Caregiver and Child Behavioral Health Service Utilization Following Pediatric Traumatic Brain Injury

Abstract

Given sparse literature examining receipt of behavioral health service in children and caregivers following traumatic brain injury (TBI), we sought to identify predictors of unmet need. We performed an individual participant data meta-analysis using generalized linear mixed-effect models to examine predictors of behavioral health service use and unmet need. We included 572 children, ages 3 to 18, who were hospitalized overnight following complicated mild to severe TBI between 2002 and 2015. Caregivers completed ratings of depression and distress, child behavior problems, family functioning, and behavioral health service utilization. For children, unmet behavioral health service need was defined as an elevation on one or more child behavior problem scales without receipt of behavioral health services. For caregivers, unmet need was defined as an elevation on either a depression or distress scale without behavioral health service utilization. Among those with behavioral health needs, rates of unmet need were high for both children (77.8%) and caregivers (71.4%). Poorer family functioning was related to more unmet need in children (F(1, 497)=6.57, p=.01; OR=1.8) and caregivers (F(1, 492)=17.54, p< .001; OR=2.7). Children with unmarried caregivers also had more unmet behavioral health service need than those with married caregivers (F(1, 497)=12.14, p<.001; OR=2.2).In conclusion, unmet needs are common after pediatric TBI and relate to family factors. The findings underscore the importance of monitoring service needs following pediatric TBI and point to disparities in service use.

Pediatric traumatic brain injury (TBI) is a leading cause of morbidity and mortality (Corrigan & Hammond, 2013; Thurman, 2016). Children with TBI are at a greater risk for impairments in cognitive functioning, behavioral and executive functioning, mental health, school performance, social participation, and quality of life (Babikian, Merkley, Savage, Giza, & Levin, 2015; Catroppa et al., 2015).

Rates of new mental health diagnoses in children with TBI are higher than in healthy children, and pre-injury mental health disorders continue or worsen following TBI (Li & Liu, 2013). The most commonly diagnosed mental health disorders after TBI are attention deficit/hyperactivity disorder (ADHD), personality change, conduct disorder, posttraumatic stress, and anxiety (Schachar, Park, & Dennis, 2015). In one study, 50% of children with TBI, compared to 13% of children with orthopedic injury, were diagnosed with a new psychiatric disorder in the first 3 months following injury (Max et al., 2012). Mental health problems may also emerge over time, especially for children injured at a younger age (Chapman et al., 2010; Crowe, Catroppa, Babl, & Anderson, 2012; Karver et al., 2012; Keenan, Clark, Holubkov, Cox, & Ewing-Cobbs, 2018). Many factors increase the likelihood of mental health diagnoses after TBI including greater injury severity, poorer family functioning, single parent status, younger age at injury, pre-injury functioning, and lower socioeconomic status (SES; Keenan et al., 2018; Li & Liu, 2013; Raj et al., 2018; Wade et al., 2020).

Despite high rates of mental health problems secondary to TBI, many children do not receive needed behavioral therapy or counseling services (Fuentes et al., 2018). One study found that among children with TBI identified as having a behavioral impairment, 50–68% were not receiving mental health services (Huebner et al., 2018). Although children and their families often need mental health services following pediatric TBI, the family, child, and injury characteristics that predict service utilization and unmet need (i.e., mental health difficulties without service receipt) are unclear. Previous studies have found that behavior problems and white race predicted increased service use, and nonwhite race, lower SES, poorer family functioning, and Medicaid insurance status predicted greater unmet need (Fuentes et al., 2018; Huebner et al., 2018; Moore et al., 2018; Slomine et al., 2006). Future research is needed to corroborate and expand on these findings.

TBI is also associated with increased caregiver burden and distress (Aitken et al., 2009; Stancin, Wade, Walz, Yeates, & Taylor, 2008). Caregivers face significant stressors during their child’s recovery, including financial strain and coping with changes in their child’s cognition and behavior (Aitken et al., 2009; Stancin, Wade, Walz, Yeates, & Taylor, 2010). One study found that 41% of caregivers of children with TBI reported clinically significant stress (Hawley, Ward, Magnay, & Long, 2003). Caregiver psychological distress is associated with poorer child adaptive functioning and worse behavior problems following pediatric TBI (Micklewright, King, O’Toole, Henrich, & Floyd, 2012; Raj et al., 2014). Specifically, caregiver psychological distress may reduce warm and responsive parenting and increase maladaptive parenting styles, which in turn may exacerbate children’s behavioral deficits following injury (Wade, Cassedy, et al., 2011). Despite the importance of caregiver wellbeing to children’s recovery, no studies have examined behavioral health service utilization in caregivers or quantitatively captured unmet needs among caregivers of children with a TBI. Given high rates of child behavior problems and parent distress following pediatric TBI, a better understanding of factors that contribute to unmet behavioral health needs in both children with TBI and their caregivers is imperative.

To address these knowledge gaps, we conducted a cross-sectional examination of the relationship between child and caregiver behavioral health needs and receipt of services and identified predictors of both service utilization and unmet need among participants from seven randomized controlled trials (RCTs) of family-centered interventions. Based on previous research, we hypothesized that poorer family functioning and nonwhite race would be associated with greater unmet needs for child behavioral health services (Moore et al., 2018; Slomine et al., 2006). We also hypothesized that, given high rates of unmet needs for child behavioral health services identified in previous studies and the caregiver stress associated with pediatric TBI, unmet needs for behavioral health service would be high among caregivers.

METHOD

The sample for the current study was drawn from seven RCTs conducted between 2002 and 2015. Each RCT involved an online family-centered skill-building intervention focused on improving child and family functioning following TBI (Antonini et al., 2014; Wade, Carey, & Wolfe, 2006; Wade et al., 2017; Wade et al., 2019; Wade et al., 2014; Wade, Taylor, et al., 2018; Wade, Walz, et al., 2011). Each study used the Trauma Registries of participating hospitals to identify potentially eligible participants. Two studies also recruited from outpatient clinics (Wade et al., 2017; Wade, Kaizar, et al., 2018).

Across studies, participants were eligible if they were hospitalized overnight for a complicated mild TBI, defined as a Glasgow Coma Scale (GCS) score of 13 to 15 with positive findings on imaging; a moderate TBI (lowest GCS score of 9 to 12); or a severe TBI (lowest GCS score of 3 to 8). Children were not required to have clinically significant behavior problems to be eligible to participate in the original studies. The original RCTs included individuals who sustained a TBI from 1 to 36 months prior to participation, with the exception of one study, which included participants beyond 36 months postinjury. We excluded those individuals to reduce heterogeneity in time since injury. After applying these criteria, we identified 572 participants for inclusion in the current investigation. Across studies, participants’ ages ranged from 3 to 18 years. Caregivers and children with TBI provided informed consent and assent for the original studies, respectively, and completed pretreatment questionnaires before random assignment to a treatment or a control condition. Data for the current study were drawn from those pretreatment questionnaires, and were deidentified before use in analyses. The current use of the data for secondary analysis was approved by the Institutional Review Board at Cincinnati Children’s Hospital Medical Center.

Family-Centered Interventions.

Two trials tested I-Interact, a parenting-skills training for managing the behavioral consequences of early TBI for children ages 3 to 9 (Antonini et al., 2014; Wade et al., 2017). Five trials tested variants of Online Family Problem-Solving Therapy, which taught caregivers, the injured child, and siblings (when available) how to target problems systematically and addressed cognitive reframing, behavior management, and family communication (Wade et al., 2019). All interventions involved 7 to 10 core, self-guided online modules and live videoconference sessions with a therapist to review module content and practice learned skills.

Measures

Background interview.

Caregivers were interviewed regarding the child’s injury, medical, psychiatric (including premorbid learning disorders, ADHD, and other premorbid behavioral or emotional disorders), and educational history. Caregivers provided information about their education and marital status.

Injury information.

Mechanism of injury, date of injury, length of hospital stay, and lowest GCS score were abstracted from medical charts.

Service utilization.

We assessed utilization of behavioral health services by asking caregivers if their child was receiving any therapy or counseling at the pretreatment assessment. We also asked caregivers if they were receiving behavioral health services for themselves. Answers reflected current service utilization at the time of the pretreatment assessment in each trial.

Family Assessment Device (FAD).

The FAD is a self-report questionnaire measuring structural, organizational, and transactional characteristics of families (Miller, Epstein, Bishop, & Keitner, 1985). Caregivers rated how well each statement described their own family (e.g., ‘we don’t get along well together,’ ‘we confide in each other’). Scores range from 1 to 4 with higher scores indicating worse functioning (Epstein, Baldwin, & Bishop, 1983). We used the 12-item General Function scale to assess global family dysfunction (Miller et al., 1985) The internal consistency of the General Functioning scale is .92. The FAD discriminates between clinical and nonclinical samples, and the General Functioning scale demonstrates good validity (Byles, Byrne, Boyle, & Offord, 1988).

Measures of child clinical need.

Child clinical need was assessed using three behavior rating scales completed by caregivers.

Child Behavior Checklist (CBCL).

The CBCL is recommended as a measure to assess behavior problems in children with TBI (McCauley et al., 2012). Depending on the age of the child, caregivers either completed the CBCL 1.5–5 or CBCL 6–18. Both versions of the measure are well-validated, caregiver-reported measures of child behavior problems with high test-retest reliability (rs = .87–.95). As recommended by the developers, we used a t-score of 63 as the cutoff for clinical need based on elevations on the internalizing and externalizing subscales (Achenbach & Rescorla, 2001).

Behavior Rating Inventory of Executive Functions (BRIEF).

Caregivers completed the BRIEF or the BRIEF-Preschool version as a measure of everyday executive function (Gioia, Isquith, Guy, & Kenworthy, 2000; Isquith, Crawford, Espy, & Gioia, 2005). The BRIEF is recommended as a measure to assess executive functioning in children with TBI (McCauley et al., 2012) and is sensitive to TBI severity and outcome (Chapman et al., 2010; Chevignard et al., 2009). Both versions of the BRIEF have high internal consistency (αs = .80–.98) and moderate to high test-retest reliability (r = .78–.90). The measure also has adequate to excellent convergent validity with other commonly used child behavior scales (Gioia et al., 2000; Isquith et al., 2005). The Global Executive Composite (GEC) provides an overall measure of executive dysfunction, and, as recommended by the developers, t-scores of 65 or greater were used to represent clinical need.

Home and Community Social Behavior Scale (HCSBS) - Social Competence.

Caregivers of children in Kindergarten through 12^th grade completed the HCSBS Social Competence scale, a measure of child prosocial adaptive behaviors (Merrell, Streeter, Boelter, Caldarella, & Gentry, 2001). The HCSBS has strong reliability as evidenced by excellent Cronbach’s alphas (.96–.98 for the total score; (Merrell & Caldarella, 1999). The HCSBS - Social Competence scale has large correlations with other social behavioral and externalizing scales (.55–.72) and small correlations with internalizing scales (.23–.39), demonstrating strong convergent and divergent validity. It yields a total score, with higher scores reflecting greater social competence. We used a t-score of 37 or below on the Social Competence scale to represent clinical need, which parallels other cut off scores used in this study and is equivalent to a z-score of −1.33. Families did not complete the HCSBS in the two studies that enrolled participants under the age of five.

Measures of caregiver clinical need.

Caregivers’ clinical need was assessed using two behavior rating scales, also completed by caregivers.

Symptom Checklist 90-Revised (SCL-90-R).

Caregivers completed the SCL-90-R, a 90-item self-report inventory on which they rated whether they have been bothered in the past week by a range of psychiatric symptoms (Derogatis & Savitz, 1999). The SCL-90-R subscales have internal consistency ranging from .77 to .90, and the Global Severity Index (GSI) has a test-retest reliability coefficient of .84. The SCL-90-R also shows excellent convergent and discriminant validity when compared to scales on the MMPI. The current study used the GSI as an overall measure of psychiatric distress. As recommended by the developers, we used a t-score of 63 as the cut-off score to identify clinically significant levels of distress.

Center for Epidemiological Studies Depression Scale (CES-D).

The CES-D is a 20-item scale that assesses symptoms of depression. Caregivers rated the frequency of their specific depressive symptoms over the past week, including depressed mood, restlessness, poor appetite, and social withdrawal. Higher scores (range 0–60) indicate more severe depressive symptoms. The internal consistency of this scale ranges from .85 to .90 across studies. The CES-D is able to discriminate between psychiatric inpatient samples and samples of the general population as well as discriminate levels of severity within groups. As recommended by the developers, the current study used raw scores of 16 and higher as a cut-off score to identify clinically significant depressive symptomatology (Radloff, 1977).

Behavioral health service unmet need.

We defined child unmet need for behavioral health service as an elevation on one or more of the child behavior rating scales (CBCL Internalizing, CBCL Externalizing, BRIEF GEC, or HCSCS) that was not accompanied by contemporaneous behavioral health service utilization, as reported by the caregiver. As such, a child could have significant elevations on 0 to 4 scales. We defined caregiver unmet need for behavioral health service as an elevation on the SCL-90-R GSI or CES-D, or both, that was not accompanied by caregiver behavioral health service utilization.

Data Analysis

Baseline data from the 7 RCTs (i.e., pre-treatment data) were used to characterize participant demographics, caregiver and child service use, and clinical need. We performed an individual participant data meta-analysis and used generalized linear mixed-effect models to examine predictors of service use and unmet need controlling for study level variation via random effect. This “one step approach” produced a pooled estimate without the need to estimate and synthesized the aggregate data from each study (Riley, Lambert, & Abo-Zaid, 2010). Predictor variables examined in the analyses included year of injury (2002–2015), injury severity (complicated mild/moderate vs. severe), time since injury, and presence of premorbid learning disability, ADHD, or other behavioral or emotional disorder. We also examined demographic characteristics as predictors, including child age at baseline, sex, race, caregiver marital status, caregiver education, and family functioning (measured by the FAD). Given previous findings that mental health problems may emerge over time for younger children, we examined age by time since injury interactions. However, the interactions were not significant in any of the models and were removed.

We ran post-hoc Chi-square analyses on significant predictors of child and caregiver behavioral health service utilization to determine rates of clinical need across levels of each significant predictor. We used the Cochran-Armitage trend test to examine the relationship between number of elevated child rating scales (0–4) and child behavioral health utilization. Finally, we used mixed models to determine the relation between 1) child service use and caregiver service use, 2) child and caregiver need, and 3) child and caregiver unmet need, controlling for injury severity, time since injury, presence of premorbid learning disability, ADHD, or other behavioral or emotional disorder, child age at baseline, sex, race, caregiver marital status, caregiver education, and family functioning. Less than 2% of data were missing, so we did not adjust for missing data in this analysis. All statistical analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC).

RESULTS

Child service utilization

Table 1 depicts sample characteristics. In our sample, 20.3% (116) of children were receiving behavioral health services. Premorbid ADHD diagnosis was positively associated with child behavioral health service utilization (F(1,496)=10.43, p=.001) and clinical need (χ²(1)=13.49, p <.001). Older age at baseline (F(1,496)=11.14, p=.001) was also associated with greater service use, but age at baseline was not associated with clinical behavioral health service need.

Table 1.

Participant Characteristics by Study; Count (%) or Mean (SD)

Characteristic	All	Online	CDC	TOPS-Orig	CAPS	TOPS-RRTC	I2 Pilot	I2 RRTC
Na	572	43	66	50	132	152	26	103
Dates	2003–15	2003–4	2005–6	2006–9	2007–11	2010–14	2009–11	2010–15
Age (y)b	3–18	5–18	5–18	11–18	12–17	11–18	3–9	3–9
Time since injury (m)c	0–36	0–24	0–24	0–24	0–7	0–18	0–36	0–24
Exclusion criteria	Abusive head trauma, insufficient recovery for child to participate, significant pre-injury intellectual impairment, pre-injury psychiatric hospitalizations, caregiver hospitalized for psychiatric reasons in past year
Male	364 (63.4)	26 (60.4)	41 (62.1)	27 (54.0)	86 (65.1)	107 (70.4)	16 (61.5)	61 (59.2)
Caucasian	455 (79.6)	32 (74.4)	56 (84.9)	44 (88.0)	106 (80.3)	122 (80.3)	15 (57.7)	80 (77.7)
Child Hisp/Latd Ethnicity	32 (5.6)	0 (0.0)	0 (0.0)	2 (4.0)	6 (4.6)	12 (7.9)	2 (7.7)	10 (9.7)
Primary Caregiver
Mother	509 (89.0)	38 (88.4)	61 (92.4)	45 (90.0)	115 (87.1)	132 (86.8)	24 (92.3)	94 (91.3)
Father	46 (8.0)	5 (11.6)	1 (1.5)	4 (8.0)	13 (9.9)	18 (11.8)	1 (3.9)	4 (3.9)
Other	17 (3.0)	0 (0.0)	4 (6.1)	1 (2.0)	4 (3.0)	2 (1.4)	1 (3.8)	5 (4.9)
Caregiver Educatione
<=HSf	251 (43.9)	23 (53.5)	39 (59.1)	19 (38.0)	61 (46.2)	62 (40.8)	16 (61.5)	31 (30.1)
>HSf	321 (56.1)	20 (46.5)	27 (40.9)	31 (62.0)	71 (53.8)	90 (59.2)	10 (38.5)	72 (69.9)
Married	341 (59.6)	20 (46.5)	43 (65.2)	33 (66.0)	82 (62.1)	89 (58.6)	14 (53.9)	60 (58.3)
Age at Baseline (y)b	12.3 (4.2)	11.3 (3.1)	12.1 (3.5)	14.6 (2.4)	14.9 (1.7)	14.9 (2.0)	5.4 (1.8)	6.2 (2.0)
Time Since Injury (m)c	6.5 (5.6)	13.4 (6.9)	4.4 (3.3)	9.4 (5.3)	3.6 (1.8)	5.8 (4.0)	14.5 (10.4)	6.1 (4.9)
TBI Severityg Severe	217(39.3)	12 (27.9)	20 (28.6)	20 (43.9)	51 (38.6)	67 (47.5)	6 (12.8)	41 (87.2)
Moderate/Complh	351 (59.9)	28 (65.1)	45 (71.4)	30 (56.1)	81 (61.4)	85 (52.5)	20 (24.4)	62 (75.6)
ADHD Premorbidi
Yes	72 (12.7)	5 (12.2)	12 (18.2)	3 (6.0)	19 (14.5)	22 (14.5)	2 (7.7)	9 (8.8)
No	496 (87.3)	36 (87.8)	54 (81.8)	47 (94.0)	112 (85.5)	130 (85.5)	24 (92.3)	93 (91.2)
Other E/Bj Premorbid
Yes	35 (6.7)	0 (0.0)	7 (10.6)	4 (8.0)	9 (6.9)	12 (8.0)	0 (0.0)	3 (2.9)
No	490 (93.3)	0 (0.0)	59 (89.4)	46 (92.0)	122 (93.1)	138 (92.0)	26 (100.0)	99 (97.1)
LDk Premorbid
Yes	52 (9.2)	4 (9.8)	8 (12.1)	7 (14.0)	16 (12.1)	11 (7.3)	0 (0.0)	6 (5.8)
No	516 (90.8)	37 (90.2)	58 (87.9)	43 (86.0)	116 (87.9)	139 (92.7)	26 (100.0)	97 (94.2)

^aN = number of participants assessed at baseline,

^by= years,

^cm =months,

^dHisp/Lat = Hispanic/Latino,

^eCaregiver Education = reported education of the primary caregiver,

^fHS=High School,

^gUnknown values omitted from display but included in percentage calculation,

^hModerate/Compl = Moderate or Complicated Mild Traumatic Brain Injury,

ⁱADHD = Attention Deficit Hyperactivity Disorder,

^jE/B = Emotional/Behavioral,

^kLD = Learning Disorder.

Child Unmet Need

A total of 365 (63.8%) children had elevated scores on one or more behavioral scales, and mean scores as a function of child behavioral health service use can be found in Table 2. The most commonly reported problems (scale elevations) included difficulties with executive functioning (33.4% on the BRIEF GEC), internalizing problems (31.9% on CBCL internalizing), and externalizing problems (29.6% on CBCL externalizing). Fewer caregivers reported difficulties with social competence in their children (11.9% on the HCSBS). Children with significant elevations on a greater number of scales were more likely to receive services (Z= −4.05, p< .001), with 11.2% of children with only one significant elevation receiving services, and 38.6% of children with four significant scale elevations receiving services. Of those needing behavioral health service, 284 (77.8%) did not have their needs met. Children whose caregivers were unmarried (F(1, 497)=12.14, p<.001) and reported poorer family functioning (F(1, 497)=6.57, p=.01) had more unmet need (Table 4).

Table 2.

Caregiver-reported mean scores on child behavior scales as a function of child behavioral health service receipt.

	CBCLa Internalizing	CBCL Externalizing	BRIEFb	HCSBSc
Yes	58.50 (10.95)	57.02 (10.78)	64.46 (12.66)	49.43 (9.67)
No	52.38 (11.32)	52.13 (11.45)	56.76 (11.74)	44.93 (9.48)
Total	53.64 (11.50)	53.14 (11.48)	58.41 (12.34)	48.40 (9.79)

^aCBCL = Child Behavi or Checklist;

^bBRIEF = Behavior Rating Inventory of Executive Function;

^cHCSBS = Home and Community Social Behavior Scale

Table 4.

Significant predictors of service utilization and unmet need

Effect	F	Odds Ratio	Confidence Interval	p-value
Child Behavioral Health Utilization
Pre-morbid ADHD	10.43	0.37	0.20–0.68	.001
Pre-morbid learning disorder	0.17	0.60	0.28–1.26	.17
Pre-morbid other emotional or behavioral disorder	1.56	0.60	0.27–1.34	.21
Sex	0.05	1.05	0.65–1.71	0.83
Age at baseline	11.14	1.14	1.06–1.24	.001
TBI Severity	3.38	1.55	0.97–2.46	.07
Race	3.00	0.56	0.29–1.08	.08
Caregiver education	0.41	1.17	0.73–1.86	.52
Caregiver marital status	0.29	1.14	0.70–1.86	.59
Family functioning	0.01	1.00	0.60–1.68	.99
Year of injury	1.80	1.09	0.96–1.23	.18
Child Unmet Need
Pre-morbid ADHD	0.10	0.90	0.48–1.70	.75
Pre-morbid learning disorder	0.54	1.32	0.63–2.75	.46
Pre-morbid emotional or behavioral disorder	2.41	0.54	0.24–1.18	.12
Sex	0.94	1.24	0.80–1.92	.33
Age at baseline	2.03	0.94	0.86–1.03	.16
TBI severity	0.16	1.09	0.71–1.67	69
Race	1.60	1.42	0.83–2.43	.21
Caregiver education	3.75	1.53	0.99–2.35	.05
Caregiver marital status	12.14	2.18	1.41–3.38	<.001
Family functioning	6.57	1.82	1.15–2.88	.01
Year of injury	1.49	0.90	0.76–1.07	.22
Caregiver Behavioral Health Utilization
Pre-morbid ADHD	0.41	0.79	0.38–1.63	.52
Pre-morbid learning disorder	0.89	0.68	0.30–1.53	.35
Pre-morbid other emotional or behavioral disorder	9.91	0.27	0.12–0.61	.002
Sex	0.25	1.14	0.68–1.92	.61
Age at baseline	0.15	0.98	0.90–1.27	.70
TBI severity	0.96	1.29	0.78–2.15	.33
Race	3.63	0.49	0.24–1.02	.06
Caregiver education	0.10	0.92	0.55–1.55	.75
Caregiver marital status	2.71	1.56	0.92–2.65	.10
Family functioning	13.33	2.78	1.60–4.82	<.001
Year of injury	0.88	1.08	0.92–1.27	.35
Caregiver Unmet Need
Pre-morbid ADHD	0.13	1.13	0.60–2.13	.71
Pre-morbid learning disorder	0.10	1.13	0.53–2.40	.75
Pre-morbid other emotional or behavioral disorder	0.00	1.03	0.46–2.32	.95
Sex	0.14	1.09	0.70–1.68	.71
Age at baseline	0.65	1.02	0.97–1.08	.42
TBI severity	7.50	1.80	1.18–2.73	.006
Race	0.01	0.98	0.57–1.67	.93
Caregiver education	2.61	1.42	0.93–2.17	.11
Caregiver marital status	2.41	1.42	0.91–2.22	.12
Family functioning	17.54	2.67	1.69–4.24	<.001
Year of injury	0.22	0.98	0.91–1.06	.64

Caregiver service utilization

Caregivers who reported better family functioning (F(1,486)=13.33, p< .001) used more services, although those who reported worse family functioning were more likely to have clinical need (χ²(1) =56.55, p< .001). Caregivers of children with a premorbid emotional or behavioral disorder (F(1,486)=9.91, p=.002) were more likely to be receiving behavioral health services. However, presence of a child emotional or behavioral disorder was not associated with caregiver need.

Caregiver Unmet Need

Among caregivers, 138 (24.1%) reported clinically significant distress on the GSI, and 183 (32.0%) reported clinically significant depression on the CES-D. However, only 16.6% reported receiving behavioral health services at the pretreatment assessment. Mean scores as a function of caregiver behavioral health service use can be found in Table 3. Of caregivers with clinical need, 71.4% were not receiving behavioral health services. Unmet behavioral health service need was more common among caregivers of children with more severe TBI (F(1,497)=7.50, p=.006), although they did not report more need than caregivers of children with less severe TBI. Unmet behavioral health service need was also more common among caregivers who reported poorer family functioning (F(1, 492)=17.54, p < .001).

Table 3.

Caregiver-reported mean scores on child behavior scales as a function of child behavioral health service receipt.

	SCL-90a	CES-Db
Yes	60.30 (11.12)	18.71 (11.87)
No	53.41 (10.83)	11.48 (8.90)
Total	54.55 (11.17)	12.71 (9.84)

^aSCL= Symptom Checklist;

^bCES-D = Center for Epidemiologic Studies Depression Scale

The relation between child and caregiver service need, utilization, and unmet need

Controlling for demographic and injury characteristics, caregiver and child service needs were associated (F(1,497)=13.89, p<.001; OR=2.62), as were caregiver and child service use (F(1,485)=41.70, p<.001; OR=6.78). However, caregiver unmet need was not associated with child unmet need.

DISCUSSION

This study described behavioral health service utilization and unmet behavioral health needs after pediatric TBI. Consistent with previous research in this population, findings revealed high rates of unmet needs for behavioral health service among children and caregivers (Huebner et al., 2018; Karver et al., 2014; Narad et al., 2019). We found that almost 80% children with TBI who needed behavioral health services were not receiving them. A previous study found no difference in unmet behavioral health service needs between children with TBI and those with orthopedic injury (Narad et al., 2019). However, unmet need among children with TBI appears to be higher than in the general population of children, which ranges from 8–20% (Derigne, Porterfield, & Metz, 2009; Simon, Pastor, Reuben, Huang, & Goldstrom, 2015). Unmet needs may also be higher in children with TBI than in other populations, such as children with autism spectrum disorders and those with special health care needs (Benevides, Carretta, & Lane, 2016; Chiri & Warfield, 2012). Lack of identification of need may be a barrier to behavioral health service receipt in this population. In qualitative studies, caregivers reported uncertainty as to whether their child’s behavior is due to their brain injury or due to the child’s stage of development (Kirk, Fallon, Fraser, Robinson, & Vassallo, 2015). Additionally, physicians and caregivers may characterize the child’s current difficulties (e.g., aggression, impulsivity) as transient symptoms of the injury rather than persistent problems (Brown, Whittingham, Sofronoff, & Boyd, 2013).

To our knowledge, this is among the first studies to explore behavioral health service needs in caregivers following pediatric TBI. The high rate of unmet need among caregivers is concerning given the association between caregiver mental health, parenting, family functioning and recovery after TBI (Micklewright et al., 2012; Raj et al., 2014). We also identified significant relationships between caregiver and child behavioral health service use and need. These findings are consistent with previous research (Raj et al., 2014; Taylor et al., 2001) and further highlight the interdependence of child and family behavioral health outcomes.

Similar to previous findings, children with more severe TBI did not have greater unmet needs for behavioral health service (Slomine et al., 2006). Greater injury severity did, however, predict more unmet need for behavioral health services among caregivers. Caregivers of children with more severe TBI may be overwhelmed by their children’s needs and less able to attend to their own functioning and distress. Qualitative studies demonstrate that caregivers of children and adults with TBI put their needs aside to provide care for their children (Brown et al., 2013; Gan, Gargaro, Brandys, Gerber, & Boschen, 2010).

Consistent with a previous study showing that families with worse global functioning were more likely to have unrecognized needs during the first year after injury (Slomine et al., 2006), poorer family functioning was associated with less caregiver behavioral health service utilization and greater unmet needs among both children and caregivers. Given the importance of family functioning and caregiver well-being for child recovery (Ryan et al. 2016), these results highlight the utility of ongoing monitoring of both child and caregiver service need following child TBI. We also found that older age at baseline was associated with greater service utilization, but unmet needs were not different between older and younger children. Despite previous literature documenting emerging behavioral health problems, especially for younger children (Chapman et al., 2010; Crowe et al., 2012; Karver et al., 2012; Keenan et al., 2018), we did not identify differences in service need or unmet need based on time since injury or any interaction between time since injury and age.

The association of race with service use has been explored in previous research (Moore et al., 2018). Despite previous findings in pediatric TBI (Moore et al., 2018) as well as studies in the general pediatric literature (Coker et al., 2009; Merikangas et al., 2011) that individuals who are not white experience greater rates of unmet behavioral health service need, we did not identify greater unmet needs among non-white participants.

By examining service use in families whose children were injured between 2002 and 2015, we were able to characterize utilization patterns cross-sectionally. In doing so, we found that unmet need did not vary based on year of injury, suggesting historical changes have not influenced unmet needs in this population. Future research could examine barriers to behavioral health service in children with TBI and their caregivers use such as lack of recognition, stigma and cost (Gulliver, Griffiths, & Christensen, 2010).

The results should be interpreted in the context of certain limitations. The study was cross-sectional and therefore unable to examine service use over the course of a child’s recovery. Both service use and child and caregiver need for services were based on caregiver report, and we were unable to identify the referral source for services received (e.g., self-referral, primary care referral). Future research could incorporate additional sources of information (e.g., child report, structured diagnostic interviews) and other measures of caregiver psychological functioning (e.g., parenting stress). We did not assess for either the quality or amount of services, or whether children were receiving psychotropic medication. This is particularly relevant in this study, as 12.7% of children in our sample were reportedly diagnosed with ADHD prior to their injury. Future studies should obtain more detail about participants’ service use, including psychotropic medication, to understand whether and in what ways children’s needs are being addressed. In using rating scales of internalizing, externalizing, executive functioning and social competence, we did not examine the full range of behavioral and emotional problems that children and adolescents may face after TBI, such as posttraumatic stress and personality change (Li & Liu, 2013). Future research could incorporate a more comprehensive battery to assess behavioral health service need. The two studies that enrolled participants under the age of five did not administer a measure of social competence. Therefore, we are unable to make conclusions about unmet social needs in children under the age of 5.

Because the sample was drawn from RCTs of family problem-solving therapy, and families may have chosen to participate in the studies because they were experiencing child behavioral, parental or family concerns our sample may not be generalizable. This limitation is particularly relevant as family functioning predicted both child and caregiver unmet need. Additionally, the sample included relatively few nonwhite participants, and so we dichotomized race into white and nonwhite, rather than examining individual racial and ethnic groups. Selection bias in addition to the small sample of nonwhite participants may have driven the lack of association between race and behavioral health use and unmet need. Future could examine behavioral health services prospectively in a larger, more heterogeneous sample in terms of race and ethnicity.

Conclusions

We characterized behavioral health problems, service utilization, and unmet behavioral health service need among a large sample of children who sustained a TBI. We also identified predictors of service use and unmet need following TBI, including injury severity, caregiver marital status, and family functioning. The high rates of unmet behavioral health service needs among children with TBI suggest significant gaps in the identification and monitoring of need in this population as well as availability of services. Given that health care needs of children with TBI change over time (Fuentes et al., 2018) and that many experience persistent behavior problems (Karver et al., 2012; Schwartz et al., 2003), a clear need exists for ongoing monitoring of behavioral health service needs. Improvements in identification and monitoring have the potential to advance delivery and utilization of behavioral health services that can improve psychosocial functioning and quality of life. Our findings also highlighted high rates of unmet need for behavioral health services in caregivers. Caregiver utilization of behavioral health services has been overlooked but is critical given the importance of caregiver wellbeing to children’s recovery and the association of poorer family functioning with greater need for behavioral health services. Future research is needed to find ways to better identify child and family needs for behavioral health services after pediatric TBI and to determine better ways to meet these needs.

Key points.

• Despite the importance of caregiver wellbeing to child recovery, no studies have characterized caregiver service utilization following TBI.

• In this study of 572 caregivers and children, we identified high rates of unmet behavioral health service need in caregivers of children following pediatric TBI.

• Unmet needs among children and caregivers were related to family factors.

• The high rates of unmet behavioral health service needs among children with TBI and their caregivers suggest gaps in the identification and monitoring of need in this population.

• A clear need exists for ongoing monitoring of behavioral health service needs after pediatric TBI.

Abbreviations:

TBI

Traumatic Brain Injury

RCTs

randomized controlled trials

ADHD

attention deficit/hyperactivity disorder

GCS

Glasgow Coma Scale

FAD

Family Assessment Device

BRIEF

Behavior Rating Inventory of Executive Functions

CBCL

Child Behavior Checklist

HCSBS

Home and Community Social Behavior Scale

SCL-90-R

Symptom Checklist 90-Revised

CES-D

Center for Epidemiological Studies Depression Scale