Intimate Partner Violence and its Association with Symptoms of Depression and Post-Traumatic Stress Disorder Among HIV-Negative Cisgender Sexual Minority Men in the United States

Abstract

We examined mediators of the association between intimate partner violence (IPV) and mental health among 500 HIV-negative cisgender sexual minority men (SMM) residing across the United States—a population disproportionately impacted by depression and post-traumatic stress disorder (PTSD). Depressive symptoms were measured with the Patient Health Questionnaire (PHQ-8) and PTSD symptoms with the Primary Care PTSD Checklist (PC-PTSD-5). Forty percent (n = 20) of the sample reported any IPV (victimization or perpetration) in the past 6 months. IPV (represented by mean number of IPV domains with at least one behavior that occurred more than once) was positively correlated with symptoms of depression and PTSD, and these associations were partially mediated (both direct and indirect effects were statistically significant) by social support, coping self-efficacy, substance use consequences, and sexual orientation discrimination; relationship communication partially mediated the association between IPV and depressive symptoms. These findings highlight the strength of the direct association between IPV and mental health for SMM. While the psychosocial and relationship characteristics studied in this analysis may help inform strategies for helping SMM to manage stress related to IPV, multicomponent interventions that address both direct and indirect pathways are likely needed.

INTRODUCTION

Prevalence rates of mental health problems are disproportionately high among sexual minority men (SMM; i.e., gay, bisexual, and other men who have sex with men) in the United States (U.S.), including rates of depression and post-traumatic stress disorder (PTSD). Systematic reviews of published research (most of which was conducted in the U.S.) show that cisgender SMM are twice as likely to experience clinical depression and four times as likely to attempt suicide (King et al., 2008; Ploderl & Tremblay, 2015), compared to heterosexual populations. Similarly, a systematic review of mostly U.S.-based trauma exposure research among sexual minority populations found trauma rates that were roughly double that of the general population (Marchi et al., 2023), and prevalence of PTSD as high as 40% (Pantalone et al., 2020). Among the many psychosocial factors that can contribute to these mental health challenges is intimate partner violence (IPV) (Calton et al., 2016; Plichta, 2004; Spencer et al., 2019; Trombetta & Rollè, 2023), and we sought to examine factors that may influence the association between IPV and mental health among SMM in the U.S.

IPV, defined as psychological, physical or sexual harm by a current or past partner (Center for Disease Control and Prevention, 2024), can involve a myriad of physical and psychological health sequalae (Calton et al., 2016; Plichta, 2004; Spencer et al., 2019; Trombetta & Rollè, 2023). IPV victimization affects roughly a third of cisgender SMM, and one in four report IPV perpetration (Liu et al., 2021), in assessments of lifetime experience, according to a meta-analysis of research conducted mostly in the U.S. About half of SMM who experience IPV report both victimization and perpetration (i.e., bidirectional IPV) (Messinger, 2018). A systematic review of mostly U.S.-based research showed that IPV was among the strongest correlates of depression among cisgender SMM (Buller et al., 2014). These findings hold true for both IPV victimization and perpetration (Sharma et al., 2021; Stults et al., 2021; Zavala, 2016). Similarly, PTSD symptoms among sexual minority populations are positively correlated with experiencing IPV victimization (Walters et al., 2013), as well as perpetration (Basting et al., 2024). Some studies of cisgender SMM suggest that the perpetrators of IPV may be more likely to experience depression and PTSD compared to victims of IPV, given the strong association between IPV perpetration and histories of childhood abuse and violence (Miltz et al., 2019). IPV is often thought of as leading to depression or distress, but the relationship between mental health and IPV may be bidirectional, as depressive and PTSD symptoms may enhance vulnerability to dysfunctional relationship dynamics, as adaptive coping mechanisms are compromised (Beck, 2008; Herek & Sims, 2007).

IPV may be a strong contributor to depression and PTSD, but several other psychosocial factors may also influence these mental health problems. Minority stress theory posits that mental health disparities between sexual minority and heterosexual populations stem from chronic exposure to distal (e.g., heterosexist discrimination) and proximal (e.g., internalized homophobia) sexual identity-related stressors, and these stressors can result in adverse behavioral manifestations (e.g., substance abuse) that also pose mental health challenges (Meyer, 1995, 2003). Consistent with this theory, sexual identity-related stressors and substance abuse have been shown to be associated with higher levels of depression and PTSD among sexual minority populations including cisgender SMM in studies mostly conducted in the U.S. (Goldbach et al., 2014; Lewis et al., 2012; Rendina et al., 2017), as well as experiences of IPV (Edwards et al., 2015; Longobardi & Badenes-Ribera, 2017). Resilience theory uses a strengths-based approach to explaining how individuals are able to cope and adapt in the face of adversity (Szymanski & Gonzalez, 2020). Social support and coping methods are resilience factors that have been associated with IPV as well as depression and PTSD among cisgender SMM in the U.S. (Argyriou et al., 2021; Brandt & Stults, 2024; Goldberg-Looney et al., 2016). Furthermore, with IPV occurring in the context of relationships, it is relevant to consider the role of relationship quality and communication, which has been associated with depression, PTSD and IPV among coupled cisgender SMM in the U.S. (Bosco et al., 2022; Sarno et al., 2022; Thomeer et al., 2015).

Prior research and the theories of minority stress and resilience highlight how the factors described above are associated with IPV and symptoms of depression and PTSD. However, what is less studied is how these factors may mediate or explain how or why IPV is associated with mental health problems among cisgender SMM (Hung et al., 2022). For example, the experience of IPV may contribute to substance abuse and reduced coping skills, which in turn lead to mental health problems. Understanding the role of these factors in how IPV relates to depression and PTSD may inform strategies for helping men manage and cope with IPV and mitigate its influence on mental health.

We conducted an online survey of 500 HIV-negative cisgender SMM residing across the U.S. who enrolled in project EROS (Empowering Relationships and Opportunities for Safety), an ongoing longitudinal observational cohort study. From these data, we examined experiences of IPV and its association with depression and PTSD, and whether measures of social support, coping self-efficacy and relationship communication are negatively correlated with IPV and symptoms of both depression and PTSD, and measures of discrimination and substance use consequences are positive correlates, as we hypothesize. We also assess whether these psychosocial variables mediate the associations between IPV and symptoms of depression and PTSD.

METHODS

Study Design

A cross-sectional analysis was conducted with data from the EROS project. Participants were recruited online and were followed over 24 months; only baseline data were used in the analysis reported here. The study protocol was approved by the Institutional Review Board at San Diego State University (HS-2022–0094). Details of the study protocol are available in an open access publication (Storholm et al., 2022).

Participants

Participants enrolled in the study between September 2022 and December 2023. Recruitment was through online advertisements placed on social networking sites (i.e., Facebook and Instagram) and mobile dating/hook-up apps (e.g., Scruff, Jack’d, Grindr), as well as flyers distributed at bars, gyms, Pride events, and LGBTQ+ community organizations. Potential participants who clicked on the online advertisements or who scanned the quick response (QR) code displayed on the physical flyers, were directed to a short self-administered Qualtrics survey that provided information about the study, consent to be screened, and screening questions to assess eligibility.

The primary objective of the larger study was to examine the influence of IPV on HIV protective behaviors among HIV-negative SMM. Therefore, to be eligible, prospective participants needed to report: (1) being cisgender male (assigned male sex at birth and currently identify as a man); (2) being in a romantic relationship with another cisgender man for a minimum of three months at the time of screening (but not necessarily at the time of the baseline survey); (3) being 18–45 years of age; (4) having HIV-negative or unknown HIV status; (5) living in one of the Center of Disease Control and Prevention (CDC)-defined U.S. Ending the HIV Epidemic (EHE) priority jurisdictions (Fauci et al., 2019); and (6) willingness to complete home-based HIV/STI test kits. To reduce risk of violence resulting from participation in the study, participants were informed that only one partner in a relationship dyad was eligible to participate. Study staff verified partner non-participation prior to enrollment by collecting the partner’s first name and last 4 digits of their phone number from potential participants and checking these data against all enrolled participants.

After completing the initial screening, participants were not considered to be fully eligible until they completed the following three steps: (1) attended a live 15–20-minute virtual onboarding session with a study coordinator, which consisted of identity verification (via identification card), orientation to the study, and obtaining written consent via electronic signature; (2) completed a 45–60 minute online, self-administered baseline survey; and (3) returned a viable HIV/STI test kit to confirm HIV-negative status at baseline. Participants were remunerated $50 (e-gift cards) for completion of all baseline tasks (virtual onboarding, virtual survey, returned home-based HIV/STI test kit). In total, 9010 unique persons responded to the study ad by completing a screener, 1042 were eligible for the study based on their screener data, 742 attended the virtual onboarding session, 690 completed the baseline survey, and 522 returned a viable test kit; 22 were not eligible due to a HIV seropositive test result, these participants were notified and immediately linked to HIV care. Therefore, a final sample of 500 SMM enrolled in the cohort and constitute the analytic sample for this analysis.

Measures

Intimate partner violence (IPV)

We measured IPV with 62 items from the validated IPV for gay and bisexual men scale (IPV-GBM; Stephenson & Finneran, 2013), including adapted items related to economic abuse (Postmus et al., 2016); high internal reliability (Cronbach’s alpha = .90) has been documented by the developers of the scale (Stephenson & Finneran, 2013), and our data revealed a Cronbach’s alpha of .84 and .90 for the perpetration and victimization items, respectively. For each item, respondents were asked to respond yes/no to whether they had experienced the behavior in the prior six months; each item was asked twice, to assess both victimization (if they experienced the abuse behavior from a male partner) and perpetration (if they did the behavior to a male partner). To address shortcomings with IPV measures used in prior research with SMM, we also asked participants to exclude any behaviors that were part of consensual sexual behavior (e.g., role play), or that were done by accident. The behaviors span across eight domains of IPV: controlling (13 items; e.g. demanded access to phone, controlled access to family), emotional (7 items; e.g. criticized appearance, told you no one else would want you), financial (3 items; e.g. racked up debt in your name to intentionally damage your credit), identity-based (5 items; e.g. threatened to “out” you, call immigration services), physical (8 items; e.g. slapped, threw objects), sexual (6 items; e.g. raped, forced to watch pornography), sexual health (12 items; e.g. prevented from taking pre-exposure prophylaxis [PrEP], pressured to take off condom), and stalking (8 items; e.g. showed up to places unwanted, used technology to monitor). For any behaviors indicated as having occurred (yes), respondents were asked if the behavior occurred once or more than once in the past six months. Respondents were asked to answer questions with respect to their current male partner (if applicable) and any other male partners in the past 6 months. Immediately following this section of the survey, all participants were provided with contact information for the national Domestic Violence Hotline. At the end of the survey, all participants received a resource guide with national- and state-level resources for IPV, mental health, substance use, and sexual health.

For analysis, the following sums of IPV behaviors and behavioral domains were computed: number of victimization behaviors that occurred more than once, number of perpetration behaviors that occurred more than once, and the total of these two sums to represent the overall number of all IPV behaviors that occurred more than once in the past six months; number of victimization abuse domains that had at least one behavior that occurred more than once, number of perpetration abuse domains that had at least one behavior that occurred more than once, and the total of these two sums to represent the total of all IPV domains with at least once behavior that occurred more than once. This approach to deriving summary measures of IPV has been used in other published research using data from the same IPV measure (Walsh & Stephenson, 2023). We chose to use the approach of counting behaviors that happened more than once, because this may be a better representation of impactful IPV exposure, compared to behaviors that happened only once.

Respondents who reported victimization and perpetration behaviors were classified into three IPV groups based on the difference between sums of victimization and perpetration domains with at least one behavior that occurred more than once: balanced bidirectional (the two sums were the same), only/predominantly victimization (victimization sum was greater than the perpetration sum), and only/predominantly perpetration (perpetration sum was greater than the victimization sum). Respondents who reported only victimization behaviors or only perpetration behaviors were included in the corresponding group.

Mental health

We measured depressive symptoms with the 8-item version (Cronbach’s alpha = 0.90) of the Patient Health Questionnaire (PHQ-8) (Kroenke et al., 2009). Each item corresponds to one of the symptoms of depression used to diagnose major depression in the Diagnostic Statistical Manual, 5^th Edition (DSM-5) (American Psychiatric Association, 2013), apart from suicidal ideation, which is omitted in the PHQ-8. Each item is scored 0 ‘not at all’ to 3 ‘nearly every day’ to represent the frequency of the symptom over the past two weeks; total score ranges from 0–24, and scores > 9 have been shown to highly correspond to Major Depression. Symptoms of post-traumatic stress disorder (PTSD) were assessed with the 5-item (Cronbach’s alpha = 0.81) Primary Care PTSD Screen for DSM-5 (PC-PTSD-5) (Prins et al., 2016); respondents were first asked if they had ever experienced a traumatic event, and if so, were then asked about the presence of five symptoms [e.g., had nightmares about the event(s); thought about the event(s) when you did not want to] in the past month using a yes/no response; sum of present symptoms was calculated (possible range: 0–5). Participants without any traumatic event exposure received a total score of “0”. Symptoms of negative affect are present in the measures of both depression and PTSD; however, the negative cognitive and affective symptoms of PTSD are directly related to the traumatic event, whereas those of depression are more generalized to the person’s overall mental state (Post et al., 2011).

Potential mediators

Social support was assessed with two subscales [received emotional (Cronbach’s alpha = 0.82) and received instrumental (Cronbach’s alpha = 0.89) support] from the Berlin Social Support Scales (Schulz & Schwarzer, 2003), the scores from which were summed (possible range: 8–32) and higher scores reflect greater support. Coping self-efficacy was assessed with the Coping Self-Efficacy Scale (Chesney et al., 2006), which measures the respondent’s confidence in their coping abilities when faced with life challenges. The scale has three subscales: problem-focused coping (Cronbach’s alpha = 0.91), stop unpleasant emotions and thoughts (Cronbach’s alpha = 0.92), and get support from friends and family (Cronbach’s alpha = 0.81). The total score (sum of the three subscales) was used in analysis (possible range: 0–130), and higher scores indicate more confidence in one’s stress coping abilities. Relationship communication was assessed with the 3-item positive interaction subscale (e.g., Both of us try to discuss the problem; Both of us suggest possible solutions and compromises) of the Communication Patterns Questionnaire-Short Form (Futris et al., 2010), which captures mutual constructive communication between the partners when discussing a conflict (possible range: 3–9; Cronbach’s alpha = 0.66). This measure was only asked if the respondent reported being in a current committed relationship. Sexual orientation discrimination was measured using a modified version of the 10-item subscale (Cronbach’s alpha = 0.68) of Multiple Discrimination Scale (MDS) (Bogart et al., 2013) to measure participants’ experienced specific discrimination events based on their sexual orientation; response options for each item was yes/no and a count of affirmative responses was calculated (possible range: 0–10). To assess substance use we used the Alcohol, Smoking, and Substance Involvement Screening Test (National Institute on Drug Abuse, 2009) to measure use frequency of alcohol and illicit drugs over the past three months. The frequency of substance use consequences was assessed with a single item that asked, “How often in the past 3 months, did the use of any of these substances lead to health, social, legal or financial problems?”; response options ranged from 1 ‘never’ to 5 ‘daily’.

Demographics

We assessed age (years), race/ethnicity (American Indian or Alaska Native, Asian, Black, Native Hawaiian or Other Pacific Islander, Hispanic/Latino, White, other), years of formal education, current employment (part- or full-time), state and city or town of residence (which was used to designate the national region of their residence, and whether they lived in an urban setting), and relationship status (single or casually seeing someone, versus in a relationship).

Data Analysis

Descriptive statistics were used to characterize the sample with regards to IPV, mental health and the constructs being examined as mediators. Bivariate statistics were used to assess correlates of three measures of IPV (victimization or perpetration) over the past 6 months (binary indicator of any reported IPV behavior, number of IPV behaviors that occurred more than once, and the number of IPV domains in which at least one behavior occurred more than once), and the two mental health measures, from among the five potential mediators using 2-tailed, independent t-tests and correlations (Pearson or point-biserial, depending on whether both measures were continuous or one variable was dichotomous). One way ANOVA was used to compare the three IPV groups on the two mental health measures.

Next, using an approach described by Hayes and Preacher (2014), we evaluated whether the proposed mediators mediated the association between IPV (measured by the total sum of IPV domains in which at least one type of event occurred more than once) and depressive and PTSD symptoms, in separate analyses. To identify our mediation models, we required there to be significant (p<.05) associations between (1) IPV and the mental health measure; (2) IPV and the potential mediator; and (3) the potential mediator and the mental health measure. If these three requirements were met, we proceeded with mediation analysis using structural equation modeling (SEM), as described by Gunzler et al. (2013). We first estimated a linear regression model with just IPV and the mental health measure in the model, excluding the potential mediator. Second, we estimated a mediation model using SEM that included the potential mediator, and assessed whether the mediator changed the magnitude of the association between IPV and the mental health measure. This stepped approach has been used successfully in several published analyses (Ghanem et al., 2020; Matovu et al., 2023; Wagner et al., 2023). Lastly, because many of the mediators were significantly correlated with each other, we conducted an additional model that included all mediators that were inter-correlated, following guidance from VanderWeele and Vansteelandt (2014). The direct effect is the association between IPV and the mental health measure, and the indirect effect represents the portion of the relationship between IPV and the mental health measure that is mediated by the mediator. R Studio version 2024.04.2+764, SPSS Statistics version 29 and STATA were used for analysis.

RESULTS

Sample Characteristics

Table 1 lists the sample characteristics of the 500 enrolled men, including demographic and mental health characteristics, and the variables being examined as mediators in the analysis. The sample resided across 33 states, with representation across the four regions (West, South, East, Midwest) of the United States. Mean age was 32.7 years (SD=6.2), 72.6% had completed at least four years of college, 81.0% were currently employed, and 88.8% were in a relationship at the time of data collection. Racial and ethnic identification was diverse, including 30.0% non-Hispanic white, 34.2% Hispanic, 16.8% Black, and 14.6% Asian or Pacific Islander.

Table 1.

Sample characteristics for the whole sample and a comparison of characteristics by presence of any form (victimization or perpetration) of IPV in past 6 months

	Total (n=500)	No IPV1 (n=298)	Any IPV1 (201)
	Mean (SD)/n (%)	Mean (SD)/n (%)	Mean (SD)/n (%)	p
Age	32.7 (6.2)	32.7 (6.3)	32.8 (6.1)	.92
Race/ethnicity				.04
Non-Hispanic White	150 (30.0%)	90 (30.2%)	60 (29.9%)
Black or Multiracial Black	84 (16.8%)	55 (18.5%)	29 (14.4%)
Hispanic or Multiracial Hispanic	171 (34.2%)	87 (29.2%)	83 (41.3%)
API or Multiracial API	73 (14.6%)	50 (16.8%)	23 (11.4%)
Other	22 (4.4%)	16 (5.4%)	6 (3.0%)
At least 4 years of college	363 (72.6%)	211 (70.8%)	152 (75.6%)	.24
Employed	405 (81.0%)	248 (83.2%)	156 (77.6%)	.12
Region of residence				.96
Northeast	118 (23.6%)	67 (22.5%)	50 (24.9%)
South	165 (33.0%)	102 (34.2%)	63 (31.3%)
West	149 (29.8%)	86 (28.9%)	63 (31.3%)
Midwest	68 (13.6%)	43 (14.4%)	25 (12.4%)
Urban location	121 (24.3%)2	70 (23.6%)3	50 (24.9%)	.75
Currently in a romantic relationship	444 (88.8%)	272 (91.3%)	174 (86.4%)	.06
Depression (PHQ-8 total)	6.0 (5.3)	4.9 (4.6)	7.6 (5.9)	<.001
Depressed (PHQ-8>9)	113 (22.6%)	44 (14.8%)	69 (34.5%)	<.001
PTSD symptoms (PCL-5)	0.8 (1.5)	0.5 (1.3)	1.2 (1.7)	<.001
Substance use consequences	0.11 (0.44)	0.08 (0.38)	0.17 (0.52)	.02
Never	460 (92.0%)	282 (94.6%)	177 (88.1%)
Once or twice	29 (5.8%)	12 (4.0%)	17 (8.5%)
Monthly	6 (1.2%)	2 (0.7%)	4 (2.0%)
Weekly	4 (0.8%)	1 (0.3%)	3 (1.5%)
Daily or almost daily	1 (0.2%)	1 (0.3%)	0
Illicit drug use at least weekly	58 (11.6%)2	31 (10.4%)	27 (13.5%)	.29
Social support	26.4 (4.7)	26.9 (4.7)	25.7 (4.6)	.01
Coping self-efficacy	88.2 (22.8)	90.4 (23.6)	84.8 (21.4)	.01
Relationship communication4	7.5 (1.4)	7.7 (1.4)	7.1 (1.4)	<.001
Sexual orientation discrimination	0.57 (1.13)	0.39 (0.88)	0.84 (1.39)	<.001

IPV = intimate partner violence; PTSD = post-traumatic stress disorder; PHQ-8 = 8-item Patient Health Questionnaire; PCL-5 = 5-item PTSD check list; SD = standard deviation; API = Asian Pacific Islander

¹499 participants provided data on IPV; any IPV reflects any report of IPV victimization or perpetration in past 6 months

²498 participants provided data

³297 participants (one case with missing data from this subgroup) provided data

⁴This measure was only assessed among the 444 who reported currently being in a relationship

IPV Prevalence

Two hundred and one participants (40.2%) reported at least one behavior of either IPV victimization or perpetration with a male partner in the prior 6 months; 185 (37.0%) reported any victimization and 125 (25.0%) reported any perpetration, including 109 (58.9% of those reporting victimization and 87.2% of those reporting perpetration) who reported both victimization and perpetration. Based on the criteria defined in the Methods, these 109 men consisted of 53 (48.6%) with balanced bidirectional IPV, 37 (33.9%) predominantly victimization, and 19 (17.4%) predominantly perpetration. Adding the latter two groups to the men who reported only victimization (n=76) and only perpetration (n=16), the 201 participants reporting any IPV were classified into these three IPV groups: balanced bidirectional (n=53; 26.4%), only/predominantly victimization (n=113; 56.2%), and only/predominantly perpetration (n=35; 17.4%).

Table 1 lists the sample characteristics of those who reported any IPV (n=201; 40.2%) and those who did not [n=298 (59.6%); one person did not provide data], with regards to demographics, mental health characteristics, and the measures of the potential mediators. Among those who reported any IPV, the mean number of IPV behaviors that occurred more than once in the prior six months was 5.8 (SD=11.1; median=2), of which most were victimization behaviors [mean (SD) = 4.0 (7.8); median=2). The mean number of IPV domains in which at least one behavior type occurred more than once was 1.7 (SD=2.1; median=1), of which most were victimization domains (mean=1.1; SD=1.4; median=1). Among those reporting any victimization (n=185), the proportion who reported at least one behavior type within the domains of abuse victimization were: emotional (54.1%), controlling (50.8%), physical (35.1%), stalking (25.5%), sexual (25.4%), identity (18.4%), sexual health (17.8%), and financial (14.6%). Among those reporting any perpetration (n=125), the proportion that reported at least one behavior type within the domains of perpetrated abuse were: emotional (46.4%), controlling (47.2%), physical (42.4%), stalking (19.4%), sexual (12.8%), identity (10.4%), sexual health (10.4%), and financial (6.4%).

Associations Between IPV and Mental Health

The mean PHQ-8 total score for the whole sample was 6.0 (SD=5.3), and 113 (22.6%) participants had scores reflective of major depression (PHQ-8 score > 9). The mean number of PTSD symptoms reported was 0.8 (SD=1.5), including 134 (33.5%) participants who reported at least one symptom. The subgroup of 201 men that reported any IPV event in the past six months had significantly higher depressive (mean = 7.6 vs. 4.9; t= 5.5, df=352, p<.001) and PTSD (mean = 1.2 vs. 0.5; t=4.4, df=345, p<.001) symptoms, compared to the men who did not report IPV. Among those who reported any IPV, mean depressive symptoms [bidirectional: 8.6; only/predominantly victimization: 7.3; only/predominantly perpetration: 6.9; F (df) =1.1 (2,197), p=.33] and PTSD symptoms [bidirectional: 1.1; only/predominantly victimization: 1.2; only/predominantly perpetration: 1.2; F (df) =0.02 (2,197), p=.99] did not differ significantly across the three IPV groups. Due to these lack of differences across the IPV subgroups, all subsequent analyses involved measures of IPV that merged reported victimization and perpetration events.

Table 2 lists the bivariate correlations between three IPV measures (binary indicator of any reported IPV behavior, total sum of victimization and perpetration IPV behaviors that occurred more than once, total sum of victimization and perpetration IPV domains in which at least one behavior occurred more than once) and the measures of depressive and PTSD symptoms. Each IPV measure was significantly positively correlated with depressive (coefficients range from .25 to .31) and PTSD (coefficients range from .20 to .26) symptoms.

Table 2.

Bivariate correlates of IPV and mental health (depression and PTSD) measures

Variable	Any IPV1	Sum of IPV behaviors2	Sum of IPV domains3	Depressive symptoms	PTSD symptoms
Mental Health
Depressive symptoms (PHQ-8 total)	.25***	.30***	.31***	--	.41***
PTSD symptoms (PCL-5)	.20***	.25***	.26***	.41***	--
Potential Mediators
Social support	-.12**	-.11*	-.14**	-.33***	-.16***
Overall coping self-efficacy	-.12*	-.13**	-.16***	-.48***	-.18***
Substance use consequences	.10*	.32***	.28***	.36***	.23***
Discrimination related to sexual orientation	.19***	.13**	.23***	.22***	.31***
Relationship communication4	-.21***	-.24***	-.29***	-.19***	.01

¹Binary indicator of any (victimization or perpetration) reported IPV behavior in past 6 months;

²Number of IPV behaviors (victimization or perpetration) that occurred more than once in past 6 months;

³Number of IPV domains (victimization or perpetration) in which at least one behavior occurred more than once in past 6 months

⁴This measure was only assessed among the 444 who reported currently being in a relationship

IPV = intimate partner violence; PTSD = post-traumatic stress disorder; PHQ-8 = 8-item Patient Health Questionnaire; PCL-5 = 5-item PTSD check list

^*p<.05;

^**p<.01;

^***p<.001

Mediators of the Associations Between IPV and Mental Health

Table 2 lists the bivariate correlations between the measures of the potential mediators (social support, coping self-efficacy, substance use consequences, sexual orientation discrimination, relationship communication) and the measures of IPV (binary indicator of any reported IPV behavior, number of IPV behaviors that occurred more than once, and the number of IPV domains in which at least one behavior occurred more than once) and mental health (depressive and PTSD symptoms). Each IPV measure was significantly associated with each of the potential mediators, with negative coefficients in the associations with social support, coping self-efficacy and relationship communication (coefficients ranged from −.11 to −.29), and positive coefficients with substance use consequences and sexual orientation discrimination (coefficients ranged from .10 to .32). Both depression and PTSD symptoms were significantly negatively correlated with social support and coping self-efficacy (coefficients ranged from −.16 to −.48), and significantly positively correlated with substance use consequences and sexual orientation discrimination (coefficients ranged from .22 to .36); relationship communication was significantly negatively correlated with depressive symptoms (r= −.19), but not PTSD symptoms. Each potential mediator was then evaluated as a mediator of the association between IPV [measured by the total sum of IPV domains (i.e., forms of IPV, both victimization and perpetration) in which at least one behavior occurred more than once] and depressive symptoms, and all but relationship communication was examined as a mediator of the association between IPV and PTSD symptoms, in separate analyses.

In analysis of the association between IPV and depression in which each mediator was examined in separate models, each potential mediator was found to be a partial mediator [beta (SE) for substance use consequences = −0.44 (.15), coping self-efficacy = −0.10 (.01), sexual orientation discrimination = 0.73 (.20), social support = −0.32 (.05), relationship communication = −0.38 (.17)]; in all models, both the indirect and direct effects were significant (see Table 3). Except for sexual orientation discrimination, each of the other four mediators was significantly correlated with each other mediator (coefficients ranges from −0.15 to 0.44; p values < .05); sexual orientation discrimination was only significantly negatively correlated with social support (r= −0.10) and significantly positively correlated with substance use consequences (r= 0.13) (see Table 2). Given the correlations among the other four mediators (social support, coping self-efficacy, substance use consequences, relationship communication), we added a model with all four of these mediators in the same model. Except for relationship communication, each of the other three mediators had a significant association with depressive symptoms [beta (SE) for substance use consequences = 1.71 (.28), which was positively correlated with depression; beta (SE) for coping self-efficacy = −0.07 (.01), and social support = −0.12 (.05), both of which were negatively correlated with depression], and together they partially mediated that association between IPV and depressive symptoms; both the direct and indirect effects were significant (see Table 3).

Table 3.

Structural equation model mediation analysis of factors that may mediate the association between IPV and symptoms of depression

	Depression [Beta (SE)]
	No mediator	Mediator: Substance use consequences	Mediator: Coping self-efficacy	Mediator: Sexual orientation discrimination	Mediator: Social support	Mediator: Relationship communication	Mediators: SU consequences Coping self-efficacy Social support Rel. communication
IPV	1.06(.15)***	1.06(.15)***	1.06(.15)***	1.06(.15)***	1.06(.15)***	1.06(.15)***	1.06(.15)***
Direct effect	-	1.15 (.15)***	0.81 (.13)***	0.94 (.15)***	0.93 (.14)***	1.08 (.16)***	0.60 (.13)***
Potential Mediators
SU consequences	-	−0.44 (.15)***	-	-	-	-	1.71 (.28)***
Coping self-efficacy	-	-	−0.10(.01)***	-	-	-	−0.07 (.01)***
Sexual orientation discrimination	-	-	-	0.73 (.20)***	-	-	-
Social support	-	-	-	-	-0.32 (.05)***	-	−0.12 (.05)**
Relationship communication	-	-	-	-	-	−0.38 (.17)*	−0.06 (.15)
Indirect effect	-	−0.09 (.04)***	0.24 (.07)***	0.12 (.04)**	0.14 (.05)**	0.10 (.05)*	0.61 (.14)***
Comparative fit index	-	1.00	1.00	1.00	1.00	1.00	1.00
% total effect mediated	-	8.3%	22.7%	11.6%	12.9%	8.8%	50.8%

IPV = intimate partner violence (measured by the total sum of IPV domains [both victimization and perpetration] in which at least one behavior occurred more than once in the past 6 months); SE = standard error; SU = substance use

^*p<0.05,

^**p<0.01,

^***p<0.001

In analysis of the association between IPV and PTSD symptoms, each of the four mediators examined was found to be a partial mediator [beta (SE) for substance use consequences = 0.35 (.08), coping self-efficacy = −0.01 (.002), sexual orientation discrimination = 0.27 (.03), social support = −0.03 (.01)]; in each of these models, both the indirect and direct effects were significant (see Table 4). When social support, coping self-efficacy, and substance use consequences were all included in the same model, all but social support was a significant correlate of PTSD symptoms [beta (SE) for substance use consequences = 0.33 (.08), which was positively correlated with PTSD; beta (SE) for coping self-efficacy = −0.01 (.003), and social support = −0.01 (.01), both of which were negatively correlated with PTSD], and together they partially mediated that association between IPV and PTSD symptoms; both the direct and indirect effects were statistically significant (see Table 4).

Table 4.

Structural equation model mediation analysis of factors that may mediate the association between IPV and PTSD symptoms

	PTSD [Beta (SE)]
	No mediator	Mediator: Substance use consequences	Mediator: Coping self-efficacy	Mediator: Sexual orientation discrimination	Mediator: Social support	Mediators: SU consequences Coping self-efficacy Social support
IPV	0.30 (.06)***	0.30 (.06)***	0.30 (.06)***	0.30 (.06)***	0.30 (.06)***	0.30 (.06)***
Direct effect	-	0.10 (.04)**	0.15 (.04)***	0.13 (.03)***	0.16 (.03)***	0.08 (.04)*
Potential Mediators
SU consequences	-	0.35 (.08)***	-	-	-	0.33 (.08)***
Coping self-efficacy	-	-	−0.01 (.002)***	-	-	−0.01 (.003)*
Sexual orientation discrimination	-	-	-	0.27 (.03)***	-	-
Social support	-	-	-	-	−0.03 (.01)**	−0.01 (.01)
Indirect effect	-	0.08 (.02)***	0.02 (.01)*	0.05 (.01)***	0.01 (.01)*	0.09 (.02)***
Comparative fit index	-	1.00	1.00	1.00	1.00	1.00
% total effect mediated	-	42.2%	13.0%	25.6%	7.9%	54.8%

IPV = intimate partner violence (measured by the total sum of IPV domains in which at least one behavior occurred more than once); SE = standard error; PTSD = post-traumatic stress disorder; SU = substance use

^*p<0.05,

^**p<0.01,

^***p<0.001

DISCUSSION

In this large, diverse sample of HIV-negative cisgender SMM recruited and surveyed online across the U.S., forty percent reported some form of IPV in the prior six months, and multiple measures of IPV were positively correlated with higher levels of both depression and PTSD symptoms. Drawing on the theories of minority stress and resilience, several psychosocial (substance use, discrimination, coping, social support) and relationship characteristics (quality of communication) were examined as potential mediators of the association between IPV and the two measures of mental health. While each of these factors were significantly associated with IPV, and symptoms of depression and PTSD, each factor only partially mediated the associations between IPV and depression and PTSD, highlighting the strength of the direct relationship between IPV and mental health in cisgender SMM in the U.S.

Our data revealed high rates of IPV and elevated symptoms of depression and PTSD. Similar to other studies of cisgender SMM in the U.S. (Liu et al., 2021; Messinger, 2018), about one third of the men in our study reported recent IPV victimization, and one fourth reported recent IPV perpetration; over half (54%) who reported any IPV experienced both victimization and perpetration. It is important to note that the measured IPV behaviors spanned a wide spectrum of forms (from verbal insults to physical and sexual assault) and frequency (singular behaviors to many behaviors across multiple domains). Also consistent with other studies of cisgender SMM (King et al., 2008; Livingston et al., 2020; Lukaschek et al., 2013; Pantalone et al., 2020; Ploderl & Tremblay, 2015), and studies of sexual minorities more broadly (Livingston et al., 2020; Pantalone et al., 2020), the self-reported measures in our study showed that one in five men had elevated depression symptoms reflective of clinical depression, and one third reported at least one PTSD symptom.

Aligned with findings from newly published systematic reviews of mostly U.S.-based studies involving cisgender SMM (Hong et al., 2025) and sexual and gender minorities more broadly (Rodrigues et al., 2025), our data revealed higher levels of IPV exposure were associated with more symptoms of depression and PTSD. While these systematic reviews only examined associations with IPV victimization, we assessed whether depression and PTSD symptoms differed amongst those experiencing predominantly IPV victimization, those who experienced predominantly perpetration, and those with balanced bidirectional IPV. Our data revealed no group differences on these mental health parameters. While there is a dearth of research on IPV perpetration and mental health among cisgender SMM, some studies have found IPV perpetrators to have higher levels of depression and PTSD compared to those experiencing IPV victimization (Hines & Saudino, 2002; Miltz et al., 2019). Researchers have hypothesized that dysfunctional aspects of these mental health problems (e.g., inability to control one’s emotions and/or heightened alertness to potential harm) (Bell & Naugle, 2008; Taft et al., 2007) may be driving this association.

One possible explanation for differing results across studies may be the complexity involved with classifying men who report both IPV victimization and perpetration. Analyses that account for one and not the other when classifying IPV experience among SMM will be confounded by the substantial overlap between victimization and perpetration, resulting in findings that may not adequately capture the experience of these men. We chose a criterion for classifying such men as either predominantly victimization, predominantly perpetration, or balanced bidirectional IPV, whereas other studies may have used other criterion or not attempted to make such a classification. Further research is needed to better understand what distinguishes these subgroups of IPV exposure, which will then enable valid methods for operationalizing the definition of these subgroups

Our use of the three-group classification of IPV resulted in relatively smaller numbers of men in each group and reduced statistical power, which likely contributed to our nonsignificant differences. The balanced bidirectional subgroup in our sample had numerically higher levels of depressive symptoms compared to the other two IPV subgroups; the difference may have been statistically significant with a larger sample and warrants further research. It is plausible that men who equally experience comparable levels of both IPV victimization and perpetration are confronted with corresponding multifaceted stressors that pose a significant challenge to mental health. While our research only included a single member of a relationship dyad, it is also plausible that the experience of both IPV victimization and perpetration reflects other relationship-level characteristics unable to be measured here (e.g., both members experiencing stressors including discrimination, unmet mental health needs, etc.). This is a potential area for future research, namely the unique and shared lived experiences and needs that each partner brings into the dyad, and the implications for how couples experience and respond to stressors inside and outside the relationship.

With the lack of evidence of significant differences between the IPV subgroups and mental health, we assessed several mediators of the associations between the measures of mental health (depressive and PTSD symptoms) and a measure of IPV that reflected the overall experience of all types of IPV victimization and perpetration. Substance use consequences and discrimination related to sexual orientation are identified by minority stress theory as the potential pathway by which IPV may influence mental health. Both have been found to be positively correlated with depression and PTSD among cisgender SMM (Goldbach et al., 2014; Lewis et al., 2012; Rendina et al., 2017), and they were as well in the SEM models in this study. However, both were only partial mediators of the associations between IPV and symptoms of depression and PTSD. Social support, coping, and relationship communication have been negatively correlated with depression and PTSD among cisgender SMM in other studies (Argyriou et al., 2021; Bosco et al., 2022; Brandt & Stults, 2024; Dawes et al., 2024; Goldberg-Looney et al., 2016; Sarno et al., 2022; Thomeer et al., 2015), and therefore as contributors to resilience these factors may serve as a mechanism by which IPV influences mental health. Each of these factors was significantly negatively correlated with both depressive and PTSD symptoms in the SEM models, but these factors also only partially mediated the associations between these measures of mental health and IPV. While these findings highlight the strength of the direct association between IPV exposure and both depressive and PTSD symptoms, they nonetheless also suggest that each of these mediating factors may help to inform the development of useful strategies to help cisgender SMM manage their emotional and psychological well-being when exposed to IPV.

Taken together, the present findings underscore the importance of examining IPV among cisgender SMM through trauma-informed and systems-based lenses. The consistent link between IPV and mental health highlights that interventions cannot focus solely on individual coping and symptom reduction. Rather, they must also address the relational and contextual dynamics that perpetuate cycles of IPV and mental distress. This requires that practitioners and researchers move towards the development and testing of interventions that target both the psychological and structural determinants of IPV-related trauma.

Study Limitations

Limitations to the study findings include the cross-sectional nature of the data, which does not allow for causal inferences, and constraints on the generalizability of the findings. The observed associations between IPV and mental health may be bidirectional, making it difficult to draw implications for how the examined mediating factors may contribute to mental health in the context of IPV exposure. Our sample does not adequately represent various important segments of the SMM population. The sample was largely recruited online, which enabled us to sample men living across the U.S.; however, this form of recruitment may have restricted our outreach to men with more resources, and indeed our sample was highly educated and lacks representation from men with low socioeconomic status. Also, those living with HIV were not enrolled and yet HIV has been shown to be a positive correlate of IPV and mental health problems among SMM (Lemons-Lyn et al., 2021). While the overall sample was large and diverse, some of the IPV subgroups (e.g., only/predominantly perpetration) were small, which limited the statistical power to assess the associations and mediation examined in the analysis. Other limitations include the lack of an assessment of whether perpetration was a response to the need for self-defense, which would enable us to better classify such behaviors—particularly among those who also reported physical forms of victimization (Messinger, 2018). Lastly, the self-report measure of IPV is limited by its susceptibility to social desirability bias and underreporting of IPV victimization and perpetration, and its lack of assessment of severity and frequency (beyond once or more than once).

Research and Practice Implications

The findings of the current study provide several implications for future research and practice. In addition to better delineating the experience of bidirectional IPV as described above, future research should prioritize longitudinal designs to better understand the causal relationships between experiences of IPV and mental health outcomes. This includes examining the impact of different forms of IPV and their multifaceted influences on mental health. Just as important is the need to examine how IPV affects mental health help-seeking behaviors. For instance, controlling IPV victimization may limit one’s ability to access healthcare services (McCloskey et al., 2007), while consistent psychological and emotional violence victimization may undermine self-esteem, hinder help-seeking behaviors, and increase engagement in sexual risk behaviors and substance use (Bosco et al., 2022; Miller et al., 2024; Santoniccolo et al., 2023), thereby indirectly contributing to mental distress.

From a clinical practice perspective, our findings point to several actionable priorities. First, social workers and mental health providers should incorporate routine IPV and trauma screening, as well as screening for other syndemic factors (e.g., substance abuse) that could serve as the mechanism by which IPV affects mental health (Kirschbaum et al., 2023), into standard assessments with SMM in safe, affirming, and confidential settings. IPV screeners should include items that capture experiences that are specific to SMM, such as threats to out someone’s sexual orientation or restricting access to affirming spaces. Second, programs should adopt trauma-informed care frameworks that recognize how repeated exposure to violence, stigma, and discrimination shape clients’ emotional regulation, trust, and attachment. Training social workers and clinicians in trauma-informed communication, safety planning, and crisis response is critical for fostering engagement and avoiding retraumatization.

Third, evidence-based treatments such as LGBTQ-Affirmative Cognitive Behavioral Therapy (Pachankis et al., 2022), Acceptance and Commitment Therapy (Stitt, 2022; Yadavaia & Hayes, 2012), Compassion Focused Therapy (Gilbert, 2010; Gilbert, 2014), and Skills Training in Affect and Interpersonal Regulation (Cloitre et al., 2020) could be adapted to address IPV-related PTSD, depression, and problematic relationship patterns among SMM in affirming ways. Interventions can be strengthened by embedding components that explicitly address minority stress, internalized stigma, and relationship dynamics. For example, cognitive behavioral therapy modules on increasing cognitive flexibility could incorporate exercises that challenge internalized homonegativity and IPV-related shame, while interpersonal skills training could be modified to improve communication and boundary setting in relationships. Practitioners can also facilitate group interventions that combines peer support with psychoeducation on health relationship behaviors, conflict resolution, and emotion regulation.

At a systems level, individual and structural interventions should be offered in tandem. For example, integrating IPV and trauma assessment and services into HIV and behavioral health programs such as substance use services. Strengthening partnerships between LGBTQ+ community organizations and IPV service providers and ensuring culturally competent referral pathways can enhance the continuity of care for SMM. Policies that expand access to IPV services for SMM, mandate provider training on IPV in same-gender relationships, and include IPV and PTSD assessment in routine health visits would help to translate this research into routine practice. Finally, structural interventions that target poverty, housing insecurity, and discrimination are essential to addressing the root causes of IPV and associated mental health dipartites among SMM.

A holistic approach to providing comprehensive trauma-informed mental health and IPV services (including enhancing coping strategies, fostering social support, and addressing risk for substance abuse) are needed to address the intersecting mental health challenges of cisgender SMM and promote resilience. By centering trauma-informed care, relationship-level interventions, and social justice advocacy within social work and mental health practice, professionals can move beyond crisis management towards prevention and empowerment for SMM experiencing IPV.

Conclusions

In conclusion, high levels of IPV and symptoms of depression and PTSD were found in this large sample of HIV-negative cisgender SMM sampled online across the U.S., and greater IPV was associated with more symptoms of depression and PTSD. Factors that could serve as mechanisms by which IPV influences mental health (e.g., social support, coping, relationship communication, sexual orientation discrimination, substance use consequences) were found to be correlated with both mental health and IPV, but only partially mediated the relationship between IPV and the measures of depressive and PTSD symptoms. These findings highlight the strength of the associations between IPV, depression and PTSD, and suggest that while the mediating factors may inform strategies for helping cisgender SMM to manage stress related to IPV, multicomponent interventions that address both direct and indirect pathways are likely needed. By integrating trauma-informed, minority stress responsive and relationship centered approaches, social workers and mental health practitioners can play a pivotal role in reducing the mental health burden associated with IPV and advance equity in care for SMM.

Availability of data, material, and code:

De-identified dataset and statistical code are available to researchers upon submission of proposal and review by the study team.