Longitudinal Analysis of Persistent Postconcussion Symptoms, Probable TBI, and Intimate Partner Violence Perpetration Among Veterans

Abstract

Objective:

To determine Veterans’ IPV perpetration following report of traumatic brain injury (TBI) and persistent postconcussion symptoms (PPCS).

Setting:

Five Veterans Affairs (VA) medical centers.

Participants:

Veterans with nonmissing data on main measures, resulting in N = 1150 at baseline and N = 827 at follow up.

Design:

Prospective cohort study with secondary data analysis of self-reported TBI, PPCS, and intimate partner violence (IPV) perpetration, controlling for common predictors of IPV, including binge drinking, marijuana use, pain intensity, and probable posttraumatic stress disorder (PTSD).

Main Measures:

VA TBI Screening Tool to assess for probable TBI and PPCS; Conflict Tactics Scale-Revised (CTS-2S) to assess for IPV perpetration.

Results:

Almost half (48%) of participants reported IPV perpetration at follow-up. Both probable TBI and higher PPCS at baseline were associated with overall IPV perpetration and more frequent IPV perpetration at follow-up. Only PPCS significantly predicted IPV perpetration after controlling for common predictors of IPV perpetration. Neither probable TBI nor PPCS predicted frequency of IPV perpetration.

Conclusion:

When considered alongside common risk factors for IPV perpetration, PPCS was uniquely associated with the likelihood of IPV perpetration in this Veteran sample. Given post-9/11 Veterans’ elevated risk for head injury, findings emphasize the distinctive value of PPCS in understanding risk for IPV perpetration. We recommend increased assessment for PPCS in clinical practice among Veterans enrolled in VA care and highlight several important areas for future research and intervention development.

Introduction

Due to the increased use of improvised explosive devices (IEDs) in post-9/11 conflicts and improvements in military protective equipment, recently returning Veterans are more likely to experience and survive traumatic brain injuries (TBIs) than previous generations.¹ In fact, blast-related TBIs are considered “the signature injury” incurred during the post-9/11 conflicts and represent the most common injury among post-9/11 service members and Veterans.² Because combat-related TBI is complex, frequent, and can lead to long-lasting physical, psychological, and social harm, continued research is needed to better understand the relationship between head injury and associated factors.

TBI and its postacute symptoms, such as persistent postconcussion symptoms (PPCS), have been widely studied in relation to intimate partner violence (IPV), including IPV perpetration.^3–6 IPV includes actual or threatened physical violence, sexual violence or coercion, or psychological aggression by an intimate partner⁷ and is associated with negative physical and mental health outcomes, including cardiovascular and respiratory problems, chronic pain, reproductive health complications, PTSD, anxiety, depression, substance use, suicidality, and other disabilities.^7–12 Post-9/11 Veterans may be at heightened risk of perpetrating IPV due to predisposing risk factors, including high rates of PTSD and TBI.⁶ Up to 60% of post-9/11 Veterans report engaging in IPV, with almost a third reporting that their relationship partner is afraid of them.⁶ Yet, the unique associations between, and mechanisms of, combat-related TBI and IPV perpetration remain unclear.

Symptoms following TBI can result in a myriad of concerns, including cognitive deficits (e.g., memory and concentration problems), affective issues (e.g., depressed mood, anxiety, and irritability), behavioral difficulties (e.g., sleep disturbance and agitation), physical problems (e.g., headaches and chronic pain), and adverse social functioning (e.g., relationship difficulties and aggressive behaviors).¹³ Even in cases of mild TBI (mTBI; i.e., concussion), these symptoms could persist, resulting from structural impairment and damage to areas of the brain.¹⁴ Furthermore, despite many symptoms related to head injury resolving with time, affective and behavioral difficulties, such as irritability and verbal aggression, have been shown to develop or increase following the acute TBI phase and persist long-term, thereby interfering with one’s relationships and interpersonal functioning.^15,16 Moreover, although the prognosis after mTBI is generally good, with typical recovery expected within six months, a significant portion of those with mTBI continue to report PPCS for months, even years, after the initial injury.^21,22

Emerging evidence suggests that PPCS may predict risk of violence perpetration among Veterans, including IPV perpetration, above and beyond risk predicted by TBI.¹⁷ In a sample of Veterans who recently received treatment for substance use and/or mental health problems, PPCS was associated with both partner violence aggression and general (nonpartner) violence aggression even after accounting for other risk factors such as age, substance use, and PTSD.¹⁷ Persistent postconcussion symptoms continue beyond the period of acute head injury and include symptoms such as headaches, irritability, difficulty concentrating, memory problems, and sleep disturbance.^18,19 These symptoms have a markedly negative influence on psychosocial and interpersonal domains, including emotional disturbances, depressive symptoms, and cognitive concerns.²⁰ The functional domains impacted in those with PPCS are areas critical to effective interpersonal and relationship functioning and overlap greatly with domains impacted by other commonly reported concerns and conditions of post-9/11 Veterans.⁶ Understanding the unique role that PPCS plays in IPV perpetration among Veterans is essential, particularly due to the lingering impact PPCS has on affective, behavioral, and social functioning such as in the areas of cognitive dysfunction, psychiatric issues, and emotional dysregulation, all of which may be risk factors for IPV perpetration.²³

The significant symptom overlap between TBI, PPCS, and PTSD can complicate the clinical presentation and management of each disorder, and PTSD symptoms can impede recovery from mTBI.¹ Elevated risk for head injury, PTSD, and IPV perpetration among post-9/11 Veterans underscores the critical need to better understand the potential unique contributions of these conditions and symptoms to IPV perpetration. Elucidating risk factors for IPV perpetration can improve screening and treatment for this population. Because research examining the distinctive roles that TBI and PPCS play in increasing risk for IPV perpetration over time is still in its infancy, studies have yet to examine TBI symptoms and PPCS independent of one another. TBI and its comorbidities can compromise self-report and complicate assessment and treatment of social, learning, neurologic, and substance related factors.²⁴ These concerns may be further exacerbated by PPCS. The current longitudinal study seeks to build on prior work by examining the relationship between TBI, PPCS, and IPV perpetration in a sample of post-9/11 Veterans. The purpose of this study is to examine the associations between probable TBI, PPCS, and IPV perpetration. We hypothesize that TBI and PPCS reported at baseline will be positively associated with IPV perpetration at follow up, controlling for demographic characteristics and common predictors of IPV perpetration.

Methods

Participants and Procedure

The present study was approved by the local institutional review board at VA Connecticut Healthcare System. We used data for this secondary data analysis from a prospective cohort study derived from the Department of Defense’s OEF/OIF/OND Roster, which includes Veterans discharged from military service and enrolled in VA healthcare between October 1, 2001 and December 31, 2018. Veterans received a recruitment letter and survey; those who completed and returned the consent and baseline survey were invited to complete one annual follow-up survey. Participants received $20 as compensation for returning each survey (please see previous publication²⁵ for detailed description of the sample and methods procedures).

Measures

IPV Perpetration

We measured past year IPV perpetration with the perpetration scale of the Conflict Tactics Scale-Revised Short Form (CTS-2S).²⁶ The CTS-2S consists of six behaviorally specific statements related to the three major IPV domains: psychological, sexual, and physical violence (e.g., “I punched, kicked, or beat-up my partner”). Participants reported their frequency of engaging in each behavior in the past year behavior on a 7-point scale from 0 (never) to 6 (more than 20 times) with two other response options (“Not in the past year, but has happened before” and “This has never happened in any relationship”). Per the recommended frequency scoring on the CTS-2S, we scored the CTS-2S by summing the midpoints of the response categories for each item (e.g., for “3 to 5 times” the midpoint is 4).

Probable TBI and Persistent Post-Concussion Symptoms (PPCS)

We measured probable TBI and PPCS with the three-item Brief Traumatic Brain Injury Screen (BTBIS).¹⁹ Veterans screened positive for probable TBI based on: (1) endorsement of a head injury during deployment (e.g., from a bullet, fall, blast, vehicular accident, fragment, or other) that (2) led to alteration of consciousness (e.g., being dazed, confused, or “seeing stars”), loss of consciousness for any amount of time, post-traumatic amnesia, or concussion-related symptoms (e.g., headache, dizziness, irritability). Persistent post-concussion symptoms were based on endorsement of item three from the BTBIS (i.e., current head injury-related headache, dizziness, memory problems, balance problems, tinnitus, irritability, and/or sleep problems at time of survey completion). We generated the PPCS severity total by summing score of the endorsed PPCS.

Control Variables

In our analyses, we controlled for demographic variables (age, gender, marital status, race/ethnicity, and income) along with common predictors of IPV perpetration, including binge drinking, marijuana use, pain severity, and probable PTSD. Binge drinking was based on the question “How often do you have six or more drinks on one occasion?” Responses ranged from never to daily or almost daily on a 5-point scale with endorsement of monthly, weekly, or daily or almost daily categorized as binge drinking.²⁷ Marijuana use was measured with endorsement of the marijuana use item on the Drug Abuse Screening Test 10 (DAST-10).²⁸ We assessed probable PTSD using the PTSD Checklist-Military Version (PCL-M),²⁹ a 17-item checklist designed and validated for use with military populations. Responses on the PCL-M range from not at all to extremely on a 5-point scale. We measured pain severity using the Brief Pain Inventory (BPI),³⁰ designed to assess pain severity and its impact on functioning on a 0 (no pain) to 10 (worst pain imaginable) scale.

Data Analysis

We included Veterans with nonmissing data on the main study outcome (i.e., IPV perpetration). Out of the N = 1150 Veterans who completed baseline, 860 (74.8%) completed follow-up. Our analytic sample consisted of the N = 827 (96.1%) who completed data on IPV at follow-up. We first conducted descriptive statistics. Next, we examined the bivariate associations between baseline TBI and PPCS with both follow-up IPV perpetration (using chi square and odds ratios) and IPV perpetration frequency among those who reported IPV perpetration (using Welch’s t-test and Kendall’s Tau correlation). For our main analyses, we used zero-inflated negative binomial (ZINB) regression to examine whether baseline TBI and PPCS predicted IPV perpetration at follow-up, controlling for demographics and other common predictors of IPV at baseline (marijuana use, binge drinking, probable PTSD, and pain severity). To account for the possibility of PPCS mediating TBI, we entered TBI and PPCS into separate models. The ZINB regression has two parts: occurrence and frequency. One logistic regression predicted whether an outcome did not occur (e.g., no IPV perpetration reported) and the other predicted the frequency of the outcome (e.g., IPV perpetration frequency). Using ZINB regression was appropriate for this study because the data were overdispersed (i.e., the variance of IPV frequency is greater than the mean), the outcome had an excessive amount of zeros (434 [37.7%] were zero for IPV frequency), and two processes created opportunity for participants to provide a zero response (e.g., participants could report no IPV perpetration because they have had no recent intimate partners and thus no opportunity to perpetrate IPV, or they could report no IPV because they were managing conflict in their relationship without perpetrating IPV). Statistically, a ZINB binomial was justified by a significant Vuong test and lower AIC values compared to the more parsimonious negative binomial model (AIC = 2636.24) for the TBI model (z = 3.25, p < .001, AIC = 2591.17) and PPCS model (z = 3.19, p < .001, AIC = 2588.75). We report models with exponentiated coefficients, incident rate ratios (IRRs) for the negative binomial portion, and adjusted odds ratios (AORs) for the logit portion. We conducted all analyses using R (version 3.6.3; 2020) using packages pscl (1.5.5) and MASS (7.3.51.4).

Results

Demographic and descriptive information is presented in Table 1. Half the sample (50.6%) was female, and half were under 45 years old (49.7%). More than half had served in the Army (61.5%) and were married at baseline (59.8%). Most participants were non-Hispanic White (78.4%). Around a third (30.2%) were making $25000 a year or less. Ten percent (9.8%) reported using marijuana in the last 12 months and 14.1% reported binge drinking on occasion. One in five participants (22.5%) met criteria for probable PTSD. Over a quarter (29.1%) of the sample reported a TBI at baseline and almost half (47.5%) reported IPV perpetration at follow up.

Table 1.

Sample Characteristics

	n (%)
Age
< 45	410 (50.3%)
>= 45	405 (49.7%)
Binary gender
Men	403 (49.4%)
Women	413 (50.6%)
Marital Status
Married	492 (59.8%)
Unmarried	331 (40.2%)
Race
Non-Hispanic White	626 (78.4%)
Other racial and ethnic identities	172 (21.6%)
Low Income Status
0 to 25k	241 (30.2%)
>25000K	558 (69.8%)
Past year marijuana use
Yes	80 (9.8%)
No	739 (90.2%)
Binge drinking (6 or more drinks)
At least occasionally	115 (14.1%)
Never	700 (85.9%)
Service branch
Army	496 (61.5%)
Air Force	140 (17.3%)
Marine Corps	58 (7.2%)
Navy	113 (14.0%)
Probable PTSD (PCL>=50)
Yes	180 (22.8%)
No	609 (77.2%)
Probable TBI
Yes	211 (29.1%)
No	514 (70.9%)
Past year IPV Perpetration (CTS2-S)
Yes	393 (47.5%)
No	434 (52.5%)
	M (SD)
Past year IPV Frequency (CTS2-S)	4.23 (8.07)
Pain severity (BPI)	3.05 (2.21)
Persistent post-concussion symptoms (PPCS)	1.82 (2.39)

Notes: N = 827, PTSD = post-traumatic stress disorder, PCL-C = PTSD checklist military version, TBI = traumatic brain injury, BPI = Brief pain inventory short form, (CTS-2S) = Conflict Tactics Scale Revised Short Form, PPCS = persistent post-concussion symptoms (PPCS)

Bivariate Analyses

Correlations are reported in Table 2. Veterans with TBI at baseline were more likely to report IPV perpetration at follow-up than Veterans without TBI at baseline (54.0% to 44.2%; OR = 1.49, 95% CI: 1.08, 2.05; χ²(1, N=725) = 5.84, p = .0195). Yet, among those who reported IPV perpetration at follow-up, TBI was not related to frequency of IPV perpetration (Welch’s t(257) = −0.36, p = .7166). Baseline PPCS was associated with both increased odds of IPV perpetration at follow-up (OR = 1.10, 95% CI: 1.04, 1.17, p = .0011) and greater frequency of IPV perpetration among Veterans who reported perpetrating IPV(rτ= .10, p = 0.0167). A Mann-Whitney U test showed PPCS scores were higher for Veterans with a TBI (Mdn = 5, IQR = 3) than those without a TBI (Mdn = 0, IQR = 1), W = 12454, p < .001. Almost all Veterans with a TBI (91.0%) reported at least one current PPCS symptom, yet only 58.9% of Veterans who reported current PPCS symptoms had a deployment-related TBI, implying the rest had PPCS from other head injuries.

Table 2.

Correlations Among Intimate Partner Violence (IPV) Use and Frequency and Predictor Variables

	1	2	3	4	5	6	7	8	9	10	11	12
1. Age >= 45 (Ref: < 45)	-
2. Women (Ref: Men)	.17***	-
3. Married (Ref: Unmarried)	.12***	.19***	-
4. Race (Ref: non-Hispanic White)	−.03	−.05	−.10**	-
5. Low-income 0–25K (Ref: >25K)	−.06	−.14***	−.04	.04	-
6. Marijuana use (Ref: None)	−.12***	.00	−.10**	.06	.08*	-
7. Binge drinking (Ref: Never)	−.10**	.16***	.03	−.03	−.04	.12***	-
8. Probable PTSD	−.07	−.01	−.03	.17***	.17***	.16***	.10**	-
9. Pain severity (BPI)	.07*	−.01	.02	.19***	.16***	.08*	.05	.43***	-
10. Probable TBI	−.05	.15***	.00	.13***	.09*	.15***	.03	.36***	.34***	-
11. PPCS	−.03	.15***	.00	.12**	.13***	.10**	.03	.46***	.46***	.66***	-
12. Past year IPV Use (CTS2-S)	.01	.04	.29***	−.06	.04	.02	.07*	.10**	.09**	.09*	.12**	-
13. Past year IPV Use Frequencya	−.15	−.01	−.01	−.03	.05	.10	.01	.19***	.15**	.05	.10	NA

Note. For all correlations other than correlations involving past year IPV use frequency, ns = 694 to 827

^awe report Kendall’s Tau for correlations involving past year IPV use frequency among Veterans who reported IPV use, ns = 341 to 393

^*p < .05

^**p < .01

^***p < .001

PTSD = post-traumatic stress disorder, PCL-C = PTSD checklist military version, TBI = traumatic brain injury, BPI = Brief pain inventory short form, (CTS-2S) = Conflict Tactics Scale Revised Short Form, PPCS = persistent post-concussion symptoms (PPCS)

Associations Between TBI, PPCS, and IPV Perpetration

We first examined the relationship between baseline levels of TBI and IPV perpetration at follow-up (Table 3). After controlling for other common predictors of IPV, TBI had no association with the occurrence of IPV perpetration (i.e., it did not predict a zero on IPV perpetration in the logit model) or with the frequency of IPV perpetration. We next examined the relationship between PPCS and IPV perpetration (Table 4). PPCS was associated with occurrence of IPV perpetration yet not the frequency. Veterans with higher PPCS scores were less likely to report no IPV perpetration (AOR = .87, 95% CI: 77, .98). In both models, the only control variable associated with occurrence of IPV was marital status, with married people significantly more likely to report IPV perpetration. Probable PTSD was the only predictor associated with greater frequency of IPV perpetration in the TBI model (yet was not in the PPCS model).

Table 3.

Association of TBI and IPV Perpetration Incident Rate Ratios for IPV Frequency and Adjusted Odds Ratios for Zeroes (no IPV Perpetration)

Variable	Count: Negative Binomial IRR (95% BCa CI)	Zero Inflation: Logit, AOR (95% BCa CI)
Age >= 45 (Ref: < 45)	0.85 (0.63, 1.17)	1.26 (0.77, 2.06)
Women (Ref: Men)	1.06 (0.77, 1.47)	0.99 (0.61, 1.60)
Married (Ref: Unmarried)	0.92 (0.66, 1.28)	0.24 (0.14, 0.41)***
Racial/ethnic minority (Ref: non-Hispanic White)	0.94 (0.65, 1.36)	1.35 (0.77, 2.37)
Low-income 0 to 25K (Ref: >25K)	0.95 (0.69, 1.31)	0.89 (0.54, 1.47)
Marijuana use (Ref: None)	1.28 (0.79, 2.07)	1.2 (0.58, 2.48)
Binge drinking (Ref: Never)	0.86 (0.60, 1.25)	0.49 (0.22, 1.09)
Probable PTSD	1.54 (1.04, 2.27)*	0.80 (0.42, 1.52)
Pain severity	1.08 (1.00, 1.16)	0.98 (0.87, 1.11)
Probable TBI	0.83 (0.58, 1.19)	0.53 (0.27, 1.02)

Note.

^*p < .05

^***p < .001

AOR = adjusted odds ratio; CI = confidence interval; IRR = incidence rate ratio, BCa = bias-corrected and accelerated, n = 616

Table 4.

Association of PPCS and IPV Perpetration Incident Rate Ratios for IPV Frequency and Adjusted Odds Ratios for Zeroes (no IPV Perpetration)

Variable	Count: Negative Binomial IRR (95% BCa CI)	Zero Inflation: Logit, AOR (95% BCa CI)
Age >= 45 (Ref: < 45)	0.84 (0.62, 1.14)	1.17 (0.73, 1.86)
Women (Ref: Men)	1.12 (0.82, 1.53)	0.97 (0.61, 1.56)
Married (Ref: Unmarried)	0.93 (0.67, 1.28)	0.25 (0.15, 0.41)***
Racial/ethnic minority (Ref: non-Hispanic White)	0.89 (0.62, 1.28)	1.24 (0.73, 2.13)
Low-income 0 to 25K (Ref: >25K)	0.93 (0.68, 1.27)	0.89 (0.54, 1.44)
Marijuana use (Ref: None)	1.23 (0.77, 1.98)	1.1 (0.54, 2.22)
Binge drinking (Ref: Never)	0.91 (0.63, 1.31)	0.53 (0.26, 1.08)
Probable PTSD	1.46 (0.99, 2.15)	0.86 (0.46, 1.62)
Pain severity	1.06 (0.97, 1.15)	1.00 (0.89, 1.13)
Persistent post-concussion symptoms (PPCS)	1.02 (0.95, 1.09)	0.87 (0.77, 0.98)*

Note.

^*p < .05

^***p < .001

AOR = adjusted odds ratio; CI = confidence interval; IRR = incidence rate ratio, BCa = bias-corrected and accelerated, n = 616

Discussion

Although prior research highlights high rates of both TBI and PPCS among Veterans following return from warzones,³¹ no longitudinal research to date has considered implications of TBI and PPCS, in tandem, on Veterans’ perpetration of IPV. Our findings indicate that PPCS was positively associated with IPV perpetration above and beyond other common IPV risk factors. As such, PPCS may be an important risk factor to identify through screening and address in treatment with Veterans. However, the patterns of findings varied depending on the model, which further demonstrates the need to disentangle unique and cumulative effects of various conditions on relationship functioning in Veteran populations.

Our study results echo prior research showing moderate to high rates (28 to 60%) of IPV perpetration among male Veterans.^32–34 We also found rates of probable PTSD^35,36 comparable to previous studies in non-treatment seeking Veteran samples. Moreover, our rate of probable TBI (29.1%), is nearly identical to that found in another large sample of post-9/11 Veterans who were screened for TBI within VA.³⁷ Taken together, these rates demonstrate the presence of several comorbidities commonly cited that contribute to Veteran impairment.

Interestingly, PPCS, but not TBI, were associated with an increased likelihood of perpetrating IPV in the subsequent year when controlling for other common IPV risk factors. This research adds to mounting evidence of the notable impact of post-concussive symptoms in the psychosocial domain of Veteran health.³⁸ This finding underscores the importance of using multivariate statistical modeling to include various risk factors when studying IPV perpetration among Veterans. Considered together, our results suggest that it is not head injury alone, but the lingering affective and cognitive effects of such injuries, that seem to contribute to the likelihood of using relationship violence. Indeed, prior work with post-9/11 Veterans has shown unique PPCS to fully mediate the relationship between TBI on mental health outcomes (i.e., PTSD, depression).³¹ Our results also align with previous cross-sectional research showing that PPCS, but not TBI, are positively associated with likelihood of perpetrating both partner and non-partner violence among post-9/11 Veterans when controlling for probable PTSD and chronic pain.³⁹

Despite accumulating evidence that PPCS may predispose Veterans toward mental health concerns, including perpetration of violence, our findings showed that Veterans with elevated PPCS were more inclined to perpetrate IPV, but did not necessarily perpetrate more frequently. This result is somewhat contrary to what was found in Morris et al.’s (2019)³⁹ study of Veterans in treatment for substance use wherein PPCS was related to increased partner-based physical aggression as well as non-partner based psychological and physical aggression. One potential reason for our discrepant findings may be our model’s inclusion of several factors shown to contribute to IPV perpetration including TBI, PTSD, pain, and substance use. This hypothesis is likely as PPCS in our sample was associated with IPV frequency among those who perpetrated IPV when not controlling for covariates. It may be that this combination of factors, or perhaps other factors not accounted for in our model, better explain repeated IPV perpetration. It may also be that our study’s oversampling of women Veterans (50.6%), who are not as often represented in studies of IPV perpetration or combat-related head injury, played a role in our pattern of findings.

Another possible explanation for our pattern of findings may be an inherent difference in PPCS-related IPV perpetration when considered over time. Ample literature supports the notion that IPV perpetration is related to difficulties with emotion regulation.^40–42 Arguably, PPCS, characterized by specific disruptions in affect and cognition, may be more likely to contribute to emotional challenges that result in episodic rather than ongoing IPV perpetration, particularly when controlling for other factors that exacerbate these challenges. Previous research has demonstrated how challenging it is to distinguish the effects of comorbid TBI, pain, and PTSD, referred to as the “polytrauma clinical triad,”^43,44 in relation to violent impulses for returning Veterans.⁴⁵ Similarly, a systematic review of the research on IPV perpetration and cognitive executive functioning revealed mixed findings across studies despite consistently poorer performance on cognitive flexibility tests by individuals who perpetrate IPV.⁴⁶ Additional studies are needed to elucidate PPCS as a mechanism contributing to IPV perpetration and to better understand the intersection between PPCS and other risk factors of IPV perpetration for Veterans and how they function uniquely, and in co-occurrence, with TBI.

Limitations

Despite our study’s noteworthy contribution to evolving literature on the relationship between head injury and IPV, our methodology was not without limitations. First, our exclusive use of self-report methodology provides only one source of information. Likewise, our method did not include any formal evaluations or assessments of our variables of interest by trained professionals. We were therefore not able to develop clear diagnoses for TBI and PTSD using gold standard and/or structured assessments (e.g., CAPS-5⁴⁷). Furthermore, although use of a longitudinal data makes a stronger case for directionality of findings (PPCS to IPV perpetration), we cannot infer a causal link. Moreover, because TBI and PPCS data were collected at the same timepoint, we are unable to test whether PPCS mediates the relationship between TBI and IPV; future research should empirically examine the question of plausible mediation by assessing TBI and PPCS at different time points.

With respect to external validity, given that our study was predominately Caucasian, married Veterans under the age of 45 and all from OEF/OIF/OND eras, the generalizability of study results to the Veteran population at large is unclear. Moreover, this study describes a secondary analysis of a pre-existing data set that did not include information on weighting to account for non-response; future studies are warranted in more representative or weighted samples. In addition, because our sample was not comprised of Veterans who were specifically seeking treatment for PPCS or TBI, results may not generalize to Veterans seeking care for distress or impairment related to this constellation of symptoms.

Future Directions

Future longitudinal research should examine the biopsychosocial impact of PPCS on Veterans’ relationship functioning and IPV perpetration. Longitudinal research that includes biomarkers and/or use of imaging,⁴⁸ for example, would strengthen our understanding neurobiological reactions and tissue changes, if any, that coincide with behavioral differences in use of violence for Veterans who have experienced head injuries and report PPCS. Such data could also be used to examine how Veterans with PPCS differ from those without these symptoms over time. Other forms of data gathering, such as ecological momentary assessment (EMA) may similarly provide more precise and frequent datapoints⁴⁹ to aid in determining the extent to which PPCS functions as a risk factor for IPV perpetration as it occurs in Veterans’ everyday lives.

Moreover, increased understanding of potential gender differences in the association between head injury and future IPV perpetration is warranted. Additionally, bidirectional IPV (i.e., both partners perpetrate IPV during the same or separate episode) is a common form of IPV among Veterans.⁵⁰ Therefore, gathering data about behaviors from both Veteran and partner perspectives would serve to build on our results. Incorporating clinician observations or assessments of Veterans would also strengthen the measurement of variables of interest and address this limitation of our research. Lastly, as Veterans who use IPV⁵¹ and report PPCS⁵² may be more inclined to access healthcare services, research documenting the broad range of healthcare needs and services accessed by Veterans with PPCS who use IPV would improve the VA and community partners’ capacity to respond. Specifically, examining healthcare practices would help to better identify Veterans experiencing PPCS who may be at risk for IPV perpetration and those who would benefit from more early intervention or prevention strategies.

Clinical Implications

Findings gathered herein highlight a potential avenue for improving IPV perpetration detection, assessment, and intervention among Veterans who report PPCS symptoms following head injury. This research also presents policy and practice implications by highlighting opportunities to incorporate a clearer understanding of complex and multifaceted relationship of PPCS and IPV perpetration into strategic planning and the development of innovative tools and practices for Veterans. Understanding IPV perpetration among Veterans in the context of their post-service injuries may help providers better conceptualize their patients holistically and serve to reduce stigma surrounding IPV perpetration. Reducing stigma can lead to enhanced opportunities for screening and referral to appropriate follow up care.

Our findings indicate a potential pathway through which head injury may elevate the risk for IPV perpetration, implicating PPCS as an etiologic component of IPV among men and women Veterans not currently, to our knowledge, being addressed within VA. As intervention development for PPCS is growing,⁵³ our results suggest a need for healthcare response to potential violence that may follow in the year following reported experiences of PPCS. Establishing a target of treatment beyond personal education and symptom management to the realm of relationship health and safety could provide a means of addressing potential misattributions of behavior related to postconcussive complaints for Veterans and their partners.⁵⁴ As post-9/11 Veterans are at heightened risk for head injuries during the course of their military service and particularly during combat,^37,55 it is imperative that future research lends itself to improved understanding and response to the sequalae of these injuries, including their apparent influence on the incidence of IPV perpetration.