal to the corresponding number, for example, RC20 is the line along which the injury burden is 20 days of time-loss per 1000 hours of exposure.

Severity and burden are neglected

When making decisions on prevention targets based on frequency and severity, we argue that both outcomes, incidence and severity, should be addressed to evaluate the effectiveness of interventions. However, injury prevention studies have traditionally focused on analysing the frequency of injuries as the primary outcome, neglecting injury severity.8 Focusing solely on incidence can result in underestimating (or overestimating) the true impact of interventions and may lead to flawed conclusions regarding their effectiveness. For argument’s sake, let us consider three measures we can calculate from an athlete cohort: injury incidence, severity and burden. Using fictitious data in table 1, we illustrate what would happen to these three measures based on different interventions and their potential effects.

Table 1

Hypothetical outcomes of three interventions on injury incidence, severity and burden

Intervention A

An exercise programme can reduce the likelihood of sustaining injuries without affecting injury severity. One example could be a study targeting anterior cruciate ligament (ACL) injuries, where there is a reduction in injury incidence (the number of injuries), but where severity (time to return to sport) is rather unchanged and dependent on the quality of medical care (eg, waiting for surgery, choice of graft, rehabilitation programme). However, with fewer injuries, the injury burden will also decrease.

Intervention B

Protective equipment, such as knee guards for volleyball players, offers a preventive strategy that mainly protects against minor injuries, such as contusions and abrasions. Assessing the effectiveness of these measures will likely show a reduction in overall injury incidence. However, the average injury severity will increase because only minor injuries are prevented, leaving the burden relatively unchanged.

Intervention C

Improved immediate injury management and rehabilitation will not prevent injuries but can accelerate return to sport. Although there is no direct impact on the incidence of injuries, a decrease in the number of days lost influences the average severity of injuries and, subsequently, will result in a lower overall injury burden.

A second look at existing data

To examine the benefits of adding burden as a main outcome (and potential pitfalls of the classic approach focusing on incidence alone), we have reanalysed the raw data from published intervention trials for which we calculated injury severity and burden (table 2).7 9–11 Figure 2 provides a visual representation of these outcomes as risk matrices. For illustrative purposes, we have only conducted basic analyses without any correction for confounding variables or other potential biases. Nonetheless, what becomes clear is that for an appropriate interpretation of the effect of the prevention programmes, we need to examine not only injury incidence but also severity and burden. Our examples highlight further that different combinations of incidence and severity define burden estimates and show that we should aim for a trifecta of outcomes when assessing interventions.

Figure 2

Visual representation into risk matrices of our secondary analysis of existing data, as presented in table 2. For illustrative purposes, we have conducted only basic analyses without any correction for confounding variables or other potential biases.

Table 2

Secondary analysis of existing data from published intervention trials for which we calculated injury severity and burden, with the respective 95% confidence intervals

The study by Verhagen et al7 positively affected ankle injury incidence and reduced the average severity and burden. In contrast, Nauta et al9 showed fewer injuries under the influence of their intervention. However, average injury severity did not show a significant change. This could be because the intervention was more effective in preventing less severe injuries. The preventive effect was, however, sufficient to reduce injury burden. Barboza et al10 found reduced injury rates for acute injuries, but the average severity was twofold, and consequently, the injury burden remained unchanged. In contrast, overuse injuries were not reduced in number nor their severity, but the burden of injuries was reduced. If these additional outcomes had not been considered, the impact of this intervention on overuse injuries might have been overlooked. Finally, consider the study by Ekstrand et al,11 who compared professional football teams implementing the Nordic Hamstring Exercise programme as part of team training to other Champions League teams, documenting a reduction in injury rates, severity and burden. In other words, focusing on injury incidence alone does not paint the whole picture; the intervention was even more effective when injury severity and burden were considered as well.

Severity-effectiveness planes

Risk matrices are not designed to plot trial outcomes and, as seen in figure 2, may quickly become messy when presenting multiple outcomes. We propose using ‘planes’ to plot incidence and severity outcomes. Such figures are also used in economic evaluations, in which two or more alternative interventions are compared regarding their effect and associated costs.12 To do so, we must compare our intervention to a control condition, usually standard practice.Figure 3A illustrates the outcomes of such a comparison, which poses an easier-to-interpret depiction of an intervention’s impact than a risk matrix. In the centre of this graph, we have a control (comparator) condition; the x-axis represents the relative injury rate, and the y-axis represents the relative severity. Each quadrant presents a potential outcome based on a combination of injury rate and severity. An outcome in the lower left quadrant would implicate an intervention that reduces the injury rate and the severity. An outcome in the upper left quadrant signifies an intervention that reduces injury rate, but injuries are more severe on average. However, remember that this could indicate that the intervention was effective for less severe injuries, as shown in our intervention B example above. This potential outcome highlights the necessity of considering incidence, severity and burden when interpreting prevention outcomes. Interventions in the other quadrants should not be considered as these would increase injury rates.

Figure 3

Severity-effectiveness planes illustrate the relative injury rate and severity of an intervention group compared with a control group. (A) An illustrative explanation of interpreting this plane. The other panels provide exemplary severity-effectiveness planes based on the secondary analyses of studies by Nauta et al9 (B) and Barboza et al10 (C).

Now what?!

The analyses we propose here need to be developed in sufficient detail, and here lies a challenge for future research. Statistical methods should be refined to account for the nature of injury data, especially their proneness to missing data, skewed nature and the methods needed to assess their uncertainty.13 On top of that, we have discussed injury severity and burden as measures of the impact injuries may have on athlete health and their availability for training and competition. Impact, however, goes beyond athlete availability. Hence, we should emphasise that although most often days of time loss are used as a severity measure, other measures like injury-related costs, performance losses or the Oslo Sports Trauma Research Centre Questionnaire′s severity score can describe the consequences of injuries (and other health problems). Such additional outcomes should be considered to fully assess the broader benefits of interventions. Definitions and measurements are available for some of these while others need to be developed.

Time to paint the whole picture

Considering severity and burden, rather than solely focusing on injury incidence, allows for a more comprehensive evaluation of the effectiveness of interventions to prevent sports injuries. Although these measures represent latent indicators rather than direct measures of interest, this information is essential for a comprehensive interpretation of the impact of an intervention. It supports a more holistic approach to managing and preventing injuries in sports. Considering this, we advocate that severity and burden will be established as standard outcome measures in prevention trials assessing intervention effectiveness in addition to injury incidence.

Ethics statements

Patient consent for publication



  • X @Evertverhagen

  • Contributors All authors contributed to developing the ideas outlined in this editorial. EV prepared the text and is the guarantor of this manuscript. All other authors provided critical feedback during the process.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

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