Postinjury Outcomes After Non–Sport-Related Concussion: A CARE Consortium Study Open Access

This study was a secondary analysis of data obtained from the National Collegiate Athletic Association Department of Defense (NCAA-DOD) Concussion Assessment, Research and Education (CARE) Consortium, which includes varsity athletes from 26 civilian universities and 4 US service academies. A comprehensive description of the CARE Consortium methods has been previously published.¹⁰  Participants were included in the analysis if they were NCAA athletes and were diagnosed with a concussion per current consensus guidelines during the 2014–2015 through 2019–2020 academic years.¹¹  They were grouped by mechanism of injury (eg, sport related versus non-sport related) based on the postinjury question, “Injury occurred during…” Injuries that occurred during competition, training, or practice were grouped as sport related, and injuries that occurred outside of sport were grouped as non-sport related. Specific injury scenarios were collected for SRCs (eg, collision with teammate; contact with ball, stick, or puck; tackling; blocking) and non-SRCs (eg, fall, slip, or trip; motor vehicle crash; being struck by an object). Only the first concussion sustained by a participant was analyzed. Demographic information and medical history were obtained during athletes’ preseason preinjury assessments using a standard case report form across all sites. Each person self-reported age, sex, race, ethnicity, sport, and health history, including concussion, headaches or migraines, psychiatric disorders, attention-deficit/hyperactivity disorder (ADHD), and learning disabilities. Sport was categorized by the research team as contact, limited contact, or noncontact.¹²  To address the second aim, only data from non-SRCs were included. The Children’s Hospital of Philadelphia institutional review board and the local institutional review board at each performance site reviewed and approved all study procedures. Individuals provided written informed consent before participation.

Injury characteristics and outcomes were obtained from all patients with diagnosed SRCs and non-SRCs and included information on who made the initial diagnosis (eg, team physician, AT, primary care physician) and if the injury was reported immediately (yes or no). Acute outcomes were delayed symptom onset, altered mental status, loss of consciousness (LOC), posttraumatic amnesia, retrograde amnesia (RGA), motor impairment, and if hospital transport was required (all yes or no). Concussion symptoms were self-reported at the first visit within 48 hours of injury using the Graded Symptom Checklist (GSC) from the Sport Concussion Assessment Tool-Third Edition (SCAT-3). The GSC is a 22-item scale that prompts participants to self-report symptom severity using a Likert scale from 0 to 6, with higher scores indicating worse symptom severity (total range = 0–132). Recovery outcomes were collected at the time of unrestricted return to play and consisted of the provider type issuing final clearance (eg, team physician, AT, primary care physician), the number of days of reported concussion symptoms, the number of days lost to injury, and whether the athlete experienced a slow recovery, defined as ≥14 days to being asymptomatic or ≥24 days to unrestricted return to play.¹³ 

Our primary analysis assessed group differences in postinjury outcomes between participants with SRCs and those with non-SRCs. Independent t tests were performed to determine group differences in continuous demographic variables (age, height, and weight), and χ² tests were performed for categorical variables (sex, race, ethnicity, sport type, and concussion history). Effect sizes for continuous variables (age, height, and weight) were characterized using the Cohen d, and effect sizes for categorial variables were characterized using the φ coefficient for 2-level variables (sex and concussion history [yes or no]) and the Cramer V for variables with 2+ levels (race, ethnicity, and sport type). Adjusted relative risks (ARRs) with 95% CIs were used to evaluate differences in categorical postinjury outcomes between those with SRCs versus non-SRCs. We calculated the ARR via log-binomial regression analysis to control for relevant factors. Separate analyses were conducted to determine the effect of possible confounding variables, including sex, age, height, weight, race, ethnicity, NCAA sport type, and concussion history. Based on this analysis, we included sex and concussion history in the final models as covariates. Due to overdispersion of our count response variables, multivariate negative binomial regressions, controlling for sex and concussion history, were computed to assess associations between injury mechanism and acute GSC symptom severity (SRC mean = 27.2, variance = 452.3; non-SRC mean = 34.1, variance = 545.44), the number of days participants reported concussion symptoms (SRC mean = 8.2, variance = 74.9; non-SRC mean = 12.5, variance = 146.4), and the number of days lost to injury (days from injury to unrestricted return to play; SRC mean = 16.4, variance = 337.2; non-SRC mean = 23.0, variance = 406.5). Dispersion parameters were reviewed to evaluate the fit for each model.

For our secondary analysis, we investigated sex differences in injury outcomes within the non-SRC group only. Independent t tests were performed to identify group differences in continuous demographic variables (age, height, and weight), and χ² tests were performed for categorical variables (sex; race; ethnicity; sport type’ history of concussion, psychiatric disorder, ADHD, and learning disability). The ARR (controlling for concussion history) with 95% CI was used to compare differences in categorical postinjury outcomes between male and female athletes with non-SRCs. Due to overdispersion of our count response variable, multivariate negative binomial regressions, controlling for concussion history, were used to characterize associations between male and female athletes with non-SRCs and acute GSC symptom severity (male mean = 29.7, variance = 481.5; female mean = 36.9, variance = 567.8), the number of days participants reported concussion symptoms (male mean = 9.9, variance = 91.4; female mean = 13.8, variance = 167.5), and the number of days lost to injury (days from injury to unrestricted return to play; male mean = 20.2, variance = 256.0; female mean = 24.3, variance = 487.6). We reviewed dispersion parameters to assess the fit for each model. For all analyses, α was set to .05 a priori. Data were analyzed using SAS (version 9.4; SAS Institute Inc).

A total of 5460 concussions from the CARE Consortium were examined for inclusion. After exclusion of nonathlete military academy cadets (n = 1283) and recurrent concussions (n = 677), 3500 athletes with concussions were included in the primary analysis (n = 555 non-SRC; 42.5% female; age = 20.1 ± 1.3 years). Participant demographics are summarized in Table 1. A greater proportion of female athletes sustained non-SRCs (23.3%) compared with male athletes (10.3%). Among the total sample, non-SRCs were evenly distributed among contact- (33.2%), limited-contact– (34.6%), and noncontact- (32.3%) sport athletes, whereas most SRCs were sustained by contact-sport athletes (76.3%). Most athletes reported no concussion history (56.8% and 67.0% for SRCs and non-SRCs, respectively). The most common injury scenario for non-SRCs was a fall, slip, or trip (28.3%), followed by being struck by an object (25.8%), motor vehicle crash (19.8%), and unintentional contact with a person (8.6%). The most frequent injury scenario for SRCs was collision with a teammate or opponent (51.9%), followed by contact with a ball, stick, or puck (20.0%); tackling or being tackled (15.5%); blocking (6.4%); heading with or without physical contact (5.0%); and checking or being checked (1.3%).

Non–sport-related concussions were less likely to be reported immediately (ARR = 0.73, 95% CI = 0.65, 0.81), and those with non-SRCs were more likely to have delayed symptom onset (ARR = 1.17, 95% CI = 1.03, 1.32), LOC (ARR = 3.15, 95% CI = 2.32, 4.28), RGA (ARR = 1.77, 95% CI = 1.22, 2.57), and motor impairment (ARR = 1.45, 95% CI = 1.14, 1.84; Figure 1 and Supplemental Table 1, available at https://dx.doi.org/10.4085/1062-6050-0181.23.S1). In the non-SRC- cohort, male athletes were more likely to present with altered mental status (ARR = 1.19, 95% CI = 1.001, 1.41) and RGA (ARR = 1.87, 95% CI = 1.01, 3.45) than female athletes (Figure 2 and Supplemental Table 2). Compared to SRCs, non-SRCs were more likely to require hospital transport (ARR = 4.02, 95% CI = 3.12, 5.18; Figure 1).

Participants with non-SRCs reported greater symptom severity within 48 hours of injury (median = 28 versus 22; P < .001), had a longer duration of symptoms (median = 9 days versus 6; P < .001), and lost more days to injury (median = 17 versus 12; P < .001; Table 2) than those with SRCs. Patients with non-SRCs were also more likely to experience slow recovery (ARR = 1.50, 95% CI = 1.31, 1.71; Figure 1). Within the non-SRC group, female athletes described greater symptom severity (median = 31 versus 24; P = .003), had a longer duration of symptoms (median = 10 versus 8; P < .001), and lost more days to injury (median = 18 versus 15; P = .03) compared with male athletes (Table 3). Additionally, among athletes with non-SRCs, male athletes were less likely to have a slow recovery (ARR = 0.61, 95% CI = 0.47, 0.79; Figure 2).

Non–sport-related concussions were more likely to be diagnosed (ARR = 2.45, 95% CI = 1.95, 3.07) and cleared (ARR = 1.58, 95% CI = 1.35, 1.85) by primary care physicians and less likely to be diagnosed (ARR = 0.73, 95% CI = 0.66, 0.82) and cleared (ARR = 0.75, 95% CI = 0.62, 0.92) by ATs than SRCs (Figure 1). In the non-SRC cohort, male athletes were less likely to be diagnosed (ARR = 0.54, 95% CI = 0.30, 0.83) and cleared by team physicians (ARR = 0.61, 95% CI = 0.42, 0.87; Figure 2).

The primary purpose of our study was to investigate concussion reporting characteristics and postinjury clinical outcomes in collegiate athletes who sustained non-SRCs compared with those who sustained SRCs and, secondarily, to compare reporting characteristics and acute clinical outcomes between male and female collegiate athletes with non-SRCs. The non-SRCs were more likely to be reported later, present with worse acute clinical outcomes (presence of LOC, RGA, motor impairment, and higher symptom burden), and have a slower recovery compared with SRCs. Among those with non-SRCs, female athletes indicated a greater acute symptom burden, more days with concussion symptoms, and more days lost to injury. Additionally, non-SRCs were more likely to be diagnosed and cleared by primary care physicians and less likely to be diagnosed and cleared by ATs. Within the non-SRC group, male athletes were less likely to be diagnosed and cleared by team physicians.

When assessing injury reporting characteristics, participants with non-SRCs were less likely to report their concussions immediately than participants with SRCs. This may have been due to the logistical limitations of an athlete sustaining a concussion outside the context of sport and, therefore, away from onsite or sideline health care providers (eg, ATs, team physicians). Athletes may seek initial care outside the university or academy-affiliated providers (eg, personal primary care physicians, ED) for concussions occurring outside athletics, outside the competitive season, or outside the academic year. Other factors may include different interpretations of “immediate” reporting among clinicians completing the postinjury report, embarrassment about or hesitation in reporting due to “mild” mechanisms (eg, falling out of bed, bumping one’s head on a cabinet), or athletes being less apt to report a non-SRC to athletics-affiliated providers if they believe the mechanism might result in team or legal disciplinary action (eg, fighting, substance related). This aligns with previous results that children with non-SRCs took longer to seek care and were less typically diagnosed than children with SRCs.^14–17  In the pediatric ED, patients aged 5 to 18 years with non-SRCs were less likely to receive a concussion diagnosis^14,15  and to be immediately diagnosed.¹⁶  Among adults, patients who sustained a concussion via motor vehicle crashes or other non–sport-related mechanisms were 8 and 5 times more likely, respectively, to not be diagnosed with a concussion in the ED compared with those who incurred sport-related mechanisms.¹⁷  In a specialty concussion clinical setting, adolescent and young adult patients involved in motor vehicle crashes took an average of 21 days to report to the clinic compared with 14 to 16 days for those with sport-related mechanisms.¹⁸  Our findings in collegiate athletes add to these data and indicate that delayed reporting or presentation for initial care of non-SRCs occurs across health care settings and providers. This may be true even in acute care settings such as EDs or urgent care clinics. The delay may, in part, also be due to the lack of recognition of concussion symptoms after a non-sport injury; efforts educating about and raising awareness of concussions from mechanisms outside sports may represent an important intervention that could reduce the time to concussion diagnosis and improve outcomes.

We found that athletes with non-SRCs were more likely to report with delayed symptom presentation, LOC, RGA, and motor impairment and that non-SRCs were associated with a greater initial symptom burden, more days with concussion symptoms, more days lost to injury, and a greater risk of slow recovery. Similarly, in the general adult population, previous research showed that non-SRCs were associated with greater symptom endorsement and severity,⁸  sleep disturbance,¹⁹  and persistent postconcussion symptoms²⁰  compared with SRCs. Poor clinical outcomes after non-SRCs may reflect several factors, including delayed reporting or variable injury mechanisms. As mentioned earlier, we identified that participants with non-SRCs less often reported their injuries immediately, in part, perhaps, because they were unrecognized. Mild non-SRC injury mechanisms (eg, falling out of bed, bumping one’s head on a car door) that occur without medical supervision may not be recognized as possible injuries and, after delayed reporting, may result in a greater symptom burden at the clinical visit. In a recent meta-analysis of 12 studies (10 included SRCs only, 2 included all injury mechanisms), the authors established that delayed reporting was associated with greater initial symptom severity scores and a longer recovery relative to concussions reported immediately.²¹  In pediatric patients seen in the ED, a delayed concussion diagnosis has been associated with more medical visits, a longer time to symptom recovery, and a higher likelihood of persistent symptoms.¹⁶  Among collegiate athletes,²²  delayed reporting and delayed removal from play have been associated with more time missed due to injury and more days with concussion symptoms, though the researchers excluded those with non-SRCs.²²  Though prolonged recovery (concussion symptoms for >28 days)²³  is commonly used as a prognostic indicator, only 3% (n = 106) of the study sample experienced prolonged recovery, and therefore, slow recovery was determined to provide a more granular description of athlete recovery in this study. We suggest that delayed reporting may similarly influence postinjury clinical outcomes for patients who sustain concussions with non–sport-related mechanisms.

Within our non-SRC cohort, male athletes were more likely to present with altered mental status and RGA compared with female athletes and female athletes reported greater initial symptoms and had a longer duration of symptoms, more days lost to injury, and a greater risk of slow recovery. Prior studies that assessed sex differences in clinical outcomes after SRCs described mixed results, with some indicating that female athletes described greater symptom burden^24–26  and a longer recovery trajectory^24,26,27  than male athletes, whereas others demonstrated no sex-based differences in symptoms or recovery.^7,25  Specifically in collegiate athletes with SRCs, Master et al⁷  observed that female athletes had higher symptom severity scores, which aligns with our results. Elevated self-reported symptoms in female athletes after SRCs is well known^7,24  and may indicate either biologically based or gender-based behavioral differences in symptom reporting. Similar to previous literature in which authors examined sex differences after SRCs in sex-comparable collegiate sports,⁷  we found no sex differences in LOC after injury. Though we noted that female athletes with non-SRC had worse recovery outcomes (more days with concussion symptoms, more time lost due to injury, and a higher risk of slow recovery), previous research of collegiate SRCs only identified sex differences in recovery metrics between female and male contact-sport athletes.⁷ 

Approximately 1 in 5 non-SRCs required hospital transport compared with 1 in 20 SRCs. This may be due to several factors, including a lack of access to onsite or sideline medical care, differences in acute clinical presentation, or comorbid injuries. In athletic-related situations at the collegiate level (practice, training, or competition), initial injury assessment is usually provided by ATs or team physicians, and hospital transport is only required when a higher level of care is deemed necessary.²⁸  Athletes who sustain a concussion outside of sport often do not receive the initial assessment by an onsite health care provider and may be more likely to require emergency services. Additionally, the variable mechanisms of non-SRCs may indicate more severe injuries, leading to increased hospital transports. Although patients who had non-SRCs were more likely to require hospital transport compared with patients who had SRCs, they represented only 20% of the overall non-SRC population, suggesting that hospital transport was still infrequent for those with a non-SRC. Also, 19 (3.4%) non-SRCs were diagnosed by ED clinicians. This small proportion of ED diagnoses may indicate missed diagnoses of non-SRCs in EDs¹⁷  or data-collection errors when reporting diagnosis information. Continued research regarding the diagnosis and management of non-SRCs as well as education or training (or both) is needed in emergency and urgent care settings.

When assessing health care provider diagnosis and clearance, we found that non-SRCs were more likely to be diagnosed and cleared by primary care physicians and less likely to be diagnosed and cleared by ATs. As non-SRCs, by definition, occur outside of sport, it is unsurprising that they are more often seen by health care providers outside of athletics-affiliated providers. Baugh et al²⁹  recently showed that greater trust in relevant stakeholders (ie, sports medicine clinicians, coaches, administrators) was associated with higher odds of intending to report a concussion in collegiate football athletes. If athletes with non-SRCs have a low level of trust in the sports medicine system, have poor rapport with clinicians, or are unwilling to disclose an injury that resulted from illegal or inappropriate activity (eg, assault, substance use), they may be less likely to report an injury occurring outside of sport to athletics-affiliated clinicians. Further, male athletes were less likely to be diagnosed and cleared by team physicians compared with female athletes. Weber et al³⁰  have demonstrated sex differences in concussion reporting among collegiate athletes: female athletes had greater indirect intentions to report (ie, “I would stop playing and report symptoms…”). Our results build on this, suggesting that sex differences in whom the athletes are reporting to after a non-SRC may exist. Future researchers should continue investigating care-seeking behaviors after non-SRCs to better guide injury management across providers.

This study was not without limitations. As a secondary analysis of a larger dataset, our sample was limited to collegiate athletes, and therefore, our findings regarding non-SRCs may not be generalizable to a nonathlete population. Reporting of acute clinical signs (LOC, RGA, posttraumatic amnesia, motor impairment, and altered mental status) by athletes with concussions occurring without immediate or sideline access to care (non-SRCs and SRCs that may have been reported late) relied on athlete self-report and may have been subject to recall bias. Immediate injury reporting was not operationally defined before the study period, and thus, the meaning of immediate was left to the discretion of the clinician completing the postinjury report. Similarly, the selection of team physician as the diagnosing or clearing provider may have been viewed as describing different clinical specialties (eg, orthopaedics, sports medicine, primary care, neurology) depending on the interpretation of the clinician completing the report. Additionally, we did not collect information regarding whether the athlete was at home on break when the non-SRC occurred or when clearance took place, which may have influenced access to health care or concussion management (or both). Lastly, other social (eg, concussion reporting and management) and biological (eg, injury threshold or severity) factors may be associated with or interact with our primary outcomes but could not be distinguished with the current study design.

In summary, athletes who sustained non-SRCs reported later, had higher symptom severity scores at initial presentation, experienced slower recoveries, and more often reported to nonathletic medical team health care providers compared with athletes who sustained SRCs. Continued investigation of non-SRC management outside of athletic environments is critical, and medical teams need to improve clinical outcomes. Within the non-SRC group, female athletes reported higher symptom severity scores and had a greater risk of slow recovery compared with male athletes. This was likely due to a combination of intrinsic factors (ie, biological, reporting behaviors) and extrinsic factors (ie, access to medical teams) but warrants further evaluation into sex differences in non-SRC management.

Contributing CARE Consortium investigators included Scott Anderson, Elsa Ermer, Stacey Harcum, Kalyn Jannace, and Scott Fransesca (Uniformed Services University of the Health Sciences); Justus Ortega (Humboldt State); Nicholas Port (Indiana University); Dianne Langford (Temple University); Thomas Kaminski (University of Delaware); Christopher Miles (Wake Forest University); Stefan Duma and Steven Rowson (Virginia Tech University); Luis A. Feigenbaum (University of Miami); Jason Mihalik (University of North Carolina at Chapel Hill); April Hoy (Azusa Pacific University); Holly Benjamin (University of Chicago); Kenneth Cameron (West Point); Anthony Kontos (University of Pittsburgh); Anthony Kontos (University of Pittsburgh); Brian Dykhuizen (University of North Carolina at Wilmington); Alison Brooks (University of Wisconsin); Christopher Giza and Josh Goldman (University of California at Los Angeles); Gerald McGinty (United States Air Force Academy); Jeff Bazarian (University of Rochester); Jessica Dysart Miles (University of North Georgia); Jonathan Jackson (United States Air Force Academy); Julianne Schmidt (University of Georgia); Megan Roach (West Point); Kevin Guskiewicz (University of North Carolina at Chapel Hill); and Laura Litner (Wake Forest University). We also thank the athletes and research and medical staffs at each of the participating sites as well as the staff of the CARE Consortium.

Research reported in this publication was supported by the National Institute of Neurological Disorders and Stroke of the National Institutes of Health (grants No. R01NS097549 and No. T32NS043126) and the Grand Alliance CARE Consortium, funded in part by the NCAA-DOD. The US Army Medical Research Acquisition Activity, 820 Chandler Street, Fort Detrick, MD 21702-5014, is the awarding and administering acquisition office. This work was supported by the Office of the Assistant Secretary of Defense for Health Affairs, through the Combat Casualty Care Research Program, endorsed by the DOD, through the Joint Program Committee 6/Combat Casualty Care Research Program—Psychological Health and Traumatic Brain Injury Program under award No. W81XWH1420151. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the DOD.

Supplemental Table 1. Adjusted Relative Risk (ARR) and 95% CI Shown in Figure 1 for All Postinjury Outcomes Between Sport-Related and Non–sport-Related Concussions

Supplemental Table 2. Adjusted Relative Risk (ARR) and 95% CI Shown in Figure 2 for All Postinjury Outcomes in Non–sport-Related Concussions Between Male and Female Athletes

Found at DOI: https://dx.doi.org/10.4085/1062-6050-0181.23.S1