Normative and psychometric characteristics of the Health and Behavior Inventory among children with mild orthopedic injury presenting to the emergency department: Implications for assessing post-concussive symptoms using the Child Sport Concussion Assessment Tool 5th edition (Child SCAT5)

Abstract

Objective:

The study sought to present normative and psychometric data and reliable change formulae for the Health and Behaviour Inventory (HBI), a post-concussive symptom rating scale embedded in the Child Sport Concussion Assessment Tool 5th edition (Child SCAT5).

Design:

Prospective cohort with longitudinal follow-up.

Setting:

Pediatric Emergency Departments (ED).

Participants:

As part of 3 studies conducted in the US and Canada between 2001 and 2019, 450 children 8 to 16 years of age with mild orthopedic injuries (OI) were recruited during ED visits and assessed post-acutely (M = 9.38 days, SD = 3.31) and 1- and 3-months post-injury.

Independent variables:

Rater (child vs parent), sex, age at injury.

Main outcome measure:

HBI.

Methods:

Children and parents rated children’s symptoms at each timepoint; parents also rated children’s pre-injury symptoms retrospectively. Normative data (mean, SD, skewness, kurtosis, percentiles) were computed for child and parent ratings. Internal consistency was assessed using Cronbach’s alpha (α), and test-retest reliability and inter-rater agreement were assessed with intraclass correlations (ICC). Reliable change formulae were computed using linear regression and mixed models.

Results:

HBI ratings were positively skewed. Mean ratings and percentiles were stable over time. Child and parent ratings demonstrated good-to-excellent internal consistency (α .76–.94) and moderate-to-good test-retest reliability (ICC .51–.76 between adjacent assessments). However, parent-child agreement was poor-to-moderate (ICC .31–.69).

Conclusions:

The HBI demonstrates acceptable normative and psychometric characteristics. Modest parent-child agreement highlights the importance of multiple informants when assessing post-concussive symptoms. The results will facilitate the use of the HBI in research and clinical practice.

Introduction

Sport-related concussions are common in children and adolescents; 1.1–1.9 million occur in the United States annually.¹ The Sport Concussion Assessment Tool (SCAT), now in its fifth edition, is the most widely used tool for assessing sport-related concussion.² The Child SCAT, also in its fifth edition, was specifically developed for use with children ages 5–12 years.³ The Child SCAT incorporates a symptom rating scale, the Health and Behaviour Inventory (HBI).⁴ The HBI, which is also contained in the NINDS Common Data Elements for traumatic brain injury (TBI),⁵ comprises 20 items that represent somatic and cognitive symptoms commonly reported after concussion. The HBI has been shown to be sensitive to pediatric mild TBI in general, and to sport-related concussion specifically,^6–9 but published normative and psychometric data for the HBI is sparse, limiting its clinical utility.

Symptom rating scales should be reliable and valid for their intended uses.¹⁰ Reliability takes several forms, including internal consistency, test-retest reliability, and inter-rater agreement. The HBI has shown good-to-excellent internal consistency in children with concussion and orthopedic injury.^9,11 Only one previous study has described the test-retest reliability of the HBI, reporting good stability in self-ratings collected on average 64 days apart.¹¹ Parent-child agreement on the HBI has ranged from poor to good in previous research.^6,8,11,12 However, only one previous study reported on all three forms of reliability, based on a sample of 155 healthy student athletes aged 5–13 years.¹¹ Moreover, no previous study has provided normative data or reliable change formulae based on a large sample to facilitate clinical use of the HBI. Normative data are needed to determine whether HBI scores are elevated relative to peers, and reliable change formulae are needed to determine if a child’s post-injury scores are significantly higher than expected based on their pre-injury scores.¹³

The current study therefore sought to present normative and psychometric data on the HBI by combining samples of children with mild orthopedic injuries (OI) recruited as part of three prospective cohort studies of concussion or mild TBI with longitudinal follow-up. These studies used children with OI for comparison to those with concussion because the groups are well-matched demographically, share similar risk factors for injury, and have both experienced traumatic injuries necessitating medical attention.^14,15 This combined sample of children with OI, recruited from multiple sites in the United States and Canada using similar inclusion/exclusion criteria, provided the basis for examining normative and psychometric data on the HBI. Normatively, we expected ratings on the HBI to be positively skewed, indicating that most children with OI and their parents report relatively low symptom levels. We explored whether child or parent-proxy ratings were associated with children’s age or sex. With regard to reliability, based on previous reports,^6–9,11,12 we predicted that the HBI would demonstrate good-to-excellent internal consistency and test-retest reliability, but only fair parent-child agreement.

Method

Participants

Participants included 450 children and adolescents with an OI, 8–16.11 years of age, who were recruited as part of three prospective cohort studies of mild TBI, two conducted in the United States (US) and one in Canada.^9,16,17 Recruitment for both studies in the United States occurred at Nationwide Children’s Hospital in Columbus, Ohio and Rainbow Babies and Children’s Hospital in Cleveland, Ohio. For the study in Canada, recruitment occurred at five children’s hospitals in the Pediatric Emergency Research Canada (PERC) network: Alberta Children’s Hospital in Calgary, Alberta; Stollery Children’s Hospital in Edmonton, Alberta; CHU Ste-Justine Hospital in Montreal, Quebec; Children’s Hospital of Eastern Ontario in Ottawa, Ontario; and BC Children’s Hospital, in Vancouver, British Columbia.¹⁸

Children with OI were eligible for inclusion if they sustained upper or lower extremity fractures (all studies) or sprains and strains (in the Canadian study) associated with a score of 4 or less on the Abbreviated Injury Scale (AIS)¹⁹ within 48 hours of presentation to the Emergency Departments (ED) at the respective sites. They were excluded for any of the following criteria: any facial/head injuries or symptoms of concussion; any surgical intervention or conscious sedation; hypoxia, hypotension, or shock during or following the injury (if known at the time of recruitment); non-English speaking child or parents (non-English or non-French speaking in Quebec or Ontario); previous TBI requiring hospitalization; previous concussion within the past 3 months; previous severe neurological or neurodevelopmental disorder (e.g., participants were excluded for a reported history of seizures, intellectual disability, or autism spectrum disorder, but were not excluded for a reported history of migraine or other headache, learning disability, or attention-deficit/hyperactivity disorder); hospitalized in previous year for psychiatric disorder; obvious alcohol or drug ingestion associated with injury; injury related to abuse or assault; or legal guardian not present or child in foster care.

Measures

The HBI consists of 20 items (see Table 1) that ask respondents to rate children’s symptoms over the past week (or, for retrospective ratings of pre-injury symptoms, the week prior to the injury) on a scale from 0–3 (0 = Never, 1 = Rarely, 2 = Sometimes, 3 = Often). The HBI generates two subscales, cognitive and somatic, derived from factor analysis.⁴ The cognitive scale includes 11 items (score range 0–33) and the somatic scale includes 9 items (score range 0–27). The child self-report and parent proxy versions differ slightly in wording to reflect first- versus third-person perspectives.

Table 1:

Health and Behavior Inventory subscale items (parent version)

Cognitive	Somatic
1. Has trouble sustaining attention	12. Has headaches
2. Is easily distracted	13. Feels dizzy
3. Has difficulty concentrating	14. Has a feeling that the room is spinning
4. Has problems remembering what he/she is told	15. Feels faint
5. Has difficulty following directions	16. Has blurred vision
6. Tends to daydream	17. Has double vision
7. Gets confused	18. Experiences nausea
8. Is forgetful	19. Gets tired a lot
9. Has difficulty completing tasks	20. Gets tired easily
10. Has poor problem-solving skills
11. Has problems learning

Procedure

All studies were approved by the research ethics boards of the participating institutions, and informed consent was obtained from parents and assent from children in writing prior to participation. Participants were recruited from consecutive presentations to the ED, after screening following the ED visit or during the ED when research staff were present. Recruitment for one study in the United States took place from 2001 to 2005. A total of 283 children with OI were determined eligible, and 99 (35%) consented to participate. Recruitment for the other study in the United States took place from 2014 to 2017. Out of 281 children with OI who were eligible, 120 (43%) consented. Recruitment for the Canadian study took place from 2016 to 2018. A total of 1,641 children with OI were eligible, and 334 (20%) consented.

Children who consented to participate were scheduled to complete a post-acute assessment within 3 weeks of the injury that included parent and child ratings on the HBI. Of the total 553 OI children who consented, 450 (81%) returned for the post-acute assessment (M = 9.42 days, SD = 3.34, range 3–21 days), and constitute the sample for the current study. Participants also provided symptom ratings remotely at 1-month post-injury, electronically or by phone, and at a face-to-face assessment at 3-months post-injury. Of the 450 participants, 349 (78%) completed ratings at 1 month and 382 (85%) at 3 months.

Parents also provided retrospective ratings of children’s pre-injury symptoms either at the first post-acute visit (for the first study in the United States and the study in Canada) or at the time of recruitment during ED visits (for the second study in the United States). Children did not provide retrospective ratings of pre-injury symptoms. Ratings were collected on paper forms in the first study in the United States and were entered directly into REDCap databases in the other two studies.^20,21 The data are not publicly available.

Data Analysis

Statistical analyses were conducted using SPSS (IBM SPSS Statistics for Mac, Version 26.0). We computed the means, standard deviation, skewness, kurtosis, and percentiles (25^th, 50^th, 75^th, 90^th) for the HBI cognitive and somatic symptom scales separately for child self-ratings and parent proxy ratings at each occasion using all available data. We computed Spearman correlations of symptom ratings with age at injury and compared symptom ratings based on children’s sex using non-parametric Mann-Whitney U tests based on the non-normal distributions of the ratings. Internal consistency of the HBI was evaluated using Cronbach’s alpha (α). We deemed values less than .70 as poor, 0.70–0.79 as acceptable, 0.80–0.89 as good, and above 0.90 as excellent.^22,23 Intraclass correlations (ICC) were computed to examine test-retest reliability and parent-child inter-rater reliability. We present test-retest ICC (two-way mixed effects, absolute agreement, single measure) for adjacent time points (retrospective pre-injury to post-acute; post-acute to 1 month; 1 month to 3 months), as well as across the 3 post-injury time points (post-acute, 1 month and 3 months) for both parent and child report. We also present parent-child ICC (two-way random effects, absolute agreement, single measure) at each timepoint. We deemed ICC values less than .50 as poor, 0.50–0.75 as moderate, 0.76–0.90 as good, and above 0.90 as excellent.²⁴ Paired t-tests were computed to compare mean parent and child ratings at each timepoint. To develop reliable change formulae, we regressed post-injury symptom scores for both parent and child ratings onto parents’ retrospective preinjury symptom scores separately for each post injury assessment for the HBI cognitive and somatic symptom scales. The resulting regression coefficients are used to compute standardized change scores by subtracting predicted scores from actual post-injury scores and dividing by the standard error of the estimate. We also used linear mixed models in a similar fashion to derive reliable change formulae that are applicable across the first three months post-injury, after first confirming that mean ratings did not change significantly over time. Standardized change scores exceeding 1.64 (i.e., in the extreme 5% of the distribution) can be considered to reflect a reliable increase in symptoms.¹³

Results

Demographics

Table 2 summarizes the demographics for the samples from each study and for the combined sample. The total sample is predominantly male and white, with a broad representation of age and maternal education. One of the studies in the United States had notably more racial diversity than the other two studies. Participants in the two studies in the United States had mothers with lower levels of education than those in the Canadian study, with 71.2% versus 45.2% having less than a bachelor’s degree.

Table 2:

Demographics of study samples and total sample

		US Study 1 (n = 99)		US Study 2 (n = 73)		Canadian Study (n = 278)		Total Sample (n = 450)
		Mean	SD	Mean	SD	Mean	SD	Mean	SD
Age (y)		11.76	2.23	12.34	2.38	12.63	2.19	12.39	2.25
		n	%	n	%	n	%	n	%
Child’s sex	Female	35	35.4	25	34.2	128	46.0	188	41.8
	Male	64	64.6	48	65.8	150	54.0	262	58.2
Child’s race	White	64	64.6	26	35.6	194	70.3	284	63.4
	Black	32	32.3	31	42.5	7	2.5	70	15.8
	Asian	0	0	0	0	22	8.0	22	4.9
	Hispanic	0	0	4	5.5	12	4.3	16	3.6
	Indigenous	0	0	0	0	3	1.1	3	0.7
	Multiracial	3	3.0	12	16.4	38	13.8	53	11.8
Maternal education	Less than high school	9	9.1	6	8.2	6	2.2	21	4.7
	High school or equivalent	29	29.3	17	23.3	33	11.9	79	17.6
	Partial college, or trade/college diploma	25	25.3	35	47.9	79	28.4	139	30.9
	Bachelor’s degree	24	24.2	9	12.3	91	32.7	124	27.6
	Graduate degree	10	10.1	6	8.2	52	18.7	68	15.1
	Unknown	2	2.0	0	0	17	6.1	19	4.2

Note: US = United States

Normative Data

Normative data for child and parent proxy ratings on all occasions are presented in Table 3 and 4, respectively. Scores are strongly right skewed for both child and parent ratings, particularly for somatic symptom ratings, which also were highly kurtotic. Mean scores for children are consistent over time, as are scores at the 90th percentile, a common cut-off for designating a clinically meaningful elevation. For parents, mean scores for retrospective ratings of pre-injury symptoms are somewhat higher than those for post-injury ratings, but scores at the 90th percentile are consistent across timepoints.

Table 3:

Child HBI symptom rating normative data

		Post-Acute		1 Month		3 Month
		Cognitive	Somatic	Cognitive	Somatic	Cognitive	Somatic
N		446	447	333	333	376	377
Mean		7.58	3.99	7.11	3.76	7.53	3.96
SD		6.94	4.50	7.09	4.61	7.35	4.70
Skewness		0.85	1.83	0.96	1.66	0.90	1.76
Kurtosis		−0.06	4.34	0.05	2.98	0.23	3.80
Percentiles	25	2.00	1.00	1.00	0.00	1.00	0.00
	50	6.00	3.00	5.00	2.00	6.00	2.00
	75	12.00	6.00	11.00	6.00	12.00	6.00
	90	18.00	10.00	19.00	10.00	18.00	10.00

Table 4:

Parent HBI symptom rating normative data

		Pre-injury		Post-acute		1 Month		3 Month
		Cognitive	Somatic	Cognitive	Somatic	Cognitive	Somatic	Cognitive	Somatic
N		450	450	446	446	347	348	382	381
Mean		8.50	2.35	7.03	1.98	7.16	2.00	7.48	2.18
SD		7.55	3.41	7.46	3.07	7.37	3.29	7.45	3.48
Skewness		0.66	2.67	1.06	2.43	1.03	2.62	0.94	2.47
Kurtosis		−0.43	10.27	0.43	7.25	0.39	8.79	0.04	6.93
Percentiles	25	1.75	0.00	0.00	0.00	1.00	0.00	1.00	0.00
	50	7.00	1.00	4.00	1.00	4.00	0.00	5.00	1.00
	75	14.00	4.00	12.00	3.00	12.00	3.00	11.25	3.00
	90	19.00	6.00	19.00	6.00	18.00	6.00	19.00	6.00

Age was significantly related only to parent retrospective ratings of pre-injury cognitive symptoms, with a very small effect size (r = −.100, p = .034). All other correlations of ratings with age were not significant (child r −.077 to .003; parent r −.101 to .087). Children’s ratings of cognitive symptoms did not vary by sex at any occasion. Girls reported significantly more somatic symptoms than boys at 1- and 3-months post-injury, although their median and 90^th percentile scores differed by only 1 point in all but one comparison. Parent proxy ratings of cognitive symptoms were significantly higher for boys than girls at all occasions, with differences of 1 to 5 points for median and 90^th percentile scores. In contrast, parent proxy ratings of somatic symptoms did not differ by sex on any occasion.

Internal Consistency

Internal consistency ranged from acceptable to excellent (.757–.943) for parent ratings and from good to excellent (.849–924) for child ratings (see Table 5). Internal consistency was consistently excellent (.918 −.943) for the cognitive scale and acceptable to good (.757–.861) for the somatic scale.

Table 5:

Internal consistency, test-retest reliability, and inter-rater agreement

Assessment	Scale	Rater	Cronbach’s alpha	Test- retest ICC	Inter-rater ICC	Parent-child mean difference
Post-Acute	Cognitive	Parent Child	.940	Retro/PA .764	.313	−0.544
			.918	-
	Somatic	Parent Child	.757	Retro/PA .619	.280	−2.02*
			.849	-
1 Month	Cognitive	Parent Child	.919	PA/1M .664	.690	0.248
			.924	PA/1M .667
	Somatic	Parent Child	.805	PA/1M .511	.471	−1.679*
			.861	PA/1M .682
3 Month	Cognitive	Parent Child	.943	1M/3M .760	.469	−0.054
			.924	1M/3M .744
	Somatic	Parent Child	.859	1M/3M .585	.305	−1.781*
			.851	1M/3M .649

NOTE: Retro = retrospective pre-injury; PA = post-acute; 1M = 1 month; 3M = 3 month

^*Mean difference significant, p < .05

Test–retest Reliability

Test-retest intervals averaged 9.38 days (SD = 3.31) from retrospective to post-acute timepoints, 21.30 days (SD = 4.06) from post-acute to 1-month timepoints, and 63.63 days (SD = 9.81) days from 1- to 3-month timepoints. Table 5 presents ICC for adjacent timepoints. Test-retest reliability was moderate to good (.664–.764) for the cognitive subscale and moderate (.511–.682) for the somatic subscale; and did not clearly vary as a function of test-retest interval. Test-retest reliability across all post-injury time points was moderate for both the cognitive (parent 0.694, child 0.708) and somatic subscales (parent 0.537, child 0.654).

Inter-rater Agreement

Inter-rater agreement ranged from poor to moderate (.313 to .690) for cognitive symptom ratings and was consistently poor (.280 to .471) for somatic symptom ratings. Child and parent mean ratings did not differ significantly at any occasion for cognitive symptoms, but children consistently reported higher mean ratings than parents for somatic symptoms.

Reliable Change

Tables 6 and 7 present formulae for computing reliable change z scores. The formulae can be used to compute reliable change z scores that compare a parent’s retrospective rating of pre-injury cognitive or somatic symptoms to a child’s or parent’s corresponding ratings of symptoms reported during the first 3 months post-injury. Any reliable change z score of 1.65 or greater indicates a significantly greater increase in symptoms than expected based on a parent’s retrospective rating of pre-injury symptoms.

Table 6:

Reliable change equations for child HBI symptom ratings

Cognitive scale
Post-acute z score = (Post-acute cognitive score − (4.965 + (0.310 × Retrospective pre-injury cognitive score)))/6.541
1 month z score = (1 month cognitive score − (2.883 + (0.519 × Retrospective pre-injury cognitive score)))/5.933
3 month z score = (3 month cognitive score − (3.819 + (0.426 × Retrospective pre-injury cognitive score)))/6.620
Post-injury z score = (Post-injury cognitive score − (4.163 + (0.401 × Retrospective pre-injury cognitive score)))/6.424
Somatic scale
Post-acute z score = (Post-acute somatic score − (3.062 + (0.399 × Retrospective pre-injury somatic score)))/4.292
1 month z score = (1 month somatic score − (2.526 + (0.515 × Retrospective pre-injury somatic score)))/4.267
3 month z score = (3 month somatic score − (2.778 + (0.491 × Retrospective pre-injury somatic score)))/4.402
Post-injury z score = (Post-injury cognitive score – 2.963 + (0.423 × Retrospective pre-injury cognitive score)))/4.324

Table 7:

Reliable change equations for parent HBI symptom ratings

Cognitive scale
Post-acute z score = (Post-acute cognitive score − (0.497 + (0.770 × Retrospective pre-injury cognitive score)))/4.694
1 month z score = (1 month cognitive score − (1.925 + (0.634 × Retrospective pre-injury cognitive score)))/5.726
3 month z score = (3 month cognitive score − (1.752 + (0.657 × Retrospective pre-injury cognitive score)))/5.548
Post-injury z score = (Post-injury cognitive score − (1.072 + (0.717 × Retrospective pre-injury cognitive score)))/5.257
Somatic scale
Post-acute z score = (Post-acute somatic score − (0.659 + (0.563 × Retrospective pre-injury somatic score)))/2.394
1 month z score = (1 month somatic score − (0.828 + (0.525 × Retrospective pre-injury somatic score)))/2.864
3 month z score = (3 month somatic score − (0.867 + (0.550 × Retrospective pre-injury somatic score)))/2.959
Post-injury z score = (Post-injury cognitive score − (0.757 + (0.554 × Retrospective pre-injury cognitive score)))/2.703

Discussion

The results indicate that the HBI demonstrates acceptable normative and psychometric characteristics. Normative analyses indicate that the distribution of symptom ratings is skewed, especially for somatic symptoms, with relatively low mean ratings, implying that parents and children give ratings of “never” or “rarely” for most individual symptoms. However, the means and percentile scores were quite consistent across time; moreover, they were not strongly related to age or sex. Thus, specific norms stratified by age and sex do not appear to be needed for the HBI, at least in this age range and population, even though recovery from concussion may vary by age and sex.²⁵ Instead, the general norms provide robust cut-offs for determining whether scores based on child self-ratings or parent proxy ratings are above average (e.g., > 90^th percentile).

When parent ratings of pre-injury symptoms are available, the reliable change formulae can be used to determine whether a child shows a significant increase in symptoms during the first 3 months post-injury. Parents’ mean retrospective ratings of pre-injury symptoms were somewhat higher than their mean post-injury ratings. This decline could reflect test-retest attenuation, whereby scores on symptom rating scales decline from the first to second administration, particularly over relatively short intervals.^26,27 In any case, the statistical procedures used to derive the reliable change formulae take the higher pre-injury ratings into account.

The internal consistency and test-retest reliability of the HBI was satisfactory. Both child and parent ratings demonstrated good-to-excellent internal consistency and moderate to good test-retest reliability. Moreover, mixed models confirmed that mean post-injury ratings did not change significantly over time. The results suggest that the two HBI subscales, originally derived through factor analysis,⁴ are internally consistent in this population, and that the subscale scores are reasonably stable over time, confirming research in smaller samples.^9,11

In contrast, inter-rater agreement on the HBI was generally limited, albeit somewhat better for cognitive than somatic symptoms. Moreover, children consistently reported higher mean somatic symptom ratings than parents. These results are consistent with previous research on the HBI,^6,8,11,12 as well as with a broader body of research showing that parent-child agreement regarding children’s adjustment is typically modest.²⁸ Future research is needed to identify predictors of discrepancies between parent and child symptom ratings (e.g., symptom type, or child age or clinical status)^8,12, and to examine the relationship of discrepancies to other outcomes, rather than treating discrepancies as simply measurement error. The modest parent-child agreement suggests that one limitation of the current study is that only parents, and not children, provided retrospective ratings of children’s pre-injury symptoms.

The study has several other limitations. The sample was limited to children who sought treatment for mild OI at a pediatric ED. They may have had somewhat more severe OI than children presenting to other health care settings or who do not seek medical care following injury, so the findings may only be generalizable to children seen in ED settings, and perhaps only those in North America. The sample also excluded children with certain comorbidities that could influence symptom ratings, such as more severe psychiatric, neurological, or neurodevelopmental disorders, which may also limit generalizability. However, the sample covers a wide age range, has broad representation of sex, race/ethnicity, and maternal education, and includes a diverse range of mechanisms of injury (i.e., both sports and non-sports injuries). Future studies are needed of children younger than age 8 given that the Child SCAT5 is intended for use with children down to age 5, and the validity of self-report measures such as the HBI in younger children is uncertain.

The three contributing studies occurred over a relatively long timespan, such that historical cohort differences could have affected ratings. Changes in racial composition in the two US samples, which were recruited at the same locations using the same inclusion/exclusion criteria, are consistent with this notion. However, although the study samples differed in various demographics, ratings were not strongly associated with age or sex. Moreover, all study samples were recruited using virtually identical inclusion/exclusion criteria, suggesting they can be combined to create a robust normative group.

Differential consent and attrition may have introduced unmeasured bias into the results. Consent rates were consistent with other longitudinal cohort studies,²⁹ though, and attrition was expected given the longitudinal nature of the studies. In each individual study, moreover, comparisons of children who consented to those who did not and of those who were and were not included in analyses, as well as analyses of missing data patterns, have suggested that significant bias is unlikely, notwithstanding some increased attrition of children with lower socioeconomic status.^9,16

Despite these limitations, the current study has important implications for both research and clinical practice. The provision of normative data and reliable change formulae will facilitate the use of the HBI in research and clinical contexts, both as a freestanding measure and as a component of the Child SCAT5. Clinicians can use the percentile cut-offs to determine whether a youth with concussion demonstrates substantially higher symptom ratings than expected normatively; and if a parent has provided ratings of a youth’s pre-injury symptoms, then the reliable change formulae can be used to determine if a child has shown a significant increase in symptoms after injury. Limited parent-child agreement highlights the importance of a multi-informant approach to assessing post-concussive symptoms, rather than relying solely on either child self-report or parent proxy report. Overall, the findings support the use of the HBI in the evaluation of children with concussion and mild TBI, consistent with current guidelines calling for the use of age-appropriate, validated postconcussive symptom rating scales to assist with diagnosis and monitoring recovery.^30,31

Clinical relevance:

Clinicians can use the presented normative data and reliable change formulae to inform their assessment of children with concussion, whether using the HBI as a freestanding symptom rating scale or as a component of the Child SCAT5.