Injury Patterns in Highly Specialized Youth Athletes: A Comparison of 2 Pathways to Specialization

Design and Study Population

This study received approval from a large academic institutional review board. We conducted a retrospective chart review of patients aged 12 to 17 years who completed a sports participation survey (Appendix) after presenting to a large midwestern sports medicine clinic with a sport-related injury between January 2015 and April 2019. Patients were excluded if they did not play any organized sports, did not fill out the survey completely, or did not have a sport-related injury (Figure 1).

View Figure

• Download as Powerpoint
• Download Figure

Measurements

Patient information collected from the electronic medical record included patient age, sex, race or ethnicity, date of injury, diagnosis, injury type (acute, overuse, or serious overuse), and the sports participation survey. Injury type was determined from the description of the injury mechanism in the electronic medical record as either acute (from a single, traumatic event) or overuse (gradual onset). Serious overuse injuries were classified based on the Jayanthi et al¹⁶ criterion of any overuse injury for which the physician recommended treatment of >1 month of rest from sports. This category contained but was not limited to injuries such as stress fractures, physeal stress injuries, elbow ligament overuse injuries, spondylolysis, and osteochondritis dissecans.

The sports participation survey included questions that classified athletes as low, moderate, or highly specialized according to the 3-point system used by Jayanthi et al,¹³ whereby 1 point each was assigned for (1) playing only 1 sport, (2) training for that 1 sport >8 months/year, and (3) quitting all other sports to focus on their main sport. However, the survey had an additional question first used by Miller et al¹⁹ that asked athletes if they had previously played any other sports.

Based on the responses to those items, we categorized participants into 3 groups: (1) exclusive highly specialized athletes, who played only 1 sport and trained >8 months/year but answered no to the question regarding quitting other sports (because they had never played any other sports); (2) evolved highly specialized athletes, who played only 1 sport, trained >8 months/year for 1 sport, and quit other sports to focus on 1 sport; and (3) low-moderately specialized athletes, who either played multiple sports or trained <8 months/year for a single sport. Low and moderately specialized athletes were combined into a single group, consistent with the methods used by Miller et al¹⁹ when developing the exclusive and evolved highly specialized distinctions.

For the highly specialized athletes, sport type was classified as either team or individual according to the same criterion used in previous studies, which defined a team sport as “a sport in which athletes play with others at the time of play.”^23,24 The 2 highly specialized athletes who indicated their sport as other were excluded from this part of the analysis.

Statistical Analysis

We performed the statistical analysis using SAS software (version 9.4; SAS Institute Inc). Bivariate analyses, including χ² tests, t tests, and 1-way analyses of variance, were conducted to compare the demographic characteristics, sport type, injury type, and injury severity among the 3 groups. For the post hoc analysis examining the relationship between sport type and months of rest per year, we calculated 1-way analyses of variance. Two multivariable logistic regression models were fit to predict the likelihood of overuse injury. The first model was predicted by the 3 sport specialization groups (reference: low-moderately specialized), age, and sex. The second logistic regression model was predicted by the 2 highly specialized groups (reference: evolved), age, sex, sport type, starting age of training >8 months per year, and number of months of rest per year. The selection of which predictor variables to include accounted for factors that have been linked to sport specialization and injury risk (age, sex, sport type),^13,17 as well as data that were available clinically from the sports participation survey that may affect injury risk (starting age of training >8 months per year, number of months of rest per year). Odds ratios (ORs) and 95% CIs were computed based on the logistic regression models. To adjust for possible collinearity and validate the second regression model, we ran bivariate analyses (χ² for discrete variables, t test for continuous variables) on each predictor variable individually. A third regression was fit to include only predictor variables determined to have a statistically significant relationship with the overuse injury proportion by the bivariate analyses (sex, age, starting age of training >8 months per year, and sport type). The significance level for all factors was set at P < .05.

Participant Characteristics

Of the 1647 patient charts identified on initial chart review, 1171 (71.1%) patients met the inclusion criteria. Of the 1171 patients, 169 (14.4%) were exclusive highly specialized athletes, 247 (21.1%) were evolved highly specialized athletes, and 755 (64.4%) were low-moderately specialized athletes.

Athlete Demographics

Overall, 700 athletes (59.8%) in our participant population identified as female (Table 1). Girls represented 77.5% of the exclusive highly specialized group, 64.8% of the evolved highly specialized group, and 54.2% of the low-moderately specialized group; the difference in the proportions of female athletes across the 3 groups was significant (P < .0001). The age at presentation to the clinic was similar across specialization groups, with an age of 15.26 years for the exclusive highly specialized group, 15.35 years for the evolved highly specialized group, and 15.23 years for the low-moderately specialized group (P = .57).

Table 1. Sex Differences Among Specialization Groups, No. (%)

View Fullscreen

The number of months per year of rest from organized sports between the 2 highly specialized groups did not differ, with the exclusive highly specialized group resting for an average of 0.71 months per year and the evolved highly specialized group resting for 0.69 months per year (P = .84). The exclusive highly specialized group began intensive training (>8 months per year) for their main sport at an earlier age than the evolved highly specialized group (8.23 versus 9.62 years, respectively; P < .0001; Table 2).

Table 2. Demographic Differences Among Specialization Groups

View Fullscreen

Sport Type

A higher proportion of the exclusive highly specialized athletes (71.4%) than the evolved highly specialized athletes (55.3%) played a sport classified as individual (P = .0009; Table 3). Gymnastics was the most common individual sport in our study (23.80% overall, 29.0% of the exclusive highly specialized group, and 20.2% of the evolved highly specialized group), and soccer was the most frequent team sport (15.4% overall, 16.6% of the exclusive highly specialized group, and 14.6% of the evolved highly specialized group). Post hoc analysis revealed that individual-sport athletes rested for fewer months per year than team-sport athletes (0.50 versus 1.19, respectively; P < .0001).

Table 3. Sport Type Differences Between High Specialization Groups, No. (%)

View Fullscreen

Injury Type Distribution

Overuse injuries accounted for 100 of the 169 injuries (59.17%) in the exclusive highly specialized group, 133 of the 247 injuries (53.85%) in the evolved specialized group, and 342 of the 755 injuries (45.30%) in the low-moderately specialized group (Figure 2). In a direct comparison, the proportion of injuries that were overuse did not differ between the exclusive and evolved specialized groups (P = .29). When directly compared with the low-moderate group, the proportions of injuries that were overuse were greater in both the exclusive highly specialized group (P = .001) and the evolved highly specialized group (P = .02; Table 4).

View Figure

• Download as Powerpoint
• Download Figure

Table 4. Injury Type Distribution Comparison Between Specialization Groups, No. (%)^a

View Fullscreen

Serious overuse injuries constituted 10 of the 100 overuse injuries (10.00%) in the exclusive highly specialized group, 14 of the 133 overuse injuries (10.53%) in the evolved highly specialized group, and 32 of the 342 overuse injuries (9.36%) in the low-moderately specialized group (Figure 3). Direct comparisons between the exclusive highly specialized and evolved highly specialized groups (P = .90), exclusive highly specialized and low-moderately specialized groups (P = .85), and evolved highly specialized and low-moderately specialized groups (P = .70) revealed no differences between the groups with respect to the proportion of overuse injuries that were serious (Table 5).

View Figure

• Download as Powerpoint
• Download Figure

Table 5. Comparison of Overuse Injury Severity Between Specialization Groups, No. (%)^a

View Fullscreen

Multivariable Regression

Multivariable regression analysis of all 3 groups (Table 6) demonstrated that, compared with the reference group of low-moderately specialized athletes, exclusive highly specialized athletes (OR = 1.56; 95% CI = 1.10, 2.21) and evolved highly specialized athletes (OR = 1.36; 95% CI = 1.01, 1.82) were more likely to have overuse injuries than acute injuries. Compared with male sex, female sex was associated with overuse injuries relative to acute injuries (OR = 1.75; 95% CI = 1.38, 2.23). The same was true for higher age (OR = 0.89; 95% CI = 0.82, 0.95).

Table 6. Multivariable Analysis of Overuse Injury Proportions in All 3 Specialization Groups^a

View Fullscreen

Regression analysis of data from the 2 highly specialized groups (Table 7) showed that the specialization group (evolved versus exclusive) was not associated with an increased proportion of overuse injuries. Athletes specializing in individual sports were at 1.95 times increased odds of having overuse injuries relative to acute injuries when compared with athletes specializing in team sports (95% CI = 1.17, 3.24). Among highly specialized athletes, the overuse injury proportion was not associated with sex, age, specialization path (evolved versus exclusive), age at specialization (when they started training for their main sport >8 months per year), or months of rest per year from organized sports. A validation regression model (Table 8), including only variables shown to be significantly associated with the overuse injury proportion in bivariate analyses (sex, age, starting age of training >8 months per year, and sport type), also indicated that only sport type had a statistically significant association with an increased proportion of overuse injuries, as individual-sport athletes were at 1.86 times greater odds of having overuse injuries relative to acute injuries (95% CI = 1.15, 3.00).

Table 7. Multivariable Analysis of Overuse Injury Proportions in the Highly Specialized Groups^a

View Fullscreen

Table 8. Multivariable Analysis of Overuse Injury Proportions in the Highly Specialized Groups Only Among Variables Individually Associated With the Overuse Injury Proportion^a

View Fullscreen

Sex Differences

As we are the first to delineate the unique characteristics of this exclusive highly specialized population, no prior research has been conducted on its demographic characteristics. The proportion of female athletes was highest in the exclusive highly specialized group, followed by the evolved highly specialized group, and then the low-moderate group (Table 1). Post et al²⁵ observed that female athletes might be more likely to be highly specialized than their male counterparts, and our study further informs this suggestion in showing that highly specialized female athletes may also be less likely to have played other sports before specializing. Gender differences in sport type may be responsible; for example, the majority of highly specialized gymnasts (78 of 102) in this study were female. Gymnastics often requires year-round training starting at a young age, and Root et al²⁶ noted that youth gymnasts often exceeded the training volume guidelines set by the American Academy of Pediatrics, which may not leave these athletes with sufficient time to participate in other organized sports. Additionally, the post hoc analysis of sport type and months of rest per year demonstrated that individual-sport athletes rested for fewer months per year than team-sport athletes, which offers another potential explanation for why these athletes may be more likely to specialize, as having less time off to rest may not lend itself to participating in multiple organized sports.

Age at Specialization

Exclusive highly specialized athletes began to train intensively (>8 months per year) for their main sport at a younger age than did evolved highly specialized athletes (Table 2). This factor may reflect the typical training schedule of the athlete’s chosen sport, as individual-sport athletes have reported greater training volume than those in team sports,¹⁷ and the authors²⁶ of 1 study determined that more specialized gymnasts had been participating in the sport for longer than their low-specialized counterparts. Thus, their ability to participate in other sports at a younger age would be limited, and those who begin their sports experience with a single sport may therefore have more time to devote to training in that sport compared with those who start their athletic experience sampling a variety of sports.

Team- Versus Individual-Sport Differences

Pasulka et al¹⁷ demonstrated that individual-sport athletes may be more likely to specialize than team-sport athletes. Our work corroborated this, as we identified a higher proportion of athletes playing individual sports in the exclusive highly specialized group compared with the evolved highly specialized group (Table 3). Hence, children in team sports may be more likely to play multiple sports before choosing to focus on one, whereas individual-sport athletes are more likely to specialize in the first sport they play at an organized level. Although high specialization is no longer unique to individual sports such as gymnastics and figure skating,²⁷ a difference in training philosophy between individual and team sports may exist: among the highly specialized groups, individual-sport athletes in our study rested for fewer months per year than team-sport athletes. As a consequence, especially in the context of individual-sport athletes reporting greater training volume than team-sport athletes,¹⁷ these sports may consume more of an athlete’s schedule and therefore lend themselves to exclusive specialization.

Serious Overuse Injuries

Unlike prior investigators, we observed no difference in the proportion of serious overuse injuries between highly specialized athletes and low-moderately specialized athletes. The risk factors for serious overuse injury may be related to the specific sport (eg, stress fractures in cross-country athletes) rather than the degree of specialization or path to high specialization.

Multivariable Regression

Among this population of injured athletes presenting to a sports medicine clinic, multivariable analysis reaffirmed that high specialization (both evolved and exclusive) was independently associated with greater odds of the injury being due to overuse.

Though prior researchers determined that injured adolescent athletes were, on average, older than their uninjured counterparts,¹³ we analyzed overuse injuries as a proportion of total injuries rather than as a proportion of the population as a whole and showed that, independent of other factors, older athletes had a lower proportion of overuse injuries than younger athletes. This differs from the findings of Stracciolini et al,²⁸ who established that 13- to 17-year-old athletes had a greater proportion of overuse injuries than 5- to 12-year-old athletes; however, these results are not directly comparable, as Stracciolini et al²⁸ categorized patients into 2 age brackets (5–12 and 13–17 years), whereas we analyzed age as a continuous variable. The regression also affirms the results of the prior analyses in this study: high specialization, regardless of the path taken (exclusive versus evolved), was associated with increased odds of sustaining an overuse injury (Table 6).

Female sex was also independently associated with greater odds, when presenting to clinic with an injury, of the injury being due to overuse, which is consistent with earlier research.^16,29–31 Jayanthi et al¹⁶ proposed that sex differences in the chosen sport type could be an explanation for this finding, suggesting that female athletes may be more likely to participate in individual sports or those with consistent repetitive motions and as a result be more susceptible to overuse injuries. Targeted studies are needed to test this hypothesis and investigate this, and potential other, mechanisms by which female athletes may be more susceptible to overuse injuries than their male counterparts.

When the analysis was restricted to the highly specialized athletes, only sport type was a significant predictor of proportion of overuse injuries: individual-sport athletes were at increased odds of sustaining an overuse injury when compared with team-sport athletes (Tables 7 and 8). The effects of age at presentation to clinic and gender were diminished when only the highly specialized groups were analyzed. Though this may seem to contradict prior outcomes of an increased risk of overuse injury in highly specialized young female athletes,^30–32 each had a different definition of highly specialized, as Leppanen et al³⁰ and Carragher et al³¹ did not use the 3-point scale, and Biese et al,³² as in all sport specialization studies before this one, did not account for the exclusively highly specialized group. The regression also supported our other analyses regarding the lack of difference in overuse injury proportions between the exclusive and evolved highly specialized groups. Future sport-specific studies are necessary to determine how sport type interacts with specialization status and other factors, such as age and sex, to influence injury characteristics and susceptibility.

Limitations

Our study was conducted at a single site, which was a tertiary care pediatric sports medicine clinic, so the results may not be generalizable to the wider population of youth athletes presenting for medical care for sport-related injury. Providers in this type of clinic tend to have more patients with overuse injuries than an urgent care or emergency room setting, whose providers see more patients with acute injuries. The lack of an uninjured control group was another limitation, as the injuries were quantified as a proportion of the total injuries at a tertiary sports medicine clinic during the study period. The effects of sport specialization, sport type, age, and sex might have been different if injuries were quantified as a proportion of total athletes rather than a proportion of injured athletes. Population-based studies with a control group of uninjured athletes are needed to further explore how the pathway to becoming highly specialized may influence the injury risk and characteristics.

Additionally, similar to any research that relies on patient questionnaires to collect data, our work was subject to recall bias.^13–18 In an effort to mitigate this as much as possible, health care providers reviewed the patient surveys at the time of completion in the clinic for completeness and accuracy.

Another limitation was that our analysis did not address the timing of quitting other sports to focus on a main sport, which could have affected the risk for overuse injury in the evolved highly specialized group. Including the age at which athletes began training for their main sport for >8 months per year explained some of this effect, but the age at which the athlete stopped playing other sports altogether was not addressed. Future examinations of evolved highly specialized athletes targeting the specific timing of quitting other sports are needed to elucidate this effect. Also, future authors could explore whether the type of sport that athletes played before specialization (individual versus team) was associated with the overuse injury proportion, as we did not account for this.

The study population consisted of many female gymnasts, and their overrepresentation may have driven some of the results. However, the regression analysis of all 3 groups demonstrated that female athletes were at increased odds of having a greater proportion of overuse injuries when other variables were controlled, which was consistent with the current literature.