Traumatic brain injuries (TBI) are injuries caused by a bump, blow, or jolt to the head, or penetration of an object that disrupts the normal function of the brain (1). Mild TBIs, commonly referred to as concussions, and TBIs of all severities have gained public health attention in recent years because of advances in the understanding of the impact of TBIs on future health outcomes. Although many TBI symptoms resolve over time and most individuals have a complete recovery, TBI is associated with long-term impairment for various health and social outcomes (2–4). According to the Centers for Disease Control and Prevention (CDC), there were over 214,000 TBI-related hospitalizations in 2020 (5). The highest rates of TBIs are among children, adolescents, and adults 75 years and older (6). Among children, motor vehicle crashes and unintentional falls are the leading causes of nonfatal TBI-related hospitalizations (7).

In the United States, national public health surveillance systems for TBI are lacking and existing data sources have inherent limitations. Administrative sources like hospital discharge data may underestimate TBI prevalence and incidence as they only capture those who receive healthcare, Data suggest that a significant portion, potentially up to half, of adults with suspected TBI do not seek medical attention for their injury (8). In addition, children and youth can receive care in a variety of settings that are not typically captured in administrative data such as from pediatrician offices, urgent care, or evaluation during organized recreational activities by trained medical professionals (9, 10).

Nationally representative population-based surveys are an additional data source used to monitor the prevalence of TBI among youth and adults. Unlike administrative data, surveys can capture injuries of varying severity and are not dependent on whether an individual used healthcare services. Further, as in the case of the National Health Interview Survey (NHIS), multiple survey questions can be used and measure experiences like TBI symptoms in conjunction with medical evaluation and subsequent diagnoses. Surveys are also useful as they can simultaneously collect a wealth of other health and demographic measures, which can be used to evaluate risk and protective factors at the population level. However, surveys face challenges due to the lack of standardized methodology for measuring TBI. Estimates vary depending on factors like questionnaire wording and context (11). Accurately estimating the prevalence of TBI helps to assess the need for services, treatments, and interventions. Monitoring prevalence over time may be useful for researchers interested in exploring the impact of prevention programs.

Prevalence estimates of TBI using survey data have been shown to vary by reporter type (e.g., parent-report or self-report) (12). Previous work comparing survey data from different sources has found that youth self-reported estimates of concussions are 2–3 times higher than that of parent-report, suggesting the value of incorporating self-reported data (12). Among youth with a history of diagnosed concussion, agreement between youth and parent dyads varied considerably according to symptomology, with high concordance among externalizing symptoms (e.g., behavioral problems) and poorest concordance among internalizing symptoms (e.g., anxiety, depression) (13–15).

Understandably among older youth, parents may underreport due to a lack of awareness about the event and the youth not communicating injuries to their parents. This may be particularly of concern among youth athletes where there may be internal and external pressures to not disclose (e.g., to continue participating in their sport), and therefore they may choose to not share this information with others (16). Less is known about injuries sustained outside of sports. Previous work comparing population-level prevalence estimates using surveys has been limited in its ability to fully understand differences by reporting source, as surveys have considerable variation in methodology, question phrasing, and participant eligibility (e.g., including youth ages >18 vs <18) (12). Since many surveys rely on parent-report, it is critical to understand if the estimates obtained via these methods are accurate.

There is limited research available to determine whether proxy reporting accurately reflects youths’ TBI experiences within a general population-based health survey. In particular, there is a need to understand how responses from youth and their parents compare when using similar survey methodology and identical questions. For this reason, we aimed to expand on this body of literature by examining nationally representative data from parent-report paired with youth-reported data to understand the concordance of TBI symptoms, TBI evaluation, and TBI diagnosis of teens aged 12–17 years old. Prevalence estimates for TBI outcomes by sociodemographic characteristics and experiences (e.g., participation in sports) are presented by reporter type (self-report vs. parent-report).Additionally, agreement between teen and parent survey responses for each TBI outcome was examined.

Using pairs of teens and their parents, our findings indicated that the prevalence of TBI symptoms is nearly three times higher when using teen self-report rather than parent-report. Further, our findings revealed poor to moderate concordance between parent-reported and teen self-reported TBI symptoms. Conversely, lifetime prevalence of TBI evaluation and TBI diagnosis produced similar estimates and had substantial agreement when comparing parent and teen self-report.

There is the potential for discordance in injury and symptom reporting between parent-report and youth self-report because youths may sustain injuries and experience symptoms of which their parents are not aware. Past research on youth-parent agreement of concussion symptoms among those who sustained injuries during sport play have found moderate to substantial agreement (31, 32). Since our sample was not limited to sports participants, but represents experiences of injuries in any setting and the recall period was lifetime, this may explain why symptom agreement was not as strong as in previous work (31). Higher concordance among samples of sports participants may be the result of increased knowledge of concussion due to past experience with the condition and thus being able to recognize symptoms as part of the injury. Athletes and their parents often have access to coaches or trainers that may provide concussion guidance information and therefore may be better reporters than the general population of youth (33).

Other research has also demonstrated that agreement between parent and youth report of severity of post-concussive symptoms (due to any cause including sports) were modest (13) with youth reporting higher mean levels of symptoms severity than parents (15). Findings from the current study underscore the utility for collection of teen self-report of TBI symptoms for understanding patterns in youth if possible. For more observable and objective measures of health, parent-report yields similar prevalence estimates. In particular, our results have wider utility as they are nationally representative of teens, and better inform the nature of parent-reported indicators of their child’s injuries.

It is challenging to compare concordance findings across studies due to a wide variation of questions used to assess undiagnosed TBI experiences. Concordance assessments may be based on a single question, series of questions, or instruments (e.g., Sports Concussion Assessment Tool Version 2), and this may also contribute to observed differences between studies (34). Additionally, variations in study populations, such as age groups or specific populations (e.g., athletes), can further complicate the comparison of concordance results across studies. A lack of standardized methodologic approaches for measuring TBI using surveys may lead to inconsistent measurement across settings and populations (11). The current study’s use of a survey of a national population sample to quantify the burden of TBI in the United States underscores the role reporter type may also play in producing significant differences for the prevalence of TBI symptoms, but not reports of healthcare utilization or diagnosis.

Results from this study represent injuries that occurred across the lifetime.

Time since injury also could have impacted agreement observed, suggesting over time, parents may be unable to recall symptoms while teens may have greater self-awareness of their symptoms (31). Among young children (aged <6 years), parent recall of injuries in general has been shown to decrease with time (35). Given that the recall period in this study was lifetime, it is possible that a shorter recall period may have revealed better agreement between reporters.

Despite poor concordance on TBI symptoms, and higher estimates of TBI symptoms from teens compared with parents, parent-report of evaluation and diagnosis produced similar prevalence estimates to that of teen self-report, with a few exceptions. Analysis revealed significant differences in the reported prevalences for TBI evaluation among some subgroups. Still, in these instances, teens reported a higher prevalence of having an evaluation compared with parents. Previous research on youth athletes has shown underreporting of concussions due to reasons like not wanting to upset parents (36, 37). It is possible that these evaluations are happening when the parent is not present, and the results of the evaluation did not warrant sharing the information with the parent. Post-hoc analysis found that around 80% of discordant teens (those where the teen reported being evaluated but the parent reported no evaluation) later reported not being diagnosed with TBI. Variations in reporting evaluation may also be attributed to differences in the underlying population participating in sports, which is more likely to be teens who are non-Hispanic White and have high family income (38). Regardless, evaluation and diagnosis are more objective measures than symptoms of TBI, measuring interactions with healthcare professionals and subsequent diagnosis that parents are likely involved in. Teens experiencing head injuries are likely unable to access healthcare settings on their own. For athletes, best practices for concussion response plans outlined for athletic trainers includes engaging parents (39). Parent perspectives may provide accurate information regarding healthcare utilization patterns and subsequent diagnosis for TBI. Further research may be helpful to understand the underlying reasons for the variations among teens and parents, especially in subgroups with significant discordance.NHIS-Teen is a nationally representative sample of teens living in the United States. Strengths of the data include linkage to parent-reported measures of health, healthcare utilization, and detailed demographics (e.g., health insurance type, family income) to understand population health by detailed subgroups. TBI outcomes examined in this study utilized identical survey questions that were reported by parents and their teens. This study highlights that estimates can vary depending on the reporter, even when using the same questions.

Despite these strengths, this study is not without limitations. The period of assessment for TBI outcomes was lifelong, whereas many sociodemographic indicators (e.g., health insurance and FPL), as well as sports participation were based on the past 12 months. All data are self-reported, or parent-reported, and therefore may be subject to recall bias. Additionally, parent-reported data are collected in the presence of an interviewer, and respondents may be reluctant to mention an injury to an interviewer. These reports are subject to social desirability bias, particularly if the circumstances where the injury implicates the parent (e.g., physical abuse) which could contribute to underestimating the prevalence of TBI. In contrast, teen surveys were self-administered via the web in the absence of an interviewer and may be less likely to be influenced by social desirability. Furthermore, the questions used in the current study have not been evaluated for psychometric properties nor compared with external sources (e.g., medical records) to assess validity. Measures were based on self-report (or parent-report) and therefore are based on the judgement and interpretation of the individual. NHIS-Teen sample size limited the ability to produce stable estimates among detailed race and ethnicity subgroups. Last, because NHIS-Teen is a general health survey, data are lacking on explanatory covariates like injury severity, mechanism of injury, and intrapersonal characteristics between parent and teen (e.g. attachment style). This limits the assessment of recall or concordance.

Parent-report for youth TBI symptoms is discordant with youths’ experiences. However, measures of TBI evaluation and diagnosis showed substantial agreement between youth and parent-report. Obtaining accurate prevalence estimates of TBI are important to inform programs designed to prevent and decrease TBI at the population level. These findings suggest that teen report of TBI symptoms may enhance surveillance.