Abstract
Introduction:
This study assessed 1) differences in mental health service use among those with and without TBI, 2) whether social supports, employment, and insurance moderated the TBI-mental health service use relationship, and 3) racial and ethnic differences in these associations.
Methods:
In an analytic sample of 1,018 participants, we used propensity score matching and inverse probability of treatment weighting to assess differences in both lifetime and past 12-month mental health service use among those with and without TBI. We used a logistic regression model for the moderation analyses.
Results:
Those with TBI reported greater depressive, anxiety, and PTSD symptoms than those without TBI, as well as higher probability of both lifetime and past 12-month mental health service use. After adjusting for the propensity score, those with TBI were about 10.5% more likely to use mental health services compared to those without TBI. Higher levels of social support were associated with a three-fold increase in the odds of using mental health services in individuals with TBI. Employment and insurance did not significantly moderate the TBI-mental health service use relationship, but insurance was independently associated with grater odds of both lifetime and past 12-month mental health service use.
Conclusion:
Greater psychiatric symptoms in individuals with TBI, along with their increased use of mental health services, highlight the need for a multidisciplinary approach to TBI care. Interventions could also target social relationships, as lower social support represents potential pathways for differences in mental health service use for those with TBI.
Keywords: Traumatic brain injury, Psychiatric disorders, Mental Health Service Use, Social Support, Employment
Introduction
Traumatic brain injury (TBI) is a significant cause of mortality and morbidity worldwide, with many adverse health outcomes such as physical disability, cognitive impairment, psychiatric disorders, and increased healthcare utilization.[1] The sequelae of TBI occurs across all levels of TBI severity, including mild TBI, and have been documented in both the short and long term. Psychiatric disorders are among the most common sequelae, also observed across the spectrum of TBI severity: up to 77% of TBI patients receive a psychiatric diagnosis in the first year post-injury,[2] an increased risk that remains for over a decade.[3] At the same time, psychiatric disorders can precede and co-occur alongside TBI and prolong recovery trajectories[4]. This comorbidity can be associated with greater health services utilization,[1] increasing the individual and societal burden of TBI. However, the interplay between TBI and psychiatric disorders—and how this interaction affects mental health service utilization—is less well studied. This is important to consider given current shortages in the mental health workforce and potential for increased strain on an already overburdened system.
Social drivers of health, like social support and employment, that facilitate access to care and specialty services, have also received insufficient attention. Both social support and employment represent potential facilitators and barriers to healthcare services for TBI and concurrent conditions, consistent with a social determinants of health framework that recognizes non-medical factors as critical to health outcomes, including TBI[5]. Prior qualitative work has identified a lack of social support and a fragmented structure of care as barriers to receiving needed care following TBI.[6] At the same time, TBI sequelae can challenge interpersonal relationships. Employment can facilitate access to employment-based insurance and, thereby, appropriate referrals and treatment following TBI, with prior studies having identified insurance as playing a mediating role in TBI rehabilitation racial and ethnic disparities.[7] Considering these social determinants of health therefore represent potential modifiable factors and opportunities for intervention that may shape the trajectory of care following TBI. However, there are also racial and ethnic disparities in return to employment and employment instability following TBI[5,8]. With racial and ethnic disparities identified across the continuum of TBI care, as well as in receipt of mental health services, considering variations across groups is important for informing whether targeted interventions are needed[8]. Notably, existing studies typically involve less racially and ethnically diverse samples, as well as clinical rather than community-based samples, which limits current understanding of TBI care[9].
We attempt to fill the literature gap by using a diverse population to ask the following questions: 1) what are differences in mental health service use among those with and without TBI, 2) what is the moderating role of social supports, employment, and insurance status on the TBI-mental health service use relationship, and 3) whether there are any racial and ethnic differences in these associations. We hypothesized that there would be increased mental health service use among all groups with TBI compared to those without, and that increased social support and employment would attenuate the relationship between TBI and mental health service use. We hypothesized these relationships because greater social support can help address mental health needs, and employment would facilitate community integration and could attenuate mental health needs. We hypothesized that these differences would be greater among racial and ethnic minoritized populations, with lesser mental health service use among minoritized groups due to decreased access and greater moderation by social support and employment.
Materials and Methods
Participants and Procedures
Data came from the Strong Minds-Strong Communities randomized controlled trial (RCT), which recruited adult participants from the community in two distinct geographic locations within the US, North Carolina and Massachusetts.[10] After assessing their capacity to consent, potential participants were screened for eligibility, which included being 18 years or older; being fluent in English, Spanish, Mandarin, or Cantonese; and having moderate to severe depression or anxiety symptoms, which were assessed using the CAT-MH, a suite of validated computer adaptive mental health tests.[11–13] Potential participants were ineligible if they reported recent psychotherapy or counseling; reported a history of severe psychosis, mania, or psychotic symptoms; or had severe alcohol or substance dependence, assessed using the CAT-MH. Participants who disclosed a suicidal plan or attempt on the Paykel Suicide Risk Questionnaire[14] were entered into an emergency protocol and re-screened for eligibility after 30 days.
After the screening process, a total of 1,044 participants who were eligible and completed a baseline assessment were randomized 1:1 to either the intervention (n = 524) or the control groups (n = 520). Randomization was stratified by site using a block randomization scheme of size two. Participants in the intervention group received an evidence-based psychoeducational intervention for up to 6 months. The intervention included ten one-hour weekly sessions, delivered individually in the participant’s language by a trained community health worker , and incorporated approaches from cognitive behavioral therapy, mindfulness, healthy habits, motivational interviewing, and behavioral activation. Participants in the control group, characterized as enhanced usual care, received an NIH booklet about anxiety and depression and were contacted 4 times by a care manager in the first six months (mimicking the administration schedule in the intervention group). Additional details regarding the intervention and control groups can be found elsewhere.[10] The Institutional Review Board approved the study procedures of all participating institutions.
In the current study, only data from the baseline assessment was employed (n = 1,044). We excluded ten participants who self-reported their race as American Indian (n = 3), mixed (n = 5), or another race (n = 2) because of the small sample size of these participants. Sixteen additional participants were excluded because of missing data in at least one of the variables included in the current study (except for employment status, as described in the measures section below). This resulted in a final analytical sample of 1,018 participants who reported their race and ethnicity as either non-Latino White (n = 90), non-Latino Black (n = 149), Asian (n = 136), or Latino (n = 643) and who had no missing data at baseline. No significant differences were observed between baseline participants excluded because of missing data (n = 16) and our final analytical sample (n = 1,018).
Measures
Outcome: Lifetime and Past 12-Month Mental Health Service Use
Lifetime mental health service use was assessed using an affirmative response to the question, “Have you ever received mental health services (such as seeing a mental health professional to discuss your depression, anxiety, or mental health concerns, receiving prescriptions or medicine, and/or hospitalizations for mental health concerns)?” Past 12-month mental health service use was assessed using affirmative responses to five items about whether a participant, in the last 12 months, (1) has had appointments with their primary care physician to discuss their depression, anxiety, or mental health; (2) has had appointments with a psychologist, counselor, or therapist (not including any appointments that were part of the RCT)?; (3) has had appointments with a psychiatrist?; (4) has had emergency room visits for their depression, anxiety, or mental health?; and (5) has had hospitalizations because of their depression, anxiety, or any mental health problems?
Exposure: Possible Lifetime Traumatic Brain Injury (TBI)
Lifetime traumatic brain injury was assessed using the HELPS TBI Screening Tool,[15] a widely used tool to determine possible TBI. HELPS is an acronym for H = Have you ever Hit your Head or been Hit on the Head?; E = Were you ever seen in the Emergency room, hospital, or by a doctor because of a head injury?; L = Did you Lose consciousness?; P = Are you having cognitive or social Problems in your daily life (such as headaches, mood changes, or difficulty concentrating since hitting your head)?; S = Did you experience a significant Sickness following your head injury? A HELPS screen is considered positive for a possible TBI when (1) an event occurred that could have caused a brain injury (yes to H, E, or S); (2) there is an indication that the injury was severe by either losing consciousness or being seen in the emergency room (yes to E or L); and (3) there was a presence of chronic problems as listed under P. As this is a screening tool and not a clinical diagnosis of TBI, this represents “possible TBI” rather than a definitive diagnosis of TBI. For simplicity, we refer to those who screen positive for a possible TBI as being in the “TBI” group hereafter but acknowledge the limitations of the screening tool.
Moderators: Social Support, Employment Status, and Insurance Status
Current social support was measured using a 7-item measure adapted from Thoits.[16] The first three items assessed the frequency with which participants talk to their family, spouse/partner, or friends when they have a problem (1 ‘never’ to 4 ‘always’). The fourth item asked participants how often they told family or friends when they were sad (1 ‘never’ to 4 ‘always’). Two more items assessed the frequency of getting together with family or friends they don’t live with and how often family members took care of them (1 ‘once a week’ to 5 ‘never’). The last item involved rating their level of satisfaction with social support received in their lives (1 ‘very satisfied’ to 4 ‘very unsatisfied’). When appropriate, responses were reverse coded so higher values were better. All items were then rescaled to a 4-point scale ranging from 0 to 3 and a final score was computed by averaging all items such that higher scores represent increased levels of current social support.
Employment status was assessed using participant self-reported information about their current employment: employed (part/full-time employment or self-employment), unemployed (not employed and seeking employment), or not in the labor force (not employed and not seeking employment). An item about employment status was added to the baseline survey after data collection had started, and about 29% of our final analytical sample had missing data in this variable. However, as shown in Table S2 of the Online Supplement, we did not observe significant differences between baseline participants with and without missing data regarding their employment status. This made it plausible to assume that the missing data mechanism behind employment status was random and not associated with other variables in the present study.
Insurance status was assessed using participant self-reported information about their current health insurance coverage. Responses were used to categorize participants as either insured (private or public health insurance) or uninsured (no health insurance).
Statistical Analysis
Inverse Probability Weighting (IPW) Using a Propensity Score
Participants with a HELPS screening considered positive for a possible TBI and participants without a positive HELPS screening might differ in unobserved characteristics that make them difficult to compare. We applied propensity score methods to create a comparison group that can provide an appropriate counterfactual for the outcome trends of participants with a possible TBI in our study. Specifically, we used an Inverse Probability Weighting (IPW) estimator using a propensity score (which we denote ) to match the TBI and non-TBI groups based on several matching variables. First, we estimated the propensity of being in the TBI group using baseline data on depressive symptoms (scores on the Patient Health Questionnaire-9, PHQ-9), anxiety symptoms (scores on the Generalized Anxiety Disorder-7, GAD-7), posttraumatic stress disorder (PTSD) symptoms (scores on the PTSD Checklist for DSM-5, PCL-5), age, female sex, having a high school diploma, being married, and race and ethnicity (non-Latino White, non-Latino Black, Asian, and Latino). We then used the estimated propensity score, , to weight observations in each group (TBI and non-TBI) to account for selection assignment differences. A weight equal to was used for each observation in the TBI group, and a weight equal to was used to weight each observation in the non-TBI group. This strategy aimed to turn participants in the non-TBI group into a representative sample of participants in the TBI group concerning the distribution of the baseline variables used to estimate the propensity score.
Mental Health Service Use and TBI Status
We began by comparing the unweighted distribution of all the variables in the present study between the TBI and non-TBI groups. Using IPW, we then estimated the weighted difference in the probability of lifetime and past 12-month mental health service use between the TBI and non-TBI groups. As described above, this weighted difference was estimated by applying a weight equal to for each observation in the TBI group, and a weight equal to for each observation in the non-TBI group. After, we examined whether social support, employment status (being employed), and insurance status (being insured) moderated the weighted difference in the probability of lifetime and past 12-month mental health service use between the TBI and non-TBI groups. Moderation analyses were conducted using a logistic regression model where the outcome was either lifetime or past 12-month mental health service use. The main independent variables were TBI status, either social support, employment status, or insurance status, and their two-way interaction. As before, these moderation analyses applied a weight of for each observation in the TBI group, and a weight of for each observation in the non-TBI group. Finally, we conducted post-hoc subgroup analyses by race and ethnicity. Given the small sample size of the TBI × race/ethnicity and the TBI × moderator × race/ethnicity interactions, these post-hoc subgroup analyses were exploratory. All analyses and descriptive statistics were performed in the Stata software version 17.
Results
Sample Characteristics
Table 1 compares the unweighted distribution of all variables included in the present study (outcomes, moderators, and matching variables) between participants in the TBI and non-TBI groups. Within our sample, 152 (14.9%) had a positive HELPS screening for a possible TBI. As shown in Table 1, participants in the TBI group had a higher probability of lifetime and past 12-month mental health service use than participants in the non-TBI group (53.9% versus 40.2% and 34.9% versus 24.6%, respectively). Compared to participants in the non-TBI group, participants in the TBI group also had lower levels of social support and higher levels of depressive, anxiety, and PTSD symptoms. Participants in the TBI group were also less likely to be married (30.9% versus 40.9%), more likely to identify as either non-Latino White or non-Latino Black, and less likely to identify as Asian or Latino compared to participants in the non-TBI group. In the last column of Table 1, we also present the P-value for the comparison in the distribution of matching variables between the TBI and non-TBI groups after IPW. After adjusting for the estimated propensity score, all significant differences between the TBI and non-TBI groups in the matching variables were no longer observed, which suggested that IPW plausibly effectively balanced these two groups.
Table 1. Unweighted distribution of baseline variables between the TBI and non-TBI groups in the final analytical sample (N = 1,018).
| Baseline Variable | non-TBI group (n = 866) |
TBI group (n = 152) |
P
|
|
|---|---|---|---|---|
| Without IPW | With IPW | |||
| Outcomes | ||||
| Mental Health Service Use, Lifetime, N (%) | ||||
| No | 518 (59.8%) | 70 (46.1%) | 0.002 | NA |
| Yes | 348 (40.2%) | 82 (53.9%) | ||
| Mental Health Service Use, Past 12 Months, N (%) | ||||
| No | 653 (75.4%) | 99 (65.1%) | 0.013 | NA |
| Yes | 213 (24.6%) | 53 (34.9%) | ||
| Moderators | ||||
| Social Support, Mean (SD) | 0.9 (0.6) | 0.7 (0.5) | 0.007 | NA |
| Employment status, N (%) | ||||
| Not Employed or Not in the Labor Force | 300 (34.6%) | 68 (44.7%) | 0.629 | NA |
| Employed | 308 (35.6%) | 51 (33.6%) | ||
| Missing* | 258 (29.8%) | 33 (21.7%) | ||
| Insurance, N (%) | ||||
| No | 295 (34.1%) | 56 (36.8%) | 0.512 | NA |
| Yes | 571 (65.9%) | 96 (63.2%) | ||
| Matching Variables | ||||
| Depression, PHQ-9, Mean (SD) | 9.3 (5.7) | 11.9 (5.9) | < 0.001 | 0.479 |
| Anxiety, GAD-7, Mean (SD) | 8.1 (5.2) | 10.4 (4.8) | < 0.001 | 0.384 |
| PTSD, PCL-5, Mean (SD) | 21.1 (15.3) | 30.5 (17.8) | < 0.001 | 0.864 |
| Age, Mean (SD) | 42.4 (13.2) | 43.7 (13.1) | 0.244 | 0.576 |
| Sex, N (%) | ||||
| Male | 135 (15.6%) | 32 (21.1%) | 0.122 | 0.691 |
| Female | 731 (84.4%) | 120 (78.9%) | ||
| Education, N (%) | ||||
| Less Than High School | 328 (37.9%) | 56 (36.8%) | 0.808 | 0.974 |
| High School or Above | 538 (62.1%) | 96 (63.2%) | ||
| Married, N (%) | ||||
| No | 512 (59.1%) | 105 (69.1%) | 0.015 | 0.712 |
| Yes | 354 (40.9%) | 47 (30.9%) | ||
| Race/Ethnicity, N (%) | ||||
| Non-Latino White | 71 (8.2%) | 19 (12.5%) | < 0.001 | 0.489 |
| Non-Latino Black | 113 (13.0%) | 36 (23.7%) | ||
| Asian | 125 (14.4%) | 11 (7.2%) | ||
| Latino | 557 (64.3%) | 86 (56.6%) | ||
Notes: NA, Not Applicable
Mental Health Service Use by TBI Status
The weighted difference in the probability of lifetime and past 12-month mental health service use between the TBI and non-TBI groups is presented in Table 2. As shown in Table 2, after adjusting for the propensity score, lifetime mental health service use in the non-TBI group was 41.1% compared to 50.9% in the TBI group, and this difference was statistically significant (difference = 9.8 percentage points, P = 0.032). That is, participants in the TBI group were, on average, 9.8 percentage points more likely to use mental health services ever in their life compared to participants in the non-TBI group. We also observed that past 12-month mental health service use was lower in the non-TBI group than in the TBI group (25.0% versus 31.0%), but this difference was not statistically significant.
Table 2. Weighted difference in the probability of lifetime and past 12-month mental health service use between the TBI and non-TBI groups in the overall sample and racial and ethnic subgroups.
| Outcome | Probability of use | Difference | Std. Err. | 95%CI | P | |
|---|---|---|---|---|---|---|
| non-TBI group | TBI group | |||||
| Lifetime Service Use | 0.411 | 0.509 | 0.098 | 0.05 | [0.01, 0.19] | 0.032 |
| Non-Latino White | 0.763 | 0.898 | 0.135 | 0.08 | [−0.02, 0.29] | 0.083 |
| Non-Latino Black | 0.537 | 0.702 | 0.166 | 0.08 | [0.01, 0.32] | 0.041 |
| Asian | 0.573 | 0.888 | 0.315 | 0.09 | [0.14, 0.49] | < 0.001 |
| Latino | 0.303 | 0.396 | 0.093 | 0.06 | [−0.02, 0.21] | 0.106 |
| Past 12-Month Service Use | 0.250 | 0.310 | 0.060 | 0.04 | [−0.02, 0.14] | 0.145 |
| Non-Latino White | 0.592 | 0.764 | 0.172 | 0.12 | [−0.06, 0.40] | 0.139 |
| Non-Latino Black | 0.269 | 0.317 | 0.047 | 0.09 | [−0.12, 0.22] | 0.589 |
| Asian | 0.498 | 0.847 | 0.349 | 0.13 | [0.10, 0.59] | 0.005 |
| Latino | 0.149 | 0.219 | 0.070 | 0.05 | [−0.02, 0.16] | 0.141 |
In post-hoc exploratory subgroup analyses, lifetime mental health service use in the non-TBI group was lower than in the TBI group among both non-Latino Black (53.7% versus 70.2%, P = 0.041) and Asian participants (57.3% versus 88.8%, P < 0.001) after adjusting for the propensity score. Further, past 12-month mental health service use was also lower in the non-TBI group (49.8%) compared to the TBI group (84.7%) among Asian participants (P = 0.005).
The Role of Social Support, Employment Status, and Insurance Status
In Table 3, we present the results of our analyses examining whether social support, employment status (being employed), and insurance status (being insured) moderated the weighted difference in the probability of lifetime and past 12-month mental health service use between the TBI and non-TBI groups. As shown in Table 3, social support significantly moderated the difference in lifetime mental health service use. In particular, higher levels of social support were associated with increased odds of lifetime mental health service use among participants in the TBI group compared to participants in the non-TBI group (Odds Ratio [OR] = 3.00; 95% CI = [1.59, 5.69]; P < 0.001). In Figure 1, we provide a graphical representation of this moderation effect by plotting the probability of lifetime mental health service use in the TBI and non-TBI groups for different levels of social support: low (mean – 1SD), average, and high (mean + 1SD). As displayed in this figure, at low levels of social support (mean – 1SD), participants in the TBI group were slightly less likely to report using mental health services ever in their lives compared to participants in the non-TBI group. In contrast, at high levels of social support (mean + 1SD), participants in the TBI group were almost twice as likely to report using mental health services ever in their lives compared to participants in the non-TBI group. No other significant moderation effects were observed (Table 3). That is, neither employment nor insurance status moderated the weighted difference in the probability of lifetime and past 12-month mental health service use between the TBI and non-TBI groups.
Table 3. Social support, employment status, and insurance status as moderators of the weighted difference in the probability of lifetime and past 12-month mental health service use between the TBI and non-TBI groups in the overall sample and racial and ethnic subgroups.
| Moderator | Outcome: Lifetime mental health service use | ||||
| Total sample OR [95% CI] |
Non-Latino White OR [95% CI] |
Non-Latino Black OR [95% CI] |
Asian OR [95% CI] |
Latino 95% CI |
|
|
| |||||
| Social Support | |||||
| TBI | 0.61 [0.33, 1.15] | 4.30 [0.57, 32.48] | 2.89 [0.55, 15.11] | 0.59 [0.09, 3.91] | 0.45 [0.18, 1.10] |
| Social Support | 0.77 [0.61, 0.97]* | 0.60 [0.21, 1.74] | 0.64 [0.31, 1.31] | 1.38 [0.67, 2.85] | 0.81 [0.60, 1.10] |
| TBI × Social Support | 3.00 [1.59, 5.69]* | 0.57 [0.10, 3.15] | 0.80 [0.16, 4.16] | 4.96 [0.18, 134.69] | 4.23 [1.86, 9.63]* |
| Employment Status | |||||
| TBI | 1.41 [0.77, 2.56] | 1.27 [0.19, 8.36] | 3.71 [0.84, 16.40] | 0.63 [0.12, 3.25] | 1.28 [0.54, 3.05] |
| Employed | 1.23 [0.89, 1.71] | 0.72 [0.19, 2.65] | 1.15 [0.47, 2.84] | 0.96 [0.42, 2.20] | 1.18 [0.76, 1.85] |
| TBI × Employed | 1.35 [0.55, 3.35] | NA1 | 0.53 [0.06, 4.43] | 2.09 [0.15, 29.57] | 1.68 [0.49, 5.69] |
| Insurance Status | |||||
| TBI | 0.98 [0.51, 1.90] | 4.24 [0.37, 48.46] | 1.36 [0.33, 5.66] | 1.17 [0.31, 4.39] | 0.71 [0.29, 1.72] |
| Insured | 1.95 [1.44, 2.63]* | 5.48 [1.59, 18.86]* | 1.06 [0.44, 2.51] | NA2 | 1.32 [0.91, 1.91] |
| TBI × Insured | 2.02 [0.88, 4.65] | 0.58 [0.02, 15.35] | 2.32 [0.37, 14.51] | NA2 | 3.57 [1.16, 10.96] |
|
| |||||
| Moderator | Outcome: Past 12-month service use | ||||
| Total sample OR [95% CI] |
Non-Latino White OR [95% CI] |
Non-Latino Black OR [95% CI] |
Asian OR [95% CI] |
Latino 95% CI |
|
|
| |||||
| Social Support | |||||
| TBI | 0.76 [0.37, 1.55] | 0.53 [0.08, 3.60] | 0.50 [0.08, 2.99] | 0.62 [0.09, 4.34] | 0.84 [0.29, 2.40] |
| Social Support | 0.76 [0.57, 1.01] | 0.30 [0.09, 0.97]* | 0.62 [0.25, 1.53] | 1.55 [0.76, 3.15] | 0.84 [0.55, 1.28] |
| TBI × Social Support | 2.00 [0.94, 4.29] | 5.17 [0.65, 40.96] | 3.68 [0.58, 23.41] | 5.12 [0.23, 115.11] | 1.92 [0.70, 5.25] |
| Employment Status | |||||
| TBI | 1.62 [0.87, 3.03] | 2.46 [0.50, 12.05] | 7.06 [1.41, 35.30] | 0.50 [0.10, 2.60] | 1.57 [0.58, 4.27] |
| Employed | 1.69 [1.16, 2.46] | 1.64 [0.50, 5.42] | 4.51 [1.19, 17.13] | 0.80 [0.36, 1.79] | 1.74 [0.96, 3.15] |
| TBI × Employed | 1.09 [0.43, 2.76] | 4.57 [0.28, 74.93] | 0.17 [0.02, 1.49] | 4.23 [0.30, 59.76] | 1.23 [0.31, 4.87] |
| Insurance Status | |||||
| TBI | 1.04 [0.43, 2.50] | 1.71 [0.20, 14.41] | 2.03 [0.35, 11.76] | 1.25 [0.33, 4.75] | 0.66 [0.18, 2.37] |
| Insured | 3.38 [2.27, 5.03]* | 3.98 [1.18, 13.41]* | 2.78 [0.87, 8.90] | NA2 | 1.90 [1.15, 3.14] |
| TBI × Insured | 1.52 [0.55, 4.19] | 2.32 [0.15, 35.20] | 0.70 [0.09, 5.33] | NA2 | 3.17 [0.72, 13.86] |
Notes:
P < 0.05
The moderating effect of employment status in the association between TBI and lifetime mental health service use among non-Latino White participants was not identified. Everyone in the TBI group who was employed reported lifetime mental health service use; thus TBI × employed × non-Latino White predicted lifetime mental health service use perfectly.
Both the main and the moderating effect of insurance status were not identified among Asian participants because all but three were insured, and the three Asian participants who were uninsured did not report either lifetime or past 12-month mental health service use (i.e., uninsured predicted no service use perfectly).
Figure 1. Weighted probability of lifetime mental health service use the in TBI and non-TBI groups for different levels of social support.
In post-hoc exploratory subgroup analyses, higher levels of social support were associated with increased odds of lifetime mental health service use among Latino participants in the TBI group compared to Latino participants in the non-TBI group (OR = 4.23; 95% CI = [1.86, 9.63]; P < 0.001). No other TBI × moderator × race/ethnicity interactions were observed (Table 3). Nevertheless, we noted that the moderating effect of employment status among non-Latino White participants was not identified because everyone in the TBI group who was employed reported lifetime mental health service use (i.e., TBI × employed × non-Latino White predicted lifetime mental health service use perfectly). Further, both the main and the moderating effect of insurance status were not identified among Asian participants because all but three were insured, and the three Asian participants who were uninsured did not report either lifetime or past 12-month mental health service use (i.e., uninsured predicted no service use perfectly).
Discussion
Findings from the present study suggest that those with TBI reported higher levels of depressive, anxiety, and PTSD symptoms, as well as a higher probability of both lifetime and past 12-month mental health service use than those without TBI. Participants in the TBI group were more likely to identify as non-Latino White or non-Latino Black and less likely to identify as Asian or Latino, and they reported lower levels of social support and were less likely to be married. This suggests one potential pathway for this population’s differences in mental health service use. We found that higher levels of social support were associated with three-fold increased odds of lifetime mental health service use among participants in the TBI group. These findings suggest that mental health and social support may be key pathways through which sequelae of TBI may persist, representing opportunities for intervention to impact the trajectory of care following TBI. In fact, studies have conceptualized TBI as a “chronic disease” with a range of problems that can persist after TBI more than 10 years following injury[17], necessitating a long-term view of post-TBI care. We did not find a significant moderating effect of employment or insurance on the TBI-mental health service use relationship, although insurance was independently associated with higher odds of both lifetime and past 12-month mental health service use, highlighting it as a potential target for interventions to improve access to mental health care more broadly.
The findings of this study are significant because they emphasize the high prevalence of mental illness comorbidity among those with TBI and how this must be screened for and integrated into the treatment plan among clinicians who manage TBI. Mental health disorders are common in many neurological disorders,[18] but especially TBI. This relationship could be due to shared biological risk or social risk factors that can interact.[19] The subsequent high prevalence of mental health service use is also consistent with prior studies, such as one national Canadian study that showed mental health care utilization as 60% higher in TBI patients compared with those who are not injured after adjusting for various sociodemographic factors.[20] The high prevalence of psychiatric disorders and mental health service use among TBI patients highlights the need for a multidisciplinary approach to their management. Previous qualitative studies with individuals who have experienced brain injury have emphasized the importance of integrated health services for improving care.[6]
However, the provision of integrated health services is not widely implemented, particularly in the overlap of psychiatric and neurological conditions. This could be due to individual and practice-level factors. At an individual level, studies indicate that neurologists often feel uncomfortable assessing psychiatric conditions in their neurological patients, resulting in under-detection or inadequate management of these conditions. One study assessing referral patterns across 33 European psychiatric consultation-liaison services found that referrals from neurology were characterized by an underestimation of psychiatric comorbidity and a late detection, comparable to other specialties.[21] Other studies have replicated this finding[22]. At the practice level, this lack of detection is compounded by a lack of partnerships between psychiatrists and neurologists, including poor training in each discipline.[23] Future steps could include the expansion of available psychotherapies for TBI, such as work showing how current treatments for psychiatric conditions such as depression can be effectively modified for the TBI population.[24]
We found racial and ethnic differences in TBI diagnoses among Latino and Asian participants. This could, in part, be related to disparities in diagnosis. One study among veterans found that Latino and Asian veterans did not differ from white veteran in screening positive for TBI but had 19% and 33% lower odds of diagnosis.[25] Although disparities have been identified across the continuum of TBI care, this remains an area of continued study. Prior research has also identified that Spanish-speaking Latinos tend to have more misconceptions about TBI compared to English-speaking Latinos and non-Latino whites,[26] raising the possibility that more mild injuries may go undetected and untreated. There could also be factors like stigma, English proficiency, or lower health literacy that interfere with people’s health-seeking behavior and obtaining a TBI diagnosis[27]. These participants may also have lower access to specialized services to obtain a diagnosis. At the same time, there may be differences in TBI risk profiles in different populations. Altogether, social drivers of health are patterned by race and ethnicity and influence whether and how a person with TBI receives appropriate diagnosis and treatment.
Another important finding of this study is the significant moderating effect of social support, with higher levels of social support associated with 3-times greater odds of lifetime mental health service use among participants with TBI compared to participants without TBI. Greater social support may facilitate tangible supports like transportation or financial help or even facilitate behavioral activation to encourage someone with TBI to seek care. Multiple prior studies have identified social supports as critical for recovery and adaptation to TBI,[28,29] and lack thereof, a barrier to TBI services.[6,29] Interventions could, therefore, include programs with family involvement[30–33] or peer mentoring programs that are helpful.[34] Future studies should explore the extent and dose of social supports needed to see benefit in TBI recovery.
We did not find the same moderating effect in employment, although there could be differences based on type of employment-accommodations, paid time leave, and other employment resources, which we did not capture in our study. For example, qualitative studies have identified that those who could change their work environment to accommodate brain-injury-related needs or had employers or colleagues who cared about their needs saw improved recovery long-term.[29] Future studies could investigate employment nuances and their role on mental health services use and TBI recovery. Notably, insurance status was independently associated with greater odds of both lifetime and past 12-month mental health service use. This highlights the importance of expanding coverage for mental health services more broadly.
Limitations
This study is not without limitations. First, due to the inclusion criteria for the RCT sample that comprised our study sample, individuals were more likely to have anxiety and depression but not have been treated recently for it. This may reflect potential confounders not captured in the study, such as differences in internal coping strategies, although we sought to reduce the impact of baseline differences in depression and anxiety through our propensity-matching design. Second, although the HELPS screening is a validated TBI screening instrument, it has not been validated in Spanish or compared with other screening measures.[35] It also does not replace a formal diagnosis of TBI, with positive screening corresponding to “possible” TBI rather than definitive TBI. Third, we did not capture other medical comorbidities that may influence mental health service use. Other studies have identified TBI of any severity as associated with higher risk of chronic cardiovascular, endocrine, and neurological comorbidities above and beyond psychiatric conditions.[3] This represents an area of future study to inform multidisciplinary screening and management of multiple diseases after TBI. Fourth, although the sample size of our study is not insignificant, larger studies involving diverse populations and longitudinal studies could better elucidate the nature of the relationships we found. For example, we were most likely underpowered to detect small racial and ethnic differences as noted in our post-hoc exploratory analysis. Further, because the screening instrument is a lifetime measure of TBI, it is plausible that individuals with more mental health challenges are at increased risk of TBI. Similarly, the time frame for the social support measure is current, meaning it is plausible that the moderation could be interpreted as people who have experienced TBI having more supportive social relationships when they have received mental health care. These issues of reverse causality can be addressed in future research leveraging longitudinal, prospective cohorts with screening instruments that capture the timing of TBI and social support in the same time frame.
Conclusion
The higher levels of depressive, anxiety, and PTSD symptoms among those with TBI, as well as their higher probability of both lifetime and past 12-month mental health service use, underscores the necessity of a multidisciplinary approach to care to facilitate timely and accurate diagnosis and treatment. The decreased rates of TBI and mental health diagnoses among Asian and Latino groups suggests that further research and efforts could be dedicated to increasing screening in these populations, among a broader commitment to addressing and promoting social drivers to health like social supports and insurance coverage for mental health service use. Understanding the relationship between TBI and mental health service use, including the influence of social drivers, facilitates appropriate targeting of health care resources and interventions to facilitate optimal care for patients with TBI. This is especially important in a strained healthcare system, where TBI already represents a significant resource burden.
Supplementary Material
Highlights:
Individuals with TBI had greater depressive, anxiety, and PTSD symptoms than those without TBI.
Individuals with TBI had higher probability of mental health service use.
Greater social support was associated with greater odds of mental health service use.
Employment did not significantly moderate the TBI-mental health service use relationship.
Acknowledgements
AS would like to thank Rodrigo Nadal and Marine Nimblette for her research assistance with this manuscript. They were compensated for their time as a paid research coordinator.
Funding:
This work was supported by the National Institutes of Health/National Institute of Mental Health grant number R01MH117247. Altaf Saadi is supported in part by grant number K23NS128164 from the National Institute of Neurological Disorders and Stroke.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declaration of Interest Statement
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- [1].To MJ, O’Brien K, Palepu A, Hubley AM, Farrell S, Aubry T, et al. Healthcare Utilization, Legal Incidents, and Victimization Following Traumatic Brain Injury in Homeless and Vulnerably Housed Individuals: A Prospective Cohort Study. Journal of Head Trauma Rehabilitation 2015;30:270–6. 10.1097/HTR.0000000000000044. [DOI] [PubMed] [Google Scholar]
- [2].Alway Y, Gould KR, Johnston L, McKenzie D, Ponsford J. A prospective examination of Axis I psychiatric disorders in the first 5 years following moderate to severe traumatic brain injury. Psychol Med 2016;46:1331–41. 10.1017/S0033291715002986. [DOI] [PubMed] [Google Scholar]
- [3].Izzy S, Chen PM, Tahir Z, Grashow R, Radmanesh F, Cote DJ, et al. Association of Traumatic Brain Injury With the Risk of Developing Chronic Cardiovascular, Endocrine, Neurological, and Psychiatric Disorders. JAMA Netw Open 2022;5:e229478. 10.1001/jamanetworkopen.2022.9478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Silverberg ND, Iverson GL. Etiology of the post-concussion syndrome: Physiogenesis and psychogenesis revisited. NRE 2011;29:317–29. 10.3233/NRE-2011-0708. [DOI] [PubMed] [Google Scholar]
- [5].Taylor S, Brayan K, Storch B, Suh Y, Walsh S, Avrith N, et al. Association Between Social Determinants of Health and Traumatic Brain Injury: A Scoping Review. Journal of Neurotrauma 2024;41:1494–508. 10.1089/neu.2023.0517. [DOI] [PubMed] [Google Scholar]
- [6].Grewal J, Kennedy CJ, Mamman R, Biagioni JB, Garcia-Barrera MA, Schmidt J. Understanding the barriers and facilitators of healthcare services for brain injury and concurrent mental health and substance use issues: a qualitative study. BMC Health Serv Res 2024;24:881. 10.1186/s12913-024-11316-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Gao S, Kumar RG, Wisniewski SR, Fabio A. Disparities in Health Care Utilization of Adults With Traumatic Brain Injuries Are Related to Insurance, Race, and Ethnicity: A Systematic Review. Journal of Head Trauma Rehabilitation 2018;33:E40–50. 10.1097/HTR.0000000000000338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Saadi A, Bannon SM, Watson E, Vranceanu A-M. Racial and Ethnic Disparities Associated with Traumatic Brain Injury across the Continuum of Care: A Narrative Review and Directions for Future Research. Journal of Racial and Ethnic Health Disparities n.d.;In Press. [DOI] [PubMed] [Google Scholar]
- [9].Cook NE, Gaudet CE, Kissinger-Knox A, Liu BC, Hunter AA, Norman MA, et al. Race, ethnicity, and clinical outcome following sport-related concussion: a systematic review. Front Neurol 2023;14:1110539. 10.3389/fneur.2023.1110539. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Alegría M, Cruz-Gonzalez M, Markle SL, Falgas-Bague I, Poindexter C, Stein GL, et al. Referrals to Community and State Agencies to Address Social Determinants of Health for Improving Mental Health, Functioning, and Quality of Care Outcomes for Diverse Adults. Am J Public Health 2024;114:S278–88. 10.2105/AJPH.2023.307442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Gibbons RD, Weiss DJ, Pilkonis PA, Frank E, Moore T, Kim JB, et al. Development of a Computerized Adaptive Test for Depression. Arch Gen Psychiatry 2012;69:1104. 10.1001/archgenpsychiatry.2012.14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Gibbons RD, Hooker G, Finkelman MD, Weiss DJ, Pilkonis PA, Frank E, et al. The Computerized Adaptive Diagnostic Test for Major Depressive Disorder (CAD-MDD): A Screening Tool for Depression. J Clin Psychiatry 2013;74:669–74. 10.4088/JCP.12m08338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Gibbons RD, Weiss DJ, Pilkonis PA, Frank E, Moore T, Kim JB, et al. Development of the CAT-ANX: A Computerized Adaptive Test for Anxiety. AJP 2014;171:187–94. 10.1176/appi.ajp.2013.13020178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Paykel ES, Myers JK, Lindenthal JJ, Tanner J. Suicidal Feelings in the General Population: A Prevalence Study. Br J Psychiatry 1974;124:460–9. 10.1192/bjp.124.5.460. [DOI] [PubMed] [Google Scholar]
- [15].Picard M, Scarisbrick D, Paluck R. HELPS: a brief screening device for traumatic brain injury. New York: Comprehensive Regional Traumatic Brain Injury Rehabiliation Center; 1991. [Google Scholar]
- [16].Thoits PA. Stress, Coping, and Social Support Processes: Where Are We? What Next? Journal of Health and Social Behavior 1995;35:53. 10.2307/2626957. [DOI] [PubMed] [Google Scholar]
- [17].Wilson L, Stewart W, Dams-O’Connor K, Diaz-Arrastia R, Horton L, Menon DK, et al. The chronic and evolving neurological consequences of traumatic brain injury. The Lancet Neurology 2017;16:813–25. 10.1016/S1474-4422(17)30279-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Carson AJ. Neurological disease, emotional disorder, and disability: they are related: a study of 300 consecutive new referrals to a neurology outpatient department. Journal of Neurology, Neurosurgery & Psychiatry 2000;68:202–6. 10.1136/jnnp.68.2.202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Li LM, Carson A, Dams-O’Connor K. Psychiatric sequelae of traumatic brain injury — future directions in research. Nat Rev Neurol 2023;19:556–71. 10.1038/s41582-023-00853-8. [DOI] [PubMed] [Google Scholar]
- [20].Kureshi N, Clarke DB, Feng C. Association between traumatic brain injury and mental health care utilization: evidence from the Canadian Community Health Survey. Inj Epidemiol 2023;10:16. 10.1186/s40621-023-00424-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Jonge PD, Huyse FJ, Herzog T, Lobo A, Malt U, Opmeer BC, et al. Referral pattern of neurological patients to psychiatric Consultation-Liaison Services in 33 European hospitals. General Hospital Psychiatry 2001;23:152–7. 10.1016/S0163-8343(01)00137-2. [DOI] [PubMed] [Google Scholar]
- [22].Dawood S, Poole N, Fung R, Agrawal N. Neurologists’ detection and recognition of mental disorder in a tertiary in-patient neurological unit. BJPsych Bull 2018;42:19–23. 10.1192/bjb.2017.7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Kanner AM. Is it time to train neurologists in the management of mood and anxiety disorders? Epilepsy & Behavior 2014;34:139–43. 10.1016/j.yebeh.2014.02.029. [DOI] [PubMed] [Google Scholar]
- [24].Gallagher M, McLeod HJ, McMillan TM. A systematic review of recommended modifications of CBT for people with cognitive impairments following brain injury. Neuropsychological Rehabilitation 2019;29:1–21. 10.1080/09602011.2016.1258367. [DOI] [PubMed] [Google Scholar]
- [25].Kysar-Moon A, Mustillo S. Race/Ethnicity and Traumatic Brain Injury: Are There Disparities in Positive Screenings and Diagnoses Among Service Members Returning From Afghanistan and Iraq? Armed Forces & Society 2019;45:155–76. 10.1177/0095327X18761852. [DOI] [Google Scholar]
- [26].Pappadis MR, Sander AM, Struchen MA, Leung P, Smith DW. Common Misconceptions About Traumatic Brain Injury Among Ethnic Minorities With TBI. Journal of Head Trauma Rehabilitation 2011;26:301–11. 10.1097/HTR.0b013e3181e7832b. [DOI] [PubMed] [Google Scholar]
- [27].Iverson GL, Maietta JE, Saadi A, Cook NE. Social Determinants of Health and Lifetime History of Parent-Reported Concussion in School-Aged Children and Adolescents in the United States. Neurotrauma Rep 2025;6:20–31. 10.1089/neur.2024.0083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Lequerica A, Edmond A. The Importance of Social Determinants of Health in Predicting Healthcare Access after Traumatic Brain Injury. Archives of Physical Medicine and Rehabilitation 2024;105:e109. 10.1016/j.apmr.2024.02.307. [DOI] [Google Scholar]
- [29].Fadyl JK, Theadom A, Channon A, McPherson KM. Recovery and adaptation after traumatic brain injury in New Zealand: Longitudinal qualitative findings over the first two years. Neuropsychological Rehabilitation 2019;29:1095–112. 10.1080/09602011.2017.1364653. [DOI] [PubMed] [Google Scholar]
- [30].Dam AEH, De Vugt ME, Klinkenberg IPM, Verhey FRJ, Van Boxtel MPJ. A systematic review of social support interventions for caregivers of people with dementia: Are they doing what they promise? Maturitas 2016;85:117–30. 10.1016/j.maturitas.2015.12.008. [DOI] [PubMed] [Google Scholar]
- [31].Couchman G, McMahon G, Kelly A, Ponsford J. A new kind of normal: Qualitative accounts of Multifamily Group Therapy for acquired brain injury. Neuropsychological Rehabilitation 2014;24:809–32. 10.1080/09602011.2014.912957. [DOI] [PubMed] [Google Scholar]
- [32].Foster AM, Armstrong J, Buckley A, Sherry J, Young T, Foliaki S, et al. Encouraging family engagement in the rehabilitation process: a rehabilitation provider’s development of support strategies for family members of people with traumatic brain injury. Disability and Rehabilitation 2012;34:1855–62. 10.3109/09638288.2012.670028. [DOI] [PubMed] [Google Scholar]
- [33].Kreutzer JS, Marwitz JH, Godwin EE, Arango-Lasprilla JC. Practical Approaches to Effective Family Intervention After Brain Injury. Journal of Head Trauma Rehabilitation 2010;25:113–20. 10.1097/HTR.0b013e3181cf0712. [DOI] [PubMed] [Google Scholar]
- [34].Hibbard MR, Cantor J, Charatz H, Rosenthal R, Ashman T, Gundersen N, et al. Peer Support in the Community: Initial Findings of a Mentoring Program for Individuals with Traumatic Brain Injury and Their Families. Journal of Head Trauma Rehabilitation 2002;17:112–31. 10.1097/00001199-200204000-00004. [DOI] [PubMed] [Google Scholar]
- [35].Lequerica AH, Arango-Lasprilla JC, Krch D, Corrigan JD. Assessing lifetime exposure to traumatic brain injury among an international sample of Spanish speakers. NRE 2021;48:109–17. 10.3233/NRE-201539. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.