Article Text

Download PDFPDF

Are prevalence measures better than incidence measures in sports injury research?
  1. Rasmus Oestergaard Nielsen1,
  2. Katrin Debes-Kristensen1,
  3. Adam Hulme2,3,
  4. Michael Lejbach Bertelsen1,
  5. Merete Møller1,
  6. Erik Thorlund Parner4,
  7. Mohammad Ali Mansournia5,6
  1. 1 Section for Sports Science, Department of Public Health, Aarhus University, Aarhus, Denmark
  2. 2 Australian Collaboration for Research into Injury in Sports and its Prevention (ACRISP), Federation University Australia, Victoria
  3. 3 Centre for Human Factors and Sociotechnical Systems, University of the Sunshine Coast, Queensland, Australia
  4. 4 Section for Biostatistics, Department of Public Health, Aarhus University, Denmark
  5. 5 Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran
  6. 6 Sports Medicine Research Center, Neuroscience Institute, Tehran University of Medical Sciences, Tehran, Iran
  1. Correspondence to Professor Mohammad Ali Mansournia, Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran; mansournia_ma{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.


Prevalence and incidence are terms commonly used to describe the number, proportion and rate of sports injury in the epidemiological and clinical literature. Indeed, scientific articles have discussed the practical implications of choosing one of these measures over another in relation to better understanding the burden of injury as well as sports injury aetiology, prevention and treatment.1–3 Recently, the use of prevalence and not incidence has been promoted to demonstrate a relationship between training load and overuse sports injury.2 However, there is a need to further elaborate on the circumstances in which prevalence and incidence are used as both measures are relevant and depend on the goals of the researcher and study. Therefore, the purpose of this editorial is to explain the differences between prevalence-based and incidence-based measures, and promote when they should be used by drawing on available data from an existing 1-year, 933-person running injury cohort study (DANORUN4).

Incidence-based measures

Why do sports injuries develop? What we can do to prevent them? Which treatment options should be offered to injured athletes? These are arguably some of the most important questions facing coaches and clinicians. Fortunately, incidence-based measures can help to provide answers given that they relate to the occurrence of new sports injury in an athletic population within a specified period or how many injured athletes recover over time.

Incident cases

The number of incident cases can facilitate the identification of when injuries occur. This allows coaches and clinicians to carefully consider whether there have been any noticeable and possibly deleterious changes to training loads prior to injury occurrence. Figure 1A demonstrates that the number of new injuries was high in weeks 2–3 with >20 new injuries occurring, whereas no injuries occurred from week 17 to 18. The number of incident cases can be cumulated to show the number of persons having had at least one injury during follow-up as seen in figure 1B.

Figure 1

Six graphs (A–F) based on data from a 1-year, 933-person prospective cohort study (DANORUN) on running-related injuries. (A) Graph revealing the number of runners sustaining an injury on a weekly basis during follow-up (incident cases). (B) Cumulative incident cases (CI) showing the number of runners having sustained an injury since baseline. (C) Cumulative incidence proportion (CIP, %) without censoring shows the proportion of runners having had at least one injury (CIP without censoring = (CI/933) * 100). Here, we wrongly assume that all participants complete the 52-week follow-up. (D) Kaplan-Meier injury survival graph showing the proportion on injury-free runners at certain time points. This calculation considers the effect of censoring. (E) Point prevalence graph showing the number of injured athletes at certain time points. (F) Point prevalence proportion (PP) graph showing the proportion of injured runners at certain time points (PP = (injured/runners at risk at that time) × 100).

Incidence proportion

The incidence proportion can be used to identify the proportion of new sports injuries during follow-up. For example, in the DANORUN study,4 the incidence proportion of athletes with new injuries was (123/933) × 100=13.2% over 11 weeks—that is, if we (wrongly) assume all 933 runners who attended baseline remain in the study during follow-up (figure 1C). Since study participants are censored over time, a more appropriate graph to visualise the proportion of injury survivors is the Kaplan-Meier graph as depicted in figure 1D.

Incidence rate

Incidence rate is another measure that estimates the rapidity with which new injuries develop by observing an athletic population over a specified period (eg, one season). The number of new injuries is counted and divided by, for instance, person-time or athletic exposure hours. In DANORUN,4 a total of 254 injuries occurred during the 36 015 weeks at risk. Based on this, the yearly injury incidence rate is 254 injuries/36 015 weeks=0.007 injury per week or equivalently 7 injuries per 1000 athlete-weeks.

In the box, the three incidence-based measures presented above have been defined.



Prevalence-based measures

Prevalent cases: The number of athletes with injuries at a certain time point, for example, at a specific time during a season, or during a certain time period.

Prevalence proportion: The proportion (in percentage) of athletes with injuries at a certain time point (point prevalence proportion) or time period (eg, lifetime injury prevalence proportion). Usually, prevalence and prevalence proportion are used interchangeably. Prevalence proportion=injured athletes/all athletes at a given time point.

Incidence-based measures

Incident cases: The number of athletes with new injuries over a certain time period, for example, during a season.

Incidence proportion: The proportion of athletes with new injuries during a follow-up period expressed as a percentage. Cumulative incidence proportion without censoring=athletes with new injuries/all athletes at risk at the start of follow-up over a certain time period. If athletes leave the study during follow-up, we recommend considering censoring by calculating a cumulative incidence proportion at certain time points during follow-up.

Incidence rate: Incidence rate measures the rapidity with which new injuries develop. The rate of new injuries during a follow-up period per person-time at risk over that period. Usually, incidence and incidence rate are used interchangeably. Incidence rate=number of new injuries/total of athletic exposure hours per, for example, 1000 athlete-hours.

Prevalence-based measures

Prevalence-based measures have recently been promoted as preferable measures in injury surveillance studies.1–3 While prevalence has its advantages, it is important to acknowledge that by using these measures researchers are unable to (i) address why injury develops, (ii) assist with the identification of appropriate preventive strategies, or (iii) suggest possible treatment options. Instead, these measures can be used to identify athlete availability and the need for medical attention and/or treatment.

Prevalent cases

In the running injury study,4 the number of injured runners was 15 by the end of the first week and 90 by the end of week 11. It is likely that the need for medical attention and/or treatment was greater in the latter time point given that the number of prevalent cases was considerably higher. Since prevalent cases is defined as the number of athletes with injuries at a certain point in time, these numbers recognise certain time points during follow-up with few or many injured athletes. Figure 1E shows the number of injured athletes at certain time points.

Prevalence proportion

Coaches and team staff members are often interested in player or athlete availability. Recently, an averaged prevalence proportion over a time period has been used in the sports injury literature.5 6 Although interesting, this number does not reveal weeks with high and low athlete availability. Accordingly, a graph like figure 1F can be helpful because it shows the prevalence proportion of team players or athletes with injuries at certain time points. To calculate the proportion of injured runners in week 11, the 90 runners who were injured at that time point and the 755 runners remaining in the study (some left the study during the first 10 weeks) at week 11 are needed. This leads to a prevalence proportion of (90/755) × 100=11.9% in week 11.


When reading scientific articles, coaches and clinicians should actively distinguish between prevalence and incidence-based measures of sports injury. Therefore, whether to report incidence or prevalence is a matter of circumstance and is dependent on the goals of the study. We recommended using incidence-based measures in studies examining injury aetiology, prevention and treatment,7 whereas prevalence-based measures should be used descriptively to reveal the number or proportion of athletes with injuries at a certain time point or period. The latter further helps to identify the treatment needs and athlete availability.


None declared.



  • Handling editor Karim M Khan

  • Contributors RON wrote the drafts to the editorial and the remaining authors revised it for important intellectual content.

  • Competing interests None declared.

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