Risk of adverse outcomes among Veterans who screen positive for traumatic brain injury in the Veterans Health Administration but do not complete a comprehensive evaluation: a LIMBIC-CENC study

Abstract

Objective:

To examine whether Post-9/11 Veterans who screened positive for mild traumatic brain injury (mTBI) but did not complete a comprehensive TBI evaluation (CTBIE) were at higher risk of subsequent adverse events compared to Veterans who screened positive and completed a CTBIE. Upon CTBIE completion, information assessed by a trained TBI clinician indicates whether there is mTBI history (mTBI+) or not (mTBI−).

Setting:

Veterans Health Administration (VHA) outpatient services.

Participants:

52,700 Post-9/11 Veterans who screened positive for TBI were included. The follow-up review period was between fiscal years 2008–2019. The three groups studied based on CTBIE completion and mTBI status were: (1) mTBI+ (48.6%), (2) mTBI− (17.8%), and (3) No CTBIE (33.7%).

Design:

Retrospective cohort study. Log binomial and Poisson regression models adjusting for demographic, military, pre-TBI screening health, and VHA covariates examined risk ratios of incident outcomes based on CTBIE completion and mTBI status.

Main Measures:

Incident substance use disorders (SUDs), alcohol use disorder (AUD), opioid use disorder (OUD), overdose, and homelessness documented in VHA administrative records, and mortality as documented in the National Death Index, three years post-TBI screen. VHA outpatient utilization was also examined.

Results:

Compared to the No CTBIE group, the mTBI+ group had 1.28–1.31 times the risk of incident SUD, AUD, and overdose, but 0.73 times the risk of death three years following TBI screening. The mTBI− group had .70 times the risk of OUD compared to the No CTBIE group within the same time period. The No CTBIE group also had the lowest VHA utilization.

Conclusions:

There were mixed findings on risk of adverse events for the No CTBIE group relative to the mTBI+ and mTBI− groups. Future research is needed to explore the observed differences, including health conditions and health care utilization documented outside VHA among Veterans who screen positive for TBI.

Traumatic brain injury (TBI) is a signature injury among Post-9/11 United States (U.S.) military Service Members (SMs) and Veterans. Since 2000, TBI has been documented for more than 465,000 SMs, with more than 80% of injuries categorized as mild TBI (mTBI).¹ Mild TBI can have negative long-term impacts on health and function, particularly if unrecognized and untreated. Therefore, in April 2007, the Veterans Health Administration (VHA) implemented a policy to identify Post-9/11 Veterans who experienced a deployment-related mTBI not previously diagnosed.² The TBI screening tool was designed to avoid false negatives, as evidenced by its high sensitivity (0.94) and moderate specificity (0.59).³ Post-9/11 Veterans who separated from active duty and present for VHA services are screened for the following deployment-related events: (1) an injury such as a blast/explosion or vehicular accident, (2) sequelae indicative of TBI, such as loss of consciousness, being dazed/confused, or not remembering the injury event, (3) symptoms that began or worsened following the injury, such as problems with memory or headaches, and (4) whether symptoms from the previous question were experienced during the past week.⁴ The screening stops once there is a negative response (negative TBI screen). An endorsement of all four questions indicates a positive TBI screen. Per VHA policy,² Veterans who screen positive should be referred to a VHA Comprehensive TBI Evaluation (CTBIE), where a TBI specialist conducts a physical examination and follows a template to thoroughly assess if a deployment-related TBI diagnosis is warranted, identifies current health conditions, and develops an individualized rehabilitation treatment plan if indicated.⁵

Veterans with “mild” TBI may not receive critical follow-up care, especially if their symptoms and sequelae are initially more subtle. However, relative to Veterans without TBI, those with mTBI have had higher rates of adverse outcomes including mental health disorders, pain, cognitive issues, substance-related behavior and substance use disorders (SUDs), overdose, homelessness, and mortality ^6,7 Therefore, it is important to identify Veterans with mTBI history to intervene and mitigate any negative downstream effects.

Substance use problems are common among Post-9/11 Veterans who use VHA care.^8–10 At-risk substance use behaviors may have started during military service where binge drinking has been normative for decades,^11,12 and prescribing opioids for pain and injuries was routine during the Afghanistan/Iraq conflicts, particularly during the first decade.^13–16 Studies have found that military members with TBI and Post-9/11 deployment history had increased risk for post-deployment alcohol misuse,¹⁷ binge drinking and heavy drinking,^18,19 and receipt of long-term opioid therapy.²⁰

Once Post-9/11 Veterans enter the VHA, SUDs are more likely to be identified and diagnosed, particularly for alcohol use disorder (AUD).^8,21,22 TBI history appears to confer additional risk; a study of Veterans using the VHA between fiscal years (FYs) 2006–2015 found that those with a TBI diagnosis had a much higher prevalence of SUD diagnoses compared to those without TBI.⁹

Military/Veteran and civilian populations with TBI history may be at heightened risk for negative effects of opioid use due to a “perfect storm” of vulnerabilities.^23–25 Post-TBI consequences may include increased prescription opioid receipt due to high comorbidity of pain²⁶ and overlapping opioids and sedative-hypnotic prescriptions,²⁷ and increased likelihood of opioid misuse or opioid use disorder (OUD) perhaps due to cognitive deficits following TBI.^23,24 Among those receiving long-term opioid therapy for chronic pain, Veterans with TBI were at increased risk for non-fatal opioid overdose²⁸ compared to those without TBI.

Homelessness continues to be a public health crisis and disproportionately affects Veterans relative to the general adult population.²⁹ Lifetime prevalence of homelessness among Veterans is approximately 10% and the first episode typically occurs within 10 years following military separation.³⁰ TBI is a risk factor for homelessness and is two³⁰ to three³¹ times more prevalent among Veterans with homelessness history than those without.

The VHA TBI screening and evaluation process is an entry point to interdisciplinary care. Without a confirmed mTBI diagnosis, Veterans and providers may not have the full context for current symptomology. This can impact level of treatment engagement and response, placing Veterans with undiagnosed/untreated mTBI at increased risk for negative outcomes. In this study, we hypothesized that Veterans who screened positive for mTBI, but did not complete a CTBIE, were at higher risk of adverse events (i.e., incident diagnoses for SUD, AUD, OUD, overdose, homelessness, and death) compared to Veterans who screened positive and completed a CTBIE.

Methods

This retrospective cohort study was part of the Long-term Impact of Military Relevant Brain Injury Consortium-Chronic Effects of Neurotrauma Consortium (LIMBIC-CENC) Phenotypes Study. The research protocol was approved by the University of Utah Institutional Review Board, which waived informed consent for this secondary data analysis. Administrative records were obtained from the Departments of Defense (DoD) and Veterans Affairs (VA), including data from the DoD and VA Infrastructure for Clinical Intelligence, which includes DoD Institutional and Tricare inpatient and outpatient encounter data, VHA data from the Corporate Data Warehouse and Veterans Services Network, TBI screening and CTBIE data from the VHA Office of Patient Care Services, and the Joint VA/DoD Mortality Data Repository, which includes data from the National Death Index (NDI). NDI records are independent of VHA utilization. VA service-connected disability (SCD) ratings, which reflect a Veterans Benefits Administration (VBA) determination of severity of injuries that were acquired or worsened because of active military service and are associated with monthly compensation, were obtained from the Veteran Service Network (VETSNET) and were current as of April 28, 2020. DoD data ranged from October 1, 1999 to September 30, 2019 (FYs 2000–2019) and VA data ranged from FYs 2002–2019. Using an encrypted identifier, datasets were merged to create a longitudinal record for each Veteran. International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) and ICD-10-CM codes from inpatient and outpatient care were used to identify clinical diagnoses.

The sample was restricted to Post-9/11 Veterans who entered VHA healthcare between FYs 2006–2014 and received a TBI screen between FYs 2008–2016 (Figure 1). Since the TBI screening and evaluation program was developed to identify undetected mTBI incurred during Post-9/11 deployment, Veterans were excluded if they had ICD-9-CM/10-CM documentation of TBI in DoD or VHA records prior to their TBI screen.³² Additional exclusion criteria were evidence of TBI following a negative TBI screen or a TBI that was diagnosed outside of the CTBIE. For the CTBIE, mTBI diagnosis was identified by self-reported loss of consciousness (≤30 minutes) and/or alteration of consciousness or post-traumatic amnesia (≤24 hours).³²

Figure 1.

Study inclusion and exclusion criteria

Note: CTBIE, Comprehensive Traumatic Brain Injury Evaluation; DoD, Department of Defense; FY, fiscal year; mTBI, mild traumatic brain injury; VHA, Veterans Health Administration

TBI screening groups.

Veterans who screened positive for TBI were categorized into three groups based on the outcomes of CTBIE completion and any CTBIE-determined mTBI status: (1) CTBIE completion, mTBI history (mTBI+) (2) CTBIE completion, no mTBI history (mTBI−), and (3) no CTBIE completion or documented TBI history, status of TBI unknown (No CTBIE). Those who screened negative with no documented TBI history (n = 414,758) were excluded from analysis.

Adverse outcomes.

The six adverse outcomes were incident (i.e. no previous diagnoses were present in DoD/VHA records) SUDs (excluding Tobacco Use Disorder), AUD, OUD, overdose, receipt of homelessness or housing insecurity services (hereafter, referred to as “homelessness),” and mortality that were documented three years subsequent to TBI screening, based on ICD-9-CM and ICD-10-CM diagnosis codes being assigned at least once or NDI documentation of mortality. VHA clinic codes were also used to identify Veterans using homelessness services.

Sociodemographic, military, and VHA characteristics.

Sociodemographic characteristics included sex, race/ethnicity, and age group at time of TBI screening. Military-related characteristics included branch (e.g., Army), component (e.g., Active), rank (e.g., Enlisted), Post-9/11 deployment history (yes/no), and deployment-related injuries based on the TBI screener. VHA characteristics were U.S. region of TBI screening (e.g., North) and the Polytrauma/TBI level (Levels 1–4; Level 1=most resourced polytrauma/TBI care) of the VHA facility where the screening occurred.

VHA health care utilization and comorbid health conditions.

For utilization, we examined VHA encounters for the following outpatient services during the three years post-TBI screen: Primary Care, Mental Health, Posttraumatic Stress Disorder (PTSD) Clinic, Pain Clinic, Physical Medicine & Rehabilitation, Polytrauma, Justice Involvement, and Emergency Department. We also extracted DoD and VHA data for health conditions that may co-occur with mTBI. Diagnoses were based on conditions documented in DoD or VHA records any time prior to or three years subsequent to TBI screening. These included: cognitive issues such as problems with attention, memory loss, or any other cognitive dysfunction, PTSD, depression, anxiety, headache, and neck, back, or other pain.

See Supplemental Digital Content Table 1 for ICD and clinic codes used in this study.

Analysis.

We conducted descriptive analyses to assess the sociodemographic, military/VHA, and health characteristics of our sample. We calculated risk ratios (RRs) and 95% confidence intervals (CI) using log-binomial regression modeling. Unadjusted and adjusted models, controlling for sociodemographic, military/VHA, and pre-TBI screening health covariates, were run for each adverse outcome. Poisson regression was used for OUD, homelessness, and mortality due to lack of convergence.

To help characterize the sample and explain any observed differences in outcomes, we performed chi-square and ANOVA with least squares means post-hoc analysis on VHA outpatient encounters of interest documented three years post-TBI screen and prevalence of diagnoses for comorbid health conditions of interest that were documented in DoD/VHA records either prior to or three years following the TBI screen. Except for pre-TBI screening diagnoses, these were not adjusted for in any modeling since these factors may have been a function of the screening outcome itself and part of the rationale for discrepant outcomes among screening groups (i.e., on the “causal pathway” of interest).

We also examined percentages of Veterans considered to be “regular VHA users,” defined as having ≥ 1 VHA inpatient or outpatient visit for each of the three years following a TBI screen. Those who died in subsequent years after being a regular VHA user for at least one year were still considered regular VHA users. For this sensitivity analysis, the purpose was to examine whether comparison groups had similar opportunity for assessment and coding of each of the five outcomes relying on ICD diagnosis codes or VHA clinic codes for measurement.

All analyses were performed with SAS Enterprise Version 8.2. (SAS Institute, Cary, North Carolina).

Results

The final analytical cohort consisted of 52,700 Post-9/11 Veterans who screened positive for TBI. Nearly half (48.6%) completed the CTBIE and had evidence of mTBI+, 17.8% were determined to not have TBI (mTBI−) following CTBIE completion, and 33.7% did not complete the CTBIE.

The sample was primarily male, between the ages of 17–39 years with an average age of 31.6 years (SD=8.3), White, served in the Army, in the Active Component, were enlisted, had Post-9/11 deployment history, and had blast/explosion exposure. Nearly half of the TBI screens occurred in the U.S. South at VHA facilities designated as level 3 Polytrauma Support Clinical Teams. Prior to TBI screening, documented health conditions of interest ranged from 5.1% (cognitive issues) to 64.7% (neck, back, other pain) (Table 1).

Table 1.

Sociodemographic, military, health, and VHA characteristics and incident adverse outcomes by TBI screening group

	mTBI+ (N = 25,588) %	mTBI− (N = 9,368) %	No CTBIE (N = 17,744) %	Total (N = 52,700) %
Sex
Female	4.4	7.7	6.5	5.7
Male	95.6	92.3	93.5	94.3
Race/Ethnicity
White	65.6	58.7	63.9	63.8
Black/African American	13.8	20.7	15.9	15.7
Asian	2.2	2.4	2.1	2.2
Native American/Pacific Islander	2.6	2.6	2.1	2.4
Hispanic	11.8	11.5	11.1	11.5
Unknown	4.0	4.2	4.9	4.3
Age (years)
17–29	54.3	45.3	53.4	52.4
30–39	27.9	30.7	29.0	28.8
40–49	14.5	18.3	13.7	14.9
50+	3.4	5.7	3.9	4.0
Branch
Army	66.7	69.9	63.8	66.3
Air Force	4.0	5.4	5.5	4.8
Navy/Coast guard	7.6	7.4	10.3	8.5
Marines	21.6	17.2	20.5	20.5
Unknown	.03	.02	.04	.03
Component
Active	71.8	67.5	72.1	71.1
National Guard	18.7	21.9	18.4	19.2
Reserve	9.2	10.3	8.5	9.2
Unknown	.27	.38	.98	.53
Rank
Officer/Warrant Officer	3.2	3.8	3.4	3.4
Enlisted	95.7	95.1	93.8	95.0
Unknown	1.1	1.1	2.8	1.7
Post-9/11 Deployment History
Yes	98.0	97.9	95.0	97.0
No	2.0	2.1	5.0	3.1
Deployment-related Injury Events per TBI screen
Blast or explosion	82.8	80.2	76.7	80.3
Vehicular accident/crash	39.9	34.2	36.9	37.9
Fragment or bullet wound above shoulders	4.1	2.5	3.4	3.6
Fall	43.7	42.7	43.0	43.3
TBI Screening Region
Midwest	19.3	17.7	21.4	19.7
Northeast	10.4	10.6	9.5	10.1
South	44.0	50.9	44.4	45.3
West	26.4	29.8	24.6	24.8
VHA Polytrauma/TBI System of Care level
Level 1: PRC	6.4	3.5	5.1	5.4
Level 2: PNS	18.9	20.4	17.2	18.6
Level 3: PSCT	61.0	64.0	62.8	62.1
Level 4: PPOC	13.8	12.1	15.0	13.9
DoD/VHA Diagnoses Prior to TBI Screen
Cognitive Issues	5.1	5.0	5.1	5.1
PTSD	46.3	42.5	41.1	43.9
Depression	32.1	34.5	32.5	32.7
Anxiety	23.5	25.8	24.6	24.2
Headache	22.5	22.3	17.9	20.9
Neck, back, other pain	65.4	67.3	62.4	64.7
Incident Adverse Outcomes Three Years Post-TBI Screen
SUD	11.8	10.0	8.8	10.5
AUD	11.0	9.5	8.0	9.7
OUD	1.7	1.2	1.8	1.7
Overdose	2.3	1.6	1.7	2.0
Homelessness	6.3	6.3	5.9	6.2
Death	0.84	0.92	1.18	0.97

Note. VA, Department of Veterans Affairs; TBI, traumatic brain injury; mTBI, mild traumatic brain injury; VHA,Veterans Health Administration; CTBIE, Comprehensive Traumatic Brain Injury Evaluation; PRC, Polytrauma Rehabilitation Center; PNS, Polytrauma Network Site; PSCT, Polytrauma Support Clinical Team; PPOC, Polytrauma Point of Contact; PTSD, posttraumatic stress disorder; SUD, substance use disorder; AUD, alcohol use disorder; OUD, opioid use disorder

For SUD, individuals who had incident AUD and OUD were counted only once.

Three years post-TBI screen, Veterans had the following incident outcomes documented: SUD (10.5%), AUD (9.7%), OUD (1.7%), overdose (2.0%), and homelessness (6.2%). The mortality rate for this same period was relatively low across groups (0.97%), but the No CTBIE group had the highest rate (1.18%) compared to the mTBI+ (0.84%) and mTBI− (0.92%) groups, p < 0.002.

Table 2 shows the regression models for each incident adverse outcome; interpretations for the unadjusted and adjusted models are similar. In the adjusted models, the two CTBIE completion groups were at increased risk of SUD (mTBI+: RR=1.28, 95% CI=1.21–1.36; mTBI−: RR=1.13, 95% CI=1.04–1.21) and AUD (mTBI+: RR=1.34, 95% CI=1.26–1.42; mTBI−: RR: 1.19, 95% CI=1.10–1.29) relative to the No CTBIE group. However, the mTBI− group was at lower risk of OUD (RR=.70, 95% CI=.56-.86), and the mTBI+ group was at lower risk of death (RR=.73, 95% CI=.60-.88), than the No CTBIE group. That is, the No CTBIE group was at increased risk of OUD (RR=1.43, 95% CI=1.16–1.79) compared to the mTBI− group and of mortality (RR=1.37, 95% CI=1.14–1.67) compared to the mTBI+ group. Risk of homelessness was similar across all groups.

Table 2.

Unadjusted and adjusted models of incident risk of adverse outcomes among Post-9/11 Veterans documented three years post-TBI screen

	mTBI+		mTBI−		No CTBIE
Incident Outcome	Model 1 Unadjusted RR (95% CI)	Model 2 Adjusted RR (95% CI)	Model 1 Unadjusted RR (95% CI)	Model 2 Adjusted RR (95% CI)
SUD	1.30 (1.23–1.37)	1.28 (1.21–1.36)	1.09 (1.01–1.17)	1.13 (1.04–1.21)	Ref
AUD	1.36 (1.27–1.44)	1.34 (1.26–1.42)	1.15 (1.06–1.24)	1.19 (1.10–1.29)	Ref
OUD	.95 (.82–1.10)	.91 (.79–1.06)	.67 (.54–.83)	.70 (.56–.86)	Ref
Overdose	1.32 (1.15–1.51)	1.31 (1.14–1.50)	.92 (.76–1.12)	.97 (.80–1.18)	Ref
Homelessness	1.04 (.97–1.13)	1.06 (.98–1.15)	1.06 (.96–1.16)	1.08 (.97–1.19)	Ref
Death	.72 (.59–.87)	.73 (.60–.88)	.78 (.61–1.00)	.82 (.63–1.05)	Ref

Note. TBI, Traumatic Brain Injury; mTBI, mild traumatic brain injury; SUD, substance use disorder; AUD, alcohol use disorder; OUD, opioid use disorder; CTBIE, Comprehensive Traumatic Brain Injury Evaluation; RR, Risk Ratio; CI, Confidence Interval; Ref, Referent.

Model 2 adjusted for age, sex, race/ethnicity, branch, component, rank, deployment history (yes/no), deployment-related injury events (per TBI screen Question 1), TBI screening region, VHA Polytrauma/TBI System of Care level where TBI screening was completed, and DoD/VHA diagnoses prior to TBI Screen

Diagnoses.

After three years post-TBI screen, the prevalence of documented health issues was 11.5% (cognitive issues) to 76.0% (neck, back, other pain) (Table 3). Among all three groups, the mTBI group had the highest prevalence of documented cognitive issues (14.9%), PTSD (72.4%), and headache (42.9%), p <0.0001, whereas the No CTBIE group had the lowest documented prevalence of each condition.

Table 3.

Prevalence of diagnoses and number of VHA outpatient encounters documented three years post-TBI screen

Prevalence of Diagnoses	mTBI+ N = 25,588 %		mTBI− N = 9,368 %		No CTBIE N = 17,744 %		Total N = 52,700 %		p-value
Cognitive Issues	14.9		9.1		7.8		11.5		<0.0001
PTSD	72.4		64.9		57.3		66.0		<0.0001
Depression	49.6		49.9		43.8		47.7		<0.0001
Anxiety	37.4		38.8		34.2		36.5		<0.0001
Headache	42.9		33.6		23.3		34.6		<0.0001
Neck, back, other pain	79.0		78.9		70.1		76.0		<0.0001
VHA Outpatient Utilization	Median	Mean (SD)	Median	Mean (SD)	Median	Mean (SD)	Median	Mean (SD)	p-value
Total Appointments	33.0	47.6 (50.6)a	30.0	44.0 (48.1)b	18.0	30.1 (40.2)c	27.0	41.0 (47.5)	<0.0001
Primary Care	5.0	6.5 (6.0)a	5.0	6.6 (6.5)a	3.0	5.0 (5.5)b	4.0	6.0 (6.0)	<0.0001
Mental Health	4.0	8.9 (14.1)a	3.0	8.6 (14.1)a	2.0	6.2 (12.8)b	3.0	7.9 (13.7)	<0.0001
PTSD Clinic	0	3.4 (9.9)a	0	3.3 (10.2)a	0	1.6 (6.6)b	0	2.8 (9.0)	<0.0001
Pain Clinic	0	.22 (1.3)a	0	.21 (1.4)a	0	.12 (.9)b	0	.19 (1.2)	<0.0001
Physical Medicine & Rehabilitation	0	.34 (1.2)a	0	.36 (1.4)a	0	.14 (.8)b	0	.28 (1.1)	<0.0001
Polytrauma	1.0	3.4 (6.9)a	1.0	1.5 (2.7)a	0	.36 (1.4)b	1	2.0 (5.2)	<0.0001
Justice Involvement	0	.20 (2.3)	0	.15 (1.8)	0	.20 (2.3)	0	.19 (2.2)	.13
Emergency Department	1.0	1.7 (3.2)a	0	1.6 (3.1)b	0	1.5 (3.0)b	0	1.6 (3.1)	<0.0001
Regular Users of VHA Care	71.2%		70.1%		57.3%		66.4%		<0.0001

Note. VHA, Veterans Health Administration; TBI, traumatic brain injury; mTBI, mild traumatic brain injury; PTSD, posttraumatic stress disorder; CTBIE, Comprehensive TBI Evaluation. Between group differences for prevalence of diagnoses was assessed by chi-square and for VHA outpatient utilization by ANOVA with post-hoc least squares means. Different subscripts within rows denote significant group differences at p < 0.05.

VHA Health Services Outpatient Utilization.

The mTBI+ group had the highest average number of outpatient encounters during the three-year follow-up period, followed by the mTBI− and the No CTBIE groups (all between-group differences, p <0.0001) (Table 3). The mTBI+ and mTBI− groups had similar mean numbers of Primary Care, Mental Health, PTSD, Pain, Physical Medicine & Rehabilitation, and Polytrauma clinic visits, which were higher than the utilization for the No CTBIE group. Justice Involvement was the same across all groups. Emergency Department use was highest for the mTBI+ group, and similar for the mTBI− and No CTBIE groups. An examination of “regular VHA users” showed that the No CTBIE group was less likely to use VHA care annually for each of the three years post-TBI screen (57.3%) compared to the mTBI+ (71.2%) and mTBI− (70.1%) groups, p < 0.0001.

Finally, VA SCD ratings analyses indicated that a higher percentage (23.1%) of the No CTBIE group had a no/0% SCD rating, whereas a higher percentage (70.0%) of the mTBI+ group had a SCD rating of 50–100%, p<0.0001 (Table 4).

Table 4.

Service-connected disability rating status post-TBI screen

Service-Connected Disability Rating	mTBI+ (N = 25,588) %	mTBI− (N = 9,368) %	No CTBIE (N = 17,744) %	Total (N = 52,700) %	p-value
					0.0001
None or 0%	13.42	15.83	23.11	17.11
10–40%	16.56	17.59	18.14	17.28
50–100%	70.01	66.58	58.75	65.61
Mean (SD)	60.0 (33.3)	57.0 (34.3)	50.0 (35.8)	56.1 (34.7)	0.0001

Note. TBI, traumatic brain injury; mTBI, mild traumatic brain injury; CTBIE, Comprehensive TBI Evaluation; SD, standard deviation.

Discussion

This study found that one-third of the 52,700 Veterans who screened positive for mTBI did not complete a CTBIE to determine final mTBI disposition within three years post-TBI screening. Our hypothesis that the No CTBIE group would be at increased risk of the six incident adverse events of interest (SUD, AUD, OUD, overdose, homelessness, and death) compared to Veterans who completed a CTBIE was only partially supported. Consistent with our predictions, the mTBI− group had .70 times the risk of OUD, and the mTBI+ group had .73 times the risk of death compared to the No CTBIE group (i.e., the No CTBIE group was at 1.43 times the risk of OUD compared to the mTBI− group, and at 1.37 times the risk of death compared to the mTBI+ group). Contrary to our predictions, Veterans who completed a CTBIE with a mTBI− or mTBI+ determination were at 1.13 to 1.34 times the risk of SUD and AUD, respectively, and Veterans with mTBI+ were at 1.31 times the risk of overdose, relative to the No CTBIE group. Homelessness was documented for 6.2% of the sample, with all groups showing similar risk.

Prior to TBI screening, VHA/DoD records indicated that the No CTBIE group generally had similar to lower prevalence of six cognitive, mental health, and physical health conditions as the CTBIE completion groups. Three years post-TBI screen, the diagnosis gap widened between the No CTBIE and the two CTBIE completion groups. The No CTBIE group had lower prevalence of each condition relative to the CTBIE completion groups, with the largest differences observed with the mTBI+ group.

This study focused on risk of adverse events among Veterans who screened positive for TBI, with a specific focus on the No CTBIE group outcomes. The No CTBIE group had significantly lower VHA outpatient utilization than the mTBI+ and mTBI− groups, and similar utilization of Emergency Department care as the mTBI− negative group. The No CTBIE group was also less likely to use the VHA services of interest at least once annually for each of the three years post-TBI screen relative to the CTBIE completion groups. It is unclear if the Veterans in the No CTBIE group had lower prevalence of diagnoses relative to the CTBIE completion groups because they perceived themselves as not needing as much health care, or whether other factors, such as non-VHA health care use or barriers to receiving any health care, contributed to lower VHA utilization and, therefore, less opportunity for VHA to document health conditions. To explore this further, we examined SCD ratings; higher SCD ratings increase eligibility for a variety of VA benefits and services. Therefore, Veterans may be motivated to have SCD conditions documented by VBA, even if not treated by VHA. Similar to VHA utilization, a comparison of the three groups showed that the No CTBIE group had lower SCD ratings relative to the mTBI+ and mTBI− groups. Approximately one-quarter had a no/0% SCD rating, which may suggest overall better health for a relatively larger percentage of the No CTBIE group.

Although death was rare, the No CTBIE group had the highest mortality of all three groups and was 1.43 times as likely to die within three-years post-TBI screen than the mTBI+ group. While the mTBI+ group had higher prevalence of documented diagnoses and was more likely to have documented SUD, AUD, and overdose, it also had higher VHA outpatient utilization than the No CTBIE group. This may represent more opportunity to diagnose, intervene, and prevent mortality in the mTBI+ group. Examining cause(s) of death, especially drug- and alcohol-related, could provide additional information on adverse events not captured by VHA data.³³ For example, a review of drug overdose mortality based on VHA use (yes/no; defined as having a VHA encounter in the calendar year or the prior year) found that in 2019, non-VHA using Veterans had a higher number of overall drug overdose deaths and opioid overdose deaths than Veterans using VHA care.³⁴ In our cohort, higher OUD and mortality among the No CTBIE group may suggest a higher risk of OUD/overdose-related death. Future research in this area is warranted, as including cause(s) of death could indicate probable conditions of interest and lead to adjustments to the incident rates of adverse outcomes reported in this study for VHA and non-VHA users.

Our results add to other literature that has examined adverse outcomes among Veterans with TBI. More than 10% of the mTBI+ group had incident AUD three-years post-TBI screen. Studies with military populations have found that those with deployment-related TBI were at increased risk for post-deployment at-risk drinking,^18,19 suggesting that early intervention while in military service following TBI may prevent downstream AUD development.¹² Specific to OUD and overdose, studies have found that SMs/Veterans with TBI are at increased risk for receiving more prescription opioids.^20,25 Our findings showed similar risk of OUD among the mTBI+ and No CTBIE groups, and increased risk of SUD, AUD, and overdose among the mTBI+ group compared to the No CTBIE group, with the mTBI+ group also showing higher prevalence of headache and non-headache pain relative to the No CTBIE group. Observed higher risk of overdose, but not OUD, provides additional support for part of the “perfect storm” theory that individuals who experienced TBI are at increased risk of developing OUD and experiencing overdose, perhaps due to multimorbidity of the “polytrauma phenotype” (TBI, mental health, pain).^7,23–25 OUD incidence three years post-TBI screen was 1.7% and 1.8% for the mTBI+ and No CTBIE groups, respectively. The extent to which the No CTBIE group may be composed of individuals with mTBI history and OUD that could be masking any differences between the mTBI+ and No CTBIE groups is unknown.

More than 7% of Veterans who screened positive for TBI in our sample had used homelessness services at least once three years post-TBI screen. In a study of Veterans seeking homelessness services, more than 80% had experienced at least one TBI prior to, and more than 40% experienced at least one TBI following, their first episode of homelessness, suggesting a bi-directional relationship between TBI and homelessness.³⁵ The 5.9% incidence of receiving homelessness services among the No CTBIE group may be an underestimation of this outcome, as Veterans experiencing homelessness tend to underuse VHA services and may be lost to system follow-up.³⁶

Increasing Access to Post-TBI Screening Follow-up

To date, among Veterans who screen positive for TBI and complete a CTBIE, approximately 60% have a confirmed TBI diagnosis.³⁷ Although suggesting a false positive rate of about 40%, the TBI screener is meant to be highly sensitive due to the potential for long-term adverse outcomes that may be associated with undetected mTBI.³ Indeed, the No CTBIE group likely includes some Veterans who met diagnostic criteria for mTBI, had they been evaluated.

VHA’s integrated system of care, electronic health records, telehealth services, and TBI interdisciplinary teams may allow for targeted identification of and outreach to the No CTBIE group. An electronic health record flag could prompt providers to contact Veterans who did not complete a CTBIE, provide brief education on the importance of follow-up, and submit an in-person or remote consult. A pilot program of telehealth versus in-person CTBIE has yielded similar rates of TBI diagnosis and recommendations for follow-up care within the VHA system,³⁸ suggesting that this approach could be useful for engaging harder-to-reach Veterans. However, Stroupe et al.³⁸ noted that relative to a remote evaluation, an in-person CTBIE allows for a physical examination and an environment that tends to be more conducive to discussing additional health concerns. This possibility would need to be assessed, and the trade-offs balanced, if leveraging telehealth to increase rates of CTBIE follow-up.

Technology can increase access to care. In one study of Veterans experiencing homelessness, a majority had a mobile phone, used the Internet and text messaging, and were amenable to appointment scheduling or reminders by phone or text messaging.³⁹ An effort to distribute video-enabled tablets to Veterans experiencing homelessness has resulted in the ability to provide virtual video visits to this population.⁴⁰ Telehealth has also been recognized as a useful tool for OUD management,⁴¹ delivering opioid overdose prevention and naloxone education to at-risk Veterans,⁴² and may be beneficial to usual SUD care.⁴³

Limitations.

This research has several limitations. First, our findings did not include diagnosis or utilization that occurred outside of DoD or VHA, and therefore may underestimate incidence of adverse outcomes or may not fully represent Post-9/11 Veterans who used non-VHA services exclusively, or in addition to, VHA care. Relatedly, we do not fully understand the health care utilization of this cohort and if similar trends are noted in other VHA areas of preventative and early identification health care screening (e.g., PTSD, depression).⁴⁴ Second, we did not have dates available for any suspected TBI-related injuries or military separation, and therefore could not account for differences in time between those events and Veterans’ TBI screening. Longer-term study of this cohort with additional data sources can further help to elucidate present findings.

Finally, we do not know the reasons for missing CTBIE data. Compared to the groups that completed the CTBIE, the No CTBIE group had higher rates of mortality and were less likely to use VHA care at least once for each of the three years post-TBI screen. The No CTBIE group may have not perceived a need for a CTBIE, experienced barriers to accessing follow-up care,⁴⁵ or had a TBI evaluation completed in the VHA or community but not documented in the CTBIE template.⁴⁶ As Veteran use of non-VHA, community-based care expands,⁴⁷ it is critical to ensure that Veterans with TBI history are being identified and provided with VA/DoD guidelines-based care.³² Analysis of VHA and community-based specialty care utilization may help to detail how frequently the No CTBIE cohort accessed specialty care following their positive TBI screen.

Conclusions

Study findings were mixed regarding risk of adverse events for the No CTBIE group relative to those who completed a CTBIE. Compared to the No CTBIE group, the mTBI+ group had increased risk of incident SUD, AUD, and overdose, but the No CTBIE group had increased risks of death and OUD relative to the mTBI+ and mTBI− groups, respectively. Future research should investigate the observed differences, including health conditions and health care utilization documented outside VHA to better understand the extent to which Veteran health needs are met among those who screen positive for TBI.