Effectiveness of an Exercise-Based Active Rehabilitation Intervention for Youth Who Are Slow to Recover After Concussion

Abstract

Objective: 

(1) To determine the impact of providing participants aged 8 to 17 years who are slow to recover after a concussion with a well-developed active rehabilitation intervention (ARI), compared with receiving standard care alone, on postconcussion symptoms (PCS) at 2 and 6 weeks after the initiation of ARI; and (2) to investigate functional recovery 6 weeks after initiation of ARI.

Design: 

A multicenter prospective quasi-experimental control group design.

Setting: 

Tertiary care pediatric trauma center and community health care providers.

Participants: 

Forty-nine youth were enrolled (experimental n = 36; control n = 13).

Procedures: 

Participants were assessed on 3 different occasions: (1) initial visit (baseline); (2) 2 weeks; and (3) 6 weeks after enrollment.

Main Outcome Measures: 

Child- and parent-reported PCS were obtained by the PCS Inventory Scale (primary outcome). Secondary outcomes included: (1) mood and anxiety; (2) quality of life; (3) energy level; (4) coordination and balance; (5) neurocognition; (6) parental anxiety; and (7) satisfaction with intervention.

Results: 

Both groups reported decrease of PCS over time (child: P = 0.01; parent: P = 0.03). Children in the experimental group presented higher quality of life (P = 0.04) and less anger (P = 0.02). A trend toward significance was observed for better tandem gait (P = 0.07) and for less general fatigue on self-reported PCS (P = 0.09) in the experimental group.

Conclusions: 

Active rehabilitation intervention does not affect the PCS beyond the usual management, but it increases their quality of life, decreases anger, and potentially increases energy level and balance.

Clinical Relevance

The findings of this study support our ongoing efforts to treat concussions with physical activation even in the presence of persisting symptoms.

INTRODUCTION

The incidence rate of concussion or mild traumatic brain injury (mTBI) in adults and children who seek care after injury ranges from approximately 100 to 300 per 100 000 and concussion represents 70% to 90% of all treated traumatic brain injuries (TBI).^1–3 In their extensive review of the literature, Cassidy et al⁴ found that a large number of mTBI cases are not treated in hospitals and that the actual rate of concussions is probably in excess of 600/100 000. Despite the commonly held view that the impact of concussion on youth is minimal, clinical reports indicate that residual problems occur in a range of skills, including intellectual ability, attention, memory, or balance.

Postconcussion symptoms (PCS) are the most frequent impairments observed after concussion in both youth and adults.⁵ The most frequently reported symptoms can be classified into 4 categories: (1) physical (eg, headache, dizziness, and fatigue); (2) cognitive (eg, memory and concentration); (3) sleep (eg falling asleep); and (4) emotional (eg, irritability, anger, and depression).^6–8 A controversy persists as to the duration and the impact of self-reported symptoms on youth's functioning,⁹ but it is usually recognized that they typically improve over the first few weeks after injury for a majority of individuals.^10–12 Nevertheless, it is reported that between 20% and 30% of all concussed youth will not follow the typical recovery profile^13–15 and continue to report PCS 1 month after injury.

The management of youth with concussion has been the object of much debate over the past 10 years. Traditional intervention guidelines for individuals with concussion, such as those published by the Concussion in Sports group,^16,17 have been effective in the management of simple cases who follow the expected recovery.^18–22 For example, the most recent recommendations from the Berlin Consensus explicitly promote symptom-limited activities (light activities that do not provoke concussion symptoms) as opposed to complete rest,¹⁷ as well as specific rehabilitation approaches for cervical or vestibular impairments early in the recovery period, which may help a substantial number of youth after concussion. However, for the group of individuals who fail to return to preinjury status after the expected initial recovery period, sometimes referred to as the “miserable minority,”^23,24 guidelines usually remain broad, do not suggest efficacious interventions, and are not very helpful for professionals who have to determine the best way to treat them.

According to the WHO International Classification of Diseases (ICD-10), an individual is considered to have persistent PCS (PPCS) if presenting 3 or more symptoms at least 4 weeks after injury.²⁵ In this context, those individuals seem to be a more severe population that requires special attention. Although considerable debate persists around the use of this terminology to describe persistent symptoms, the chosen period reflects the atypical and protracted nature of the recovery process, as well as the urgent need to design and evaluate interventions that are theoretically sound, clinically feasible, and ecologically valid.

This article proposes to evaluate the effectiveness of an exercise-based intervention, by comparing the recovery of youth receiving it with that of youth following standard rest-based/symptom-limited activities' recommendations. We hypothesized that (1) youth aged 8 to 17 years who had self-reported PPCS 4 weeks after the initial injury, and received individualized active rehabilitation intervention (ARI), would present with (1) decreased PCS and (2) more complete functional recovery 6 weeks after rehabilitation initiation, as compared to those in the control group receiving standard care.

METHODS

A multicenter prospective quasi-experimental control group design was used for this study. Although not as robust as a randomized trial design, constraints imposed by the fact that the proposed intervention was already the standard care at the experimental site made it ethically impossible for youth to be randomly allocated to a control intervention. To alleviate potential biases associated with the lack of randomization, youth from another institution in Canada, where this intervention was not delivered but where standard care followed usual rest/symptom-limited activity-based recommendations and return to learn strategies, were recruited as the control group. Ethical approval from both research ethics boards was obtained before initiation of the study.

Participants

Youth in the experimental group were recruited between May 2014 and February 2015 from the Concussion Clinic of the Montreal Children's Hospital, McGill University Health Centre, a tertiary care Pediatric Trauma Center. Youth were included if they presented with a confirmed diagnosis of concussion (as per referring physician, based on current operational definitions); were 8 to 17 years of age; were able to speak English or French; and presented with slow recovery 4 weeks after injury, defined as presenting with at least one PCS, reported at least once a week (eg, headache, anxiety, and fatigue), interfering with daily activities. Participants were excluded if neck pain was the only PCS, if they had sustained another concussion in the 6 months before the current injury, if they had coexisting injuries, including any previous moderate and severe TBI, or comorbidities in the past 6 months as well as diagnoses preventing participation in the intervention or assessment of standing balance/gait.

Participants were screened and enrolled, on a consecutive and voluntary basis, during their initial visit to the Concussion Clinic, by the Research Coordinator. During the course of planning and before study initiation, the control center ceased to run a concussion follow-up program, forcing the recruitment of all youth directly from community physicians' offices and the local Children's Hospital Emergency Department. Youth in the control group were thus recruited, between September 2014 and October 2015, by a university-based research team from community health care providers. Inclusion and exclusion criteria for the control group were the same as for the experimental group.

Assessment and Interventions Procedures

Participants were assessed on 3 different occasions over the follow-up period. During their initial visit (T1), before the initiation of the intervention, data were obtained combining routine clinical assessments performed by the health care team and additional measures specific to this study administered by the research personnel. A blind evaluator who was unaware of the specific content of the intervention received, or the detailed objectives of the study, assessed the participants 2 weeks (T2) and 6 weeks (T3) after initiation of the intervention. To minimize disruptions to youth's lives and promote participation, assessments for T2 were conducted over the phone, whereas T3 assessments were performed in the participant's home.²⁶ The evaluators did not have contact with the physiotherapists delivering the interventions to these youth. Parents and youth were instructed not to discuss the nature of the treatment they were receiving with the evaluator. Once having agreed to participate in the study, participants received 1 of 2 interventions, according to their institution: standard care (control group) or standard care + ARI (experimental group). All interventions were delivered weekly by the clinical team, as per usual procedure. Clinicians were not blinded to treatment allocation due to the nature of the intervention, but were not involved in the assessment of outcome once the intervention had been initiated.

Standard care consisted mainly of rest/light symptom-limited activities, general education, academic adaptations, and gradual return to school, as well as watchful waiting with youth, restricting them from participation in vigorous physical activities/sport until complete symptom resolution. It is the usual approach promoted by various associations and consensus groups.^17,27,28

The experimental approach used was based on a theoretical model anchored in current neuroscience evidence, as well as on clinical expertise of professionals involved in the care of youth with concussion.²⁶ The ARI comprised 5 components that the participants were instructed to perform daily (Table 1).

TABLE 1. - Active Rehabilitation Intervention's Description

Components	Description	Stopping Rule	Progression
Aerobic activity	Using a portable heart rate monitor, children are instructed to initiate exercise at a maximal heart rate corresponding to 50%-60% of their maximal capacity or level 2 or 3 on the Pictorial Children's Effort Rating Table (PCERT)29 for a maximum of 15 minutes. Exercise can be fast-paced walking/light jogging on a treadmill or cycling on a stationary bicycle.	The activity is terminated if a new PCS appears or if one that is reported before the exercise increases. The duration of tolerated activity is recorded.	Duration and/or intensity are increased weekly until child is symptom-free at rest and with exertion.
Coordination/sport-specific activity	Coordination exercises are tailored to the child's favorite sport or physical activity and last a maximum of 10 min. Heart rate continues to be monitored throughout this phase of the program.	As “aerobic activity” component.	As “aerobic activity” component.
Mental imagery	Visualization and imagery techniques are introduced for children familiar with the activity to reinitiate positive experiences in relation to physical activity participation. The child is asked to choose a motor component of his/her sport or favorite activity that is usually successful for them and that is finite in duration (5 min).	None.	Visualization of more advanced drills or activities.
Education	During aerobic and coordination activities, children and their families are provided reassurance and engaged in a discussion that reviews information regarding: (1) the likely time course of recovery; and (2) general coping strategies that were provided during the period of recovery	None.	As indicated.
Home program	Home-based program allowing for continued training outside the clinic, thus facilitating school attendance and minimizing disruptions to the child's daily life. The home program includes all components of the intervention, lasts approximately 20-30 min daily (monitored with the use of an activity diary), and continues until the next planned weekly visit to the clinic for reassessment.	The child and family are instructed to interrupt the home program and contact the concussion clinic if any worsening of symptoms occurs over 2 consecutive days.	As per other components.

Outcome Measures

The primary outcome of this study was child- and parent-reported level of PCS over the follow-up period. Postconcussion symptoms remain an important clinical marker of recovery after a concussion and have been identified by the Berlin consensus as the most clinically relevant criteria to track recovery after the initial few days after injury.¹⁷ The Post-Concussion Symptom Inventory (PCSI), a standardized, reliable, and valid scale, was used to document the symptoms of youth at each visit.³⁰ This scale allows the individual to report both preinjury levels of specific symptoms, as well as his/her symptoms within the past 24 hours. In this study, total score for the PCSI is thus the sum of the differences between pre- and post-injury reported levels of PCS for each of the 4 clusters: (1) physical; (2) cognitive; (3) emotional; and (4) fatigue. Also, the PCSI allowed us to look at the proportion of children who met the ICD-10 diagnostic criterion for PPCS, of 3 or more symptoms, 4 or more weeks after injury, at each time point.

There are 2 versions of the PCSI based on age. The adolescent version (13-18 years of age) is a 20-item scale designed to measure the difference in the presence and severity (on a scale of 0-6) of symptoms before injury and after injury, for a maximum score of 120. The child version (8-12 years of age) is scored using a 3-point scale (0-2) to rate 17 items. Total maximum score is 34. The parent version of the PCSI uses the same scale as the adolescent version.

Secondary outcome measures, designed to track overall recovery, were chosen based on previous work according to: (1) the logic model of the ARI,²⁶ which stated that we could have an effect on fatigue, cognition, attention, and mood; and (2) on the recommendations from the International Common Data Elements working group for pediatrics.^31,32 This first comprised the collection of participants' characteristics using a case report form (eg, age, level of sport participation, concussion, and medical and psychiatric history). Table 2 summarizes the different domains assessed and outcome measures used at each time point.

TABLE 2. - Outcome Measures by Domains and Assessment Times

Domains	Outcome Measures	Assessment Times
		T1	T2	T3
Primary outcome
PCS	PCSI (child-reported)30	X	X	X
	PCSI (parent-reported)30	X	X	X
Persistent PCS (ICD-10 diagnostic criterion)	PCSI (child- and parent-reported)30	X	X	X
Secondary outcome
Mood and anxiety	The BYI for children and adolescents—Second Edition33	X		X
	Child Behavior Check List (CBCL)34
Energy level/fatigue	Pediatric Quality of Life (PedsQL) Multidimensional Fatigue Scale35,36	X		X
Quality of life	Pediatric Quality of Life (PedsQL) Generic Module35,37	X		X
Balance and coordination	Balance component of the Sport Concussion Assessment Tool (SCAT-3)38	X		X
	Bilateral coordination (BC) and balance subtests (BS) of the Bruininks-Oseretsky Test of Motor Proficiency-2nd Edition (BOT-2).39
Parental anxiety	State Trait Anxiety Inventory (STAI)40	X		X
Cognitive function	ImPACT41,42	X		X
Physical activity participation	Physical Activity Questionnaire for Older Children (PAQ-C) and the Physical Activity Questionnaire for Adolescents (PAQ-A).43	X		X
Satisfaction with intervention	Pediatric Quality of Life (PedsQL) Health Care Satisfaction Generic Module44			X

Statistical Analysis

Baseline differences in sociodemographic characteristics (eg, age, sex, mechanism of injury, and presenting PCS) between both study groups were examined using either χ² test for categorical variables or Wilcoxon rank–sum test for continuous variables, to determine the need for adjustments in the main analysis. A 2-way mixed-design analysis of variance was used to investigate the changes in mean total and cluster PCSI scores between the 2 study groups (control and experimental) and over 3 time points (baseline—T1, 2 weeks—T2, and 6 weeks—T3 after initiation of intervention). All possible interactions between time and other variables were considered in the analysis. A 2-sample Wilcoxon rank-sum Mann–Whitney test was used to investigate the changes in mean functional recovery scores between the 2 study groups (control and experimental) over 2 time points (baseline—T1, and 6 weeks—T3 after initiation of intervention).

RESULTS

Participants Demographics

Figure 1 shows the flow of participants during the course of the study. Fifty youth in the experimental group and 18 in the control group were initially recruited in the study and received the intervention. To ensure equivalent groups regarding age and sex, youth from the experimental group were matched to those of the control group with a 2 to 3:1 ratio based on age and sex.

Figure 1.: Flow of participants during the course of the study.

Exclusion of participants for whom an intervention-control pair could not be found led to the retention of 36 participants in the experimental group and 13 participants in the control group for analysis. Youth in both groups were then comparable at baseline for all variables, except for premorbid self-reported developmental and mental health problems, where the control group had significantly more. Table 3 summarizes participant characteristics at presentation, whereas Table 4 presents child- and parent-reported PCSI scores at each time point.

TABLE 3. - Sociodemographic Variables

Sociodemographic Variables	Study Population
	Control Group (N = 13)		Experimental Group (N = 36)		P
Sex: male, n (%)	8	61.5	15	41.6	0.21
Age: yr, mean (SD)	13.2	2.6	14.0	1.9	0.11
PCS score at initial assessment
Child-reported, total*, mean (SD)	20.89	18.51	16.92	10.59	0.60
Parent-reported, total*, mean (SD)	16.77	14.54	14.42	11.85	0.74
Premorbid comorbidities:
Developmental disorders (eg, ADHD, learning disabilities), n (%)	8	66.7	0	0.0
Psychiatric disorders (eg, anxiety, depression, sleep), n (%)	4	33.3	0	0.0	0.003
Previous no. concussions
0, n (%)	10	76.9	21	58.3
1-2, n (%)	2	15.3	13	36.1
3+, n (%)	1	7.6	2	5.5	0.38
Time since injury: days, mean (SD)	40.9	15.5	37.2	13.8	0.12
Mechanisms of injury
Sports/recreational play, n (%)	8	61.5	31	86.1
Non–sport-related injury/non–sport-related fall, n (%)	3	23.0	4	11.1
Motor vehicle collision, n (%)	1	7.6	0	0.0
Assault, n (%)	1	7.6	1	2.7	0.17
Type of sport
Hockey, n (%)	1	7.6	4	11.1
Football, n (%)	0	0.0	3	8.3
Soccer, n (%)	1	7.6	10	27.7
Other (eg, basketball, gymnastics, baseball), n (%)	6	46.1	14	38.8	0.27
Type of non–sport-related injury
Slipped/fell/tripped on floor/ground, n (%)	3	23.0	3	9.6
Struck head against wall/door, n (%)	0	0.0	1	3.2	0.83
Fall impact of the surface
Grass, n (%)	1	7.6	0	0.0
Concrete, n (%)	2	15.3	0	0.0
Ice, n (%)	0	0.0	1	3.2
Other, n (%)	0	0.0	3	9.6	0.17

P-values for all categorical variables were taken from the χ² tests and P-values for continuous variables were taken from the Wilcoxon rank–sum tests.

^*Postconcussion symptom inventory total score is the sum of the differences between pre-injury and post-injury child- and parent-reported levels of PCS.

SE, standard error.

TABLE 4. - Child- and Parent-Reported PCSI Scores at Each Time Points

PCSI Scores	T1		T2		T3
	Control	Experimental	Control	Experimental	Control	Experimental
	Mean (N = 13)	Mean (N = 36)	Mean (N = 13)	Mean (N = 36)	Mean (N = 13)	Mean (N = 36)
Parents
PCSI total score (SD)	16.77 (14.54)	14.42 (11.85)	15.69 (20.88)	11.54 (13.45)	5.56 (8.97)	5.30 (10.42)
Physical cluster (SD)	7.46 (7.49)	7.97 (6.81)	6.92 (10.57)	5.11 (6.46)	3.25 (4.90)	2.83 (5.23)
Fatigue cluster (SD)	4.62 (3.94)	3.58 (3.91)	3.69 (4.64)	3.11 (3.46)	1.42 (1.88)	1.33 (2.96)
Emotional cluster (SD)	4.15 (5.24)	2.69 (2.21)	4.46 (6.41)	2.19 (4.03)	1.38 (2.06)	1.17 (3.07)
Cognitive cluster (SD)	5.00 (4.16)	3.81 (4.74)	4.85 (6.55)	4.39 (5.03)	1.54 (3.23)	1.92 (3.26)
Children
PCSI total score (SD)	20.89 (18.51)	16.92 (10.59)	13.40 (14.47)	10.93 (11.49)	8.11 (14.02)	4.70 (10.18)
Physical cluster (SD)	4.85 (5.89)	3.03 (3.09)	3.69 (7.54)	1.50 (2.60)	2.69 (6.82)	0.14 (2.65)
Fatigue cluster (SD)	2.62 (2.39)	2.36 (2.24)	1.62 (1.44)	1.53 (1.81)	1.77 (3.00)	0.47 (1.29)
Emotional cluster (SD)	3.00 (2.91)	2.17 (2.06)	2.38 (2.72)	1.56 (2.83)	1.23 (2.05)	1.14 (2.82)
Cognitive cluster (SD)	6.23 (6.31)	6.28 (4.74)	4.85 (5.77)	4.56 (5.09)	2.08 (3.90)	2.33 (5.12)

Post-concussion Symptoms

Because all youth included in the final analysis were older than 12 years, only the adolescent version of the PCSI was used. As presented in Figure 2, there were no significant interaction between group × time (P = 0.91), or differences between groups (P = 0.33), for the child-reported PCSI scores. However, PCSI scores were found to decrease over time for both groups (P = 0.01). Post hoc analysis revealed that recovery took place over the whole follow-up period (T1 and T2: P = 0.01) (T2 and T3: P = 0.01).

Figure 2.: Boxplots of the PCSI mean total score at each time point for parents and youth. Total PCSI score is obtained by adding ratings of each individual symptom. Each symptom rating represents the difference between pre- and post-injury levels of the specific symptom (delta).

Analysis of the parent-reported PCSI scores revealed similar findings with no group × time interaction (P = 0.50), no group effect (P = 0.59), and a significant time effect (P = 0.03) where both groups were found to improve over the 6-week period. More specifically, post hoc analysis on time effect showed that parents reported little recovery over the first 2 weeks, (T1 and T2: P = 0.4), some difference between T2 and T3 (P = 0.001), and a strong difference between T1 and T3 (P = 0.0001).

An analysis of differences between groups on the specific symptom clusters (eg, physical, fatigue, cognitive, and emotional) was performed. Child- and parent-reported PCSI scores showed no differences between study groups in any cluster, as well as no group × time interaction. However, there was a significant time effect for physical (child: P = 0.004; parent: P = 0.09), fatigue (child: P = 0.02; parent: P = 0.02), and cognitive (child: P = 0.07; parent: P = 0.06) clusters, but not for the emotional cluster (child: P = 0.24; parent: P = 0.14) (Table 4).

A descriptive analysis was performed at each time point in regards to the proportion of youth who met the ICD-10 diagnostic criterion for PPCS. At T1, 11 participants in the control group and 36 participants in the experimental group were experiencing PPCS. We found a decrease of 18% at T2 in the control group (n = 9) compared with 17% in the experimental group (n = 30), and a decrease of 45% at T3 in the control group (n = 6) compared with a decrease of 56% in the experimental group (n = 16).

Secondary Outcome Measures

Table 5 presents the descriptive statistics for the secondary outcome measures. Youth in the experimental group presented higher quality of life (P = 0.04) as well as scored lower on the anger scale of the Beck Youth Inventory (BYI) (P = 0.02) at T3. Furthermore, a trend toward significance in favor of the ARI was observed for tandem gait of the SCAT-3 (P = 0.07) and for general fatigue as per youth self-report (P = 0.09). Youth in the experimental group had faster speeds in tandem gait and reported more energy. No significant differences between groups were found for level of physical activity, balance, coordination, cognition, parental anxiety, or overall satisfaction.

TABLE 5. - Secondary Outcome Measures

Secondary Outcome Measures	T1		T3		P
	Control	Experimental	Control	Experimental
	Mean (N = 13)	Mean (N = 36)	Mean (N = 13)	Mean (N = 36)
Fatigue—parents report (PedsQL Multidimensional Fatigue Scale)*
General component (SD)	50.97 (18.90)	62.14 (10.76)	76.60 (18.21)	76.62 (15.12)	NS
Sleep/rest component (SD)	54.17 (19.38)	60.71 (22.20)	73.72 (17.63)	74.42 (19.26)	NS
Cognitive component (SD)	61.22 (17.79)	66.07 (19.44)	75.95 (21.58)	74.07 (21.01)	NS
Total fatigue score (SD)	55.45 (12.57)	62.98 (17.29)	75.42 (15.59)	74.38 (16.55)	NS
Fatigue—children self-report (PedsQL Multidimensional Fatigue Scale)*
General component (SD)	62.87 (27.28)	62.05 (17.97)	70.18 (20.47)	80.55 (16.36)	0.09
Sleep/rest component (SD)	64.03 (23.07)	58.45 (20.64)	66.67 (19.68)	70.37 (20.01)	NS
Cognitive component (SD)	60.21 (31.81)	57.74 (27.26)	65.39 (24.13)	70.46 (23.97)	NS
Total fatigue score (SD)	62.37 (26.63)	59.89 (18.42)	67.41 (19.89)	73.80 (16.79)	NS
Quality of life—children self-report (PedsQL Generic Module)†
Physical functioning (SD)	77.39 (23.33)	71.65 (17.72)	81.01 (20.31)	82.90 (13.99)	NS
Emotional functioning (SD)	69.58 (18.64)	78.79 (16.10)	69.23 (20.39)	83.47 (17.59)	0.02
Social functioning (SD)	82.92 (13.56)	90.00 (13.16)	81.15 (13.56)	90.97 (13.40)	0.01
School functioning (SD)	51.25 (21.65)	58.28 (19.45)	61.54 (23.03)	70.15 (20.29)	NS
Total quality of life score (SD)	70.29 (15.60)	74.69 (12.12)	73.24 (15.57)	82.25 (12.91)	0.04
Mood—children self-report (BYI)
Self-concept—T-Score (SD)	48.92 (9.34)	52.39 (8.54)	47.77 (9.67)	53.08 (9.43)	NS
Anxiety—T-Score (SD)	48.25 (7.73)	45.33 (8.39)	47.08 (9.27)	42.58 (7.87)	NS
Depression—T-Score (SD)	47.17 (10.52)	44.28 (7.81)	47.46 (10.66)	42.50 (6.31)	NS
Anger—T-Score (SD)	41.67 (5.19)	39.31 (6.80)	42.46 (5.83)	38.25 (6.33)	0.02
Disruptive behavior—T-Score (SD)	43.08 (3.57)	41.58 (5.49)	42.46 (5.83)	41.03 (4.89)	NS
Child anxiety—parent report (Child Behavior Check List)
Competence scales—T-Score (SD)	47.69 (7.31)	56.92 (11.91)	49.23 (7.22)	56.33 (11.62)	NS
Syndrome scales—T-Score (SD)	52.31 (8.39)	48.78 (8.70)	49.92 (10.83)	45.64 (9.49)	NS
Parental anxiety (State Trait Anxiety Inventory)
Total state anxiety score—percentile (SD)	28.00 (25.35)	44.69 (28.31)	25.00 (23.84)	29.50 (25.99)	NS
Total trait anxiety score—percentile (SD)	26.85 (23.36)	48.71 (22.28)	27.08 (23.69)	42.53 (7.87)	NS
Balance (SCAT-3)
Sub-tests—total no. of errors (SD)	5.77 (4.40)	3.31 (4.00)	3.92 (2.84)	4.14 (4.17)	NS
Tandem gait—in seconds (SD)	15.64 (7.86)	17.66 (5.85)	22.41 (1.16)	18.04 (5.27)	0.07
Coordination and balance (BOT-2)
Bilateral coordination—scale score (SD)	13.83 (5.26)	17.18 (5.78)	15.60 (3.78)	16.74 (2.53)	NS
Balance—scale score (SD)	12.31 (4.87)	15.03 (4.96)	13.85 (4.52)	13.00 (4.44)	NS
Cognition (ImPACT)
Memory composite verbal score (SD)	79.40 (8.47)	79.97 (11.44)	82.82 (9.02)	80.39 (12.15)	NS
Memory composite visual score (SD)	68.80 (13.74)	66.91 (13.28)	67.73 (14.77)	67.94 (15.16)	NS
Visual motor speed composite score (SD)	32.66 (5.65)	30.35 (4.93)	33.46 (4.92)	34.36 (6.27)	NS
Reaction time composite score (SD)	0.68 (0.07)	0.71 (0.12)	0.65 (0.06)	0.65 (0.09)	NS
Impulse control composite score (SD)	11.70 (12.36)	8.63 (5.63)	12.45 (8.84)	6.85 (4.02)	NS
Cognitive efficiency index (SD)	0.28 (0.11)	0.25 (0.15)	0.29 (0.13)	0.31 (0.13)	NS
Level of physical activity (PAQ-C and PAQ-A)
Physical activity—total score (SD)	1.84 (0.88)	1.62 (0.64)	2.38 (0.88)	2.26 (0.86)	NS
Satisfaction with intervention—parents score (PedsQL Health Care Satisfaction Generic Module)
Total score (SD)			91.46 (11.51)	91.07 (11.10)	NS

^*A PedsQL Multidimensional Fatigue Scale Pediatric Scale's higher score means more fatigue.

^†A PedsQL Generic module Scale's higher score means better quality of life.

DISCUSSION

Our group had published 2 case series^26,45 providing proof of concept evidence, although limited in scope by the nature of the research design, for the effectiveness of an exercise-based ARI performed in the presence of symptoms (not symptom-limited) to facilitate return to function in youth who were slow to recover after concussion. This study was a quasi-experimental clinical trial, the purpose of which was to determine whether providing ARI would lead to a greater decrease of PCS than standard, education-based intervention alone. We also wanted to monitor recovery in domains other than symptoms, which had also showed positive outcomes in the previous case series, and those were explored as secondary outcomes in this study.

As per previous findings, youth who were still symptomatic more than 4 weeks after injury and received ARI did show a decrease in their PCS scores over time, but they did so in a similar manner to those youth receiving standard care alone. All parents and youth reported similar patterns of improvement over the 6-week follow-up period, and did so for most symptom clusters. However, the proportion of youth who met diagnostic criterion for PPCS according to the ICD-10 (presence of 3 or more symptoms beyond preinjury levels) seemed less in the experimental group at T3. Interestingly, although scores on the PCSI may be similar, the burden of symptoms seems to be less in the group of children receiving ARI. To the best of our knowledge; no other research has studied the impact of burden of symptoms.

This decreased burden is further illustrated by the significant difference in patient-reported quality of life between our groups at T3 in favor of the intervention group. Quality of life is a patient-reported outcome, which is comprehensive and represents the youth's overall status. This finding strengthens the fact that engaging in supervised aerobic and coordination exercises is better than just rest for the youth with PPCS.^46–49 Quality of life scores at T1 for both groups were comparable with those reported by Novak et al⁴⁷ for youth with PPCS. Scores at T3 for youth receiving the ARI were similar to those reported for youth without symptoms,³⁷ whereas those of youth in the control group remained stable over time. The lack of correspondence between PCSI scores and quality of life scores needs to be further investigated. Symptom reporting is important to monitor functional recovery; however, it may not be the best marker because PCS are nonspecific and commonly reported by children without pathology or with a variety of other conditions. Despite symptoms, youth can be quite functional; quality of life indicator may better capture the youth's functional level.

Mental health outcomes after concussion are of increasing concern in the young adult population, with a particular focus on those individuals sustaining repeated concussions.^50,51 Because one of the ARI's goals is to improve mood in slow-to-recover youth, we used the BYI, a scale targeting emotional and social impairments, in our participants at T1 and T3. Youth who received the ARI reported lower levels of anger than youth in the control group, while presenting with similar levels for the other scales of the BYI. Anger is commonly reported in individuals with concussion.^52,53 Indeed, rehabilitation interventions have shown positive effects on emotional reactions of athletes who sustained a concussion.⁵⁴ In addition, aerobic and coordination exercises seem to have positive effects on psychological functioning within this population,⁴⁸ supporting our findings that ARI could have a positive effects on youth's behaviors when they are slow to recover after a concussion. Twelve participants in the control group had either developmental disorders or psychiatric conditions, whereas the experimental group did not have any of these conditions. These conditions were not exclusion criteria because we wanted a more representative sample in the study. It is of note that none of the developmental and mental health issues were self-reported and not formally diagnosed.

Cognition, balance, and coordination did not seem to improve over the follow-up period in our 2 groups. This could be explained by the fact that T1 performance was well within normal limits and most of the recovery had most likely taken place in the weeks immediately after the injury.^55–57 It could also be hypothesized that domains such as balance and coordination seem to require more sophisticated outcome measures (force plates and complex tasks) than those included in this study. Not surprisingly, neurocognitive deficits seem to resolve within the first 4 weeks after concussion.^31,32

Limitations

The results of our quasi-experimental clinical trial have to be interpreted in the context of the local service organization of our 2 sites, which lead to the groups being less comparable than was originally anticipated. In an effort to make the groups comparable, adjustments were made accordingly, thus causing a decrease of participants to be analyzed in each group. Despite adjustments, some measures presented problems of comparability between groups at baseline, which limits some of the data interpretation. More specifically, this led to a difference between groups regarding premorbid self-reported comorbidities (eg, attention-deficit/hyperactivity disorder [ADHD], learning disabilities, anxiety, and depression). Participants' compliance in the experimental group regarding the ARI was monitored clinically because the delivery of the intervention was not part of the research procedures. To achieve this, youth were provided with a detailed description of their home program as well as with a log to be filled with the content, duration, and response to the daily ARI. Compliance in the control group regarding inclusion of physical activity in their management protocols was not formally monitored with such logs, and it is possible that despite aerobic exercise not being part of standard care, they could have engaged in some physical activity. This could be a limitation of the study in that it could mask some of the impact of exercise in the experimental group.

CONCLUSIONS

This quasi-experimental clinical trial showed that all participants improved over time, but that there was no differential recovery of PCS for youth receiving the ARI when compared with those receiving standard education-based interventions. However, a much smaller proportion of youth in the experimental group still met the ICD-10 definition of PPCS, 6 weeks after enrollment in the study, than in the control group. Concurrently, youth who received the ARI reported higher quality of life and less feelings of anger. These findings support our ongoing efforts to treat concussions with physical activation even in the presence of persisting symptoms, and further work is needed to determine the optimal timing to initiate such interventions as well as to investigate what components or types of exercise yield the best results for youth after a concussion.