Correlates of sport‐related concussion in male junior rugby union: A concurrent analysis of biopsychosocial factors

Abstract

Objective

Sport‐related concussion (SRC) is a multi‐faceted issue that should be considered in context, with consideration for the biological, psychological, and sociocultural (i.e., biopsychosocial) factors which influence the incidence of injury in contact sports such as rugby union (rugby). Through the concurrent assessment of individual variables within a multi‐measure cross‐sectional research design, the current study aimed to contextualise the individual factors associated with SRC in male junior (ages 11–17-years) rugby.

Method

Self‐ and parent‐report measures were used to assess athletes' psychological and behavioural functioning. Sociocultural influences were considered in terms of duration of participation, involvement of immediate family, and participation in other sports. Biological measures included athlete age, BMI, aerobic fitness, and in‐utero testosterone exposure (2D:4D).

Results

Athlete age was positively correlated with concussion incidence, with adolescent (14–17‐year‐old) athletes 1.4 times more likely to report a history of SRC than pre‐adolescent (11–13‐year‐old) athletes. Multi‐sport participation and immediate family participation were found to positively predict SRC incidence. No psychological measures were identified as significant correlates to concussive injury, however, this may be due to the homogeneity of the sample scores.

Conclusions

The concurrent assessment of biopsychosocial factors associated with SRC presents the opportunity for a comprehensive analysis of the injury context. The findings from this study suggest that SRC in junior rugby cannot be predicted using individual variables. Future research directions are discussed.

Key words:

• biopsychosocial
• concussion
• individual differences
• regression analysis
• sports injuries
• youth sports

WHAT IS ALREADY KNOWN ABOUT THIS TOPIC

• The current literature exploring the aetiology of SRC has suggested that concussion incidence may be modulated by a range of biological, psychological, and social factors.

• Previous research in junior SRC has been limited by confounding variables and minimal consideration for contextual factors.

• Sport‐related concussion is a multi‐faceted issue; thus objective assessment is needed to ensure balanced prevention of sport‐related concussion, without losing sight of the many benefits of sport involvement.

WHAT THIS TOPIC ADDS

• Correlates to concussion incidence was largely unpredictable in junior (11–17-years old) rugby union players.

• The absence of a clear predictive model, despite the concurrent assessment of the impact of a comprehensive spread of biopsychosocial risk factors, suggests that SRC in junior rugby cannot be predicted using individual variables.

• The findings from this study offer a shift in focus towards improving SRC management overall, rather than simply focusing on its prevention at the individual level.

INTRODUCTION

Participating in any contact sport comes with an obvious contextual risk of injury. Previous research has estimated that in Australia, sport‐related concussion (SRC) is the third highest cause of mTBI hospitalisations, after accidental trips/falls and transport‐related injuries (Helps, Henley, & Harrison, 2008). In the context of community‐based amateur rugby union (rugby) concussion has been found to account for 10% of all sport‐related injuries (Chalmers, Samaranayaka, Gulliver, & McNoe, 2012), and up to 10% of players sustain at least one concussion per season (which equates to almost eight concussions per 1,000 playing hours; Hollis et al. (2009)). However given the complex, often dynamic nature of sport‐related injuries in fast‐paced contact sports such as rugby, concussions often go undiagnosed, underreported, and under treated (Gardner, Iverson, Huw Williams, Baker, & Stanwell, 2014), all of which complicate the calculation of an accurate incidence rate of concussion. Past literature has suggested possible links between individual factors and SRC incidence, however various methodological discrepancies (e.g., oversight of comorbidities and confounding factors; focus on adult/elite samples) have led to a somewhat scattered understanding of the possible correlates to SRC in junior sport.

The current literature exploring the aetiology of SRC has suggested that concussion incidence may be modulated by a range of factors, including an athlete's age (Gessel, Fields, Collins, Dick, & Comstock, 2007; Hollis et al., 2009); physical composition (e.g., physical fitness, musculature, prenatal testosterone exposure) (Buzzini & Guskiewicz, 2006; Hönekopp, Manning, & Müller, 2006; Kempel et al., 2005); aggressive or impulsive playing style (Castanier, Le Scanff, & Woodman, 2010; Osborn, Blanton, & Schwebel, 2009); and premorbid psychological functioning (Grady, 2010). The impact of each of these intrinsic variables on concussion incidence and possibly experiencing ongoing post‐concussion complications remains inconclusive (Abrahams, McFie, Patricios, Posthumus, & September, 2014). Sociocultural factors, such as social modelling and sporting culture, should also be considered in terms of their potential impact on concussion attitudes, knowledge, and behaviour (Register‐Mihalik et al., 2013). In an environment known for “playing tough,” as is Australian community rugby union (rugby), it is important to consider that social modelling may have on a potentially negative impact on playing behaviours, as well as a player's attitudes towards concussion (e.g., not taking the injury seriously; Kroshus, Garnett, Hawrilenko, Baugh, and Calzo (2015)).

Although previous literature has identified potential risk factors to SRC, these have not been looked at concurrently, thus are often limited by confounding factors and minimal consideration for the full injury context. SRC is a multi‐faceted issue and should be considered in context, with consideration for the biological, psychological, and sociocultural (i.e., biopsychosocial) factors which influence the incidence of injury in contact sport.

AIMS AND DESIGN

Through the concurrent assessment of individual variables within a multi‐measure cross‐sectional research design, the current study aimed to contextualise the individual factors associated with SRC in junior rugby (ages 11–17-years). Potential correlates to be measured in this study were selected based on risk factors identified in previous literature (see Abrahams et al., 2014 for review). These included biological variables (e.g., body mass index, testosterone exposure, aerobic fitness), psychological variables (e.g., executive function, mood/emotion), and social variables (e.g., family involvement, sporting engagement). Self‐ and parent‐report questionnaires were utilised in this study to reduce reporting bias. Regression analyses were undertaken using IBM Statistical Package for Social Sciences (SPSS), version 22.0 (SPSS, 2013), to evaluate the predictive value of a biopsychosocial model of SRC correlates. It was hypothesised that by using a biopsychosocial approach and concurrently investigating potential correlates, a predictive model of SRC risk could be developed.

METHODS

Participants were male junior (11–17–year‐old) rugby players from community clubs throughout regional southeast Queensland, Australia (M = 13.0, SD = 1.8). This cross‐sectional cohort study consisted of 259 participants who were enrolled and regularly playing in age‐defined teams ranging from under 11‐ to under 17‐years during the 2015 rugby season (Feb–Sept). Eight of the 11 local clubs agreed to participate in the present study, within which there were 33 eligible teams registered.

Items included in the questionnaire battery were selected to measure potential biopsychosocial correlates that were identified in previous literature. Specifically, athletes' demographic factors, concussion history, social and cognitive development, mood disorders, executive function, and biological and physical attributes.

Demographic information was collected using both self‐report and parent‐report questionnaires. Items included the athlete's age, level of schooling, average grades, involvement in other sports, participation history, training hours, and immediate‐family involvement (i.e., whether another immediate family member [e.g., sibling/parent] also actively participated in rugby). Inter‐rater reliability was assessed for self‐ and parental‐report data. No significantly divergent responses occurred in this sample, with most responses being strongly correlated between respondents (r ranging from .84 to .91, p < .01). Parents were also asked about their child's psychological and behavioural health (e.g., learning difficulties, ADHD) and whether they took regular medication.

Details of any previous concussions were asked in both the athlete‐ and parent‐report surveys. Questions included; how many times the athlete has sustained a concussion (overall and in the past 12-months); who diagnosed the concussion; which symptoms were experienced; and how long they were removed from activity.

Social‐cognitive development was assessed using the self‐report Strengths and Difficulties Questionnaire (SDQ) (Goodman, Meltzer, & Bailey, 2003). The SDQ is a brief screening measure for children and adolescents in clinical and non‐clinical samples for behavioural and emotional problems on five 5‐item sub‐scales; conduct problems, hyperactivity, emotional problems, peer problems, and pro‐social behaviour. The Revised Children's Anxiety and Depression Scale (RCADS) (Chorpita, Yim, Moffitt, Umemoto, & Francis, 2000) was used to evaluate participants' emotional functioning, which corresponds to selected diagnostic criteria for anxiety disorders and depression. The RCADS measure includes sub‐scales for separation anxiety, social phobia, generalised anxiety, panic disorder, obsessive compulsive disorder, and major depressive disorder. Attentional, motor, and non‐planning impulsiveness was assessed using the self‐report, brief version of Barratt's Impulsiveness Scale (BIS‐Brief) (Steinberg, Sharp, Stanford, & Tharp, 2012).

Executive function was measured using the parent report version of the Behaviour Rating Inventory of Executive Function (BRIEF‐PR) (Gioia, Isquith, Guy, & Kenworthy, 2000). This questionnaire evaluated the emotional, behavioural, and metacognitive skills (broadly described as executive abilities).

Height and weight measures were taken for each athlete to calculate their body mass index (BMI; weight in kilograms divided by squared metres of height (NHMRC, 2013)). Height measurements were taken using a standard; a digital weight scale was used to weigh each participant. Height and weight measures were both recorded to two decimal points. Participants' aerobic fitness was assessed using a multi‐stage 20-m shuttle run (i.e., Beep Test), which is widely used and valid measure of VO₂ max (Leger, Mercier, Gadoury, & Lambert, 1988). Each participant's scores were compared against age‐banded national averages (NHMRC, 2013; Olds, Tomkinson, Léger, & Cazorla, 2006).

A correlate of prenatal testosterone is the ratio measurements of the second (index finger) and the forth (ring finger) digits (2D:4D) (Hönekopp et al., 2006). The 2D:4D measure has been found to be associated with many factors that may influence an athletes risk of concussion, including visual–spatial ability (Kempel et al., 2005; Manning & Taylor, 2001); physical fitness (Hönekopp et al., 2006); risk taking behaviour and aggression (Bailey & Hurd, 2005; Hönekopp, 2011); cognitive ability and problem solving (Austin, Manning, McInroy, & Mathews, 2002); and perceived dominance and masculinity (Neave, Laing, Fink, & Manning, 2003). To capture participant's 2D:4D ratio in the present study, a high‐resolution portable flatbed scanner was used to take digital scans of both the left and right hands of each participant. Digital callipers were used to measure the length of the second and forth digit (i.e., from the tip of the finger to the crease proximal to the palm). These measurements were calculated as an average ratio, with lower ratio scores indicating greater in‐utero testosterone exposure.

Ethical approval for this study was granted by the Sunshine Coast University Human Research Ethics Committee. Participants were approached during their regular training sessions, where the aim of the study and participation details was explained. Athletes who wished to participate in the study were given an information pack to take home; this included an information sheet, consent form, and the demographic and psychological questionnaires (i.e., SDQ, RCADS, BRIEF‐PR, BIS‐Brief). Questionnaires completed by both the athlete and their parent/guardian were returned to the research team either in person or via post in the reply paid envelopes that were provided. Questionnaires that were at least 80% complete were included in analysis; respondents who did not return both the athlete and parent report questionnaires were excluded from the study (n = 2). Athletes who returned their consent forms had their biometrics measured (i.e., BMI, 2D:4D) and participated in the Beep Test the following week. Once both physical and questionnaire data had been collected it was de‐identified prior to analysis.

RESULTS

Athlete and parent reports of concussion were strongly correlated (r = .91, p < .01). Twenty‐three percent (n = 61) of the sample reported that they had sustained a medically diagnosed concussion at some stage in their lives while playing sport. On average, participants with a history of SRC had sustained a median of 1 concussion in their lifetime (range 1–5). Fifty‐nine percent were diagnosed by team medical aides (aka. “medics”); 25% were diagnosed by hospital/emergency room doctors, and 15% were diagnosed by family doctors/general practitioners). Over half of this sample (62%; n = 38) reported sustaining a SRC within the last 12-months; 53% (n = 20) were within the past 6 months.

Biopsychosocial model

The biopsychosocial model comprises a large array of variables in this study, including athletes' age, BMI, aerobic fitness, prenatal testosterone ratio, impulsivity, behavioural and emotional functioning, depression and anxiety, executive functioning, duration of participation in rugby, their involvement with other rugby teams as well as in other sports, and whether other members of their immediate family also participate in rugby. To assess the predictive value of the entire model a logistic regression was performed. This model explained 22.7% of the variance in sport‐related concussion risk (Nagelkerke R2 = .227). The Hosmer and Lemeshow was non‐significant indicating goodness of fit and hence no significant discrepancy between observed and predicted concussion risk (χ (8) = 15.0, p = .06). Overall the model was non‐significant (χ (17) = 25.9, p = .08). Assessing all the elements of the entire model at once means that the number of included variables in the model was excessive given the sample size and hence assessment of the model's significance was affected by restricted power. In order to more fully understand the most important elements of each domain of the model (i.e., biological, psychological, sociocultural) it was decided to run separate logistic regressions for each domain separately to circumvent these power issues.

Biological factors

To allow participant demographics to be compared against national norms, participant data was categorised into two age groups, pre‐adolescent (11–13-years) and adolescent (14–17-years). Controlling for age, there were no statistically significant difference in biological factors between participants with a history of SRC and those without. Demographic statistics for each age group are shown in Table 1.

Table 1. Independent samples t test for biological factor differences between athletes with a history of SRC and no history of SRC

	Population mean (SD)	SRC history^1 M (SD)	No history M (SD)	t	df	p	95% confidence interval of the difference
							Lower	Upper
Pre‐adolescents	n = 159	n = 31	n = 128
Height (cm)	153.9 (10.8)	155.2 (8.6)	153.7 (11.3)	1.1	157	0.27	−1.7	6.2
Weight (kg)	46.9 (10.3)	47.7 (8.3)	46.8 (10.9)	0.5	157	0.65	−2.9	4.7
BMI^2	19.6 (3.0)	19.4 (2.8)	19.6 (3.0)	−0.7	157	0.52	−1.4	0.7
Beep score^3	7.7 (2.0)	7.5 (1.8)	7.8 (2.0)	0.3	157	0.79	−0.6	0.8
2D:4D^4	.93 (.11)	.95 (.03)	.93 (.13)	−1.1	157	0.29	−0.1	0.0
Adolescents	n = 100	n = 30	n = 70
Height (cm)	173.2 (7.5)	173.0 (8.4)	173.2 (7.2)	−0.6	98	0.53	−4.5	2.3
Weight (kg)	69.4 (15.6)	72.1 (14.6)	68.4 (15.9)	0.1	98	0.92	−6.7	7.4
BMI	23.0 (4.0)	24.0 (4.0)	22.6 (4.0)	0.8	98	0.44	−1.2	2.6
Beep score	9.1 (2.1)	9.1 (2.3)	9.1 (2.0)	1.4	98	0.16	−0.3	1.6
2D:4D	.95 (.11)	.92 (.19)	.97 (.03)	−1.7	98	0.09	0.0	0.0

Only concussions sustained while playing sport, as diagnosed by a medical professional, included in this group.

BMI scores between 15 and 22 for pre‐adolescents and 16–24 for adolescents are considered “average” (NHMRC, 2013).

Beep test scores between 6.5 and 8.8 for pre‐adolescents and 8.3–11.3 for adolescents are considered “average.”

2D:4D ratios between .87 and 1.0 considered average for Australian males 2D:4D (Loehlin, McFadden, Medland, & Martin, 2006).

A logistic regression was conducted to examine the interaction between biological variables (i.e., age, BMI, aerobic fitness, and 2D:4D) and the incidence of concussion (history of SRC versus no history of SRC). In combination, the model for athletes' biological variables explained 14% of variance in SRC incidence (Nagelkerke R² = .141). The Hosmer and Lemeshow was non‐significant indicating goodness of fit and hence no significant discrepancy between observed and predicted SRC incidence (χ (8) = 2.8, p = .95). Overall the model was significant (χ (4) = 19.7, p = <.01). This model had a stronger negative predictive value (i.e., this model more strongly predicted the absence of SRC), achieving an overall percentage accuracy of 80.7%. Increasing age was the only biological variable to significantly contribute to the model, with adolescent athletes 1.35 times more likely to report SRC than pre‐adolescent athletes (ExpB [1.01, 1.81] = 1.35, SE = .15, Wald = 3.96, p = .04).

Psychological factors

Mean participant scores for all psychological measures were within normal ranges (see Table 2). Means for the BIS‐Brief followed a normal distribution (skewness = .18; kurtosis = −.27). SDQ sub‐scale means were mostly within the “normal” range, except for “Conductivity Difficulties,” which was slightly higher than the national average for this age group (M = 2.0–2.4, SD = 1.9) and Emotional Difficulties and Hyperactivity, which were both lower than the national averages (M = 2.0–2.1, SD = 2.0, and M = 3.2–4.0, SD = 2.4, respectively) (Kremer et al., 2015). All of the RCADS and BRIEF‐PR means fell within one standard deviation for all sub‐scales. There was no statistically significant difference in psychological factors between participants with a history of SRC and those without (see Table 2).

Table 2. Independent samples t test for psychological factor differences between athletes with a history of SRC and no history of SRC

	Population M (SD)	SRC history^5 M (SD)	No history M (SD)	t	df	p	95% confidence interval of the difference
							Lower	Upper
SDQ
Emotional	1.2^6 (1.3)	1.1 (1.1)	1.2 (1.4)	−.8	257	.40	−.5	.2
Conduct difficulties	4.5^7 (2.1)	4.9 (1.9)	4.4 (2.2)	1.9	257	.06	.0	1.1
Hyperactivity	2.1^8 (1.7)	1.8 (1.4)	2.2 (1.9)	−1.6	257	.11	−.8	.1
Peer problems	1.9 (1.6)	1.9 (1.7)	1.9 (1.6)	.2	257	.85	−.4	.5
Total difficulties	9.6 (4.7)	9.7 (3.9)	9.6 (5.1)	.2	257	.87	−1.1	1.3
Prosocial behaviours	7.7 (1.7)	7.8 (1.6)	7.7 (1.8)	.3	257	.78	−.4	.5
RCADS
General anxiety	41.2 (10.0)	40.7 (9.1)	41.4 (10.4)	−.5	254	.60	−3.3	1.9
Panic disorder	46.7(9.9)	45.4 (9.8)	47.4 (9.9)	−1.5	254	.14	−4.6	.6
Social phobia	43.7 (10.5)	43.4 (10.9)	43.9 (10.3)	−.3	254	.75	−3.2	2.3
Obsessive compulsive disorder	42.7 (9.3)	41.8 (9.1)	43.1 (9.4)	−1.1	254	.27	−3.8	1.1
Major depressive disorder	44.2 (10.6)	43.8 (10.3)	44.5 (10.8)	−.5	254	.63	−3.4	2.1
Total anxiety score	42.5 (10.3)	41.7 (8.6)	42.9 (11.1)	−.9	254	.39	−3.9	1.5
BRIEF‐PR
Behaviour regulation index	50.3 (10.1)	49.1 (11.0)	50.6 (10.5)	−1.0	242	.31	−4.4	1.4
Meta‐cognition index	52.3 (10.4)	52.5 (9.6)	52.1 (10.9)	.3	242	.79	−2.4	3.2
Global executive function	51.7 (10.6)	51.0 (8.9)	51.9 (11.4)	−.6	242	.52	−3.8	1.9
BIS‐brief	16.6 (3.8)	17.1 (3.5)	16.5 (3.7)	1.3	257	.19	−.3	1.6

Only concussions sustained while playing sport, as diagnosed by a medical professional, included in this group.

Population and sub‐group means in this scale were lower than the national average for males in this age group (2.0–2.1, SD = 2.0).

Population and sub‐group means in this scale were higher than the national average for males in this age group (2.0–2.4, SD = 1.9).

Population and sub‐group means in this scale were lower than the national average for males in this age group (3.2–4.0, SD = 2.4). See Kremer et al. (2015) for national norms.

Pre‐existing diagnoses were reported by 17.4% (n = 45) of participants (e.g., ADHD, learning difficulties, anxiety, epilepsy, diabetes); 31.1% (n = 14) of these participants required regular medication (e.g., Concerta, Ritalin). A logistic regression was conducted to examine the interaction between participants' psychological variables (i.e., impulsivity, behavioural and emotional functioning, depression and anxiety, and executive functioning) and the incidence of concussion. Model analysis of the measure totals explained 14.6% of variance (R² = .146), predicting 81.4% of concussion incidence. Hosmer and Lemeshow was non‐significant indicating goodness of fit and hence no significant discrepancy between observed and predicted concussion incidence (χ (8) = 12.3, p = .14). Overall the model was non‐significant (χ (23) = 22.1, p = .51), with none of the measure totals nor the sub‐scale totals significantly contributing to the model (p = > .05).

Social factors

Athletes who played rugby for more than one team (n = 109, 42%), played for either their school team (21%) or the region's representative team (20%). The majority of the participants (78%; n = 200) also participated in other forms of sport, namely other codes of football (e.g., rugby league, Australian Football League, soccer; 27%), or another team ball sport (e.g., basketball, volleyball, cricket; 17%). Over half (55%) of the participants had an immediate family member who also participated in rugby, such as a sibling (84%) or a parent (12%).

A logistic regression was conducted to examine the interaction between the incidence of SRC and athletes' sociocultural variables (i.e., duration of participation, their involvement with other rugby teams as well as in other sports, and whether other members of their immediate family also participate in contact team sport). This model explained 11% of the variance in concussion incidence (R² = .11; χ (5) = 19.5, p = <.01), and correctly predicted 78.9% of reported SRCs. Hosmer and Lemeshow test was non‐significant, indicating this model had goodness of fit (r ² (8) = 3.09, p = .92). The sociocultural variables that significantly contributed to this model were duration of participation in rugby (ExpB [1.00, 1.02] = 1.01, SE = .01, Wald = 5.15, p = .02); participation in other teams (ExpB [.25, .88] = .46, SE = .32, Wald = 5.96, p = .02); and immediate relation to a family member who also plays rugby (ExpB [1.02, 1.82] = .96, SE = .32, Wald = .02, p = .04). Analysis of variance (ANOVA) showed that SRC incidence was significantly higher for athletes who had played rugby for longer (F (2) = 8.28, p = <.01) and who played for more than one rugby team (F (2) = 8.38, p = <.01).

DISCUSSION

The current study aimed to concurrently investigating the biopsychosocial factors associated with SRC in junior male rugby union. Despite the investigation of a large inclusion of evidence‐based factors, the findings from this study supported the null hypothesis, indicating that individual variation in SRC incidence was largely unpredictable in junior rugby union players (11–17‐year‐old). Nevertheless, components of the investigation did elicit interesting results. For example, athlete age was positively correlated with concussion incidence, which is in support of previous literature (Gessel et al., 2007; Hollis et al., 2009; Makdissi et al., 2013). Although this may be simply due to increased exposure (i.e., they have been playing longer), consideration should also be given to athletes' physical characteristics. That is, increased concussion incidence in older athletes may, in part, be due to increases in athletes' physical size, fitness, and game speed associated with development (Makdissi et al., 2013). As shown in the present study, older athletes more frequently reported a history of SRC, and had substantially larger physical characteristics compared to their pre‐adolescent counterparts. Although the mean BMI of 23.0 for adolescent athletes was considered within “average” range of the national norms for adolescents, it was 3.4 points above the mean BMI of the pre‐adolescent players (i.e., BMI = 19.6). Contact sports, including rugby, are typically played in age‐banded teams, however, puberty and the physical size‐related changes it incurs develop over time. Consequently, young athletes at different stages of development may be playing within the same age‐banded team. This is illustrated in the current study in the substantial standard deviations for mean height and weight within the pre‐adolescent (SD of nearly 11 cm and 10-kg) and adolescent groups (SD of nearly 8-cm and 16-kg). Additionally, the mean “beep” score for adolescents was 1.4 points above the mean for the younger age group, which represents a one kilometre per hour increase in speed. In combination, these physical differences suggest that adolescent athletes are not only bigger and heavier than younger athletes, but they are also faster, presenting a heightened risk of injury in fast‐paced contact sports due to their increased momentum and impact potential. This was supported in the findings of the present study as adolescent athletes were 1.4 times more likely to report a history of SRC than pre‐adolescent athletes.

The finding that none of the psychological factors measured correlated to SRC incidence is also interesting as it largely contradicts the majority of previous research (O'Jile, Ryan, Parks‐Levy, Betz, & Gouvier, 2004; Wiese‐Bjornstal, 2010). While this may be due to the homogeneity within the participants (i.e., the mean scores on all psychological measures were within “normal” ranges), it may also be due to notable methodological issues in this area of research. In a critical review, Junge (2000) highlighted that existing research concerning psychological factors and sport‐related injury employs heterogeneous designs, different evaluation strategies, and non‐discrete samples. This, along with a strong focus on treatment of SRC rather than the prevention of the injury, may explain the inconsistencies in research and variable understanding of how psychological functioning impacts on SRC incidence in junior athletes.

The finding that duration of participation was a positive predictor of incidence of reported SRC is consistent with previous research (Hollis et al., 2009; Quarrie et al., 2001; Turbeville, Cowan, Owen, Asal, & Anderson, 2003). It is difficult, however, to discern whether this is due to environmental or sociocultural exposure. It can be suggested that it may well be a combination of both extra playing exposure as well as sociocultural impact, although future research should consider evaluating this in more detail. The present study also found that multi‐team participation also increased an athlete's incidence of reporting a concussion. Over 40% of participants played for either their school or the regional representative team in addition to their club team. On one hand, it can be proposed that the greater skill development through more frequent and structured training may have a protective effect against sustaining a concussive injury; on the other, the higher level of competition requires athletes to be more competitive and aggressive, which may lead them to engage in more risky game play (Gessel et al., 2007; Hollis et al., 2009).

The psychosocial impact of acculturation on incidence of reported SRC has not been widely researched, however some studies have demonstrated that athletes feel a pseudo‐cultural pressure to “play tough” and downplay injury based on the norms learnt through socialisation with their teammates and modelling from authority figures (Kerr, 2014; McCrea, Hammeke, Olsen, Leo, & Guskiewicz, 2004; Schinke, Hanrahan, & Catina, 2009; Wiese‐Bjornstal, 2010). Previous literature has identified the potential impact of this kind of sporting culture on reporting behaviours and treatment adherence (Clacy, Goode, Sharman, Lovell, & Salmon, 2016; Clacy, Goode, Sharman, Lovell, & Salmon, 2017; Hollis, Stevenson, McIntosh, Shores, & Finch, 2012). In the present study, immediate family member participation along with duration of sport participation did impact reported SRC incidence, which may suggest some sociocultural modelling effect in the incidence of SRC in junior athletes. This interaction needs to be explored more directly in future research.

LIMITATIONS

The findings of this study were limited by the relatively small sample size and the uneven number of concussed versus non‐concussed participants (n = 61 and 198, respectively), which may have restricted the effect size. The largely homogenous characteristics of the sample may also have limited the results of this study. Future research should aim to broaden the sampling criteria to include other team sports and should also consider potential gender differences. With the notable growth in women's sport, including contact sport, there is a growing need to better understand potential gender differences in concussion risk to ensure management practices reflect the needs of the female sport context. As with many studies that aim to investigate injuries in sport, the responses of the current sample may not include the characteristics of players who withdrew from sport participation due to injury. Future investigations of sport injury risk factors should aim to include this population. It would also be of benefit to include more objective factors into research, such as the neurobiology of concussion recovery and how it impacts cognitive, emotional, and physical health following injury. Finally, the cross‐sectional retrospective design of this study makes discerning direct causal links between biopsychosocial factors and SRC incidence impossible. Future research would benefit from a prospective design that evaluates athletes biopsychosocial factors pre‐season, post‐injury, and at follow‐up intervals. This study design would address many of the limitations highlighted by Junge (2000) and would offer insight into the potential consequences of SRC and how management practices might be adapted to reduce SRC‐related complications and support healthy recovery.

CONCLUSION AND FUTURE DIRECTIONS

Overall, the findings of this study failed to clearly identify a profile for “at‐risk” athletes, instead results suggest that concussion in junior rugby is largely unpredictable based on the concurrent analysis of a number of biopsychosocial variables. The absence of a clear predictive model, despite the concurrent assessment of the impact of a comprehensive spread of biopsychosocial risk factors, illustrates the dynamic nature of concussion incidence in junior sport. Although the biopsychosocial model used in this study was not able to identify a predictive model of SRC in junior athletes, it does encourage a shift in focus towards improving SRC management, rather than simply focussing on its prevention. That is, while it may not be possible to predict SRC risk, the findings from this study should encourage policy makers and researchers to consider the wider factors that may impact SRC management and recovery. SRC may well be an inevitable aspect of team sport participation, thus management guidelines need to consider individual differences as well as context‐specific limitations. Concussion management practices therefore need to reflect the resources and limitations of the junior sporting context. It should also be noted that the sample of youth athletes in this study were largely homogenous in both their physical and psychological characteristics (i.e., the vast majority of participants were physical and mentally healthy), which may reflect some of the positive benefits of sport participation during adolescence. This offers an important reminder to consider both the positive and negative consequences of sport participation.

CONFLICT OF INTEREST

The authors declare no potential conflict of interest.