Traumatic brain injury (TBI) in children has a relatively high rate of emergency department (ED) visits1 and risk for long-term adverse effects,2-5 creating a large public health concern. Traumatic brain injury in children can alter cognitive, language, and behavioral development, which can affect learning at school,6 and has also been linked to other comorbid health conditions, such as neurologic disorders (eg, motor difficulties and epilepsy), attention-deficit/hyperactivity disorder (ADHD), mental disorders, and sleep disorders.7 These conditions can further affect social and emotional development. In 2013, there were approximately 640 000 TBI-related ED visits, 18 000 TBI-related hospitalizations, and 1500 TBI-related deaths among children aged 14 years and younger.1 Children aged 0 to 4 years had the highest rates of TBI-related ED visits of any age group except those older than 75 years.1 The leading causes of TBI-related ED visits, hospitalizations, and deaths for individuals aged 0 to 14 years were unintentional falls and being struck by or against an object, whereas for those aged 15 to 24 years, the leading causes were motor vehicle crashes and falls.1 Another common cause of TBI is sports and recreational activity, which accounted for an estimated 325 000 TBI ED visits among children and adolescents in 2012.8
Mild TBI, a brief change in mental status or consciousness that is commonly called concussion,9 accounts for 70% to 90% of TBI ED visits.10,11 Children with mild TBI or concussions may seek care at clinical locations other than the ED or may not seek care, making it difficult to accurately estimate the true incidence, which is a critical factor for understanding the total public health outcome of TBI in children.10,12,13 More recent research examining point of entry for an initial visit for concussion care found that in a large, urban, pediatric health care system, 48.8% of children with TBI visited primary care; 27.2%, specialty care; and 20.2%, ED care.14 Because most pediatric TBIs are mild, only a fraction of children seen for emergency medical care for a TBI are hospitalized—a metric that has been used as an indicator of injury severity.15 In a cohort study reporting severity in children seeking emergency medical care from hospitals (N = 2940), 84.5% had mild TBI, 13.2% had moderate TBI, and 2.3% had severe TBI.16 Traumatic brain injury hospitalizations for children have decreased in recent years.1 Because most incidence reports of mild TBI are based on ED care, current estimates may be underestimating the outcome of pediatric TBI.
Traumatic brain injury in children has been linked to other childhood health conditions, including ADHD,17-20 seizures,21,22 behavior problems,23-26 mental disorders,27 learning problems,2,23,28 and hearing problems.29,30 Furthermore, children who have sustained a TBI tend to use speech and language services after their injury, indicating concerns about acquired cognitive and communication disorders.4,31 There are a number of health outcomes associated with TBI in children, supporting the need for more detailed estimates of conditions associated with TBI.
We currently have an understanding of the incidence of TBI from estimates of TBI-related ED visits and hospitalizations among children. However, these data do not allow for estimates of the number of children seen in other clinical settings or those who do not receive care and provide limited data on comorbid health conditions. The purpose of this study was to estimate both the national and state-specific prevalence of lifetime TBI based on parents’ report of a health care professional diagnosis of brain injury or concussion. Additional aims were to describe the association between TBI and other health conditions in this population and examine associations between state-level prevalence of TBI and state-level insurance coverage.
Data were analyzed from the 2011-2012 National Survey of Children’s Health32 (NSCH), conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC) as a module of the State and Local Area Integrated Telephone Survey, to determine a lifetime TBI history. Data analysis was conducted from February 1, 2016, through November 1, 2017. The NCHS Research Ethics Review Board and the institutional review board of NORC at the University of Chicago approved all study procedures and modifications. Participants provided verbal informed consent.
The NSCH is a cross-sectional telephone survey of US households using a random digit dial sample of landlines and cell phones designed to allow for both national and state estimates of children’s health.33 A parent or guardian answered questions in relation to 1 randomly selected child (N = 95 677) aged 0 through 17 years. The interview completion rate (the proportion of households who completed the interview among households known to include children) was 54.1% (landline) and 41.2% (cell phone). The overall response rate—including all nonresponses, such as households that were never successfully screened for the presence of children and telephone lines that rang with no answer—was 23.0%.
For these analyses, TBI was defined by a positive response to the question: “Has a doctor or other health care provider ever told you that [your child] had…a brain injury or concussion?” A help screen was provided for the interviewer that further defined a concussion and indicated that developmental conditions, such as autism, cerebral palsy, or brain tumors, should not be considered as brain injuries. Questions framed in a similar manner were asked for other health conditions, for example, ADHD, hearing problems, anxiety and behavioral or conduct problems, and epilepsy or a seizure disorder, to indicate a child’s lifetime history of the specific condition. The household respondent also reported sociodemographic characteristics including the child’s sex, age, race, Hispanic ethnicity, household income (later categorized to reflect the federal poverty level), health insurance status (private, public, or uninsured), and highest educational level achieved by the parent/adult respondent living with the child. Within the survey, parents were asked questions related to whether their child had health insurance, the type of insurance, and their perception of the adequacy of their child’s current health insurance. Data are weighted to represent all noninstitutionalized children living in the United States.
The weighted prevalence of a lifetime history of TBI was calculated by using SUDAAN’s crosstab procedure with 95% CIs derived from Taylor linearization. In addition, prevalence estimates were stratified by sociodemographic characteristics. Differences in TBI prevalence estimates by sociodemographic group were tested using the Wald F test. We also examined the trend in TBI prevalence by age using the stratum-adjusted Cochran-Mantel-Haenszel trend test (Wald χ2 = 257.2490, P < .001).
Crude TBI prevalence rate ratios were estimated by using logistic regression. To examine the association between TBI and specific health conditions, the prevalence of each health condition was calculated among children with and without TBI; prevalence rate ratios were adjusted by age, sex, race/ethnicity, and household income in logistic regression models. To reduce the outcome of different age distributions, age-adjusted lifetime TBI prevalence estimates were computed for each state. Logistic regression models were used to compute the odds ratios (ORs) for the associations between TBI prevalence by state with private health insurance type and with parent report of health care adequacy separately. In the logistic regression models, TBI prevalence rates were grouped hierarchically in 4 groups, and an individual’s private insurance type (private insurance type, 1/0) or report of adequate insurance (insurance adequate, 1/0) was modeled as an outcome variable as a function of state TBI prevalence (the lowest TBI category was used as the reference level).
The relative SE, a measure of an estimate’s statistical reliability (100 times SE of an estimate divided by the estimate), was calculated for each estimate and cell sizes reviewed. All relative SEs were less than 30%, and all sample sizes were greater than 20. SAS, version 9.3 (SAS Institute Inc) and SAS-callable SUDAAN, version 11.0 (Research Triangle Institute) were used for statistical analysis. All analyses were performed by incorporating complex sample design features (including stratified sampling and weighting for unequal sample selection probabilities and nonresponse). Statistical significance of all 2-sided, unpaired P values was set at .05.
Prevalence of TBI in Children
Among US children aged 0 to 17 years, the lifetime estimate of a parent-reported TBI diagnosis (ie, ever being told by a health care professional that the child had a brain injury or concussion) was 2.5% (95% CI, 2.3%-2.7%); this figure represents over 1.8 million US children (Table 1). Parent report of TBI in this survey was more common among children who were non-Hispanic white, boys, and those with private health insurance (Table 1). A trend test showed that TBI prevalence (percentage) increased with increasing age: 0 to 4 years (0.6%; 95% CI, 0.5%-1.0%), 5 to 11 years (1.7; 95% CI, 1.4%-2.0%), 12 to 14 years (3.9; 95% CI, 3.3%-4.5%), and 15 to 17 years (5.9; 95% CI, 5.3%-6.6%) (P < .01).
TBI and Childhood Health Conditions
After adjustment for sociodemographic characteristics of age, sex, parent educational level, and household income, children with a lifetime history of TBI were more likely to have received a parent-reported diagnosis of 13 of the 14 health conditions examined compared with children without a history of TBI (Table 2). For example, prevalence ratios ranged from 1.6 for ADHD to 7.3 for epilepsy or seizure disorder. Additional mental disorders (depression, anxiety problems, or behavioral and conduct problems), neurologic conditions (Tourette syndrome, epilepsy), chronic health conditions (hearing, vision), and intellectual disability were also significantly more likely to be reported by parents of children with a history of TBI. The conditions most commonly reported among children with a history of TBI were learning disorders (21%), ADHD (20%), speech and language problems (19%), developmental delay (15%), and bone, joint, or muscle problems (14%). Although learning disorders, ADHD, and speech and language problems were also the conditions most commonly reported by parents of children without a history of TBI, their prevalence was lower than in the TBI group.
Lifetime History of TBI by State
State estimates of parent-reported TBI ranged from an age-adjusted prevalence of 1.20% in Mississippi to 5.30% in Maine (Table 3, Figure 1). In examining health insurance type, states with a higher prevalence of childhood TBI were more likely to also have higher estimates of private health insurance compared with public health insurance (OR, 1.36; 95% CI, 1.27-1.46). Examples of states with higher proportions of private insurance include Maine, Vermont, Pennsylvania, Washington, Montana, Wyoming, North Dakota, South Dakota, Colorado, and Michigan. States with a higher prevalence of childhood TBI also had higher estimates of parent-reported adequate insurance compared with those reporting less-adequate insurance (OR, 1.18; 95% CI, 1.16-1.36) (Figure 2).
Based on parent report of a health care professional’s diagnosis among a nationally representative sample in 2011-2012, 2.5% of US children aged 0 to 17 years have experienced a TBI in their lifetime. To our knowledge, this is the first lifetime estimate across the developmental age span for children. Previous national estimates were based on ED visit rates or a report of visit location for concussion in a large pediatric health care system.14 In 2013, children aged 0 to 4 years (1591 per 100 000 population) had the highest rate of pediatric ED visits followed by children in the age 15- to 24-year population (1080.7 per 100 000 population).1
The estimate provided by this study does not capture children who did not receive medical care for their injury, so it is likely an underestimate of childhood TBI in the United States. In addition, the type of clinician and the location of the health care visit where the TBI was diagnosed was not included as part of the survey. Consequently, this study cannot provide information on where parents sought treatment for a TBI sustained by their children.
Findings from this nationally representative sample indicate a higher occurrence of other health and developmental conditions in children who experienced a lifetime TBI, including learning disorders, ADHD, speech and language problems, developmental delay, anxiety, depression, and behavior problems. Although many of these findings align with previous research reporting on the association of TBI with other childhood health conditions,3,17-26,34 to our knowledge, this is the first report to examine associations in a national sample and include speech and language conditions and developmental delay. Previous reports have provided prevalence data on these specific conditions but have not considered an association with TBI. The presence of these conditions may be a compounding factor affecting the likelihood of experiencing cognitive, social, and health challenges following TBI. However, due to the cross-sectional nature of the NSCH, we cannot distinguish whether the conditions occurred before or after the TBI and therefore cannot determine whether these conditions contributed to or were a consequence of the TBI. Knowing the association of other childhood health conditions with TBI suggests that health care professionals should inquire about TBI history at the time of diagnosis of childhood health conditions and ask parents about a child’s total health history at the time of TBI diagnosis. Understanding a more comprehensive picture of a child’s health status at the time of TBI diagnosis facilitates optimal management to improve recovery and outcomes.
The map (Figure 1) illustrates differences in state estimates of the age-adjusted prevalence of parent-reported TBI diagnosed by a health care professional. Estimates of uninsured populations in 2011-2012 were previously reported by region rather than state, with higher estimates in the South (18.6%) and West (17.0%) compared with the Northeast (10.8%) and Midwest (11.9%).35 Recent research indicates that insurance status among children with TBI is associated with better health outcomes. Children with TBI who had private health insurance had lower mortality rates and better quality of care following TBI than those with public insurance or those who were uninsured,36 suggesting an association between insurance type and health care quality for children with TBI. In this study, individuals in states with higher levels of TBI were more likely to report private insurance and adequate insurance coverage, suggesting that insurance coverage may explain some of the differences in lifetime TBI estimates found between states.
Some studies of pediatric TBI report high rates of private insurance,14 suggesting that coverage by this insurance type may support seeking care at the time of the TBI. Also, parents who believe that they have adequate insurance may be willing to seek care. However, further investigation is needed to determine additional factors related to seeking care for TBI in children, including parents’ views on the need for a medical assessment of TBI.
This study has limitations. The NSCH relies on parent report of health care professionals’ diagnosed brain injury or concussion, other health conditions and insurance, and does not examine medical records. Health care professionals’ diagnoses were inferred by parents’ response to survey questions; however, these estimates may have been affected by difficulties in recall as well as any challenges related to communication of the diagnosis between the health care professional and the parent/guardian. This survey did not capture those who experienced a TBI but did not seek a medical assessment by a health care professional. In addition, other types of acquired brain trauma, such as anoxic brain injuries (ie, due to lack of oxygen), may have been included in affirmative responses provided by respondents resulting from question wording (ie, a brain injury or concussion). The response rate for this survey was low, increasing the likelihood that the results were influenced by factors associated with the decision to participate. However, the data were weighted to adjust for nonresponse in an effort to account for differences between the sample and population.
The brain injury question includes help text informing parents not to consider developmental conditions, such as cerebral palsy and autism; however, it is possible that some parents missed that instruction and incorrectly included those conditions as brain injuries. To the extent that this misunderstanding occurred, it would increase the report of brain injuries and contribute to the observed association between autism and brain injury. Directionality between TBI and associated health conditions cannot be inferred owing to the cross-sectional nature of the study. Parent report of health insurance indicated current status, so it is possible that one’s insurance status was different at the time of the injury.
Children of all backgrounds may be affected by TBI in their lifetime, highlighting the importance of inquiring about a history of TBI during well-child health care visits. The combination of TBI and the health conditions associated with a TBI can have a significant outcome on a child’s overall health, learning, and behavior. Although it is unclear whether the conditions existed prior to or are a consequence of the TBI injury, improved care for children can be better achieved if pediatric health care professionals offer medical guidance to parents in the context of a child’s overall health history, including history of lifetime TBI. For children with diagnosed TBI, health care professionals can assess for conditions identified in this analysis. To produce more comprehensive estimates of TBI in children, nonmedical data sources will need to be expanded to capture children who do not or cannot seek treatment. A proposed system, the National Concussion Surveillance System, holds the potential for obtaining more comprehensive prevalence estimates of TBI in children.37