Introduction
Mild traumatic brain injury (mTBI or concussion) is especially prevalent in children aged 5 years and under (Dewan et al., Reference Dewan, Mummareddy, Wellons III and Bonfield2016; Taylor et al., Reference Taylor, Bell, Breiding and Xu2017). Compared to the substantial research in middle childhood and adolescence, few studies have focused on outcome after early childhood mTBI despite the potential for such injuries to disrupt functioning (Séguin et al., Reference Séguin, Gagner, Tuerk, Lacombe Barrios, MacKay and Beauchamp2022), and the importance of ensuring a sound cognitive, social and behavioral basis for lifespan development.
MTBI symptoms and outcomes
Symptoms of pediatric mTBI typically manifest through affective (e.g., anxiety), cognitive (e.g., poor concentration), or physical (e.g., vomiting, headaches) signs and post-concussive symptoms (PCS) (Ewing-Cobbs et al., Reference Ewing-Cobbs, Cox, Clark, Holubkov and Keenan2018; Todd et al., Reference Todd, Rebekah, Ken, Keith Owen, Gurinder, Emma, Darcy, Alexander, Isabelle, Jocelyn, Stephen, William, Kathy, Martin, Gerard and Roger2022; Zemek et al., Reference Zemek, Barrowman, Freedman, Gravel, Gagnon, McGahern, Aglipay, Sangha, Boutis, Beer, Craig, Burns, Farion, Mikrogianakis, Barlow, Dubrovsky, Meeuwisse, Gioia and Meehan III2016). While similar effects have been documented in young children (Podolak et al., Reference Podolak, Chaudhary, Haarbauer-Krupa, Metzger, Curry, Kessler, Pfeiffer, Breiding, Master and Arbogast2021), loss of consciousness, vomiting, drowsiness, and headache seem to be less common (Crowe et al., Reference Crowe, Rausa, Anderson, Borland, Kochar, Lyttle, Gilhotra, Dalziel, Oakley, Furyk, Neutze, Bressan, Davis and Babl2024) and PCS appear to be characterized primarily by behavioral manifestations such as increased moodiness, irritability and comfort seeking (Dupont et al., Reference Dupont, Beaudoin, Désiré, Tran, Gagnon and Beauchamp2021; Dupont et al., Reference Dupont, Tang, Beaudoin, Dégeilh, Gagnon, Yeates, Rose, Gravel, Burstein, Stang, Stanley, Zemek and Beauchamp2024; Suskauer et al., Reference Suskauer, Rane, Reesman and Slomine2018). Parent report can result in both over- and under-estimations of PCS (Stevens et al., Reference Stevens, Penprase, Kepros and Dunneback2010), and in young children, discrepant results concerning the extent and nature of PCS are likely to be associated with challenges in identifying mTBI due to their limited cognitive and verbal abilities (Beauchamp et al., Reference Beauchamp, Anderson, Ewing-Cobbs, Haarbauer-Krupa, McKinlay, Wade and Suskauer2024). There is some evidence for elevated behavior problems (Gagner et al., Reference Gagner, Dégeilh, Bernier and Beauchamp2020; Gornall et al., Reference Gornall, Takagi, Morawakage, Liu and Anderson2021) and social cognitive difficulties in the long-term after early mTBI (Bellerose et al., Reference Bellerose, Bernier, Beaudoin, Gravel and Beauchamp2015, Reference Bellerose, Bernier, Beaudoin, Gravel and Beauchamp2017; Séguin et al., Reference Séguin, Gagner, Tuerk, Lacombe Barrios, MacKay and Beauchamp2022).
Although disruptive, PCS are usually transient and about two thirds of school-aged (5–18 years) children recover by one-month post-injury (Zemek et al., Reference Zemek, Barrowman, Freedman, Gravel, Gagnon, McGahern, Aglipay, Sangha, Boutis, Beer, Craig, Burns, Farion, Mikrogianakis, Barlow, Dubrovsky, Meeuwisse, Gioia and Meehan III2016). Severe neuropsychological deficits are also not common (Beauchamp et al., Reference Beauchamp, Aglipay, Yeates, Désiré, Keightley, Anderson, Brooks, Barrowman, Gravel and Boutis2018). Given the reasonably encouraging prognosis, there is increasing focus on the large proportion of individuals who recover completely and rapidly after pediatric mTBI (e.g.; Beauchamp et al., Reference Beauchamp, Tang, Yeates, Anderson, Brooks, Keightley, Désiré, Boutis, Gagnon and Gravel2019). However, no study has explored this topic in children with early mTBI. Determining what predicts good outcome is useful for identifying targets to promote optimal functioning and well-being.
Optimal functioning and well-being after pediatric mTBI
Optimal outcome lacks a clear definition, and its operationalization varies across studies and disciplines. Well-being (or wellness) is defined by the World Health Organization (2023) as « a positive state experienced by individuals and societies », and is associated with related constructs such as quality of life, resilience, mental health, physical health, personal satisfaction, and healthy habits (Center for Disease Control and Prevention 2018). Hawley and Joseph (2008) suggested that well-being after TBI is characterized by an individual reaching their full potential (i.e., beyond their pre-morbid state). The current study draws on these varied conceptions and uses the broader term “optimal functioning” in reference to performance in spheres typically disrupted by pediatric mTBI, namely PCS, behavior, cognition, and quality of life (Gagner et al., Reference Gagner, Dégeilh, Bernier and Beauchamp2020; Moran et al., Reference Moran, Taylor, Rusin, Bangert, Dietrich, Nuss, Wright, Minich and Yeates2012; Séguin et al., Reference Séguin, Gagner, Tuerk, Lacombe Barrios, MacKay and Beauchamp2022; Zemek et al., Reference Zemek, Barrowman, Freedman, Gravel, Gagnon, McGahern, Aglipay, Sangha, Boutis, Beer, Craig, Burns, Farion, Mikrogianakis, Barlow, Dubrovsky, Meeuwisse, Gioia and Meehan III2016).
Only one study has specifically focused on the concept of optimal functioning (referred to as wellness in the article) and its associated factors after pediatric mTBI. Beauchamp et al. (Reference Beauchamp, Tang, Yeates, Anderson, Brooks, Keightley, Désiré, Boutis, Gagnon and Gravel2019) evaluated 311 children (6–18 years) one and three months post-mTBI. Optimal functioning was defined as the presence of good quality of life, no persisting PCS, and no neuropsychological difficulties. The proportion of children with optimal functioning increased post-injury, though not significantly, from 41.5% (1-month) to 52.2% (3 months). Participants were more likely to have optimal functioning if they were younger (6 – 8 years vs. 9 years or older at injury), were injured in a sports context (versus a fall or motor vehicle accident), had no history of pre-injury problems (e.g. neurodevelopmental or mood disorders), and had better acute working memory. While innovative in its approach focusing on positive outcomes, this work was limited by the absence of a comparison group. Furthermore, studies exploring optimal outcome in older individuals are informative (Hanks et al., Reference Hanks, Rapport, Waldron-Perrine and Millis2014; Vos et al., Reference Vos, Poritz, Ngan, Leon-Novelo and Sherer2019), but do not represent the unique aspects of early childhood mTBI, such as differential mechanisms of injury (mainly falls rather than sports injuries), and important influences at young ages, such as parental factors (Beauchamp et al., Reference Beauchamp, Séguin, Gagner, Lalonde and Bernier2021).
Child and parent factors
Family and relational factors, such as high-quality parent-child interactions and secure attachment, are associated with better cognitive and social development in typically developing children (Devine & Hughes, Reference Devine and Hughes2019; Madigan et al., Reference Madigan, Atkinson, Laurin and Benoit2013; Szpak & Białecka-Pikul, Reference Szpak and Białecka-Pikul2020). These factors also predict behavioral adjustment and social competence in young children who sustain complex to severe mTBI (Yeates et al., Reference Yeates, Taylor, Walz, Stancin and Wade2010). After early mTBI, parent-child interaction quality is reduced (Lalonde et al., Reference Lalonde, Bernier, Beaudoin, Gravel and Beauchamp2018) and parental stress is associated with worse child quality of life (Tuerk et al., Reference Tuerk, Gagner, Dégeilh, Bellerose, Lalonde, Landry-Roy, Séguin, de Beaumont, Gravel and Bernier2020), behavior problems (Yumul et al., Reference Yumul, McKinlay, Anderson and Catroppa2024), and more PCS (Bernard et al., Reference Bernard, Ponsford, McKinlay, McKenzie and Krieser2016).
Child temperament dimensions (Rothbart et al., Reference Rothbart, Ahadi, Hershey and Fisher2001) such as negative affectivity (the propension to experience unpleasant emotions such as sadness) are associated with more externalizing behavior problems (Delgado et al., Reference Delgado, Carrasco, González-Peña and Holgado-Tello2018). In adults, positive affectivity (the propension to experience pleasant emotions such as joy) is associated with post-TBI subjective outcomes (e.g. life satisfaction), while negative affectivity is associated with post-TBI objective outcomes (e.g. physical health; Hanks et al., Reference Hanks, Rapport, Waldron-Perrine and Millis2014). No such associations have been explored after early mTBI; however, elevated internalizing and externalizing behavior problems are commonly reported after pediatric mTBI (Brooks et al., Reference Brooks, Plourde, Beauchamp, Tang, Yeates, Keightley, Anderson, Désiré, Barrowman and Zemek2019; Gagner et al., Reference Gagner, Dégeilh, Bernier and Beauchamp2020; Gornall et al., Reference Gornall, Takagi, Morawakage, Liu and Anderson2021). Given the link between temperament and behavior (Delgado et al., Reference Delgado, Carrasco, González-Peña and Holgado-Tello2018), these factors could be associated with optimal functioning after early mTBI.
Objectives and hypotheses
No study has explored optimal functioning after early mTBI or included family factors as potential predictors, and previous work in older children is limited by the absence of a comparison group. This study aimed to (1) document the evolution of optimal functioning after early mTBI compared to orthopedic injury (OI) and typically developing (TDC) comparison groups at 6 and 18-months post-injury, and (2) identify predictors of early mTBI optimal functioning among sociodemographic (parent education, child age, sex), injury (number of acute symptoms), child (temperament), and parent (parenting stress, parent-child interaction quality, attachment) variables. It was expected that: (1) a lower proportion of children with early mTBI would have optimal functioning compared to both comparison groups over 6 and 18 months; (2) sociodemographic (higher parent education, lower child age, being a girl; Arambula et al., Reference Arambula, Reinl, El Demerdash, McCarthy and Robertson2019), injury (fewer acute symptoms), child (lower negative affectivity traits), and parent (less parenting stress, better parent-child interaction quality, secure attachment) factors would be associated with optimal functioning.
Methodology
Design
Participants were recruited between 2011 and 2015 as part of a longitudinal study investigating cognitive and social outcomes of early TBI (LION project). Participants were followed at 6, 18, 30, and 60 months post-TBI. The current study includes the 6 and 18-months timepoints, was approved by the Ste-Justine Hospital research ethics committee and was conducted in accordance with Helsinki Declaration.
Participants were 190 children from three groups: mTBI (n = 69), OI (n = 50), and TDC (n = 71). Inclusion criteria for mTBI, assessed by research staff based on presentation and medical charts were (1) presentation to a pediatric Emergency Department (ED) within 48 hours of injury; (2) age between 18 and 60 months at injury; (3) closed head injury with a Glasgow Coma Scale (GCS) score between 13 and 15; (4) the presence of at least one of the following: loss of consciousness, excessive irritability, persistent vomiting (two or more times), confusion, headache, fatigue, dizziness, motor or balance problems, blurred vision, hypersensitivity to light, and/or seizures. Children with complex mTBI (GCS 13 to 15 with intracranial lesion) were included (n = 9). For the OI, inclusion criteria were (1) presentation to the pediatric ED; (2) aged between 18 and 60 months at injury; (3) limb injury with a final diagnosis of simple fracture, sprain, contusion, or unspecified trauma. Inclusion criterion for the TDC was to be aged between 24 and 66 months (to be comparable to the injury groups six months later).
Exclusion criteria for all groups were: (1) diagnosis of a congenital, neurological, developmental, psychiatric, or metabolic condition; (2) birth before 36 weeks gestation; (3) parent or child cannot communicate in English or French; (4) history of a previous TBI requiring an ED visit; (5) non-accidental injury for children with mTBI or OI.
Procedure
Children with mTBI or OI were recruited at the ED by a research nurse or assistant. Families who agreed to participate were contacted within a week of their ED visit to complete the consent form and questionnaires. A case report form was completed by research staff to document acute signs and symptoms and injury characteristics. The TDC were recruited through flyers in daycares from diverse neighborhoods. Parents expressing interest were contacted by phone by a research assistant who verified inclusion and exclusion criteria. At 6 months (T1) and 18 months (T2) post-injury, parents from all groups were asked to complete questionnaires about their child’s functioning, and children participated in a direct assessment.
Measures
Sociodemographic and injury characteristics
Sociodemographic information was documented at the time of recruitment including child age, ethnicity, and sex. Maternal education was used as a proxy for socioeconomic status. Injury characteristics included: Glasgow Coma Score, mechanism, and acute symptoms: loss of consciousness, alteration of consciousness (e.g. confusion), post-traumatic amnesia, headache, excessive irritability, vomiting, hematoma, drowsiness, dizziness, convulsions, visual symptoms (e.g. blurry vision) and balance.
Definition of optimal functioning
Children were considered to have optimal functioning at T1 or T2 if they met four criteria: presence of (1) good quality of life, and absence of (2) persisting PCS, (3) clinically significant behavior problems, and (4) cognitive difficulties.
Quality of life: The Pediatric Quality of Life Inventory 4.0 (PEDSQL; Varni et al., Reference Varni, Burwinkle, Seid and Skarr2003; Varni et al., Reference Varni, Limbers and Burwinkle2007) parent report was used to document child quality of life (physical, mental, social, academic) according to caregivers using 23 questions on a five-point scale (total score on 100). Higher scores indicate better quality of life. Published norms (Varni et al., Reference Varni, Burwinkle, Seid and Skarr2003) were used such that a score of 65.4 or more (equal or higher than -1 SD from the population mean) was considered to represent good quality of life (as in Beauchamp et al., Reference Beauchamp, Tang, Yeates, Anderson, Brooks, Keightley, Désiré, Boutis, Gagnon and Gravel2019).
Post-concussion symptoms (PCS): Due to the absence of a validated PCS measure for early TBI at the time of data collection, the Postconcussive Symptom Interview (PCS-I; Mittenberg et al., Reference Mittenberg, Wittner and Miller1997) was used by parents to document 15 cognitive, physical, emotional, and sleep symptoms using yes/no answers. Persisting PCS were considered absent if children had fewer than three symptoms at a given timepoint (Zemek et al., Reference Zemek, Osmond and Barrowman2013).
Behavior : The Child Behavior Checklist (CBCL 1.5–5 years or 5–18 years; Achenbach & Rescorla, Reference Achenbach and Rescorla2000) consists of 100 questions completed by parents on a three-point scale. Standardized clinical cutoff thresholds were used such that T scores below 65 on both internalizing and externalizing subscales were required for behavior problems to be considered absent.
Cognition: In keeping with the goals of the broader study, cognition was measured using six tests of executive functioning and theory of mind. Since severe cognitive deficits are rare after pediatric mTBI (Ware et al., Reference Ware, McLarnon, Lapointe, Brooks, Bacevice, Bangert, Beauchamp, Bigler, Bjornson, Cohen, Craig, Doan, Freedman, Goodyear, Gravel, Mihalov, Minich, Taylor and Zemek2023), Beauchamp et al.’s (Reference Beauchamp, Aglipay, Yeates, Désiré, Keightley, Anderson, Brooks, Barrowman, Gravel and Boutis2018) rule for defining cognitive inefficiency was used. This method allows identification of subtle cognitive difficulties, rather than significant impairments. Participants scoring one standard deviation or more below average on two or more direct assessment measures were identified as having suboptimal cognitive functioning, and conversely, those not meeting this definition were considered to have no cognitive difficulties. Due to the wide age range of the sample, age was controlled by generating z scores for each task by age group (T1: 2, 3, 4 years old; T2: 3, 4, 5, 6 years old).
Conflict Scale: This flexibility task consists of four levels of six trials that represent different measures of conflicting executive functions. Younger children have to categorize items according to rules that change throughout the task (e.g. sorting by one color, then switching the rule to the shape instead of the color). Older children must sort the cards according to the presence (color game) or the absence (shape game) of a black border around the card. The rule is changed in a post-switch phase if the child succeeds on five trials (Beck et al., Reference Beck, Schaefer, Pang and Carlson2011; Zelazo, Reference Zelazo2006). There are 12 trials per level for a maximum of 48 points.
Spin the Pots: This working memory task requires children to find 6–10 stickers placed separately in 8–12 boxes on a tray depending on their age. After each trial, the boxes are covered with a cloth, the tray is rotated 180 degrees, and the child has to find one sticker (Beck et al., Reference Beck, Schaefer, Pang and Carlson2011). The final score is the number of stickers found divided by the number of rotations required.
Delay of Gratification: This inhibition task requires children to make a series of decisions during which they must choose between a small immediate reward or a larger reward received later (e.g. one sticker now or five stickers after the task; Beck et al., Reference Beck, Schaefer, Pang and Carlson2011). One point is given for every larger reward chosen for a total of nine points.
Shape Stroop: This inhibition and flexibility task first consists of showing images of bigger and smaller fruits. Children are asked to point to the image of a fruit for six trials. Then, three cards of big fruits with a smaller fruit in the middle are shown. Children are asked to point to the little fruits (e.g. point to the little apple) and not to the bigger ones. Children receive one point for every correct answer for a total of three points (Carlson, Reference Carlson2005; Kochanska et al., Reference Kochanska, Murray and Harlan2000).
Desires Comprehension Tasks: Children aged from 24 to 35 months performed a desires and emotion reasoning task in which they have to choose between two snacks, one typically liked by children (e.g., cookie) and one typically disliked (e.g., brocoli; Bellerose et al., Reference Bellerose, Bernier, Beaudoin, Gravel and Beauchamp2015; Repacholi & Gopnik, Reference Repacholi and Gopnik1997). The experimenter expresses their preference for the less liked food and asks the child to give them a food item to see if they will offer the experimenter’s preferred food. A total of four food combinations are presented for a total of four points, and this score was used in the analyses. For older children (> 35 months), a more advanced desires reasoning task (Desire task; Pears & Moses, Reference Pears and Moses2003) was used to document understanding of how fulfilled and unfulfilled desires can affect a character’s feelings through stories (Bellerose et al., Reference Bellerose, Bernier, Beaudoin, Gravel and Beauchamp2015; Pears & Moses, Reference Pears and Moses2003). The stories describe a character’s search for a desired object with three possible endings, each presented twice: (1) the character finds the desired object, (2) they find nothing, or (3) they find a different object. The child is asked to determine whether the character is happy or sad. Each possible ending is presented twice, for a total of six stories with a score out of six points, which was used in the analyses.
False Belief Understanding Task: Children are presented with a picture book that incorporates a deceptive element and are asked to recall their own initial belief about what they saw (e.g., children are made to believe that they see an eye through a peep hole, but they find out that it is a spot on a snake). Then, they must predict the belief of a puppet who never saw the book: 1) “what does the puppet think it is?” and 2) “what is it really?” (Bellerose et al., Reference Bellerose, Bernier, Beaudoin, Gravel and Beauchamp2015; Hughes et al., Reference Hughes, Ensor and Marks2011). A control question is also included: “What is it really, an eye or a snake?” For both scenarios, children receive a point only if they can answer the corresponding control question, for a maximum of two points and this variable was used in analyses.
Candidate Predictors
Sociodemographic and acute symptoms: Child age and sex, maternal education, and acute symptoms were included.
Mutually responsive orientation scale (MRO): This measure of parent-child interaction quality is based on the dyadic nature of parent-child exchanges (Aksan et al., Reference Aksan, Kochanska and Ortmann2006; Kochanska et al., Reference Kochanska, Aksan, Prisco and Adams2008). Two 10-minute sequences (snack and free play with toys) were videotaped. Trained research assistants coded the parent-child interactions in each video according to harmonious communication, mutual cooperation, and emotional ambiance, with a final MRO score of 15 (mean of the two interaction scores). Inter-rater reliability was satisfactory (r = 0.74 to 0.97, 17% of the sample).
Attachment Q-sort short version: This measure of attachment security consists of 30 cards representing child attachment behaviors towards a parental figure (Waters, Reference Waters1995). After viewing the same two video sequences used to score the MRO, a research assistant places the 30 cards in five rows of six cards. Each row represents a score from 1 (does not represent the child’s behavior) to 5 (represents the child’s behavior perfectly) such as: “child easily becomes angry at mother.” The research assistant sort is then correlated with a criterion sort provided by the authors of the instrument, representing the prototypically securely attached child. The prototypically securely attached child represents a fluid balance between reliance on caregiver when support is needed and exploration of the environment, with low scores on items such as “Child rarely asks mother for help,” and high scores on items such as “If mother reassures him, child will approach or play with things that initially made him cautious or afraid.” The final score ranges from -1.00 (very insecure) to 1.00 (very secure). Three research assistants (different from those who scored the MRO) coded attachment security and the inter-rater reliability score was excellent (r = 0.82; 20% of the sample).
Parenting Stress Index - Short Form: This parent questionnaire measures the level of distress experienced with regard to their parenting role with their child (e.g., their perceived competence) (Abidin, Reference Abidin1995). Each item was rated on a five-point scale and the total score from two 12-item subscales (Parenting Distress and Parent-Child Dysfunctional Interaction) was used. A higher score indicates higher stress.
Early Childhood Behavior Questionnaire (ECBQ; 18 to 36 months) or Childhood Behavior Questionnaire (CBQ; > 36 months) (Putnam et al., Reference Putnam, Gartstein and Rothbart2006; Rothbart et al., Reference Rothbart, Ahadi, Hershey and Fisher2001). The short 36-item versions based on Rothbart’s model (Putnam et al., Reference Putnam, Helbig, Gartstein, Rothbart and Leerkes2014; Putnam & Rothbart, Reference Putnam and Rothbart2006) were completed by parents. Each item is rated on a seven-point scale ranging from zero (extremely false) to seven (extremely true), yielding three dimensions (Reactivity, Negative Affectivity, Effortful Control). In preschool children, higher scores of negative affectivity have been linked to more behavior problems (Henderson & Wachs, Reference Henderson and Wachs2007), therefore this dimension was used in analyses.
Statistical analyses
Missing data
Only children who completed both assessment timepoints were included (N = 188). Some children had missing data on some of the measures administered at either timepoint. Multiple imputation was used to handle missing data (Enders, Reference Enders2010). Rates of missing data varied from 0 to 13%, far below the recommended maximum 50% for multiple imputation (Collins et al., Reference Collins, Schafer and Kam2001; Graham, Reference Graham2008). To correct for bias and maximize the precision of imputed data, demographic information and results on the CBQ, ECBQ and PSI from T2 were included in the imputation model (Enders, Reference Enders2010).The pattern of missing data was analyzed using Little’s MCAR test, which indicated that data were missing completely at random (χ2(1) = 2238.47, p = 0.126). Since Little’s test has low statistical power (Enders, Reference Enders2010), complete and incomplete cases (for variables with 5% or more of missing data) were also compared to investigate whether they differed on any sociodemographic variables or on the main outcomes. Children who had missing data on the Attachment Q-Sort (n = 25) were older at both timepoint and had higher scores on the Shape Stroop at T1 (all ts between -2.7 and -2.3, ps < 0.05). Children who had missing data on Spin the Pots at T1 (n = 15) were younger at both timepoint (all ts between 2.8 and 2.1, ps < 0.05). Children who had missing data on the Conflict Scale at T1 (n = 19) were younger at both timepoint and had higher scores on Spin the Pots at both timepoints and higher score on the Conflict Scale at T2 (all ts between 3 and -2.9, ps < 0.05). Children who had missing data on Delay of Gratification at T1 (n = 12) had lower Parenting Stress at T2 (t = 2.7, p = 0.019). Children who had missing data on the Desire Tasks at T1 (n = 17) were younger on both timepoint and had lower negative affectivity at T2 (all ts between 8.4 and 2.1, ps < 0.05). Children who had missing data on the Conflict Scale at T2 (n = 12) were younger at both timepoint and had lower scores on Spin the Pots at T2 (all ts between 6.0 and 2.9, ps < 0.05). Children who had missing data on the PSI at T1 (n = 12) had higher CBCL internalizing behaviors at T1 (t = -3.9, p = 0.003). Finally, children who had missing MRO data (n = 24) were older at both timepoints and had higher scores on Shape Stroop and higher Negative Affectivity at T1 (all ts between -2.6 and -2.1, ps < 0.05). Missing demographic information were not imputed. Maternal years of education was missing for two participants (mTBI = 1; TDC = 1).
Missing values were imputed using the Markov Chain Monte Carlo procedure in SPSS (Charles, Reference Charles1992). Twenty imputations were applied according to recommendations, and missing data estimated from all other data available (cognition, temperament, behavior, quality of life, attachment, parent-child interactions, parenting stress). Sociodemographic data were also included to maximize algorithm precision (Enders, Reference Enders2010; Graham, Reference Graham2008). Analyses were then run on each imputed data set and results were averaged (Schafer, Reference Schafer1997). Descriptive statistics were calculated to examine variable distributions.
Main analyses
Participants meeting all four criteria (no clinically significant behavioral problems; no cognitive difficulties; no persisting PCS; average quality of life or above) were considered to have optimal functioning. For the first objective, a Generalized Estimating Equation (GEE) was used to verify group and time differences between T1 and T2. For the second objective, a logistic regression with a Cox and Snell variance analysis was conducted for the mTBI group to see if predictors at T1: sociodemographic (maternal education, child sex, age); parent and child factors (parent-child interaction, attachment, parenting stress, child negative affectivity) and injury factors (number of acute symptoms), predict optimal functioning at T2. Main analyses were performed with and without participants with complex mTBI.
Results
Participant characteristics and information on imputed data are presented in Table 1.
Table 1. Participant sociodemographic characteristics, results on markers of optimal functioning, and predictor variables
Note: Statistics presented include imputed data.
PEDSQL = Pediatric Quality of Life Inventory; CBCL = Child Behavior Checklist; PCS-I = Post-Concussive Symptoms Interview; PSI = Parenting Stress Index; ECBQ = Early Childhood Behavior Checklist; CBQ = Childhood Behavior Checklist; ED = Emergency Department.
Mechanisms of injury were mainly falls (88.40%). GCS was homogenous and high (89.90% = 15). No significant differences were found between the three groups on any sociodemographic variable, but maternal years of education (Table 2) was on the cusp (X 2 (12, N = 188) = 20.918, p = 0.052).
Table 2. Maternal years of education
Two mothers had missing data on years of education (TBI = 1; TDC = 1). CEGEP = College of General and Professional Teaching. In Quebec, Canada, students attend CEGEP after high school for 2 (preparation for university) or 3 (preparation for job market) years.
Correlations between the main study variables are presented in Table 3. Descriptive statistics for the percentage and number of participants meeting optimal functioning criteria are presented in Table 4.
Table 3. Correlations matrix including predictors, sociodemographic variables, and optimal functioning for all groups
*p < 0.05; **p < 0.01.
Table 4. Percentage of participants meeting optimal functioning criteria
The results of the GEE analysis (Figure 1) showed a significant group effect (p < .001) for the proportion of children with optimal functioning. Significantly fewer children with mTBI had optimal functioning compared to OI (mean difference = −0.158, SE = 0.066, p = 0.008, 95% Wald interval between -0.290 and −0.029) and TDC (mean difference = −0.241, SE = 0.060, p < 0.001, 95% Wald interval between −0.358 and −0.121). There was no significant group difference between the TDC and OI groups with a mean difference of 0.082 (SE = 0.064, p = 0.106). No effect of time (p = 0.711) or group × time interaction (p = 0.771) was found. When children (n = 9) with complex mTBI were removed, the results were similar (Table 6; supplementary material).
Figure 1. Evolution of the proportion of children with optimal functioning. This figure presents the percentage of children meeting all four optimal functioning criteria (Y axis) in each group for both timepoints (X axis). Each symbol represents a group: the triangle for the typically developing children, the square for children with orthopedic injury, and the diamond for children with mild traumatic brain injury. Lines within the same oval indicate no significant difference, while lines not within the same oval indicate a significant difference. **ps < 0.001.
The binary logistic regression model (Table 5) was significant (X 2 (7, N = 68) = 15.387, p = 0.043), explaining 20.20% of the variance. Higher quality of parent-child interactions (B = 0.480, p = 0.025) and lower child negative affectivity temperament (B = −0.615, p = 0.045) independently predicted optimal recovery at T2. The model remained significant when children with complex mTBI were removed (X 2 (7, N = 60) = 15.222, p = 0.048, R 2 = 0.245). Higher quality of parent-child interactions was still an independent predictor (B = 0.516, p = 0.032), but negative affectivity temperament was no longer significant (B = -0.601, p = 0.089).
Table 5. Regression analysis for mild traumatic brain injury group to predict optimal functioning at T2 (n = 68)
Group effect is significant (p = 0.001).
Discussion
This study aimed to document group differences and factors associated with optimal functioning after early mTBI to identify potential targets that could promote positive outcome. The findings indicate that fewer children with mTBI have optimal functioning compared to OI and TDC at 6 (mTBI = 47.8%; OI = 66.0%; TDC = 76.0%) and 18 (mTBI = 53.6%; OI = 68.0%; TDC = 73.2%) months post-injury. Beauchamp and colleagues (Reference Beauchamp, Tang, Yeates, Anderson, Brooks, Keightley, Désiré, Boutis, Gagnon and Gravel2019) used three similar criteria (absence of PCS and cognitive inefficiency, good quality of life) and found that 52.2% of older children (6–18 years) were classified as having optimal functioning (“being well”) three months after their injury. A similar proportion was found in the present study 18-months post-injury (53.6%) and contrasted with the proportion found in typically developing children (73.2%) and children with OI (68.0%). While not directly comparable, these results suggest that younger children may have a more protracted pattern of long-term recovery. However, no data were collected between the ED visit and T1, therefore it is unclear what the rate of children meeting all four criteria of optimal functioning was during the earlier stages of recovery.
The inclusion of child behavior as a component of optimal functioning may have led to the identification of more children with poorer functioning. Previous work in the same cohort found that early mTBI was associated with more clinically significant behavior problems 6 and 18-months post-injury (Gagner et al., Reference Gagner, Landry-Roy, Bernier, Gravel and Beauchamp2018). School-age children (5–18 years) also present more behavior problems in the first months post-injury, which lessen after 3 months (Gornall et al., Reference Gornall, Takagi, Clarke, Babl, Davis, Dunne, Anderson, Hearps, Demaneuf, Rausa and Anderson2019). Other work indicates that infants and toddlers exhibit more behavioral manifestations of PCS than children 3–8 years, likely due to their inability to verbalize what they are experiencing or due to limited cognitive abilities (Dupont et al., Reference Dupont, Beaudoin, Désiré, Tran, Gagnon and Beauchamp2021). These behavioral manifestations may contribute to lower optimal functioning, as well as affect the quality of parent-child interactions (Beauchamp et al., Reference Beauchamp, Séguin, Gagner, Lalonde and Bernier2021).
Lower child negative affectivity temperament was an independent predictor of optimal functioning. Higher negative affectivity temperament and negative beliefs and attributions are associated with internalizing behavior problems in typically developing children (Crawford et al., Reference Crawford, Schrock and Woodruff-Borden2011), which can lead to a more negative view of life events (Campbell & Fehr, Reference Campbell and Fehr1990; Witthöft et al., Reference Witthöft, Basfeld, Steinhoff and Gerlach2012). Children with more negative affectivity tend to react to situations with more fear, anger, discomfort, sadness or anger (Putnam & Rothbart, Reference Putnam and Rothbart2006; Rothbart, Reference Rothbart1981). Thus, children with higher negative affectivity traits who sustain early mTBI might feel more distress in relation to the potentially traumatic experience, contributing to lowering their overall functioning. Interestingly, negative affectivity was no longer an independent predictor when participants with complex mTBI were removed from the analyses. This could be due to reduced statistical power, but evidence from the same cohort also suggests that more severe forms of early TBI (mild complex, moderate, severe) alter the trajectory of temperamental traits (Séguin et al., Reference Séguin, Dégeilh, Bernier, El-Jalbout and Beauchamp2020), thus, the inclusion of a few children with more severe TBI might explain these results.
Parent-child interaction quality was also linked to optimal functioning after early mTBI. Better parent-child interaction quality is characterized by a positive emotional ambiance (e.g. displays of affection, smiles, complicity, and joy), mutual cooperation (e.g. presence of an open posture by the parent and the child), and harmonious communication (e.g. communication initiatives by the child and the parent; Kochanska et al., Reference Kochanska, Aksan, Prisco and Adams2008). High parent-child interaction quality is associated with better sociocognitive skills in early childhood (Aubuchon et al., Reference Aubuchon, Libenstein, Moënner, Séguin, Bellerose, Bernier and Beauchamp2023; Licata et al., Reference Licata, Kristen and Sodian2016) and low parent-child interaction quality is associated with more externalizing behaviors in middle childhood (Dubois-Comtois et al., Reference Dubois-Comtois, Moss, Cyr and Pascuzzo2013). In the context of early mTBI, children with higher negative affectivity traits might feel more distress following the injury, which in turn could elicit reactions from the parents who must respond and adjust their child’s needs.
The findings did not support the idea that attachment and parenting stress predict optimal functioning after early mTBI. It may be that parent-child interaction accounts for more variance in optimal functioning because of the stimulating aspect of the interactions, which may be more salient in bolstering child recovery. While being securely attached to a parent with low stress may certainly be expected to enhance functioning, these factors do not necessarily ensure that the child has access to high-quality stimulation through verbal exchange and active play. Also, while parenting stress was not a significant independent predictor of optimal functioning, significant correlations were found with parent-child interaction quality, negative affectivity and optimal recovery, suggesting that associations are nonetheless present. Methodological factors may also be at play. Both the Parenting Stress Index and Attachment Q-Sort versions used were brief versions of the original measures. It is possible that more exhaustive, or other more extensive assessment tools, may have included elements with more predictive power.
The study findings can be interpreted in light of the Perception, Attribution, and Response after Early Non-inflicted Traumatic Brain Injury (PARENT) model (Beauchamp et al., Reference Beauchamp, Séguin, Gagner, Lalonde and Bernier2021). The model posits that when a child sustains mTBI during early childhood, parents’ accurate perception of the child’s symptoms (present or absent), their correct attribution of the symptoms to the injury, and their own behavioral adjustment contribute to child recovery. Having a child who sustains TBI can be disruptive, especially if the injured child manifests behavioral changes that require the parent to manage and respond to the child’s needs (Gagné et al., Reference Gagné, Goulet and Tremblay2012; Gagner et al., Reference Gagner, Dégeilh, Bernier and Beauchamp2020). Previous studies reported poorer parent-child interaction quality in this population, with interactions between mTBI dyads characterized by less communication and cooperation, and more conflicts and reprimands (Lalonde et al., Reference Lalonde, Bernier, Beaudoin, Gravel and Beauchamp2018). Enhancing parent-child interaction may be a key intervention target to optimize functioning after early mTBI. A meta-analysis by Thomas and colleagues (Reference Thomas, Abell, Webb, Avdagic and Zimmer-Gembeck2017) showed that Parent-Child Interaction Therapy, a program for parents that focuses on enhancing parent-child relationships and reducing challenging child behaviors, helps reduce externalizing problems and parent stress in typically developing children. Maggard and colleagues (Reference Maggard, Gies, Sidol, Moscato, Schmidt, Landry, Makoroff, Rhine and Wade2023) explored the feasibility and acceptability of online parenting-skills programs for caregivers of children with early complex mTBI and showed that the majority of participants found the program helpful. Since positive parenting contributes to child development (Knauer et al., Reference Knauer, Ozer, Dow and Fernald2019), intervention programs focusing on this aspect may be beneficial for promoting optimal functioning after early mTBI.
Strengths, limitations, and future directions
This study contributes to emerging efforts to study optimal functioning after mTBI and is the first to focus on early childhood. The inclusion of two comparison groups allowed us to tease out general injury versus brain injury effects. Nonetheless, some limitations need to be considered. First, the results are correlational and do not inform on causality. Second, participants were mainly Caucasian with educated parents, limiting generalization. Third, the PCS questionnaire was not validated for young children and may not have fully represented the reality of young, pre-verbal children. Since the study inception, developmentally appropriate methods for tracking PCS in young children have emerged (Dupont et al., Reference Dupont, Tang, Beaudoin, Dégeilh, Gagnon, Yeates, Rose, Gravel, Burstein, Stang, Stanley, Zemek and Beauchamp2024; Yumul et al., Reference Yumul, Crowe, Catroppa, Anderson and McKinlay2022). Fourth, parent report was used to document PCS, behavior, and temperament; as with any third party questionnaire, inherent subjective biases may have been introduced (Huynh et al., Reference Huynh, Gagner, Bernier and Beauchamp2023). Fifth, the number of candidate predictor variables was limited to avoid overfitting the model. Given the scant literature on temperament and TBI, it is possible that other dimensions could be interesting to explore. In this study, optimal functioning was conceptualized according to four criteria (behavior, cognition, PCS, quality of life). Future studies should investigate optimal functioning using even more comprehensive approaches, and before 6 months to obtain a more complete portrait.
Conclusions
Children aged 0–5 years constitute an understudied group despite the high prevalence of mTBI during early childhood, and clinical management and treatment strategies are often not adapted to their unique characteristics (Beauchamp et al., Reference Beauchamp, Anderson, Ewing-Cobbs, Haarbauer-Krupa, McKinlay, Wade and Suskauer2024). Given the predominant role of parents during the first years of life, the potential for early mTBI to impact parent-child relationships, and the strong association between the quality of parent-child interactions and optimal functioning, interventions and strategies to enhance communication, cooperation, and emotional connections after early mTBI appear to be good targets for promoting optimal recovery and function. Potential difficulties related to behavior and the family environment or parent-child relations could be documented when children present with early mTBI in order both to reduce risk of poor outcome and to optimize recovery.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1355617724000572.
Acknowledgements
The 24 items of the English and French versions of the PSI/SF were initially used by the study team without permission, however this has now been rectified with PAR. The PSI/SF is a copyrighted instrument and may not be used or reproduced in whole or in part, in any form or language, or by any means without written permission of PAR.
Funding statement
Project funded by the Canadian Institute of Health Research (Grant number MOP111036). Olivier Aubuchon is supported by a scholarship from the Canadian Institute of Health Research (FBD187569).
Competing interests
None.
Open access
