Long-Term Health-Related Quality of Life After Sport-Related Injury in Youth Athletes: A Longitudinal Cohort Study

Study Design

A multicenter longitudinal cohort study was conducted between November 2018 and March 2020 at Emory University, Department of Orthopaedics, and between July 2019 and March 2020 at Boston Children’s Hospital, Division of Sports Medicine, Department of Orthopaedics and Lurie Children’s Hospital of Chicago. Patients were evaluated by fellowship-trained attending-level sports medicine physicians. Each site received institutional review board approval.

Inclusion and Exclusion Criteria

Patients 8 to 18 years of age with an acute or overuse musculoskeletal injury or sport-related concussion who were evaluated in a sports medicine clinic were eligible for study enrollment, regardless of participation in competitive athletics. Nonathletes were included in the study as the initial intent was to compare athlete and nonathlete injury outcomes. However, due to low nonathlete enrollment, the study method was modified to achieve the stated objectives. Patients who were unable to read or write in English or were participants in another research study were excluded.

Recruitment and Enrollment

Eligible patients were recruited at the time of clinic visit with consent and/or assent obtained electronically via Research Electronic Data Capture (REDCap; Vanderbilt University). Recruitment ended March 17, 2020.

Patient Characteristics

Data extracted from the electronic medical record included age, sex, and International Classification of Diseases, 10th Revision (ICD-10) code for the patient’s primary injury at the time of enrollment.

Outcome Measures

Patients used a clinic laptop or quick response code using a personal electronic device to complete the Player Development Survey adapted from Jayanthi et al. ¹⁶ This survey determined a patient’s classification as athlete or nonathlete based on participation in organized sport.

Patients were emailed a unique link 1 month after enrollment to complete the Patient-Reported Outcomes Measurement Information System (PROMIS) Pediatric-25 Profile v2.0, ⁴ which measures HRQoL via 6 domains: depression/sadness, anxiety/fear, mobility, pain interference, fatigue, and peer relationships. The 1-month follow-up was modeled after a previous study that evaluated HRQoL among patients with different injury types. ³³ Patients completed the same survey at 12- and 24-month follow-up.

Up to 3 weekly reminders were sent via email or text message if patients had not completed the surveys at each follow-up time point. The patient’s unique link remained active for completion anytime.

Patients were assigned to an injury cohort based on ICD-10 code and clinic note in accordance with a previously published injury classification system. ¹⁷ Concussions were diagnosed per Berlin guidelines. ²⁴ Acute injuries were defined as related to a single traumatic event, whereas overuse injuries were attributed to gradual onset. ¹⁷ Serious overuse injuries included spondylolysis, stress fracture to spine or extremity, physeal stress injuries, elbow ligament overuse injuries, and osteochondritis dissecans. ¹⁵

Power Analysis

The sample size was fixed based upon power analysis performed for a related study from the same data set evaluating short-term HRQoL outcomes. ⁴²

Statistical Analysis

A mixed model for repeated measures was used to analyze study outcomes (T scores for 6 HRQoL domains). For PROMIS data, T scores were calculated using the HealthMeasures Scoring System (Northwestern University). T scores have a mean of 50 and standard deviation of 10 for each HRQoL domain based on a reference population of patients ages 8 to 17 years.^11,14 Patients were significantly different than the reference if their T score varied by >0.5 SD from the mean.^18,30,32

The SAS Mixed Procedure (Version 9.4, mixed linear models) was used to fit the model providing estimates of the T score means for each HRQoL domain by time point in study (1 month, 12 months, and 24 months) and injury type (acute, concussion, overuse, and serious overuse). The model included 4 predictors: injury type, time point in the study (categorical), the statistical interaction between injury type and time point in the study, and study site. The correlation among the repeated measurements across time within a participant was assumed to follow the compound symmetry covariance structure, and empirical estimates of the standard errors of parameters were used to perform statistical tests and construct 95% CIs. ⁷ The linear mixed-model analysis is valid under the assumption of missing data at random (ie, the missing data mechanism is dependent only on observed data). A P value ≤.05 was considered statistically significant for the main effects (injury type and time point in study) and for the interaction between injury type and time point in the study from the repeated-measures analysis. Specific statistical tests were done within the framework of the mixed-effects linear model. A baseline adjusted analysis was also performed for each T score HRQoL domain. All statistical tests were 2-sided.

Because the temporal patterns of the HRQoL domain T scores were similar by injury type (ie, no statistical interaction between injury type and time point in the study), the time-averaged HRQoL domain means and 95% CIs for each HRQoL domain were reported. The time-averaged mean is the mean pooled across the 3 time points estimated by the mixed-model analysis of repeated measures.

Ethical Considerations

Consent to participate was provided by parent or legal guardian for all patients ages 8 to 17 years, along with assent from the patient if 12 years or older. Only the patient provided consent if 18 years old. For patients younger than 12 years, a parent or guardian was asked to complete surveys in the presence of the patient. For patients age 12 to 18 years, a parent or guardian was asked to supervise while patients completed surveys.

Study Participants

Of 720 patients enrolled, 357 (36% male; mean age, 14.2 years) completed the 1-month PROMIS survey, comprising 119 (33%) acute injuries, 42 (12%) concussions, and 196 (55%) overuse injuries (Figure 1). Of the 357 respondents, 336 (94%) played at least 1 sport competitively and sustained SRI (Table 1). The top 5 most frequent acute and overuse injuries are listed in Appendix 1 (Table A1, b and c; available in the online version of this article).

OPEN IN VIEWER

Figure 1.

CONSORT (Consolidated Standards of Reporting Trials) diagram: observational study.

OPEN IN VIEWER

Table 1

Patient Demographic Characteristics (Completed PROMIS Survey at 1 Month After Enrollment) ^a

Characteristic	Total(N = 357)	Athlete(n = 336)	Nonathlete(n = 21)
Age, y	14.2 ± 2.5	14.2 ± 2.4	13.9 ± 3.4
Sex
Male	129 (36)	122 (36)	7 (33)
Female	228 (64)	214 (64)	14 (67)
Injury cohort
Acute	119/357 (33)	112/336 (33)	7/21 (33)
Concussion	42/357 (12)	41/336 (12)	1/21 (5)
Overuse (total)	196/357 (55)	183/336 (54)	13/21 (62)
Serious overuse	53/357 (15)	51/336 (15)	2/21 (10)

a

Data are presented as mean ± SD or n (% of total). Note: 200 patients had overuse injuries, but 4 of 200 did not report athletic participation status.

Time of baseline survey completion was approximately 1 month after enrollment (mean 38 days, median 31 days). Of the 357 respondents, 228 (63.8% retention) completed surveys at 12 months after enrollment, and 126 (35.3% retention) completed surveys at 24 months after enrollment (Appendix 1, Table A1, d and e, available online). Age, male to female ratio, and injury type ratio remained similar at 1-, 12-, and 24-month time points (Appendix 1, Table A1, d and e).

HRQoL by PROMIS Domain, Injury Type, and Time

Appendix 1 Table A2 (available online) and Figure 2 present the findings of HRQoL in relation to injury type and time after injury. With 1 exception, none of the PROMIS domains, regardless of injury type or time after injury, had worse HRQoL than NDHY. The exception was that patients with overuse and serious overuse injuries had worse mobility compared with NDHY at 1 month after injury (mean 44.6 [95% CI, 43.1-46.1] and 43.7 [95% CI, 41.3-46.2], respectively), with resolution at 12 and 24 months after injury (overuse injury: mean 50.8 [95% CI, 48.9-52.7] and 53.1 [95% CI, 50.4-55.8]; serious overuse injury: mean 51.8 [95% CI, 49.4-54.2] and 49.8 [95% CI, 46.1-53.6], respectively). Notably, no significant difference was found in mobility between patients with overuse injuries and serious overuse injuries at any time point.

OPEN IN VIEWER

Figure 2.

Model-based means and 95% CIs for Patient-Reported Outcomes Measurement Information System (PROMIS) health-related quality of life domain T scores by type of injury and time point in study.

HRQoL by PROMIS Domain, Sex, and Time

Appendix 1 Table A3 (available online) and Figure 3 present the findings of HRQoL in relation to sex and time after injury. Female patients had worse mobility than NDHY at 1 month after injury (mean 44.7 [95% CI, 43.5-45.9]), which proved to be no different than NDHY at 12- and 24-month follow-up. No differences were noted in any other PROMIS domain at any time point compared with NDHY for both male and female patients.

OPEN IN VIEWER

Figure 3.

Model-based means and 95% CIs for Patient-Reported Outcomes Measurement Information System (PROMIS) health-related quality of life domain T scores by sex and time point in study.

Appendix 1 Table A3A (available online) shows pairwise analyses. Compared with male patients, female patients showed worse mobility at 1 and 12 months after injury (P = .015 and P = .045, respectively), anxiety/fear at 1, 12, and 24 months after injury (P = .002, P≤ .001, and P≤ .001, respectively), fatigue at 12 and 24 months after injury (P = .022 and P = .023, respectively), depression/sadness at 1, 12, and 24 months after injury (P = .001, P = .001, and P = .001, respectively), and pain interference at 1, 12, and 24 months after injury (P = .021, P = .009, and P≤ .001, respectively). Anxiety/fear worsened over 2 years after injury in female patients and improved in male patients.

HRQoL by PROMIS Domain, Age, and Time

Appendix 1 Table A4 and Appendix 2 Figure A4 (available online) present the findings of HRQoL in relation to age and time after injury. Patients age 15-16 and 17-19 showed worse mobility than NDHY at 1 month (mean 44.9 [95% CI, 43.2-46.6] and 44.0 [95% CI, 41.9-46.1], respectively), with no difference demonstrated at 12- and 24-month follow-up. At no other domains or times points did any age-group demonstrated HRQoL that was significantly different from that of NDHY.

Pairwise analysis showed that patients in the oldest age quartile (17-19 years), compared with the youngest age quartile (8-12 years), showed worse fatigue at 1, 12, and 24 months after injury (P≤ .001, P≤ .001, and P = .003, respectively) and worse mobility at 1 and 12 months after injury (P = .0174 and P = .0302, respectively).

HRQoL by PROMIS Domain, Sport Attrition Status, and Time

Appendix 1 Table A5 and Appendix 2 Figure A5 (available online) present the findings of HRQoL in relation to sport attrition and time after injury. Of the 357 patients, 53 patients (14.8%) quit an organized sport at 1 month after injury; 38 of the 357 (10.6%) quit an organized sport at 1 month after injury specifically due to injury. Sport attrition data were not available at 12- and 24-month follow-up.

Patients who quit sport had worse mobility at 1 month after injury compared with NDHY (mean 40.8 [95% CI, 38.4-43.2]). However, no difference was seen at 12 and 24 months after injury. At no other domains or time points was a difference seen between injured young athletes and NDHY, regardless of sport attrition status.

Pairwise analysis showed that patients who quit sport, compared with those who did not quit sport, demonstrated worse pain interference at 1 and 12 months after injury (P = .0061 and P = .0378, respectively) and worse mobility at 1 month after injury (P≤ .001). Both mobility and pain interference differences resolved at 24-month follow-up.

HRQoL by PROMIS Domain, Surgical Status, and Time

Appendix 1 Table A6 and Appendix 2 Figure A6 (available online) present the findings of HRQoL in relation to surgical status and time after injury. Of 357 patients, 22 patients (6.2%) required surgery. Twelve surgeries (of 357 patients; 3.4%) were for acute injuries and 9 (2.5%) were for overuse injuries. One patient with concussion underwent surgery for reasons unrelated to concussion.

Patients undergoing surgery showed worse pain interference and mobility (mean 57.5 [95% CI, 53.5-61.5] and 34.3 [95% CI, 30.6-37.9], respectively) compared with NDHY at 1 month after injury. However, pain interference and mobility each improved such that there was no difference between surgical patients and NDHY at 12 and 24 months after injury. No significant difference was seen in any other domain between patients undergoing surgery and NDHY at 1, 12, and 24 months after injury.

Pairwise analysis showed that patients requiring surgery, compared with those not requiring surgery, had worse mobility at 1 month after injury (P≤ .0001) and worse pain interference at 1 and 12 months after injury (P = .0061 and P = .0378, respectively). No mobility or pain interference differences were seen at 24 months after injury.

Time-averaged means indicated there was no difference in pain interference, peer relationships, depression/sadness, fatigue, or anxiety/fear based upon surgical status (P = .146, P = .129, P = .105, P = .364, P = .112, respectively).

Key Study Finding

Our study found no long-term differences in HRQoL between injured young athletes and age-matched NDHY at 12 and 24 months after injury, regardless of injury type, sex, age, surgical status, or sport attrition status. The subsequent sections compare our study results to those of previous published manuscripts and detail the proposed mechanisms for our study findings.

Overuse Injuries, Time, and Health-Related Quality of Life

Our study showed that patients with overuse and serious overuse injuries had worse mobility compared with NDHY 1 month after injury. However, the difference resolved at 12 months after injury and remained resolved at 24 months after injury. We previously reported that the mechanism of repetitive submaximal forces occurring over a period with insufficient recovery may explain the initial HRQoL deficit in young athletes with overuse and serious overuse injuries. Multiple weeks or months of decreased training volume may be needed to resolve an overuse injury, in addition to the potential requirement of durable medical equipment such as a walking boot or crutches to promote healing. The recovery of mobility may be explained by most overuse injuries typically resolving in a period of 8 to 12 weeks. For instance, a previous study suggested that 8 to 12 weeks of activity restriction is a typical recovery time for patients with stable juvenile osteochondritis dissecans lesions. ¹ It is therefore reasonable that the longer time for recovery of mobility after overuse injury is a function of the inherent time to heal an overuse injury.

Acute Injuries, Time, and Health-Related Quality of Life

Our study showed no HRQoL domains at short-term or long-term follow-up in which young athletes had worse HRQoL compared with NDHY. Russell et al ³³ reported worse physical and overall HRQoL in athletes with sport-related extremity fracture compared with NDHY initially after injury but no difference in overall HRQoL at 1 month after injury, although decreased physical HRQoL persisted. Our study results are similar in that there was no difference in short-term overall HRQoL outcomes. However, our study results were in opposition regarding short-term physical HRQoL outcomes. Our acute injury classification was broader as opposed to including only patients with a fracture. Only 16 of 112 patients with acute injury in our study required surgery, so injury severity was likely lower compared with the study by Russell et al.

A novel finding of our study is there were no long-term HRQoL deficits in any HRQoL domain after acute injury. This may be explained by most nonsurgical acute injuries resolving within a few weeks, such as ankle sprains or contusions. The time to injury resolution would explain the preservation of domains such as pain interference and mobility. The resolution of symptoms also implies return to sport or activity, so potential effects on domains such as peer relationships may be lessened.

Concussions, Time, and Health-Related Quality of Life

Our study results showed no HRQoL domains at short-term or long-term follow-up in which young athletes had worse HRQoL compared with NDHY. Russell et al ³³ evaluated HRQoL in young patients with sport-related concussion and found worse initial physical and overall HRQoL compared with NDHY but no difference in any HRQoL domain when followed to physician-documented recovery at approximately 26 days after injury. A qualitative study by Iadevaia et al ¹³ of 7 parent-child pairs via interviews by high school athletic trainers at 1 year after concussion found negative influence on physical and emotional function, academics, and interpersonal interactions. Moran et al ²⁸ studied 186 patients, 8 to 15 years of age, who had mild traumatic brain injury (TBI). Those investigators showed that mild TBI did not have a strong, persistent influence on HRQoL at 1 year after concussion but that children who experienced higher levels of postconcussion syndrome after mild TBI were at risk for significant and lasting declines in HRQoL.

Our results differ from the findings of Iadevaia et al ¹³ and are in overall agreement with the findings of Moran et al. ²⁸ Our study was conducted at 3 tertiary academic referral centers, so the number of concussion cases with risk factors for prolonged symptoms may have been higher than in the general population. Also, our study compared HRQoL to NDHY, so the ultimate outcome measured was different. The finding of no significant difference in any HRQoL domain compared with NDHY at 12- or 24-month follow-up may be explained by most concussion injuries resolving within 3 weeks and likely only a small proportion of patients, if any, having symptoms at 12 or 24 months after concussion. ¹⁰

Sex, Time, and Health-Related Quality of Life

Our previous study showed that female patients were at higher risk of reporting worse pain interference, depression/sadness, fatigue, anxiety/fear, and mobility than their male counterparts at 1 month after injury. In addition, female patients had worse mobility than NDHY. ⁴² Our long-term findings show recovery of mobility such that female patients had no HRQoL domain that was worse than NDHY at 12- and 24-month follow-up. This reassuring finding suggests that female patients, compared with NDHY, are not at increased risk for worse HRQoL after injury.

However, we noted persistently worse pain interference, depression/sadness, fatigue, anxiety/fear, and mobility in female patients compared with male patients. These findings are similar to those of Dipnall et al, ⁸ who evaluated HRQoL in 2334 injured patients (624 female) ages 5 to 17 and found that girls had higher risk of reporting problems with mobility, activity, pain, and anxiety/depression 1 month to 2 years after injury.

Multiple studies have reported worse HRQoL for adolescent girls compared with boys, even in the absence of injury. In a 3-year longitudinal study of 403 Australian adolescents aged 12 to 15, Meade and Dowswell ²⁵ found that girls reported worse scores than boys for physical and psychological well-being, autonomy, and relationships with parents. ²⁵ In a cross-sectional study of 652 Malaysian adolescents aged 13 years, Loh et al ²¹ showed that girls reported significantly lower emotional functioning scores.

The exact mechanisms by which adolescent female patients have worse HRQoL compared with male patients are unclear. In a cross-sectional study of 295 students from middle and high schools in Korea, Shin et al ³⁴ found that menstrual health was an important factor that mediated the effects of eating, sleeping, psychological health, and social support on HRQoL. A study by Albert ² suggested that depression was more prevalent in female athletes than male athletes and that depression increased in female athletes after puberty. Depression may affect physical and psychological domains of HRQoL and have a persistent effect longitudinally. ² Overall, although female patients demonstrated worse HRQoL compared with male patients at short-term and long-term follow-up in our study, the difference may be independent of injury and caused by other factors. Further evaluation of sex difference mechanisms is needed.

Age, Time, and Health-Related Quality of Life

Our previous study showed that at 1 month after injury, peer relationships, fatigue, and mobility worsened with increasing age. ⁴² Our current study shows that patients in the oldest age quartile had worse fatigue than the youngest age quartile at 12-month and 24-month follow-up and worse mobility at 12-month follow-up. Importantly, other than patients in older age quartiles showing worse mobility than NDHY 1 month after injury, there were no other domains or time points at which any age group demonstrated worse HRQoL compared with NDHY.

A review by Martin-Hertz et al ²³ showed that older age was associated with lower HRQoL after traumatic injury. The 12- and 24-month outcomes of our study are similar in that some domains (fatigue and mobility) worsened with increasing age while others did not.

Of note, Meyer et al ²⁷ evaluated 530 healthy youth (mean age, 13.2 years) over the course of 15 months and found that HRQoL declined with increasing age. Increasing age and maturation, rather than injury status, may therefore be the factors driving the worsening of fatigue and mobility scores. One possible explanation may be the development of independence and self-conceptualization, which may increase not only as a student-athlete reaches higher levels of sport but also with increasing age in general. Further evaluation of age and HRQoL, both in the presence and absence of injury, is warranted.

Sport Attrition, Time, and Health-Related Quality of Life

Our study showed that 10.6% of patients quit an organized sport specifically due to injury 1 month after injury. Other than showing that patients who quit sport had worse mobility compared with NDHY, we found no difference in any HRQoL domain compared with NDHY at any time point, regardless of sport attrition status. However, patients who quit sport had worse pain interference and mobility 1 month after injury than patients who did not quit sport, with recovery at 12- and 24-month follow-up.

Jindo et al ¹⁹ evaluated early attrition from school sports clubs among male senior high school students in Japan. Of 331 students enrolled, 41 (15%) experienced sport attrition over 2.4-year follow-up. Risk factors for attrition included low body weight and low body mass index, no experience of injury or disability that affected daily life, low individual athletic achievement, and short duration of experience of the sport. ¹⁹ Our study results suggest that worse mobility and pain, presumably due to injury, play a key role in a patient’s decision to quit a sport.

Our study and the results of the study by Jindo et al ¹⁹ together suggest that sport attrition may be a multifactorial event associated with dynamics beyond physical pain and may include matters of sport-related achievement, sport experience, and the social aspects of sport participation. Further evaluation of the mechanisms of sport attrition is warranted to provide guidance on attaining positive sport participation experiences.

Surgical Status, Time, and Health-Related Quality of Life

Of 357 patients in our study, 22 patients (6.2%) underwent surgery. Of the 22 patients, 9 (2.5%) sustained overuse injuries, whereas 12 (3.4%) sustained acute injuries. One patient with concussion underwent surgery for reasons unrelated to concussion.

A paucity of literature is available regarding long-term HRQoL outcomes after orthopaedic surgeries in pediatrics, particularly in the context of sport participation. We believe our study is one of the first to evaluate long-term HRQoL outcomes after orthopaedic surgery in a population of young athletes.

Our results indicate that mobility and pain interference were worse for patients undergoing surgery compared with NDHY at 1 month after injury. However, mobility and pain interference deficits resolved at 12 and 24 months after injury. Our results also indicate that patients undergoing surgery had worse mobility compared with patients treated nonsurgically at 1 month after injury, but this difference also resolved at 12- and 24-month follow-up. Importantly, we found no difference between patients undergoing surgery, patients treated nonsurgically, and NDHY in any other PROMIS domains or follow-up time points. The results may be due to the body’s inherent ability to heal after surgery and the use of rehabilitation exercises to improve both mobility and pain.

Rabbits et al ³¹ evaluated HRQoL in a population of 915 patients ages 2 to 18 years (mean 9.6 years) before and after undergoing inpatient surgery. Outcomes were measured to 1 month after surgery. Of the 915 surgeries performed, 195 (21.2%) were orthopaedic. The results showed that although most patients recovered to their baseline HRQoL at 1 month after discharge, 23% of the patients had significant decline in HRQoL. Risk factors for lower HRQoL included increasing age and moderate-severe postoperative pain.

Although our study did not include a presurgical HRQoL score, our results are similar such that patients undergoing surgery had worse mobility and pain interference compared with NDHY at 1 month after injury. The results may be due to relatively longer periods of immobilization after an orthopaedic surgery using devices such as crutches or walking boots. Requirement of surgery implies an increase in injury severity, which may be directly proportional to initial pain interference and mobility deficits.

Limitations

We noted the following limitations of the study. (1) Our study was not population-based, so selection bias may be present given that patients sought care at sports medicine clinics. Several potential patients were not approached as they were enrolled in other studies. The number of patients approached was not recorded. (2) We did not include an uninjured control group. However, we believe that a comparison of injured patients with age-matched normative data for youth provides a valid comparison to fulfill the study objectives. (3) The time from injury occurrence to enrollment was not recorded. We believe that most symptoms began around the time of enrollment. However, the duration of symptoms may have influenced PROMIS scores. (4) Patients with multiple injuries and previous injuries to the affected body area were included in the analysis. We believe that the presence of multiple injuries is unlikely to significantly affect the study results given that only 2.8% of enrolled patients reported multiple injuries. (5) Injury severity was not recorded, although injuries that required surgery may be inferred as more severe than nonsurgical injuries. (6) Patients with concussion may have had decreased enrollment due to concomitant enrollment in other concussion studies at 1 study site. (7) Non-English speakers were excluded, which may limit generalizability to certain clinical settings. (8) Although the study recruitment period ended before the onset of the COVID-19 pandemic, longitudinal follow-up occurred during the pandemic and may have been affected due to sport participation factors. (9) The number of patients undergoing surgery was low, resulting in analysis that may be underpowered when comparing outcomes to patients treated nonsurgically. (10) Our study had a follow-up rate of 64% at 1 year and 35% at 2 years. As such, bias from low follow-up may have affected the results if the characteristics of the patients who dropped out were systematically different from those of patients who completed follow-up. (11) Although we suspect that HRQoL outcomes at 2-year follow-up in our study population will be durable, more definitive long-term conclusions may be obtained from additional longitudinal follow-up. (12) Sport attrition and resumption data were not available at 12- and 24-month follow-up.

Significance

Although some short-term HRQoL deficits exist, patients and their families may find reassurance in learning that young athletes generally recover from short-term HRQoL deficits and do not experience long-term HRQoL deficits following all injury types, attrition, and surgery. Further evaluation of the relationship between after injury HRQoL and sex differences is warranted.