For decades, suicide has been one of the leading causes of death [1,2] and a major health concern among veterans in the United States [3]. The most recent Department of Veterans Affairs (VA) Office of Mental Health and Suicide Prevention’s National Veteran Suicide Prevention Annual Report stated that the annual suicide rate for veterans (31.6 per 100,000) was twice the rate of that for nonveteran US adults (16.8 per 100,000) [4].

Suicide prevention is a top clinical priority for the VA and its nationally integrated health care system, overseen by the Veterans Health Administration (VHA) [4]. Central to this effort is the identification of individuals at risk for suicide or suicidal behavior [5,6]. A substantial body of research has used electronic health record data to identify individuals likely to make suicide attempts and other forms of intentional self-harm [7-10]. Research has shown that adults and youths with a history of nonfatal self-harm are at a markedly elevated risk of suicide [11-13].

Other important risk factors for suicidal behavior include sociodemographic, psychological, and physical health characteristics [14-17]. In particular, two health conditions relatively common among veterans—traumatic brain injury (TBI) and posttraumatic stress disorder (PTSD)—have been identified in multiple studies as risk factors for suicide [2,18-20]. For example, Bullman and Kang [21] reported that Vietnam veterans with PTSD were at elevated risk of suicide death when compared with those without PTSD. Another study by Jakupcak et al [22] observed similar patterns for post-9/11 veterans, in which veterans with PTSD were reported to have a four times greater likelihood of suicidal ideation as compared to veterans without PTSD. In a 2011 study, Brenner et al [18] reported that PTSD was associated with 2.8 times greater odds of a documented suicide attempt compared to veterans without PTSD; the authors suggested that future analyses compare those with and without comorbid TBI history. In another study, Brenner et al [20] conducted an analysis to examine the associations between TBI diagnosis and suicide among individuals receiving care at VHA and reported that veterans with TBI were 1.55 times more likely to complete suicide as compared to those without TBI. The mechanisms by which TBI or PTSD lead to an elevated risk of suicide or suicidal behaviors are not well understood, but the associations are consistent with constructs of risk described in the interpersonal theory of suicide (ie, thwarted belongingness, perceived burdensomeness, and acquired capability) [23]. Notably, individuals with TBI or PTSD often have co-occurring physical health (eg, chronic pain), mental health (eg, depression and anxiety), and behavioral conditions (eg, substance use disorder) that can lead to functional limitations and other vulnerabilities (eg, unemployment and homelessness) [24-28], which together or individually, can elevate the factors comprising each of these constructs [29].

Previous studies examining associations between TBI and PTSD diagnoses and suicidal behavior have mostly been conducted with relatively small or select samples or with veterans from a single war era [2,18,19]. These studies have suggested that TBI and PTSD are important risk factors for suicidal behavior. Moreover, previous studies have tended to only analyze the associations of TBI [30] and PTSD [31] separately with suicidal behavior and have not examined the potential interactive effects of comorbid TBI and PTSD.

In this study, we analyzed data from approximately 6.7 million veterans who received VHA health care between 2008 and 2017 to examine population-level variations in the prevalence of intentional self-harm among those with or without prior TBI and PTSD diagnosis. We focused on suicide attempts as well as other intentional self-harm events because previous research [11,12] has shown that adults and youths were at a markedly elevated risk of suicide completion after nonfatal self-harm events. Our analysis was stratified by level of VHA utilization to account for differential access to VHA care and the opportunity to receive the diagnoses of interest. Our objectives were to evaluate variation in the prevalence of intentional self-harm by veterans’ prior TBI and PTSD diagnosis status across different VHA utilization levels, and to examine the odds of veterans’ intentional self-harm by prior TBI and PTSD diagnosis status across different VHA utilization levels while controlling for potentially confounding variables. The overarching goal of this work is to inform suicide and other intentional self-harm prevention efforts.

Our analysis of approximately 6.7 million VHA-using veterans found that those with a documented diagnosis of TBI or PTSD were significantly more likely to have subsequent VHA documentation of a suicide attempt or other form of intentional self-harm relative to those with neither diagnosis between 2008 and 2017. For all three VHA utilization groups, the odds of intentional self-harm for veterans in the comorbid TBI/PTSD group were more than three times as high as those with neither condition, even after controlling for multiple potential confounders. For those in the PTSD only group, the adjusted odds of intentional self-harm were approximately three times that of those with neither condition. For the TBI only group, the adjusted odds of intentional self-harm were nearly double relative to those with neither condition. For all three VHA utilization groups, the prevalence of intentional self-harm for veterans in the comorbid TBI/PTSD group was higher than in the TBI only and PTSD only groups.

For the medium and low VHA utilization groups, adjusted odds were highest for the PTSD only group closely followed by the comorbid TBI/PTSD group. However, for the high VHA utilization group, adjusted odds for the comorbid TBI/PTSD group were consistently higher than those of the PTSD only group. Previous studies have suggested that symptom severity may play an important role in this association [35,36]. Our study did not consider symptom severity because this measure is difficult to ascertain in a large administrative data set such as the VHA’s CDW. Future research should consider how TBI and PTSD symptom severity drive functional outcomes and, in turn, may lead to intentional self-harm, as addressing functional limitations may be an important focus for prevention programs.

Psychological and functional impairments can be common for veterans with co-occurring PTSD and TBI [37]. PTSD is a known contributor to disability [38], resulting in more than US $40 billion in costs related to direct health care and unemployment among military populations [39]. Despite experiencing long-term symptoms of PTSD and the availability of various therapies, many individuals with PTSD wait years to decades before seeking professional treatment [40,41]. Patients may be skeptical toward therapy [42] and believe treatment to be ineffective, harmful [43], and only for extreme problems. Some may also believe treatment involves a loss of control or autonomy, or as something for people who are “weak,” “crazy,” or “incompetent” [40]. Some individuals with PTSD do not seek help at all, as only approximately one-quarter of individuals with PTSD symptoms receive mental health services [44,45]. A significant number of veterans drop out of their PTSD treatment altogether due to reasons such as work interference, stigma related to the ailment, confidentiality concerns, and perceived treatment ineffectiveness, thus contributing to a population of veterans with unresolved symptoms related to PTSD [46,47].

Similarly, veterans with TBI, regardless of severity, may have a variety of co-occurring psychological and physical health challenges for years following injury [48], including affective, cognitive, somatosensory, and vestibular postconcussive symptoms [49]. Other impairments, especially documented for post-9/11 veterans, may include social/family dysfunction, unemployment, and general difficulty with community reintegration [25]. Individuals with TBI also may not be completely aware of their limitations [50] and, therefore, unable to fully engage in treatment [51]. Given the observed difficulties of veterans with documented PTSD, TBI, or both, VHA suicide prevention services should increase efforts to educate veterans on the importance and nature of therapeutic options to better engage at-risk veterans with VA health care services.

Veterans with TBI or PTSD also face different challenges in seeking treatment such as problems with access, sociocultural environment, past medical experiences, and illness burden [42,45,52]. Access barriers are mainly organizational, due to lack of knowledge about VHA, and logistical, and include time-consuming and complex enrollment processes in VHA, veterans not knowing which VHA services they are eligible for, and time- and distance-related expenses [53]. VHA has made extensive efforts to mitigate these logistical problems, including the implementation of the community care program, which allows veterans to seek care in the community if VHA facilities are unable to furnish services or if long travel distances or appointment wait times pose a burden [54].

A study performed by McCarthy et al [5] suggests that the inclusion of the Recovery Engagement and Coordination for Health–Veterans Enhanced Treatment (REACH VET) program at the VHA has resulted in reduced prevalence of nonfatal suicide attempts along with other greater treatment engagement and care provided to veterans. The algorithm used in REACH VET to identify individuals in the facilities’ top 0.1% suicide risk tiers uses TBI and PTSD diagnoses as variables among the 381 measures used for prediction [6,17]. Our study strengthens the case for using TBI and PTSD as risk factors for suicide risk prediction in the algorithm. Our findings can further help in the identification of high suicide risk veterans by flagging patients with comorbid TBI/PTSD diagnosis using the already available identifiers used by REACH VET [2,18,55].

There are several limitations to take into account in the context of these results. First, this study only included veterans who received health care in VHA settings. Motives for veterans not seeking VHA care are complex, and thus these veterans likely have different TBI-, PTSD-, and intentional self-harm–related profiles. Second, this study relied on diagnoses in VHA administrative data to identify both TBI and PTSD diagnosis status and intentional self-harm outcomes. The high VHA utilization group, while comprising only 25.19% (1,686,066/6,693,275) of the study population, comprised 65.77% (56,986/86,644) of those with detected suicide attempts and other forms of intentional self-harm, 47.76% (44,833/93,866) of the TBI only group, 49.07% (437,947/892,420) of the PTSD only group, and 56.85% (58,295/102,549) of the comorbid TBI/PTSD group. It is likely that underdiagnosis was present [56]. For example, veterans within the low and medium VHA utilization groups may have experienced TBI, PTSD, or both but received care for them outside the VHA or not at all. Such misclassification of true exposure or outcomes status can bias results. However, we observed similar patterns across all three levels of VHA health care utilization, suggesting that results may not be explained entirely by informed presence bias. Third, our study used a broad range of diagnosis codes for intentional self-harm, per Patrick et al [32], to potentially decrease the likelihood that a true self-harm event would be missed (eg, if only suicide attempt codes were used). Future work that further examines and establishes the validity of more ICD codes to detect intentional self-harm would strengthen this line of research. Fourth, the analysis represented TBI and PTSD simply as present or absent according to diagnosis codes. A more refined analysis would incorporate the nature of trauma exposure, symptom severity, barriers to care, and social risk factors. Fifth, our work studied the associations of different TBI and PTSD diagnosis groups with intentional self-harm while adjusting for age, sex, race, ethnicity, marital status, VA service-connection status, and different comorbidities. This analysis did not directly examine the effects of these covariates on intentional self-harm. Further in-depth analysis is required to understand the associations between these sociodemographic and health variables and intentional self-harm while controlling for confounder sets relevant to those specific associations.

Our study adds to the literature on associations of TBI and PTSD diagnoses with intentional self-harm. It provides generalized results across a large veteran population of varied demographics and VHA utilization. Our findings indicate that veterans with a TBI and/or PTSD diagnosis are more likely to have an intentional self-harm event than veterans without either of these diagnoses. These results may contribute to VHA clinical decision-making regarding the frequency and intensity of follow-up or treatment for veterans with PTSD and TBI diagnoses.

Future research on the risk of intentional self-harm among veterans with TBI and/or PTSD can benefit from both qualitative and quantitative efforts. Qualitatively exploring veterans’ experiences and perceptions of health and functional status, as well as facilitators and barriers to health care services, can help to better identify health and access needs. Quantitatively exploring TBI and PTSD severity based on the nature of trauma exposure, current or previous treatment, and the intensity of posttraumatic symptoms can provide more granular associations between TBI and PTSD diagnosis and intentional self-harm. These efforts may help to inform clinicians in targeting preventative and treatment strategies toward a high-risk veteran cohort.

This study examined 6.7 million veteran VHA users and found that those with TBI and/or PTSD diagnoses were more likely to have documented intentional self-harm compared to veterans with neither diagnosis. Veterans with TBI, PTSD, or comorbid TBI/PTSD should be considered high-priority recipients of VA suicide prevention services and may need special engagement efforts given various access and treatment barriers experienced among those with these conditions.