The use of intimate partner violence (IPV; i.e., perpetration) is widely considered to be a global health concern spanning a large range of acts including psychological (e.g., threats of violence), physical (e.g., punching or slapping), and sexual violence (e.g., forced sexual contact or coercion; (Breiding et al., 2015). Understanding the primary contributing factors in the use of violence remains a key focus of IPV research, and mental and physical health problems present a promising avenue for predicting the likelihood of using violence. In meta analytic research, IPV use positively associated with mental health conditions across type including mood, personality, and trauma and stressor related disorders (Spencer et al., 2024). Of course, not all individuals with mental and physical health concerns use IPV, and the relative impact of mental and physical health, as well as potential mechanisms that may predict IPV use, remains understudied. One population that may particularly benefit from this research is military veterans who often present with a high prevalence of both IPV use and health problems (Kwan et al., 2020). However, veteran populations remain understudied in the IPV literature specifically within mental health correlates of IPV (Trevillion et al., 2015).

Following discharge from the military, veterans can encounter distinct psychosocial relationship challenges such as loss of purpose, identity confusion, and disconnection from civilian loved ones (Romaniuk and Kidd, 2018) as well as consequences related to their service that further exacerbate relationships (e.g., geographic mobility, service-related injuries, mental health problems). These challenges can increase difficulty reintegrating into family life and heighten interpersonal conflict with partners and IPV use (Sayers et al., 2009). Given these issues with community reintegration, it is perhaps not surprising that veterans also report worse overall mental health and more adverse mental health days compared to their civilian counterparts (Hoglund and Schwartz, 2014). Indeed, research suggests 63-92% of veterans have reported psychological, 26% physical, and 12–40% sexual IPV use in the past year (Kwan et al., 2020).

However, while these IPV rates speak to the veteran population overall, it is increasingly important to address gender differences in the use of IPV regarding prevalence rates, salient risk factors, and relative impact of IPV use. While psychological IPV use appears relatively consistent across binary gender, in military populations, men report higher rates of physical IPV use (27%) vs. women (22%) in recent meta analyses (Kwan et al., 2020). Additionally, the gender specific impact of IPV remains disproportionate, with a higher rate of injury for partners when physical IPV is used by military samples of men (Kwan et al., 2020). Relatedly, while any gender can sustain injury from IPV use, data from IPV-related violent deaths has shown that women are significantly more likely to be killed as a result of IPV use, while men are more likely to die by suicide, commit homicide and then die by suicide, or die by legal intervention (Caldwell et al., 2012; Kafka et al., 2021). Individual risk factors for IPV use such as common mental and physical health concerns also differ in strength depending on gender. For instance, while alcohol use remains a significant risk factor for IPV use across gender, it presents as a stronger risk factor for men (Spencer et al., 2022). As such, consideration of gender differences within theory-driven research examining how mental and physical health concerns correlate with IPV use overall and by type among veterans is needed.

Many common physical and mental health conditions appear to be correlated with IPV use among veterans (e.g., LaMotte et al., 2017; Martin et al., 2010; Crane and Easton, 2017). Specifically, healthcare utilization research has demonstrated that veterans who use IPV are most likely to access Veterans Health Administration (VHA) services for symptoms related to PTSD, chronic pain, sleep problems (e.g., insomnia), and physical conditions related to alcohol use (e.g., liver problems; Relyea et al., 2023). These four conditions likewise translate into increased likelihood of IPV use among veterans: veterans endorsing PTSD symptoms above the clinical diagnostic cut-off, hazardous alcohol use, chronic pain, or sleep problems have all demonstrated increased likelihood of using IPV in the past year (LaMotte et al., 2017; Crane and Easton, 2017; Portnoy et al., 2023). This relationship between IPV use and these issues is particularly concerning given the elevated rates of these conditions among veterans vs. the general population (Lehavot et al., 2018; Taylor et al., 2024; Panza et al., 2022; Grant et al., 2015; Nie et al., 2024; Bai et al., 2023).

Although use of medical record data and diagnosis-based predictive models for IPV use is vital, less is known about how symptom-specific and sub-clinical levels of these conditions, as well as managing multiple different conditions, may impact IPV use. For instance, when PTSD and alcohol use were considered together in a logistic regression in a veteran sample, only drinking days were associated with recent aggression, suggesting some conditions may contribute more variance to aggression than others when considered together (Flanagan et al., 2014). Findings such as these emphasize the importance of replicating and extending existing research through the consideration of multiple conditions in a single model to begin to parse their nuanced relationships to IPV. This is particularly important for the increasing amount of work extending this line of inquiry into understudied physical health disorders like sleep and chronic pain in IPV research, as both conditions have shown strong preliminary support for their influence on IPV use in the context of PTSD (LaMotte et al., 2017). Given the paucity of research, a theory-driven approach to hypothesizing the relationships between multiple conditions and IPV use is particularly indicated.

I3 theory, which defines violence risk as a combination of (1) instigating, (2) impelling, and (3) inhibiting factors, provides a helpful framework to conceptualize the impact of mental and physical health risk factors on IPV use (Finkel and Hall, 2018). (1) Instigation occurs when someone is exposed to a situation that can elicit aggression as a behavioral response option. For instance, the perception of an immediate threat to safety may afford an aggressive behavioral response. This area of risk can be heightened by increasing the frequency of threatening events (e.g., conflict in a relationship) or the mis-attribution of events as threatening (e.g., perceiving a threat from a neutral stimulus). However, instigation does not operate alone, and violence risk is further amplified by impelling and instigating factors. (2) Impelling factors influence the psychological state a person is in when they experience instigation. They are not inherent to the instigating event but are instead the stable factors which impact vulnerability to use aggression. For instance, trait level anger may impact how intensely an individual responds to instigation. Therefore, instigating and impelling factors may work together to exacerbate the likelihood of violence. (3) Inhibiting factors can override or contribute to the likelihood of violence depending on their strength. Inhibiting factors can be impacted by low trait levels of self-control, the exhaustion of mental resources, or a biological impact on inhibition (e.g., the consumption of alcohol). Instigating, impelling, and inhibiting factors are highly interactive and may encompass state and trait level variables. Existing theoretical and empirical research are utilized to categorize risk factors within this model and, frequently within IPV literature, predictors of IPV use are conceptualized to operate across I3 risk factors (Massa et al., 2020), particularly since IPV often occurs as a dynamic cycle in which consequences can become subsequent predictors of violence and/or internalized as trait-like dispositions. Therefore, it is helpful to consider risk factors that operate across multiple areas of vulnerability or present with a particularly strong effect. For instance, mental and physical health conditions are comprised of different symptoms which aggregate to diagnostic criteria and, when taken together, these symptoms can represent areas of risk across I3 theory. Applied to empirically supported mental and physical health concerns associated with IPV use for veterans, I3 theory illustrates how multiple conditions may have transdiagnostic underpinnings that function across IPV use risk factors. PTSD, alcohol use, chronic pain, and sleep problems cause vulnerabilities across instigating, impelling, and disinhibiting factors. There are a number of possible ways in which these factors might interact to increase risk for the use of IPV. For example, alcohol use, sleep deprivation, and hypervigilance symptoms are associated with misinterpreting neutral stimuli as threatening, thereby increasing the propensity for aggression (Lanius et al., 2017; Giancola et al., 2011; Krizan and Herlache, 2016). Impelling factors present within these conditions such as experiencing low mood or affective dysregulation, which are common across these conditions, further increase the likelihood for aggression (Minkel et al., 2012; Turk et al., 2016). These individuals are then more likely to see threat (instigation), be impelled to violence given enduring states central to their physical and mental health conditions, and experience disinhibition which decreases the likelihood of non-aggressive conflict resolution. Individuals experiencing symptoms of PTSD, alcohol misuse, chronic pain, and sleep problems may also experience disinhibition whether through biological impairment (e.g., alcohol use or sleep deprivation) or self-regulatory fatigue (e.g., PTSD and chronic pain) which can interfere with their ability to effectively manage instigating and impelling factors (Solberg Nes et al., 2009; Sadeh et al., 2018).

As a result, veterans managing alcohol use, sleep problems, PTSD, and/or chronic pain may encounter more situations where they are inclined to use violence (instigating), experience vulnerability to aggressive responses (impelling), and have a lower capacity to inhibit their responses (disinhibiting). However, despite frequent endorsement of multiple mental and physical health problems and their overlap within prevalent theoretical models such as I3 theory, these conditions are rarely considered together. More research is needed to delineate how much variance is contributed by each condition in multiple regression models. Further, past research is limited by examination of risk factors across types of IPV. Large-scale meta-analyses define distinct mental and physical risk factors for physical vs. psychological IPV (Spencer et al., 2024, 2022) while sexual IPV is left largely unstudied. More research is needed to examine the distinct correlations of mental and physical health correlates of IPV across IPV subtypes.

Further extensions of previous IPV work should include examination of potential mechanisms. Per I3 theory, increased risk for IPV occurs when instigating and impelling factors are high while the ability to inhibit behavior is low. This combination of risk factors is referred to as a “perfect storm” for violence (Finkel and Hall, 2018). Within this “perfect storm,” inhibitory factors may provide insight into key mechanisms driving aggressive responses. Using the I3 framework, difficulties with emotion regulation would be considered a particularly critical disinhibiting factor that increases risk of aggression in the context of instigating and impelling factors. Researchers have described deficits within emotion regulation to include non-acceptance of negative emotions, difficulty with impulse control when distressed, limited access to effective regulation strategies, and deficits in emotional awareness and clarity (Gratz and Roemer, 2004). Individuals with poor emotion regulation are more likely to respond aggressively to perceived threats (i.e., instigating factors; Shorey et al., 2011; Watkins et al., 2016). A recent literature review on dating violence among college students found that emotion regulation deficits consistently predict both physical and psychological IPV in men and women (Neilson et al., 2023). Additionally, emotion regulation can moderate the link between negative affect and physical IPV; that is, when emotion regulation is low, negative affect is more strongly associated with IPV behaviors (Shorey et al., 2015).

While overall poor emotion regulation has been associated with an increased risk of IPV use, difficulties with impulse control, in particular, are especially relevant within veteran populations. Difficulties with impulse control refers to the inability to regulate or restrain behavior when experiencing negative emotions (Gratz and Roemer, 2004). Impulse control has been strongly linked to general aggression (Bresin, 2019), IPV use (Bresin et al., 2022), can mediate reactive aggression (Gagnon and Rochat, 2017), and has been positively associated with IPV use in both court-mandated cases (Grigorian et al., 2019) and university samples (Gildner et al., 2021). Veterans with poor impulse control in the face of negative emotions may be particularly likely to use IPV if they are feeling unwell from pain, sleep disturbance, and PTSD symptoms. For example, researchers have found that emotion regulation, as a broad construct, fully accounted for the relationship between PTSD and impulsive aggression in a veteran sample (Miles et al., 2016) and that men with more trauma exposures and higher negative urgency were more likely to use IPV (Gildner et al., 2021). Yet, impulse control difficulties specifically have rarely been considered as a risk factor for IPV use for veterans or included as a potential mechanism to explain the impact of chronic conditions such as chronic pain, alcohol use, sleep problems, and PTSD severity on IPV use among veterans.

The current study aimed to build on prior IPV use research by examining the association among mental health conditions, physical health problems, and impulse control difficulties in a community based, non-clinical veteran sample. The hypothesized model first defines the bivariate relationships between common mental and physical health concerns and impulse control difficulties and IPV use. Following this, we examine the relative contributions of PTSD, alcohol use, insomnia, and chronic pain on IPV use overall and by type. Finally, as PTSD is one of the most well-studied mental health correlates of IPV use and interventions within the VHA (Creech et al., 2018) and well outlined in I3 theory, we extended this current literature through the examination of impulse control difficulties as a potential moderator in the relationship between PTSD and IPV use. To advance this understanding of IPV use, we investigated the following hypotheses:

The current study extends the current literature by investigating the I3 theoretical framework to further explore the relationship between multiple risk factors for IPV use in a community sample of veterans (Finkel, 2014). Military veterans may be at heightened risk of IPV use due to the frequent endorsement of multiple chronic conditions such as PTSD, alcohol misuse, chronic pain, and sleep disturbances—each strongly linked to IPV risk factors (Irizar et al., 2021; Folmer et al., 2020; Qureshi et al., 2023; Trivedi et al., 2015). While PTSD, alcohol misuse, chronic pain, and sleep problems are independently associated with IPV use in veteran samples (LaMotte et al., 2017; Martin et al., 2010; Crane and Easton, 2017), the reporting multiple conditions with symptoms across clinical thresholds and within unrestricted demographic samples (e.g., open recruitment to all service eras, genders, and marital statuses) remains understudied. Additionally, existing research examining chronic conditions rarely consider underlying mechanisms which may drive IPV use such as impulse control difficulties.

Consistent with Hypothesis 1, findings indicated that PTSD symptoms, alcohol use, chronic pain severity, sleep problems, and impulse control difficulties were positively associated with overall IPV use, psychological aggression, and physical assault in the past year. Findings from the current study indicate that increasing levels of these symptoms were associated with higher levels of IPV when considered continuously (i.e., not dichotomized at clinical thresholds). This extends the prior literature base which focused on the relationship between IPV and conditions that were diagnostic or based on screening cut-offs. With further corroborating research, these findings may indicate the utility of IPV prevention and intervention programs aimed at assisting veterans with multiple issues across a board range of symptom severity.

However, inconsistent with previous studies (e.g., Taft et al., 2010; Shorey et al., 2012) and Hypothesis 1 predictions, sexual coercion use was positively associated with all variables except chronic pain. While this finding is contrary to some existing research (Taft et al., 2010), there is a significant lack of literature examining sexual coercion and IPV use among veterans experiencing chronic pain. Specifically, the assessment and interpretation of sexual IPV is broadly underassessed and frequently excluded from IPV studies further warranting examination within this study. In fact, sexual IPV data are so limited and heterogenous in military samples that they are frequently unable to be analyzed within systematic or meta analyses (e.g., Kwan et al., 2020). Within chronic pain studies specifically, research typically only examines physical or psychological IPV (Spencer et al., 2024; Crane and Easton, 2017; Singh et al., 2014) or includes brief (e.g., 3-item; Taft et al., 2010) measures of sexual IPV use. The current study is novel in that it includes the sexual coercion subscale of the CTS-2 as a distinct outcome. While this finding could point to differences within IPV typology, replication is necessary to warrant full interpretation.

To test Hypothesis 2 and examine both the combined and unique impacts of our model variables, our study examined these risk factors in a single model with supported controls (i.e., age, gender, income, and social desirability) on different types of IPV use (overall, sexual, psychological, and physical). Although the full model explained a significant amount of the variance in IPV use overall, no condition-specific model variables (e.g., PTSD, insomnia symptom severity) were uniquely and significantly associated with overall IPV use. However, impulse control difficulties were associated with psychological aggression, physical assault, and sexual coercion above all other empirically supported risk-factors.

These findings further emphasize the vital importance of impulse control difficulties in the annual frequency of IPV use across psychological, physical, and sexual sub-types, above and beyond chronic condition symptom severity or the endorsement of multiple health difficulties. I3 theory emphasizes the importance of the relative strength of these factors, proposing that violence is most likely when instigating and impelling factors are high while inhibiting factors are low (Finkel, 2014). Experimental paradigms have revealed that IPV use is higher for participants with high provocation (instigation), impelling factors (trait anger), and inadequate inhibitory strategies such as attempts at thought suppression (Birkley and Eckhardt, 2019). However, impulse control difficulties as a mechanism of IPV use in the context of mental and physical health conditions should not be over-interpreted in the current study. Contrary to Hypothesis 3, impulse control difficulties were not found to moderate the association between PTSD symptoms and IPV (overall, psychological, physical, or sexual). While PTSD severity did not uniquely predict IPV use among individuals with poor impulse control compared to those with better impulse control, difficulties with impulse control did emerge as a significant predictor of IPV use across the full sample. This suggests that impulse control challenges may contribute to aggression and PTSD severity broadly, rather than being confined to specific high-risk groups. Notably, prior research has indicated that emotion regulation may fully explain the link between PTSD and impulsive aggression in veteran populations (Miles et al., 2016). This is contrary to current study findings which did not support impulse control difficulties as a moderator between PTSD symptoms and IPV and suggests further examination and replication.

Study findings and implications should be understood in the context of a number of limitations. First, this study utilized a cross-sectional design; further experimental and longitudinal studies are needed to delineate the possible causal role of impulse control difficulties on IPV use in the context of chronic clinical conditions. This is particularly important given the assessment timelines across measures utilized within the current study. While alcohol and IPV are assessed within the past year, PTSD, chronic pain, and insomnia measures utilize timelines ranging from the past 2 weeks to the past 30 days. Therefore, the current study is unable to make direct claims about the co-occurrence of symptoms. Temporal studies can better address this important area of IPV research. Another limitation of this study was the lack of nationally representative veteran sample. While we utilized a sample without clinical cut-offs or population restrictions, data are not nationally representative in terms of sociodemographic variables for the national veteran population. For instance, 87% of study participants identified as Caucasian in the current study compared to only 76% in the total veteran population (VetPop2023, 2023), limiting generalizability. Additionally, in consideration of participant burden and data quality, we utilized brief screeners to assess mental and physical health problems. The selection of these measures was directly informed by Veteran Affairs screening procedures and informed measures used for alcohol misuse, PTSD symptoms, and sleep problems (Prins et al., 2016; Bush et al., 1998; Morin, 1993). Therefore, wherever possible, measures representing typical screening procedures for veterans accessing VHA care were used. However, the brevity of measures also precludes comprehensive assessment of conditions or the examination of particular symptom profiles, an important area for future inquiry. This may be particularly helpful for PTSD in which certain symptom clusters like hyperarousal have been isolated as specific risk areas for IPV use (Birkley et al., 2016).

Findings about IPV subtypes should be considered in the context of psychometric limitations. Sexual IPV measured via the “sexual coercion” subscale of the CTS-2 demonstrated poor reliability and low endorsement. Given this, overinterpretation of these findings is cautioned against and future research should consider present limitations to improve data collection. It is probable that differences in sexual coercion findings reflect overall limitations in the measurement of sexual IPV, particularly self or partner reports such as the CTS-2SF (Straus and Sugarman, 1996; Strauss and Warren, 2003). This measurement issue presents as a common barrier within IPV literature. In consideration of the absence of sexual coercion data within extant literature examining chronic conditions and IPV use, other studies may have experienced similar reliability issues and foregone the use of sexual coercion data to preserve data quality. Future studies should discuss their choice for inclusion or exclusion of sexual coercion data and provide reliability data on subsequent measures for use in scoping analyses and reviews. When included, sexual coercion measures may benefit from more extensive measures of social desirability than used in the current study. Previous findings demonstrate that under-reporting “undesirable” sexual behavior is particularly vulnerable to the influence of social desirability within surveys (King, 2022). Future studies may consider the use of longer, widely use measures of social desirability (e.g., Crowne and Marlowe, 1960) to better examine impact and facilitate comparisons across studies.

Future studies may also wish to expand the investigation of IPV and chronic conditions in veteran populations to additional, population-specific risk factors such as head injuries including traumatic brain injuries (TBIs). According to the National Academy of Sciences (Institute of Medicine, 2013), symptoms of TBIs can include problems with executive functioning, mood, physical impairment, and social dysfunction. These symptoms may operate across areas of IPV risk; for example, within longitudinal veteran samples, persistent post-concussion symptoms were found to predict IPV use above and beyond controls including binge drinking, pain severity, and PTSD (Portnoy et al., 2022). Examining head injury in the context of other chronic conditions and underlying risk factors (e.g., impulse control difficulties) would build on the current study and existing research to better define the relative importance of these factors on the likelihood of IPV use.

Future research should continue to integrate multiple risk and protective factors of IPV use within a cohesive theoretical framework rather than examining them in isolation. The I3 theory provides a valuable lens for understanding how predictors both interact and operate independently to heighten the risk of IPV use. Increasingly, research has highlighted emotion dysregulation as a key disinhibiting process linked to both PTSD (Seligowski et al., 2015; McLean and Foa, 2017) and IPV use (Maloney et al., 2022). Further investigation is needed to clarify which aspects of emotion dysregulation contribute most to poor mental health and aggression. For instance, research suggests that deficits in emotion regulation strategies, emotional clarity, and emotional acceptance are particularly associated with PTSD severity (Christ et al., 2021). Consequently, these may be prime targets for not only additional measurement in research, but also areas to target in future clinical interventions.

Findings suggest that difficulty inhibiting behavior when experiencing negative affect may be a driving factor for IPV use above and beyond the presence of multiple, common mental and physical health issues. Study findings should be understood in the context of limitations related to cross-sectional design, social desirability, and psychometric concerns. The role of impulse control on IPV use risk in veterans should be thoroughly examined in longitudinal and experimental designs, using nationally representative samples, and in consideration of related constructs (e.g., TBIs). Given the scarcity of research on underlying mechanisms of IPV use in veterans, as well as tailored intervention options for veterans who use IPV, additional research is needed to further delineate veteran-specific risk and predictive factors for IPV informed by the findings and limitations of the current study.