|Home | About | Journals | Submit | Contact Us | Français|
ObjectivesMild traumatic brain injury (TBI) and injury-related outcomes such as postconcussive symptoms (PCS) may influence health-related quality of life (HRQOL) in children.MethodsWe evaluated HRQOL in 186 8- to 15-year-old children with mild TBI and 99 children with orthopedic injuries (OI). Parents rated the frequency and severity of PCS at an initial assessment within 2-weeks postinjury and rated HRQOL at 3- and 12-months postinjury.ResultsThe mild TBI and OI groups did not differ in psychosocial HRQOL, but the mild TBI group showed lower physical HRQOL at the 12-month follow-up. Somatic PCS were a significant predictor of physical HRQOL over time, and both cognitive and somatic PCS were significant predictors of psychosocial HRQOL over time. Children with higher PCS at the initial assessment had lower HRQOL scores at later time points.ConclusionsEffective management of PCS may be associated with improvements in HRQOL following pediatric mild TBI.
Over the past decade, research on pediatric illness and injury has increasingly focused on patient and family burden. This shift in focus has been attributed to improvements in healthcare and treatment that have lead to decreased mortality and a shift from more acute to chronic illness (Eiser & Morse, 2001a, 2001b; Winthrop, 2010). With increased survival, the definition of “health” expands beyond the absence of illness to include the physical, mental, and social well-being of the individual (World Health Organization, 1948). As such, examining quality of life has become a priority and the field has seen an increase not only in the number of measures designed to assess health-related quality of life (HRQOL) but also in the frequency with which these measures are included in the assessment of health outcomes and as part of clinical trials (Davis et al., 2006; Drotar, 2004; Eiser & Jenney, 2007; Eiser & Morse, 2001b; Matza, Swensen, Flood, Secnik, & Leidy, 2004; Raat, Mohangoo, & Grootenhuis, 2006).
Traumatic brain injury (TBI) is one of the leading causes of morbidity and mortality in children (Schneier, Shields, Hostetler, Xiang, & Smith, 2006). Neuropsychological deficits following pediatric TBI are well documented; however, children also experience disruptions in school and family functioning and display long-term behavioral difficulties following injury (Taylor, 2010; Yeates, 2009). Recent studies have documented deficits in HRQOL among children with TBI. These deficits are most noticeable within the first year postinjury and tend to improve over time (Aitken et al., and the CHAT Study Group 2009; McCarthy et al., for the Children's Health After Trauma Study Group 2006; Tilford et al., 2007), although impairments in HRQOL can be detected several years to decades following injury (Anderson, Brown, Newitt, & Hoile, 2011; Erickson, Montague, & Gerstle, 2010; Limond, Dorris, & McMillan, 2009; Stancin et al., 2002). These long-term impairments tend to be restricted to children who suffered a severe TBI, whereas deficits in HRQOL beyond the first year are less consistent for those with mild to moderate injuries (Anderson et al., 2011; Stancin et al., 2002).
Few studies have specifically examined HRQOL among children with mild TBI, despite the fact that approximately 80–90% of all TBI are classified as mild injuries (Kraus, 1995). One study found no HRQOL deficits at a 3-month follow-up, and no significant changes from estimated preinjury ratings were noted, despite 30% of the sample being cognitively impaired and showing significantly lower scores on a measure of general health and functional status (Petersen, Scherwath, Fink, & Koch, 2008). Findings from a separate study suggest deficits in HRQOL may become more apparent over time. A study conducted at 1–5 years postinjury found that significantly more children with mild TBI were rated below average on all dimensions of HRQOL as compared to normative data (Limond et al., 2009). Of note, however, sample size in both studies was fairly small and neither utilized a comparison group to control for the impact of traumatic injury on HRQOL.
The presence of mediating variables may explain the lack of consistent findings in children with mild TBI. For instance, several studies have determined that noninjury-related factors (e.g., unhealthy family functioning, socioeconomic status) are related to HRQOL after childhood TBI (Anderson, Brown, & Newitt, 2010; McCarthy et al., for the Children's Health After Trauma Study Group 2006). In contrast, injury-related factors, such as severity of injury, are less consistently related to HRQOL (McCarthy, 2007). The impact of other injury-related outcomes, such as postconcussive symptoms (PCS), on HRQOL has yet to be examined in pediatric samples. PCS are cognitive, behavioral, emotional, and somatic symptoms commonly reported following head injury. Complaints most often reported include headaches, dizziness, and poor concentration or memory. These symptoms are more common and persistent in children with mild TBI than children with other types of injuries (Mittenberg, Miller, & Luis, 1997; Mittenberg, Wittner, & Miller, 1997; Yeates et al., 1999, 2009) and may have a disruptive impact on children and families. A study of family functioning following pediatric mild TBI found a link between increased PCS shortly after injury and ratings of family burden and distress at 3-month follow-up (Ganesalingam et al., 2008). Two studies of mild TBI in adults, one conducted over the first year postinjury and the other approximately 6 years postinjury, found negative correlations between HRQOL and number of PCS (Emanuelson, Andersson Holmkvist, Björklund, & Staalhammar, 2003; King & Kirwilliam, 2011). A third study examining recovery over the first year postinjury found that participants continued to report persistent PCS, but that HRQOL and everyday functioning improved quickly (Heitger, Jones, Frampton, Ardagh, & Anderson, 2007).
In the current study, we used data from a larger prospective, longitudinal project focused on the natural history of PCS following mild TBI (Yeates & Taylor, 2005) to elucidate the relationship between mild TBI, PCS, and HRQOL. We have previously shown that children with mild TBI demonstrate more PCS than children with orthopedic injuries (OI) and that the magnitude of group differences is related to the severity of injury (Taylor et al., 2010; Yeates et al., 2009). However, we have not examined group differences in HRQOL or the relationship between PCS and HRQOL. We hypothesized that children with mild TBI would display poorer HRQOL than children with OI. We also predicted that PCS would be a significant predictor of HRQOL even when controlling for group differences.
Participants were recruited by monitoring admissions for blunt head trauma or OI to the Emergency Departments of two Midwestern children's hospitals. Only children who were 8–15 years of age at the time of injury were enrolled. Younger children were excluded in order to ensure participants would be able to complete an MRI without requiring sedation. Children over the age of 15 years were not commonly seen at the children's hospitals where the study occurred, and instead are usually treated at other adult hospitals. Mild TBI was defined by the presence of one of the following: an observed loss of consciousness (LOC); a Glasgow Coma Scale (GCS; Teasdale & Jennet, 1974) score of 13 or 14; or at least two of the following acute signs or symptoms of concussion: posttraumatic amnesia, vomiting, nausea, headache, diplopia, dizziness, disorientation, or any other indications of mental status change (i.e., dazed, foggy, slow to respond, lethargic, confused, sleepy). Criteria for exclusion from the mild TBI group included an observed LOC lasting more than 30min, any GCS score of less than 13, any delayed neurological deterioration (i.e., a decline in GCS score to below 13 following admission or any emergent neurosurgical intervention), or any medical contraindication to MRI. Children were not excluded from the study if they demonstrated intracranial lesions or skull fractures, as long as they did not require neurosurgical intervention.
Children with OI were eligible to participate if they presented with fractures to the upper or lower extremities with an Abbreviated Injury Severity (AIS) score no higher than 3 (American Association for Automotive Medicine, 1990). The AIS is used to rate severity of injury to six anatomical regions (i.e., head/neck, face, chest/thorax, abdomen/pelvis, extremities, external) on a scale from 1 to 6 (from minor to maximal). Children were ineligible if they displayed any evidence of a head trauma or acute symptoms of concussion.
General exclusion criteria applying to both groups included any surgical interventions; any associated injury with an AIS score greater than 3; hypoxia, hypertension, or shock during or following the injury; injury resulting from child abuse or assault; injuries that would interfere with neuropsychological testing (e.g., fracture of preferred upper extremity); previous head injury requiring medical treatment; history of severe psychiatric illness resulting in hospitalization; or premorbid neurological disorders or intellectual disability.
Participation rates for those who met criteria were 47% and 35% for the mild TBI and OI groups, respectively. Participants and nonparticipants did not differ significantly in age, sex, or ethnic/racial minority status; they also did not differ in census tract measures of socioeconomic status (SES; i.e., mean family income, percentage of minority heads of household, and percentage of households below the poverty line). The final sample included 186 children with mild TBI and 99 with OI. The groups did not differ on age, sex, minority status, SES, or estimates of preinjury physical or psychosocial HRQOL (Table I). Children with mild TBI and those with OI did not differ in the severity of preinjury cognitive symptoms, although the mild TBI group had higher preinjury somatic symptoms. At an initial postinjury assessment, children with mild TBI evidenced more somatic and cognitive PCS than children in the OI group (see below descriptions of measures).
Institutional review board approval and informed parental consent and child assent were obtained prior to participation. An initial assessment typically occurred within 2 weeks of injury (M=11.35 days, SD=3.42), and follow-up assessments occurred at 3- and 12-months postinjury. All 285 families completed the initial visit, with 94% retention at the 3-month assessment and 89% of participants completing the 12-month follow-up. The proportion of children who completed all follow-up assessments did not differ by group. Children who dropped out of the study were more likely to be of minority ethnic status and were of lower SES than those who completed all assessments, but did not differ in age, sex, preinjury symptoms, preinjury quality of life, or PCS at the initial assessment.
Ratings of PCS were obtained via parent report at the initial assessment using the Health Behavior Inventory (HBI; Yeates et al., 1999). Parents initially rated preinjury symptoms retrospectively, before rating current symptoms. The HBI is a 50-item inventory rating of the frequency of occurrence of PCS using a 4-point Likert-type scale ranging from never to often. The HBI was developed from previous research on children with moderate and severe TBI (Barry, Taylor, Klein, & Yeates, 1996; Yeates et al., 2001), as well as mild TBI (Ganesalingam et al., 2008; Yeates et al., 1999). Factor analyses of the HBI yielded cognitive (11 items) and somatic (9 items) factors that are robust across raters and time (Ayr, Yeates, Taylor, & Browne, 2009). Both the cognitive and somatic summary scores demonstrated high internal consistency within both groups (Cronbach's α=.89–.95).
Parents also completed the 50-item Child Health Questionnaire (CHQ-PF50; Landgraf, Abetz, & Ware, 1996), which is a multidimensional generic measure of HRQOL. Parents rate children on 11 scales, which can be summarized to produce two standardized scales that measure physical and psychosocial HRQOL (M=50, SD=10). The CHQ-PF50 has been factor analyzed and reliability and validity are well established across a range of pediatric health conditions (McCarthy, 2007; Palermo et al., 2008). Higher scores indicate better HRQOL. The CHQ-PF50 was completed at the 3- and 12-month follow-up assessments. In addition, parents completed retrospective ratings of preinjury HRQOL at the initial assessment.
General linear mixed-model analysis, also known as hierarchical linear modeling (Leeuw & Meijer, 2008; Raudenbush & Bryk, 2002), was used to examine group differences in HRQOL across follow-up, as well as the relationship of postinjury PCS at the initial assessment to later HRQOL. Data analysis was conducted using the PROC MIXED procedure in SAS (SAS Institute, Cary, NC, USA). The mixed-model approach allowed us to use data from all participants, even those not seen at every assessment, and did not require equal intervals between assessments. Further, categorical variables can be included as predictors with this method. Separate analyses were conducted for physical and psychosocial HRQOL. Models included group and time (centered at the 3-month follow-up assessment) as predictors, in addition to sex, race (White vs. Nonwhite), and SES as covariates. Preinjury HRQOL was also included as a covariate in the analyses to control for individual differences prior to injury. Therefore, the dependent measures reflect changes in HRQOL relative to preinjury levels. Finally, both somatic and cognitive PCS measured at the initial assessment were included as predictors. Models initially included interactions between predictors (e.g., triple interactions of PCS, group, and time). Nonsignificant higher- and then lower order interactions were trimmed to identify the most parsimonious models and reduce risks of over-fitting. For all significant effects, the percentage of parameter variance explained by each effect was estimated by using the method described by Singer (1998).
In unconditional models (i.e., models without any predictors or covariates), the estimated mean intercepts at 3 months postinjury differed significantly from zero for both physical and psychosocial HRQOL. The time (or slope) parameter was not significantly different from zero in the analysis of physical HRQOL. For psychosocial HRQOL, the time parameter had an estimated mean different from zero (estimate=1.38; p=.03), indicating psychosocial HRQOL improved over time on average across the entire sample. The estimated true parameter variance was significant for both intercept and slope parameters. Thus, children displayed significant variance in HRQOL at 3 months postinjury and in the rate of change in HRQOL from 3 to 12 months, so that both intercept and slope parameters were included in conditional models of both types of HRQOL.
The analyses of physical HRQOL showed a significant group×time interaction, accounting for a small amount of variance (Table II). As shown in Figure 1, the mild TBI group had lower physical HRQOL at 12 months postinjury as compared to the OI group, after controlling for premorbid HRQOL. No interactions between PCS and group or time were significant. The trimmed model, displayed in Table II, accounted for 14.3% of the variance in postinjury ratings of physical HRQOL. Parents’ ratings of somatic PCS demonstrated a significant main effect, such that higher ratings of somatic PCS were associated with lower physical HRQOL at both 3 and 12 months postinjury. Preinjury ratings of physical HRQOL accounted for significant variance in postinjury ratings.
The analyses of psychosocial HRQOL did not reveal any main effect of group or group×time interaction. Further, no significant interactions between PCS and group or time were detected. The trimmed model accounted for 47.2% of the variance in postinjury ratings of psychosocial HRQOL (Table III). Both cognitive and somatic PCS were significant predictors across time in both groups, such that higher ratings of PCS predicted lower psychosocial HRQOL. Preinjury ratings of psychosocial HRQOL accounted for a substantial proportion of the variance in postinjury HRQOL.
Although the literature on HRQOL following pediatric TBI has increased in recent years (Limond et al., 2009; McCarthy et al., for the Children's Health After Trauma Study Group 2006; Petersen et al., 2008; Winthrop, 2010), relatively little research has focused on children with mild TBI. This may be in part because of past research suggesting that children with mild TBI do not demonstrate reliable impairments in HRQOL (Limond et al., 2009; Petersen et al., 2008). Previous research, however, has been characterized by a variety of methodological shortcomings. Using more rigorous methods, children in the current study with mild TBI showed deficits in physical HRQOL, but not psychosocial HRQOL, when compared to children with OI. These mixed findings are consistent with the previous literature, which fails to indicate a strong relationship between injury severity and HRQOL in children with TBI (McCarthy, 2007), and instead suggests that other mediating factors may determine HRQOL in this population.
A key goal of the current study, therefore, was to examine the relationship between PCS and HRQOL. The findings show that early postinjury PCS are a significant predictor of both physical and psychosocial HRQOL as long as 12 months postinjury, in both children with mild TBI and those with OI. Children whose parents reported more severe PCS were more likely to demonstrate lower HRQOL at both 3- and 12-month follow-up assessments. Further, the findings reflected some degree of specificity, such that only somatic PCS predicted physical HRQOL. Physical HRQOL tends to reflect bodily pain, physical functioning, and general health, so it is perhaps unsurprising that physical symptoms, but not cognitive symptoms, are related to physical HRQOL. On the other hand, both cognitive and somatic PCS had a negative relationship with psychosocial HRQOL. A possible explanation for the influence of somatic symptoms on psychosocial functioning is that children with more somatic symptoms like headaches, dizziness, and nausea may be less able to participate in rewarding activities, which results in less time spent in social interactions and has a negative influence on mood.
Thus, the findings indicate that mild TBI per se does not have a strong, persistent influence on HRQOL, but that children who experience higher levels of PCS after mild TBI are at risk for significant and lasting declines in HRQOL. This, in turn, suggests that PCS may be a potential target for improving HRQOL following pediatric mild TBI. Our previous work suggests that a substantial minority of children with mild TBI demonstrate persistent increases in PCS after mild TBI (Yeates et al., 2009), and outcomes could be improved by identification of and intervention with this subgroup. Several psychotherapeutic interventions have been developed to reduce PCS following mild TBI in adults and these techniques could be easily adapted for use with pediatric populations. Generally, the interventions include cognitive restructuring regarding symptom etiology and severity and the graded resumption of physical activity to minimize failure experiences and lower stress (Mittenberg, Canyock, Condit, & Patton, 2001). A review and meta-analysis by Mittenberg et al. (2001) examined the effectiveness of interventions targeting PCS reduction following mild TBI in adults. Studies showed consistent reductions in PCS following treatment (d=.32), even following early, single session treatments. Psychoeducational interventions may also be effective. In children with mild TBI, early education can reduce PCS and overall stress at 3-month follow-up (Ponsford et al., 2001). Future studies are needed to examine whether reduction in PCS as a function of intervention also promotes postinjury HRQOL.
Another important finding is that preinjury HRQOL was a stronger predictor of postinjury HRQOL scores than either mild TBI or PCS. For both physical and psychosocial HRQOL, preinjury ratings accounted for more variance in postinjury HRQOL than either group membership or level of PCS. These findings are in line with recent research documenting the significant influence of preinjury functioning on outcomes following mild TBI in children (Fay et al., 2010) and imply that clinicians should carefully evaluate preinjury functioning as part of case conceptualization and treatment planning. They also emphasize the need to assess preinjury functioning as soon after an injury as possible to reduce retrospective biases and obtain more accurate estimates. Within the context of the current study, we believe that the postinjury parent ratings are valid because they were obtained shortly after the injury, minimizing demands on retrospective recall. The mild TBI and OI groups were relatively comparable on retrospective measure of preinjury functioning, suggesting that parents in the mild TBI group were able to distinguish pre- and postinjury symptoms and were not minimizing any premorbid difficulties their children might have had.
One potential limitation of this study is the use of parent report measures for the assessment of both PCS and HRQOL. This not only raises the possibility of shared rater variance, but also omits children's self-reports. Parent–child agreement regarding frequency and severity of PCS is only moderate (Ayr et al., 2009; Gioia, Schneider, Vaughan, & Isquith, 2009; Hajek et al., 2011), as is parent–child agreement about HRQOL (Eiser & Morse, 2001c; Theunissen et al., 1998; Upton, Lawford, & Eiser, 2008). Therefore, children's and parents’ ratings of PCS and HRQOL may provide different perspectives. Future research should include self-report measures of PCS and HRQOL, to complement the parent ratings reported here.
Another potential limitation of the study is that recruitment rates for the mild TBI and OI groups were both below 50%, and may, therefore, limit the generalizability of the results. Relatively greater attrition of children of lower SES and ethnic minority status may also have reduced generalizability. However, nonparticipants did not differ from participants demographically, SES and minority status were taken into account in data analyses, and the mixed-model analyses incorporated all available data, including that collected from children who missed later assessments. A final caveat about the study is that the findings may not necessarily be applicable to younger children and older adolescents, given the age range of the sample. Future research is needed to determine if the results generalize to those age ranges.
In sum, mild TBI appears to have a delayed impact on physical HRQOL, but no effect on psychosocial HRQOL. However, higher levels of PCS shortly after injury predict lower physical and psychosocial HRQOL up to 12-months postinjury for children with mild TBI. Moreover, the relationship is somewhat specific, such that somatic symptoms predict physical HRQOL and both somatic and cognitive symptoms predict psychosocial HRQOL. These results suggest that intervention is indicated for children who display significant PCS following mild TBI, and that such interventions may result in improvements in HRQOL in children with such injuries.
National Institutes of Health (grants HD44099 and HD39834 to K.O.Y.).
Conflicts of interest: None declared.