Introduction

Injuries entail major challenges for sports teams and athletes, as they have a negative impact on athletes’ health, sports performance, and success [5, 6, 12, 26]. Structured long-term injury surveillance is fundamental to better understand the injury burden and pattern, to generate ideas for injury prevention, and to evaluate the effects of interventions [24]. Long-term surveillance programmes have been established among elite athletes in sports such as football, rugby, and athletics and have provided increased knowledge of the injury burden in these sports [7, 8, 11, 14, 27]. For instance, based on FIFA injury surveillance there is evidence of reduced incidence of head injuries in professional football after introduction of a rule change penalizing direct and deliberate elbows to head with a red card [2].

For the 2008 Olympic Games, the International Olympic Committee implemented a health surveillance programme which has been conducted ever since [9, 10, 15, 19,20,21,22]. During the summer Olympic Games of 2016 and 2020 and the winter Olympic Games of 2014 and 2018, 8–12% of athletes suffered an injury [19,20,21,22]. This surveillance programme, however, only captures injuries occurring during the Olympic Games, while the injury rate and pattern outside this period is largely unknown. Injuries occurring outside the Olympic Games period could have an impact on athletes’ preparations for the Olympic Games, and more knowledge about injuries may help support athletes in their preparations, potentially reducing the injury burden and allowing better performance. Additionally, surveillance programmes may act as a first step to inform injury preventive measures. Many National Olympic Committees have support programmes for their Olympic candidates. In Norway, for example, a health monitoring programme was implemented in 2011 for the national Olympic and Paralympic candidates, which has provided further insights about injuries outside the Olympic Games period [4].

Insurance registries have been used to collect injury surveillance data [13, 16, 29, 30]. Most athletes in Sweden are insured by the same insurance company, where acute sport injuries have been registered consistently for over 20 years [30]. Previously this registry has been used to study sports injuries at a national level [29], but it has not been used to study injuries specifically occurring among elite Olympic athletes. The aims of this study were to: (1) describe the injury incidence and analyse time trends in injury incidence; and (2) describe injury types and locations, in different age groups, sexes, and sports among Swedish Olympic athletes based on insurance data.

Materials and methods

Study design

This was a cohort study based on insurance registry data from the Swedish insurance company Folksam Insurance Group (Folksam). All data were handled confidentially in line with the Declaration of Helsinki. The study was approved by the Swedish Ethical Review Authority (reg. no 2022-00438-01). The study was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology and the Sport Injury and Illness Surveillance extension [1, 25].

Population

The study population includes Swedish elite male and female athletes, covering a wide range of sports, within the Swedish Olympic Committee (SOC) support programme Top and Talent (Table 1). Inclusion criteria were all athletes that had participated in Top and Talent between January 1999 and December 2020. No exclusion criteria were applied. Top and Talent was initiated in 1998 and is an individually customised support programme for athletes with potential medal capacity in the Olympic Games, with a plan for the athlete to reach top international positions within 3–6 years of entering the programme. Athletes are accepted to the programme based on either top performance in international competitions or a talent selection [23]. Approximately 30% of the athletes in Top and Talent reach medal capacity in the Olympic Games or world championship after 4–6 years of entering the programme. Athletes in Top and Talent are provided with, among other things, medical, coaching, training, and nutritional support, and all athletes have health insurance from Folksam [23]. A total of 762 athletes (412 males, 350 females) in 38 sports participated in Top and Talent between 1999 and 2020 (Table 2). Mean ± SD age of the included athletes was 26.5 ± 5.9 years. On average, 172 ± 23 athletes (range 121–217) participated in the programme each year during the observation period. Duration of participation was median 4 years (IQR 2–6), with total of 3427 athlete-years included.

Table 1 Data collection and definitions
Table 2 Number of athletes in top and talent by sport category, sport, and sex

Data collection

Injury insurance claims were reported to the registry prospectively between 1999 and 2020 by a physician or physiotherapist from SOC via the athletes’ health insurance, after a median of 5 days (IQR 0–22) from date of injury. Only injuries occurring while the athletes participated in Top and Talent were registered. An administrator from Folksam was responsible for recording the injury in the registry. Injuries were diagnosed according to an insurance-specific coding system. The registry contained data about the insured athletes and about each injury. Details on the collected data are presented in Table 1. Injury data were extracted from the insurance registry in September 2022. Initial data processing and database construction was done by Folksam, and the data received for research and analysis were pseudonymised. The database was checked for errors by the first author, and any mismatching information (e.g. mismatched injury location and injury type) were checked against the original insurance claim record by Folksam, and corrected where possible.

Definition of injury

The health insurance covers both injuries and illnesses, but only injuries were included in this study. An eligible injury, as per the insurance policy, was defined as any physical damage with a sudden or gradual onset occurring while the athlete participated in Top and Talent. Pre-existing injuries present prior to start of the insurance were not included. Multiple injuries sustained by an athlete in a single event were recorded as one injury with multiple injury types or locations. Injury type was classified into eight different categories. Due to irregular reporting in the insurance registry, injuries reported as joint/ligament sprains, dislocations, and muscle/tendon strains and ruptures were combined by the authors to the same injury type category for descriptive statistics, named ‘soft tissue/joint injury’. The other injury type categories were contusion, fracture/bone injury, inflammation, concussion, laceration/abrasion, multiple injury types, and ‘unspecified’. Injuries were classified to a body region and location based on the reports made to the insurance registry.

Statistical analysis

Data were summarised using descriptive statistics, with mean, median, frequencies, standard deviations (SD), interquartile ranges (IQR) and/or 95% confidence intervals (CI), as appropriate. Incidence and characteristics (i.e. type and location) of injuries were presented for the total population and stratified for subgroups based on sex (male, female), competitive season (summer and winter sports), sport category (skill, power, mixed, endurance) [17], and age group (≤ 20, 21–25, 26–30, 31–35, 36–40, ≥ 41 years). Injury incidence and characteristics were presented only for individual sports with more than five athletes due to confidentiality reasons. Injury incidence was calculated as an incidence rate: number of injuries divided by number of athlete-years multiplied by 100 and expressed as number of injuries per 100 athlete-years. ‘Athlete-years’ was defined as the number of years an athlete participated in Top and Talent. If an athlete only participated half a year (spring or autumn) they were given a value of 0.5 years. All injuries (new and recurrent) and athlete-years were aggregated per athlete within each subgroup. This means that for age group specific injury numbers the same athlete can be included in multiple age groups depending on the athletes’ age at the exposure (athlete-year) and time of injury. Time trend analyses were performed for the 2001 to 2020 period with Poisson regression using generalized linear models with number of injuries as the dependent variable, year as covariate, and with natural log of the exposure variable (athlete-years) as offset variable. The model-based estimator was used for the covariance matrix and Wald statistics was used to calculate the p-value and 95% CI. The first 2 years (1999–2000) were excluded in the time trend analysis due to suspected low awareness about the health insurance when it first was released. Time trend analyses were performed for the full 20-years period for the total population and split by sex, as decided a priori. In addition, statistical time trend analyses were conducted post hoc for the first and second 10-years periods separately since ocular analysis indicated varying trends in the two decades. Time trends in injury incidence are expressed as the annual percent change between years with a 95% CI. The significance level was set at p < 0.05. All statistical analyses were performed using IBM SPSS Statistics (version 29.0, Armonk, New York, USA).

Results

Injury incidence

In total, 1635 injuries sustained by 468 athletes (61.4% of all athletes) were registered (median 2 injuries per athlete (IQR 1–5; range 1–21)). The overall incidence was 47.7 injuries/100 athlete-years, with 46.4 and 49.3 injuries/100 athlete-years among male and female athletes, respectively. The highest incidence per individual sport was seen in gymnastics, tennis, and athletics (100.0, 99.3, and 93.4 injuries/100 athlete-years, respectively). Injury incidence in the total population and for subgroups is presented in Tables 3 and 4, and Online Resource 1, and incidence by injury type and location is presented in Table 5.

Table 3 Injury incidence and number of injuries in the total population and subgroups
Table 4 Injuries, injury incidence and most injured locations by sport
Table 5 Injury incidence and number of injuries by injury type and location

Time trends in injury incidence

The time trend analysis showed that the overall injury incidence increased by 3.0% annually over the full 20-years period between 2001 and 2020 (95% CI 2.1–3.9%, p < 0.001). Males had an 2.5% annual increased incidence (95% CI 1.4–3.7%, p < 0.001) and females an 3.7% increase (95% CI 2.3–5.1%, p < 0.001). Noteworthy was that the injury incidence increase was evident only in the first 10-years period of 2001–2010 (annual change 6.0%, 95% CI 3.3–8.8%, p < 0.001), while the second 10-year period of 2011–2020 showed no such trend (annual change 0.4%, 95% CI − 1.9 to 2.7%, (n.s.)) Year-by-year injury incidences for subgroups (sex, competitive season, sport category, and age group) are presented in Fig. 1.

Fig. 1
figure 1

Time trend in injury incidence by sex (a), competitive season (b), sport category (c) and age group (d). Symbol next to the year illustrates whether the summer (sun symbol) or winter (snowflake symbol) Olympic Games was held in the current year. Due to small sample size in the age groups 31–35, 36–40 and ≥ 41 years, these age groups were combined to one group ≥ 31 years for visual presentation in graph. Data on age is missing for 14 athletes

Injury characteristics

The lower limb region, and specifically the knee, had the highest injury frequency, and soft tissue/joint injury was the most common injury type category among the specified injury types. Injury characteristics are presented in Tables 4 and 5, and Online Resource 2 and 3.

Discussion

The principal findings of the study were that the overall injury incidence among Swedish Olympic athletes increased by six percent annually between 2001 and 2010, after which the curve flattened with no change between 2011 and 2020. Athletes report approximately one injury to the insurance registry every second year, where most injuries affect the knee, foot/ankle, and spine/pelvis. There is a paucity of research on the year-round injury incidence and characteristics among Olympic athletes and the present study is, to the best of our knowledge, the first to describe injuries over a two-decade long period using insurance data.

Injury incidence and characteristics in the total population

The injury incidence was 47.7 injuries/100 athlete-years. Compared to previous studies in Olympic athletes, this incidence rate is low. During the Olympic Games of 2014 to 2020, the injury incidence was between 5.4 and 7.8 injuries/1000 athlete-days (equal to 197.1–284.7 injuries/100 athlete-years) [19,20,21,22], and among UK Olympic summer sport athletes, an injury incidence of 1.32 injuries/athlete-year (equal to 132.0 injuries/100 athlete-years) has been reported [18]. This discrepancy may be explained by differences in data collection methods (e.g. prospective study with athlete self-report versus medical record or insurance-based registry) and injury definitions (e.g. all injuries irrespective of time-loss or medical attention versus medical attention or insurance claims), and the fact that the insurance registry used in the present study likely captures primarily more severe injuries (i.e. ‘tip of the iceberg’). The true injury incidence in this study population is probably higher, and a direct comparison to previous studies is difficult. Overall, the lower limb was the most injured body region, and the knee, followed by foot/ankle, and spine/pelvis the most injured locations, which is in line with previous studies in Olympic athletes [18,19,20].

Injury incidence and characteristics in subgroups

Injury incidence was highest among power and mixed sports, possibly due to the nature and demands of these sports, which often involve high forces and athlete-to-athlete contacts. It should be noted, however, that there was a variation in injury incidence among single sports within the four sport categories. For instance, in mixed sports, the injury incidence ranged between 16.3 and 99.3 injuries/100 athlete-years, and among power sports between 45.0 and 100.0 injuries/100 athlete-years. All sports that had a higher injury incidence than the total population average belonged to the mixed or power sport categories. Previously, Ranson et al. reported that the injury incidence was highest in combat sports [18], and combat sports were included in the mixed or power sport category in the present study. Injuries to the knee and foot/ankle were prevalent among athletes in the power and mixed sport categories, indicating that these locations should be specifically targeted with injury prevention measures. In many skill and endurance sports, in addition to the knee, injuries were also frequent in the spine/pelvis and shoulder, indicating a need for other preventive measures in these types of sports.

In the present study the overall injury incidence was comparable between males and females, which is in line with previous reports during the Olympic Games [19,20,21,22]. In team sports, however, the overall injury incidence is higher among males [28], and this difference may be due to the fact that the present study mainly included individual sports. While certain knee injuries, such as anterior cruciate ligament injuries, are more common in female than in male team sport athletes [28], overall knee injuries were equally common in males and females in the present study (detailed diagnoses were not available for all knee injuries).

Comparing age groups, the highest overall injury incidence was seen among athletes ≤ 20 years, with a trend that the injury incidence decreases with higher age. This age–injury relationship was evident in the mixed and skill sport categories, but not in power and endurance sports. It should be noted, however, that comparison within single sports were not possible due to small samples, and large variations may exist. Further, there were differences in the number of athlete-years in different sports within the sport categories and age groups (Online Resource 4), and the reason for the observed age-injury relationship is unclear.

The observed variation in injury incidence and characteristics between sports and subgroups emphasises the importance of considering the contextual factors—such as environmental factors and sport and athlete characteristics—when implementing injury prevention interventions [3].

Time trends

A significant time trend was noted with a three percent annual increase in injury incidence over the full 20-years period, which was evident in both males and females. However, when analysing time trends for the first and second 10-years periods separately, the increasing trend was evident only during the first decade of 2001 to 2010, after which the injury incidence remained fairly stable between 2011 and 2020. This indicates that there is currently no trend of increasing injury incidence in the Swedish Top and Talent athlete population. Looking at the time trends year-by-year, some variations could be seen within subgroups. For example, in winter sports, there was a tendency towards a high injury incidence the year prior to each Winter Olympic Games, and low incidence the year of the Olympic Games. While not as obvious, this tendency was also seen in summer sports in relation to the Summer Olympic Games. Based on the data it is not possible to explain the observed increasing time trend or year-by-year fluctuations in injury incidence with any certainty. It may be the result of sporting factors; e.g. changes in sporting/physical demands, competition and training schedules, or rules or equipment, but also due to methodological factors such as varying population characteristics or reporting patterns.

Strengths and limitations

Strengths of the study include the long time period of data collection, and a large sample of athletes who had participated in Top and Talent between 1999 and 2020. In contrast to most previous research of injuries in Olympic athletes [19,20,21,22], but similar to Ranson et al. [18] and Clarsen et al. [4], the present study included also injuries sustained outside the Olympic game period, taking into consideration seasonal differences in injury rates and better reflecting the full athlete year cycle. Furthermore, the same inclusion criteria to Top and Talent were used over the 22 years, and the injury definition remained consistent over the years, allowing for comparable data over time. Another strength is the inclusion of sport-specific injury data, also providing data on injuries in sports with very little, or non-existent, previous research. To allow for meaningful comparison between groups with sufficient samples, sports and athletes were classified into subgroups. A limitation of this approach is that sports and athletes within the same subgroup may differ, and certain sports or athletes may be responsible for the majority of injury cases within the group, thus skewing the data. For example, within the mixed sport category in the present study, athletics was responsible for approximately 59% of injuries. Also, some sports, such as athletics, consist of different disciplines with varying injury burdens and patterns.

Another limitation is that many injuries were registered in the insurance registry as unspecified for injury type. Information about these injuries is therefore limited, and it is possible that the proportion of injuries in specific injury types would be different if all injuries had more specified information. More specific injury registration is important to increase the quality of the registry and to allow more definitive conclusions about injury rates and patterns.

Registry data from 1999, shortly after the inception of Top and Talent in 1998, was included. Due to low awareness of the health insurance in the initial years and likely some missing injury data, this results in underestimation of the injury incidence in these years, and to a smaller extent also the overall injury incidence. It is likely that insurance registry data includes primarily more severe injuries. Minor injuries—for example certain gradual onset injuries, or injuries not resulting in time-loss or affecting athlete performance—may go unreported to the insurance company. Given the insurance registry-based data collection, the generalisability of the findings to other populations and surveillance methods is unclear. Gradual onset injuries and non-time loss injuries dominate the injury pattern in Olympic sports [4, 19, 20], and gathering information about these injuries in a structured and systematic manner is important to get a complete understanding of injuries in Olympic athletes.

Conclusions

The overall injury incidence increased by six percent annually between 2001 and 2010, after which the curve flattened with no change between 2011 and 2020. Swedish Olympic athletes report approximately one injury to the insurance registry every second year, where most injuries affect the knee, foot/ankle, and spine/pelvis. The highest injury incidence was seen in gymnastics, tennis, and athletics, and for specific subgroups, in mixed and power sports and younger age groups. Injury prevention interventions should be a high priority in these sports and subgroups. The findings may support health and performance teams working with Olympic athletes in their work to improve athlete health, by highlighting incidence rates and common injury locations in different subgroups and sports.