Journal of Pediatric Psychology, Vol. 26, No. 2, 2001, pp. 79-91
© 2001 Society of Pediatric Psychology
Neurobehavioral Symptoms in Childhood Closed-Head Injuries: Changes in Prevalence and Correlates During the First Year Postinjury
1 The Ohio State University, 2 Columbus Children's Hospital, 3 Case Western Reserve University, 4 Rainbow Babies and Children's Hospital, 5 University of Cincinnati, 6 Children's Hospital Medical Center, 7 MetroHealth Medical Center
All correspondence should be sent to Keith Owen Yeates, Department of Psychology, Children's Hospital, 700 Children's Dr., Columbus, Ohio 43205. E-mail: yeates.1{at}osu.edu .
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionReferences |
|---|
Objective: To examine changes in the prevalence and correlates of neurobehavioral symptoms during the first year following childhood closed-head injuries (CHIs).
Methods: Participants included 31 children with severe CHIs, 38 with moderate CHIs, and 53 with orthopedic injuries (OIs). Children and their families were assessed shortly after injury and at 6- and 12-month follow-ups. Parents rated 15 symptoms classified as either cognitive/somatic (C/S) or emotional/behavioral (E/B).
Results: Both kinds of symptoms were more common in the CHI groups than in the OI group. C/S symptoms declined in the CHI groups over time, whereas E/B symptoms became relatively more common. Measures of injury severity, children's premorbid behavioral adjustment, and concurrent cognitive functioning predicted C/S symptoms. E/B symptoms were predicted by injury severity, concurrent cognitive functioning soon after the injury, and concurrent parent and family functioning later in time. Both types of symptoms contributed to the prediction of perceived family burden, with the relationships strengthening over time.
Conclusions: The findings indicate that the prevalence and correlates of neurobehavioral symptoms in childhood CHIs vary as a function of symptom type and time since injury.
Key words: closed-head injury; children; neurobehavioral symptoms.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Childhood closed-head injuries (CHIs) result in significant morbidity, including neurocognitive deficits, poor school performance, and declines in psychosocial adjustment and adaptive functioning (Yeates, 2000
). Childhood CHIs
also frequently initiate somatic, cognitive, and behavioral symptoms
(Barry, Taylor, Klein, & Yeates,
1996; Rivara et al.,
1994). These symptoms, commonly referred to as postconcussive in
children with mild CHIs (Yeates et al.,
1999), are also characteristic of children with more severe CHIs.
Ratings of these symptoms may provide a more sensitive and valid outcome
measure following childhood CHIs than traditional behavior rating scales,
which were not constructed specifically to assess children with brain
impairment or physical illness (Drotar,
Stein, & Perrin, 1995).
Differences in the prevalence of neurobehavioral symptoms as a function of
injury severity have been described by Rivara et al.
(1994) in a prospective study
of childhood CHIs. In that study, parents reported that cognitive and
behavioral symptoms occur more frequently among children with severe CHIs than
those with mild CHIs at both 3 and 12 months postinjury. In a previous study,
we also found that neurobehavioral symptoms were more common and persistent
following moderate to severe CHIs than following traumatic injuries not
involving the head (Barry et al.,
1996). Specifically, parents reported that children with moderate
to severe CHIs displayed a fourfold increase in the total number of symptoms
over children with orthopedic injuries (OIs), both at a baseline assessment
shortly after injury and again 6 months postinjury. In addition, we found that
the total number of symptoms predicted lower performance on cognitive and
achievement tests, more general adjustment problems, declines in adaptive
behavior, and poorer parent and family functioning.
Our previous study (Barry et al.,
1996) had two major shortcomings. First, most of our analyses
relied on a summary score representing the total number of neurobehavioral
symptoms, as opposed to measures of more specific classes of symptoms.
Research on postconcussive symptoms in mild CHI suggests that it is important
to distinguish between different kinds of symptoms. For example, factor
analytic studies indicate that somatic (e.g., fatigue, headache), cognitive
(e.g., inattention, forgetfulness), and psychological (e.g., moodiness,
irritability) symptoms tend to cluster independently
(Axelrod et al., 1996;
Cicerone & Kalmar, 1995).
The summary score used in our previous study might have obscured significant
variation in the prevalence of different symptom types.
The second major shortcoming of our previous study was its limited time
frame. The study concerned only the first 6 months postinjury and relied on
data from only two assessment occasions. The total number of symptoms did not
change significantly during this interval in children with CHIs or those with
OIs, and baseline ratings were correlated significantly with those at 6 months
postinjury. However, a closer inspection revealed changes in the prevalence of
individual symptoms over time, with declines in some cognitive and somatic
symptoms and increases in some emotional and behavioral symptoms. Rivara et
al. (1994) described a similar
pattern in children with severe CHIs from 3 to 12 months postinjury but also
relied on data from only two assessments and did not conduct statistical
analyses to determine whether changes in prevalence rates were significant.
Thus, it is unclear if the prevalence and correlates of neurobehavioral
symptoms vary across time following childhood CHIs.
In this article, we address these shortcomings using data collected from the same ongoing, multisite prospective study on which our earlier research was based. We extend our previous work by focusing on specific classes of symptoms and by including data from three assessments extending across the first year postinjury. Based on previous factor analytic studies of neurobehavioral symptoms, we classify symptoms a priori as either cognitive/somatic (C/S) or emotional/behavioral (E/B). Examples of C/S symptoms include headaches, dizziness, fatigue, forgetfulness, and distractibility. Examples of E/B problems include anxiety, moodiness, impulsivity, withdrawal, depression, and aggression. We use data from three assessments extending across the first year postinjury to examine changes in the prevalence and correlates of the two symptom types.
Based on previous research (Barry et
al., 1996; Rivara et al.,
1994), we expected that children with CHIs would show more
neurobehavioral symptoms than children with OIs. However, we hypothesized that
C/S symptoms would become relatively less prevalent over time in the CHI
groups, as compared to the OI group, and that E/B symptoms would become
relatively more prevalent. We further hypothesized that the two classes of
symptoms would show different patterns of predictors: C/S symptoms were
expected to be predicted more strongly by indicators of injury severity (i.e.,
group membership; concurrent cognitive functioning), whereas E/B symptoms were
expected to be predicted more by measures of parent and family functioning.
This hypothesis was based on previous research suggesting that neurocognitive
outcomes depend primarily on injury severity, whereas behavioral outcomes
depend more on child and family adjustment
(Rivara et al., 1994;
Taylor et al., 1999;
Yeates et al., 1997). Finally,
we hypothesized that both types of symptoms would predict the amount of burden
perceived by families as a result of the children's injuries. This hypothesis
was based on studies showing that neurobehavioral symptoms and
neuropsychiatric disorders are an important predictor of family stress and
adaptation following childhood CHIs (Max,
Castillo, Robin, et al., 1998; Rivara et al.,
1992,
1996). The current research
extends previous studies by comparing the contributions of C/S and E/B
symptoms to family burden at multiple points in time following childhood
CHIs.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Participants
The sample for this study consisted of 69 children with CHIs and 53 with OIs not involving CHIs. The sample was recruited from consecutive admissions to three hospitals in northeastern Ohio. (The larger project from which the data for this study were drawn included participants from a fourth hospital in central Ohio. However, parents at this site were asked only to indicate whether the child displayed the symptoms, not whether the symptoms they endorsed were new or had worsened since the injury. Because of this procedural error, data from the fourth site is not included in this report.) All children were between 6 and 12 years of age at the time of injury and used English as their primary language at home. Children were excluded if they had a history of child abuse, previous neurological disorder, or mental retardation.
Children were eligible for the CHI group if they sustained a blunt head
trauma and their lowest postresuscitation Glasgow Coma Scale (GCS;
Jennett & Bond, 1975)
score was 12 or less, or if the GCS score was between 13 and 15 but was
associated with an intracranial lesion on neuroimaging, skull fracture,
neurological deficits, or documented loss of consciousness for more than 15
minutes.
Children were eligible for the OI group if they sustained a noncranial fracture that required at least an overnight hospitalization but did not demonstrate any evidence of loss of consciousness or other indication of possible brain injury. The OI group was selected for comparison with the CHI group to equate the groups in terms of the experience of a traumatic injury and subsequent medical treatment. The selection of the OI group also helps to control for premorbid characteristics that increase a child's risk of sustaining a traumatic injury.
Following established conventions
(Yeates, 2000), the CHI group
was divided into two groups based on injury severity. Children whose lowest
postresuscitation GCS scores were 8 or less were considered to have severe
injuries, and children with scores of 9 or more were considered to have
moderate injuries. Many of the children in the moderate injury group had GCS
scores ranging from 13 to 15, but they all demonstrated additional
complications indicative of a more severe injury (e.g., intracranial lesion on
neuroimaging, skull fracture, neurological deficits, or sustained loss of
consciousness). Thus, their injuries were considered moderate rather than
mild.
Descriptive information about the three groups is summarized in
Table I. The groups did not
differ in gender, race, family structure, maternal education, annual family
income, or Duncan occupational status index
(Stevens & Featherman,
1981). Virtually all of the minority participants were African
American. The groups also did not differ on children's age at injury or their
pre-injury behavioral adjustment or academic performance, based on
retrospective parent and teacher ratings. As anticipated, the groups differed
in injury severity. The Injury Severity Score ISS;
(Mayer, Matlack, Johnson, & Walker,
1980) presented in Table
I is based on all injuries the children sustained, whereas the
partial ISS is calculated based only on injuries unrelated to CHI. The severe
CHI group had the most severe injuries overall but did not differ from the OI
group in the severity of injuries not involving the brain.
|
Procedure
The study was approved by the institutional review boards at all
participating institutions. All age-appropriate hospital admissions were
monitored for potential eligibility. Once children meeting entry criteria were
deemed medically stable, their parents or legal guardians were invited to
participate in the study. After informed consent was obtained, the children's
primary caregivers were asked to provide demographic information. They also
provided retrospective ratings of pre-injury family functioning and children's
premorbid behavioral adjustment. In almost all cases (93%), the child's
biological mother was the primary caregiver and respondent throughout the
study.
Consent was obtained at the time of hospitalization to request ratings of the child's classroom behavior and school performance from school personnel, and arrangements were made to complete an initial, or baseline, postinjury assessment. During baseline assessments, caregivers were interviewed regarding children's neurobehavioral symptoms. They also completed ratings of postinjury parent adjustment and family characteristics, as well as family burden associated with the children's injuries.
The baseline assessments included neuropsychological testing of the
children. Prior to testing, children in the CHI groups were screened for
post-traumatic amnesia using the Children's Orientation of Amnesia Test (COAT;
Ewing-Cobbs, Levin, Fletcher, Miner, &
Eisenberg, 1990). Children were tested only after they scored
within broad normal limits on the COAT on two consecutive days. No children
were omitted from the study based on this criterion, but assessment was
delayed somewhat for some of the children in the severe CHI group. As a
result, the groups differed significantly in the interval between the injury
and the initial postinjury assessment, although the mean difference between
the severe CHI and OI groups was only about 9 days. In almost all cases,
baseline assessments occurred within 4 weeks of injury.
Baseline assessment procedures were repeated approximately 6 and 12 months later. All follow-up assessments of child and family variables were based on concurrent child and family status.
Measures
Neurobehavioral Symptoms. Neurobehavioral symptoms were assessed
by asking the child's caregiver to complete the 30-item Post-Injury Symptom
Checklist (Barry et al., 1996).
The authors developed the checklist specifically for this study. Potential
symptoms were selected based on our clinical experience and the existing
literature (e.g., Rivara et al.,
1994). In previous research
(Barry et al., 1996), the total
score on the checklist demonstrated acceptable reliability over a 6-month
period (mean within-group correlation =.55) and predicted a variety of child
and family outcomes.
The interviewer read the symptoms to the caregiver one at a time and asked him or her to indicate whether the child displayed the symptom. For each symptom endorsed by the caregiver, the interviewer then asked whether the symptom was new or had worsened since the injury. Only symptoms reported as changed relative to the child's pre-injury status were counted as present.
For this study, we focused on 15 symptoms that the two senior authors (KOY,
HGT) classified as either cognitive/somatic (i.e., headaches/dizziness,
fatigue, memory problems, attention problems, difficulty concentrating,
difficulty following directions, confusion) or emotional/behavioral (i.e.,
anxious, quick to anger, moody, impulsive, over- or underactive, withdrawn,
depressed, aggressive). The classification followed previous factor analytic
studies of neurobehavioral symptoms
(Axelrod et al., 1996;
Cicerone & Kalmar, 1995).
The remaining 15 symptoms on the Post-Injury Symptom Checklist did not fall
clearly in either of the two categories and generally reflected physical
problems likely to arise as a result of severe CHIs (e.g., poor bladder
control, difficulties swallowing, problems in coordination and clumsiness).
They were rarely reported in either the CHI or OI groups.
Premorbid Behavioral Adjustment. Children's premorbid behavioral
adjustment was assessed during the initial hospital interview by obtaining
retrospective ratings on the Child Behavior Checklist (CBC). The CBC is a
well-known rating scale standardized on a large sample of community and
clinic-referred children between the ages of 4 and 18. It has demonstrated
satisfactory reliability and validity in previous research
(Achenbach, 1991), although it
has not always been sensitive to the effects of childhood CHIs
(Fletcher, Ewing-Cobbs, Miner, Levin,
& Eisenberg, 1990;
Fletcher et al., 1996). The
total T score from the CBC was used to measure premorbid behavioral
adjustment. As Table I shows,
the CHI and OI groups did not differ on the total T score.
Postinjury Cognitive Functioning. Cognitive outcomes were assessed
at all three occasions using a neuropsychological test battery. We selected
two measures of cognitive functioning from the larger test battery
(Taylor et al., 1995) as
possible predictors of neurobehavioral symptoms.
The first measure was an estimated Verbal IQ (VIQ) derived from a short
form of the third edition of the Wechsler Intelligence Scale for Children
(WISC-III; Wechsler, 1991).
The short form included the Similarities and Vocabulary subtests, from which
the estimated VIQ was derived (Sattler,
1992). The estimated VIQ has a validity coefficient of.88. In
previous research, we found that the estimated VIQ distinguished between the
CHI and OI groups (Taylor et al.,
1999) and that it was correlated with the total number of
neurobehavioral symptoms at the baseline assessment
(Barry et al., 1996). We also
chose the estimated VIQ because many children were unable to complete the
subtests used to estimate Performance IQ (i.e., Block Design, Object Assembly)
at the baseline assessment because of upper extremity fractures.
The second measure of cognitive functioning was the total number of words
recalled across five learning trials on a shortened, preliminary version of
the California Verbal Learning Test (CVLT;
Delis, Kramer, Kaplan, & Ober,
1994). Age-adjusted z scores for total recall were
computed using the OI group for normative purposes. We have found that total
recall on the CVLT discriminates between the CHI and OI groups
(Taylor et al., 1999) and is
correlated with the total number of neurobehavioral symptoms at the baseline
assessment (Barry et al.,
1996).
Premorbid Family Adjustment. Premorbid family adjustment was
assessed during the initial hospital interview by obtaining retrospective
ratings on the General Functioning scale from the McMaster Family Assessment
Device (FAD). The FAD is a 60-item rating scale that has shown satisfactory
reliability and validity in previous research
(Byles, Bryne, Boyle, & Oxford,
1988; Miller, Bishop, Epstein,
& Keitner, 1985). It is designed to assess family functioning
across a variety of domains and generates scores on seven subscales. The
12-item General Functioning Scale provides an overall measure of family
functioning. The CHI and OI groups did not differ at baseline on the General
Functioning Scale (Wade et al.,
1998).
Postinjury Parent Adjustment and Family Characteristics.
Postinjury parent adjustment was assessed at all three occasions using the
General Severity Index from the Brief Symptom Inventory (BSI;
Derogatis & Melisaratos,
1983). The BSI is a 53-item rating scale with satisfactory
reliability and validity designed to assess a range of psychiatric symptoms.
The General Severity Index is a global index that reflects the total score
divided by the total number of items. We have found previously that it
distinguishes between the CHI and OI groups
(Wade, Taylor, Drotar, Stancin, &
Yeates, 1998).
Postinjury family characteristics were assessed at all three occasions
using the Life Stressors and Social Resources Inventory (LSSRI;
Moos & Moos, 1988). The
LSSRI (Moos, Fenn, Billings, & Moos,
1989) is an interview measure that generates standard scores for
stressors and resources across a variety of domains. It has demonstrated
satisfactory reliability and validity in prior research
(Wade et al., 1996). For this
study, we averaged standard scores across domains to provide global measures
of family stressors and resources. Child stressor and resource subscales were
omitted because they included ratings of behavior and were potentially
confounded with the neurobehavioral symptom measures. We omitted the negative
life events stressor subscale because it was potentially confounded with
injury severity.
Postinjury Family Burden. The families' postinjury perceptions of
burden were measured with the Family Burden of Injury Interview (FBII;
Burgess et al., 1999). The FBII
was developed for this study to assess the unique burdens that families face
following childhood CHIs. During the interview, parents are asked if they are
experiencing a variety of possible sources of burden or distress. For any item
that they endorse, they are asked to rate an associated level of stress on a
5-point scale. The FBII includes questions relating to concerns about the
injured child, about the spouse's reaction to the injury, and about reactions
of extended family and friends. We used a total score from the FBII derived by
averaging the ratings on these questions. We found previously that scores on
the FBII distinguished between the CHI and OI groups
(Burgess et al., 1999;
Wade et al., 1998).
Postinjury family burden also was measured with the Impact on Family scale,
Version G (IOF-G; Stein & Jessop,
1985). This 34-item questionnaire assesses parental perceptions of
the negative impact of the child's health on the family. Because the children
in this study had all sustained a recent injury, the perceived effect of their
injury, rather than of the child's general health, was assessed using the
IOF-G. The scale generates a total score that summarizes the negative impact
of the injury. The IOF-G has been validated in previous studies of children
with chronic health conditions (Stein
& Jessop, 1985), and we have found it to differentiate between
the CHI and OI groups (Wade et al.,
1998).
Attrition and Missing Data
Of a total of 122 children, 102 completed all three assessments and 7
others completed two of the three assessments. Because of a delay in the
administration of the Post-Injury Symptom Checklist, data regarding
neurobehavioral symptoms were unavailable for the first 24 children who
entered the study, all of whom had complete data otherwise. Eleven children
completed only the baseline assessment. One child had symptom ratings
available at all three occasions but participated in neuropsychological
testing only at baseline, and one child had only baseline symptom ratings
available and did not complete any neuropsychological testing. In this study,
cross-sectional analyses at each assessment included data from all available
participants. Longitudinal analyses included only participants with the
requisite data from all three assessments.
The OI and CHI groups did not differ in the proportion of families completing at least two of the three assessments. Dropouts demonstrated lower socioeconomic status than participants remaining in the study and also were more likely to be African American. These differences did not vary across the three groups. Among the severe CHI group, dropouts had more severe injuries (i.e., lower GCS scores) than participants remaining in the study. Thus, attrition and missing data may have reduced the generalizability of the findings but did not differentially affect the OI and CHI groups.
Data Analysis
The first set of analyses focused on the prevalence of neurobehavioral
symptoms. Chi-square analyses were used first to examine group differences in
the prevalence of each individual symptom at each occasion. Repeated-measures
multivariate analyses of variance (MANOVA) were then used to examine changes
in the prevalence of individual symptoms over time, with group and occasion as
independent variables. The dependent variable in these analyses was the
presence/absence of each symptom (scored 0 or 1). Significance tests were
adjusted to reflect the use of dichotomous dependent variables
(Myers, DiCecco, White, & Borden,
1982). Finally, we also used repeated-measures MANOVA to examine
changes in the total number of C/S and E/B symptoms, again with group and
occasion as independent variables. In tests for the effect of time, the
single-degree-of-freedom linear polynomial and the interactions involving it
were examined for significance rather than the main effect for occasion and
its interactions, because study hypotheses specified different group trends in
linear change. Statistical power for the analyses of variance, based on a
p of.05, was approximately.50 for medium-effect sizes and.90 for
large-effect sizes (Cohen,
1988).
The next set of analyses focused on the prediction of neurobehavioral symptoms. We used regression analyses to examine the prediction of the total number of C/S and E/B symptoms at each occasion. Predictors included injury severity (i.e., group membership) and concurrent child cognitive functioning (estimated WISC-III VIQ, CVLT total recall z score), both potential indicators of underlying brain impairment; premorbid child adjustment (CBC total T score) and family functioning (FAD General Functioning Scale total score), both possible preinjury determinants of symptom complaints; and concurrent parent psychological adjustment (BSI General Severity Index) and family stressors and resources (LSSRI total stressors and resources scales), possible postinjury determinants of neurobehavioral symptoms. All predictors were entered simultaneously into the regression equations. Measures of injury-related family burden were not included in these analyses because they were considered more likely to result from, rather than give rise to, neurobehavioral symptoms. Statistical power for these regression analyses was approximately.75 for medium-effect sizes and.90 for large-effect sizes.
The final set of analyses focused on the prediction of family burden from neurobehavioral symptoms. We used hierarchical regression analyses to examine the prediction of the FBII total score and the IOF total score at each occasion from the total number of C/S and E/B symptoms. The two symptom scores were entered into the regression analyses simultaneously after first entering two dummy variables that represented group membership, to control for injury severity. Statistical power for these regression analyses was greater than.90 for both medium- and large-effect sizes.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Prevalence of Neurobehavioral Symptoms
Table II summarizes group differences in the prevalence of individual symptoms at each assessment. Chi-square analyses indicated that the three groups differed significantly in the prevalence of all seven C/S symptoms at all three occasions. The groups differed in the prevalence of three out of eight E/B symptoms at baseline, seven out of eight at 6 months postinjury, and all eight at 12 months postinjury. In all instances, significant differences reflected more prevalent symptoms in the CHI groups than in the OI group. In most cases, the severe CHI group also displayed more symptoms than the moderate CHI group.
|
In repeated-measures MANOVAs, group main effects were significant for all
C/S symptoms and six out of eight E/B symptoms (i.e., not for withdrawn or
depressed). Significant group by linear trend interactions were obtained for
two C/S symptoms (i.e., fatigue, confusion) and three E/B symptoms (i.e.,
quick to anger, moody, impulsive). Group differences became less pronounced
over time for the two C/S symptoms, because of declines in the two CHI groups.
In contrast, group differences became more pronounced for the three E/B
symptoms, because of increases in the severe CHI group and decreases in the OI
group over
time.,
|
|
A repeated-measures MANOVA with the total number of C/S symptoms as the dependent variable revealed a significant group main effect, F(2, 78) = 31.85, p <.001. As Figure 1 shows, the CHI groups displayed more C/S symptoms than the OI group, although the severe CHI group showed a modest decline in C/S symptoms over time. A repeated-measures MANOVA with the total number of E/B symptoms as the dependent variable revealed a significant group main effect, F(2, 78) = 11.08, p <.01, as well as a significant group by linear trend interaction, F(2, 78) = 3.16, p <.05. Figure 2 shows that the CHI groups displayed more E/B symptoms than the OI group and that the differences increased over time because of a modest increase in symptoms in the severe CHI group and a larger decrease in the OI group.
Prediction of Neurobehavioral Symptoms
In regression analyses predicting C/S symptoms, the total model was
significant at all three occasions, F(9, 77) = 12.27, p
<.001, at baseline; F(9, 88) = 5.88, p <.001, at 6
months; and F(9, 92) = 8.76, p <.001, at 12 months. The
total amount of variance accounted for was 59% at baseline, 38% at 6 months,
and 68% at 12 months. Injury severity (i.e., group membership) accounted for
unique variance in C/S symptoms at all three occasions, consistent with prior
analyses. Concurrent memory functioning accounted for unique variance at
baseline and 12-months postinjury, with more C/S symptoms related to lower
CVLT total z scores. Premorbid child adjustment accounted for unique
variance at baseline only, with more C/S symptoms related to higher CBC total
T scores. None of the other predictors contributed independently to
the prediction of C/S symptoms at any occasion.
The total model also was significant at all three occasions in regression analyses predicting E/B symptoms, F(9, 77) = 3.94, p <.001, at baseline; F(9, 88) = 8.75, p <.001, at 6 months; and F(9, 92) = 10.00, p <.001, at 12 months. The total amount of variance accounted for was 32% at baseline, 47% at 6 months, and 49% at 12 months. Injury severity (i.e., group membership) accounted for unique variance in E/B symptoms at all three occasions, again consistent with prior analyses. Concurrent child intellectual functioning accounted for unique variance at baseline and 6 months, with more E/B symptoms related to lower estimated VIQ. Concurrent family resources were a significant predictor at 6 months, such that fewer E/B symptoms were related to higher scores on the LSSRI resources scale. Concurrent parent psychological adjustment was a significant predictor at 12 months, with more E/B symptoms related to higher BSI General Severity Index scores. None of the other predictors independently contributed to the prediction of E/B symptoms at any occasion.
Prediction of Family Burden
In regression analyses predicting injury-related family burden, the total
model was significant for both dependent variables at all three occasions,
with the relationships strengthening over time: for the FBII, F(4,
92) = 7.64, p <.001, at baseline; F(4, 102) = 40.64,
p <.001, at 6 months; and F(4, 103) = 40.62, p
<.001, at 12 months; and for the IOF-G, F(4, 93) = 2.67,
p <.05 at baseline; F(4, 102) = 10.84, p
<.001, at 6 months; and F(4, 101) = 17.00, p <.001, at
12 months. As Table III shows,
the number of C/S symptoms accounted for unique variance in both the FBII and
IOF-G at 6 and 12 months postinjury but not at baseline, even after accounting
for group membership. The number of E/B symptoms accounted for unique variance
in the FBII at all three occasions and did so for the IOF-G at 12 months
postinjury but not at baseline or 6 months. The amount of variance in family
burden accounted for collectively by C/S and E/B symptoms increased
substantially over time. At both 6 and 12 months, neurobehavioral symptoms
collectively accounted for more variance in family burden than did group
membership.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
During the first year postinjury, parents report the onset or exacerbation of neurobehavioral symptoms more often among children with CHIs than among those with OIs. These results are consistent with previous studies (Barry et al., 1996
;
Rivara et al., 1994). These
findings extend prior results by showing that prevalence rates vary over time
depending on symptom type. Specifically, C/S symptoms become somewhat less
common over time among children with severe CHIs, while E/B symptoms become
relatively more common because of modest increases among children with severe
CHIs and decreases among those with OIs. Postinjury decreases in cognitive
symptoms and increases in behavioral symptoms were described in another large
study of childhood CHIs (Rivara et al.,
1994), but this is the first time that such trends have been
examined across more than two occasions using appropriate statistical
analyses.
Decreases in C/S symptoms likely reflect the recovery in neuropsychological
functioning frequently documented following childhood CHIs
(Yeates, 2000). Persistent or
even increased E/B symptoms, on the other hand, are less easily explained.
Previous studies have not consistently found sustained postinjury behavioral
disturbance following childhood CHIs (Fletcher et al.,
1990,
1996;
Rivara et al., 1994; but see
Kinsella, Ong, Murtagh, Prior, &
Sawyer, 1999; Taylor et al.,
1999). However, many previous studies used measures of behavioral
adjustment designed for mental health populations, such as the CBC, that may
not be sensitive to the effects of brain impairment. Studies like ours, which
assess E/B symptoms of the sort associated with specific neuropsychiatric
disorders (Max, Castillo, Lindgren, &
Arndt, 1998), may be more likely to document persistent behavioral
problems following severe CHIs.
Persistent E/B symptoms could reflect the reaction of family members and
their impact on children following CHIs. We have shown that severe childhood
CHI is associated with increases in parental maladjustment
(Wade et al., 1998), which
could in turn maintain a heightened rate of children's E/B symptoms following
CHI. Consistent with this notion, the OI group showed a decline in E/B
symptoms that could reflect less sustained distress among children and parents
(Wade et al., 1998).
Persistent E/B symptoms among children with CHIs also could reflect the direct
effect of brain injuries. The latter alternative is supported by the presence
of group differences on E/B symptoms across the first year postinjury even
after controlling for other child and family risk factors.
This study may help to account for changes in the prevalence of C/S and E/B symptoms by revealing differences in the predictors of the two types of symptoms. C/S symptoms were predicted primarily by injury severity (i.e., group membership), concurrent neuropsychological functioning, and premorbid child adjustment, with the latter accounting for unique variance only at the baseline assessment. In contrast, although E/B symptoms also were predicted by injury severity and neuropsychological functioning, the latter influence declined over time, while concurrent parent adjustment and family characteristics increased in importance as predictors.
C/S symptoms thus appear to depend primarily on the brain trauma associated
with CHIs, as indexed by injury severity and residual neuropsychological
deficits, and are not predicted by concurrent parent adjustment and family
functioning. In contrast, E/B symptoms are related less strongly to injury
severity than C/S symptoms but are predicted by concurrent parent and family
functioning. These findings are consistent with previous research showing that
neurocognitive outcomes depend more on injury severity, whereas behavioral
outcomes depend more on child and family adjustment, as well as
sociodemographic factors (Rivara et al.,
1994; Taylor et al.,
1999; Yeates et al.,
1997).
Although C/S and E/B symptoms were predicted by different factors, both types of symptoms independently contributed to predicting injury-related family burden. Indeed, C/S symptoms predicted injury-related family burden despite not being related to concurrent parent adjustment or family functioning. This finding provides support for our conceptualization of family burden as a potential outcome of neurobehavioral symptoms and of the other parent and family measures as predictors. More important, it suggests that C/S and E/B symptoms may differ in their relationships to parent and family functioning following childhood CHIs. C/S symptoms do not appear to be related to parent and family risk factors but do predict postinjury family burden. On the other hand, E/B symptoms are predicted by concurrent parent and family risk factors and also predict family burden. C/S symptoms therefore may have a unidirectional relationship with parent and family functioning, whereas E/B symptoms and parent and family factors may be related bidirectionally.
Neurobehavioral symptoms became stronger predictors of family burden over
time, accounting for more of the variance in family burden than group
membership at both 6 and 12 months post-injury. There are a variety of reasons
why neurobehavioral symptoms may become increasingly important determinants of
family burden (Brooks, 1991;
Demellweek & O'Leary,
1998; Rivara,
1994). Parents may be more tolerant of neurobehavioral symptoms
immediately after a CHI but become increasingly distressed by both C/S and E/B
symptoms as time passes, especially if they expect complete recovery. Parents
might find neurobehavioral symptoms increasingly burdensome because of the
cumulative difficulties of coping with the symptoms and the associated
disruption in family life. The increasing burden associated with C/S and E/B
symptoms also might reflect a decrease in the amount of external support that
the families have available to help manage the symptoms, as well as an
increase in the demands made on children as they return to school and are
expected to resume normal activities. Whatever the reason, the findings
suggest that clinicians working with families following childhood CHIs must be
alert to the stresses imposed by neurobehavioral symptoms and implement
interventions intended both to reduce the child's symptoms and to assist
parents in coping with them.
This study is characterized by several potential shortcomings, including
limitations in the procedure used to assess neurobehavioral symptoms. The
format of the ratings, which involved a simple present/absent distinction, was
likely less sensitive than Likert ratings of symptom severity. In addition,
parents were required to decide whether symptoms represented changes from
pre-injury functioning. Such retrospective judgments are subject to a variety
of biases (Mittenberg, DiGiulio, Perrin,
& Bass, 1992). To avoid reliance on parents' perceptions of
behavioral change, future studies should collect pre-injury ratings of
neurobehavioral symptoms soon after injury and then obtain ratings of current
symptoms at various times postinjury with which the pre-injury ratings can be
compared. Future studies also should obtain ratings from children as well as
parents and might use structured interviews as well as formal ratings scales
(Mittenberg, Wittner, & Miller,
1997). The use of multiple raters and of both interview and rating
scale formats will likely result in more reliable and valid estimates of the
prevalence and severity of neurobehavioral symptoms. Similar considerations
apply to most of the other interviews and questionnaires used in the study,
because they also depend almost exclusively on parent report. The inclusion of
information from other sources, such as siblings, and using other methods,
such as direct observations, would lend additional validity to the results
reported here.
In summary, these findings indicate that the prevalence and correlates of neurobehavioral symptoms in childhood CHI vary as a function of symptom type and time since injury. The results highlight the need to distinguish between different types of neurobehavioral symptoms, to study their onset and persistence over time, and to consider both neurological and environmental correlates when attempting to explain the origin and functional consequences of these symptoms. The study also underscores the need for appropriate comparison groups, which help to control for factors that might contribute to the occurrence of neurobehavioral symptoms for reasons other than residual brain injury.
Future research that takes these considerations into account is needed in
children with mild CHIs, because we are uncertain whether these findings can
be generalized to them. Mild CHIs account for the majority of pediatric head
trauma, but their outcomes remain in dispute. Children with mild CHIs rarely
show persistent deficits on standardized cognitive testing but are reported to
display postconcussive symptoms more often than uninjured children or those
with injuries not involving the head
(Satz, Zaucha, McCleary, Light, &
Asarnow, 1997; Yeates et al.,
1999). Research using methods similar to those in this study but
focusing on mild CHIs may help to resolve the ongoing controversy regarding
the clinical significance of postconcussive symptoms.
|
Acknowledgments |
|---|
The research presented here was supported by grant MCJ 390611 from the Maternal and Child Health Research Bureau (Title V, Social Security Act), Health Resources and Services Administration, Department of Health and Human Services, and by grant 5 RO1 NS36335 from the National Institute of Neurological Disorders and Stroke. We acknowledge the contributions of Matt Diamond, Nori Mercuri Minich, Madeline Polonia, Barbara Shapero, and Elizabeth Shaver. We also thank the Rainbow Pediatric Trauma Center at Rainbow Babies and Children's Hospital, the Trauma Program at Children's Hospital of Columbus, and the Trauma Center at MetroHealth Medical Center. We also acknowledge the participation of the Children's Hospital Medical Center of Akron and the collaboration of Duane Bishop, Timothy Mapstone, Scott Maxwell, George Thompson, G. Dean Timmons, and Dennis Weiner. A preliminary version of this article was presented at the annual meeting of the International Neuropsychological Society in Boston, February 1999.
Received August 19, 1999; revision received January 22, 2000; accepted April 11, 2000
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4-18 and 1991 profile. Burlington: Department of Psychiatry, University of Vermont.
Axelrod, B., Fox, D. D., Less-Haley, P. R., Earnest, K., Dolezal-Wood, S., & Goldman, R. S. (1996). Latent structure of the Postconcussion Syndrome Questionnaire. Psychological Assessment, 8, 422-427.
Barry, C. T., Taylor, H. G., Klein, S., & Yeates, K. O. (1996). Validity of neurobehavioral symptoms reported in children with traumatic brain injury. Child Neuropsychology, 2, 213-226.[Web of Science]
Brooks, D. N. (1991). The head-injured family. Journal of Clinical and Experimental Neuropsychology, 13, 155-188.
Burgess, E. S., Drotar, D., Taylor, H. G., Wade, S., Stancin, T., & Yeates, K. O. (1999). The Family Burden of Injury Interview (FBII): Reliability and validity studies. Journal of Head Trauma Rehabilitation, 14, 394-405.[Web of Science][Medline]
Byles, J., Byrne, C., Boyle, M. H., & Oxford, O. R. (1988). Ontario Child Health Study: Reliability and validity of the General Functioning Scale of the McMaster Family Assessment Device. Family Process, 27, 97-104.[Web of Science][Medline]
Cicerone, K. D., & Kalmar, K. (1995). Persistent postconcussion syndrome: The structure of subjective complaints after mild traumatic brain injury. Journal of Head Trauma Rehabilitation, 10, 1-17.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum.
Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (1994). Manual for the California Verbal Learning Test for Children. New York: Psychological Corporation.
Demellweek, C., & O'Leary, A. (1998). The impact of brain injury on the family. In R. Appleton & T. Baldwin (Eds.), Management of brain-injured children (pp. 207-230). New York: Oxford University Press.
Derogatis, L., & Melisaratos, N. (1983). The Brief Symptom Inventory: An introductory report. Psychological Medicine, 13, 595-605.[Web of Science][Medline]
Drotar, D., Stein, R. E. K., & Perrin, E. C. (1995). Methodological issues in using the Child Behavior Checklist and its related instruments in clinical child psychology research. Journal of Clinical Child Psychology, 24, 184-192[Web of Science]
Ewing-Cobbs, L., Levin, H. S., Fletcher, J. M., Miner, M. E., & Eisenberg, H. M. (1990). The Children's Orientation and Amnesia Test: Relationship to severity of acute head injury and to recovery of memory. Neurosurgery, 27, 683-691.[Web of Science][Medline]
Fletcher, J. M., Ewing-Cobbs, L., Miner, M. E., Levin, H. S., & Eisenberg, H. M. (1990). Behavioral changes after closed head injury in children. Journal of Consulting and Clinical Psychology, 58, 93-98.[Web of Science][Medline]
Fletcher, J. M., Levin, H. S., Lachar, D., Kusnerik, L., Harward,
H., Mendelsohn, D., & Lilly, M. A. (1996). Behavioral
outcomes after pediatric closed head injury: Relationships with age, severity,
and lesion size. Journal of Child Neurology,
11, 283-290.
Jennett, B., & Bond, M. (1975). Assessment of outcome after severe brain damage. Lancet, 1, 480-487.[Web of Science][Medline]
Kinsella, G., Ong, B., Murtagh, D., Prior, M., & Sawyer, M. (1999). The role of the family for behavioral outcome in children and adolescents following traumatic brain injury. Journal of Consulting and Clinical Psychology, 67, 116-123.[Web of Science][Medline]
Max, J. E., Castillo, C. S., Lindgren, S. D., & Arndt, S. (1998). The Neuropsychiatric Rating Schedule: Reliability and validity. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 297-304.[Web of Science][Medline]
Max, J. E., Castillo, C. S., Robin, D. A., Lindgren, S. D., Smith, W. L., Sato, Y., Mattheis, P. J., & Stierwalt, J. A. G. (1998). Predictors of family functioning after traumatic brain injury in children and adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 83-90.[Web of Science][Medline]
Mayer, T., Matlack, M., Johnson, D., & Walker, M. (1980). The Modified Injury Severity Scale in pediatric multiple trauma patients. Journal of Pediatric Surgery, 15, 719-726.[Web of Science][Medline]
Miller, I. W., Bishop, D. S., Epstein, N. B., & Keitner, G. I. (1985). The McMaster Family Assessment Device: Reliability and validity. Journal of Marital and Family Therapy, 11, 345-356.[Web of Science]
Mittenberg, W., DiGiulio, D. V., Perrin, S., & Bass, A. E.
(1992). Symptoms following minor head injury: Expectation as
etiology. Journal of Neurology, Neurosurgery, and
Psychiatry, 41,
611-616.
Mittenberg, W., Wittner, M. S., & Miller, L. J. (1997). Postconcussion syndrome occurs in children. Neuropsychology, 11, 447-452.[Web of Science][Medline]
Moos, R., & Moos, B. (1988). Life Stressors and Social Resources Inventory: Preliminary manual. Palo Alto, CA: Stanford University Medical Center.
Moos, R. H., Fenn, C. B., Billings, A. G., & Moos, B. I. (1989). Assessing life stressors and social resources: Applications to alcoholic patients. Journal of Substance Abuse, 1, 135-152.
Myers, J. L., DiCecco, J. V., White, J. B., & Borden, V. M. (1982). Repeated measurements on dichotomous variables: Q and F tests. Psychological Bulletin, 92, 517-525.[Web of Science]
Rivara, J. B. (1994). Family functioning following pediatric traumatic brain injury. Pediatric Annals, 23, 38-43.[Web of Science][Medline]
Rivara, J. B., Fay, G. C., Jaffe, J. M., Polissar, N. L., Shurtleff, H. A., & Martin, K. M. (1992). Predictors of family functioning one year following traumatic brain injury in children. Archives of Physical Medicine and Rehabilitation, 73, 899-910.[Web of Science][Medline]
Rivara, J. B., Jaffe, J. M., Polissar, N. L., Fay, G. C., Liao, S., & Martin, K. M. (1996). Predictors of family functioning and change 3 years after traumatic brain injury in children. Archives of Physical Medicine and Rehabilitation, 77, 754-764.[Web of Science][Medline]
Rivara, J. B., Jaffe, J. M., Polissar, N. L., Fay, G. C., Martin, K. M., Shurtleff, H. A., & Liao, S. (1994). Family functioning and children's academic performance and behavior problems in the year following traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 75, 369-379.[Web of Science][Medline]
Sattler, J. M. (1992). Assessment of children. 3rd ed. San Diego: Jerome M. Sattler.
Satz, P., Zaucha, K., McCleary, C., Light, R., & Asarnow, R. (1997). Mild head injury in children and adolescents: A review of studies (1970-1995). Psychological Bulletin, 122, 107-131.[Web of Science][Medline]
Stein, R. E. K., & Jessop, D. T. (1985). PACTS papers/AECOM: Tables documenting the psychometric properties of a measure of the impact of chronic illness on a family. Bronx, NY: Department of Pediatrics, Albert Einstein College of Medicine.
Stevens, G., & Featherman, D. L. (1981). A revised socio-economic index of occupational status. Social Science Research, 10, 364-395.[Web of Science]
Taylor, H. G., Drotar, D., Wade, S., Yeates, K. O., Stancin, T., & Klein, S. (1995). Recovery from traumatic brain injury in children: The importance of the family. In S. Broman & M. E. Michel (Eds.), Traumatic head injury in children (pp. 188-216). New York: Oxford University Press.
Taylor, H. G., Yeates, K. O., Wade, S. L., Drotar, D., Klein, S., & Stancin, T. (1999). Influences on first-year recovery from traumatic brain injury in children. Neuropsychology, 13, 76-89.[Web of Science][Medline]
Wade, S. L., Taylor, H. G., Drotar, D., Stancin, T., & Yeates, K. O. (1996). Childhood traumatic brain injury: Initial impact on the family. Journal of Learning Disabilities, 29, 652-661.
Wade, S. L., Taylor, H. G., Drotar, D., Stancin, T., & Yeates,
K. O. (1998). Family burden and adaptation during the initial
year after traumatic brain injury in children. Pediatrics,
102, 110-116.
Wechsler, D. (1991). Manual for the Wechsler Intelligence Scale for Children. 3rd ed. New York: Psychological Corporation.
Yeates, K. O. (2000). Closed-head injury. In K. O. Yeates, M. D. Ris, & H. G. Taylor (Eds.), Pediatric neuropsychology: Research, theory, and practice (pp. 92-116). New York: Guilford.
Yeates, K. O., Luria, J., Bartkowski, H., Rusin, J., Martin, L., & Bigler, E. D. (1999). Post-concussive symptoms in children with mild closed-head injuries. Journal of Head Trauma Rehabilitation, 14, 337-350.[Web of Science][Medline]
Yeates, K. O., Taylor, H. G., Drotar, D., Wade, S. L., Klein, S., Stancin, T., & Schatschneider, C. (1997). Preinjury family environment as a determinant of recovery from traumatic brain injuries in school-age children. Journal of the International Neuropsychological Society, 3, 617-630.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G A Gioia, J C Schneider, C G Vaughan, and P K Isquith Which symptom assessments and approaches are uniquely appropriate for paediatric concussion? Br. J. Sports Med., May 1, 2009; 43(Suppl_1): i13 - i22. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Ganesalingam, K. O. Yeates, M. S. Ginn, H. G. Taylor, A. Dietrich, K. Nuss, and M. Wright Family Burden and Parental Distress Following Mild Traumatic Brain Injury in Children and its Relationship to Post-concussive Symptoms J. Pediatr. Psychol., July 1, 2008; 33(6): 621 - 629. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. W. Butler and R. K. Mulhern Neurocognitive Interventions for Children and Adolescents Surviving Cancer J. Pediatr. Psychol., January 1, 2005; 30(1): 65 - 78. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. O. Yeates, H. G. Taylor, S. E. Woodrome, S. L. Wade, T. Stancin, and D. Drotar Race as a Moderator of Parent and Family Outcomes Following Pediatric Traumatic Brain Injury J. Pediatr. Psychol., June 1, 2002; 27(4): 393 - 403. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||