Gender Moderates the Association Between PTSD and Mutual IPV in an Emergency Department Sample

Abstract

Introduction:

Patients in emergency departments (EDs) constitute a diverse population with multiple health related risk factors, many of which are associated with intimate partner violence (IPV). This paper examines the interaction effect of depression, PTSD, impulsivity, drug use, adverse childhood experiences, at-risk drinking, and having a hazardous drinker partner with gender on mutual physical (IPV) in an urban emergency department sample.

Methods:

Research assistants surveyed 1037 married, cohabiting, or partnered patients in face-to-face interviews (87% response rate) regarding IPV exposure, alcohol and drug use, psychological distress, adverse childhood experiences (ACEs), and other sociodemographic features. IPV was measured with the Revised Conflict Tactics Scale. Interaction effects were examined in multinomial and logistic models.

Results:

Results showed a significant interaction of gender and PTSD (OR=3.06; 95%CI=1.21-7.23; p<.05) for mutual IPV. Regarding main effects, there were also statistically significant positive associations between mutual physical IPV and at-risk drinking (OR:1.73; 95%CI:1.07-2.77; p<.05), having a hazardous drinker partner (OR:2.19; 95%CI:1.35-3.55; p<.01), illicit drug use (OR:2.09; 95%CI:1.18-3.71; p<.01), adverse childhood experiences (OR:1.23; 95%CI:1.06-1.42; p<.01), days of cannabis use past in the 12 months (OR:1.003; 95%CI:1.002-1.005; p<.001), and impulsivity (OR:2.04; 95%CI:1.29-3.22; p<.01).

Conclusion:

IPV risk assessment in EDs will be more effective if implemented with attention to patients’ gender and the presence of various and diverse other risk factors, especially PTSD.

INTRODUCTION

Physical IPV is a major public health problem in the U.S. Results from the 2016-2017 National Intimate Partner and Sexual Violence Survey (NISVS), a random digit dialing survey by the Centers for Disease Control and Prevention, showed 12-month rates for physical IPV of 4.5% among women and 5.5% among men.¹ Emergency medicine departments (ED) have higher IPV prevalence rates, ranging from 9% to 37% for a 12-month timeframe, and as high as 46% for lifetime.^2–6 A previous analysis of the data herein showed a rate of 23.1% for physical IPV among men, and 23.3% among women.⁷

The higher rate of IPV among ED patients compared to household populations has several potential explanations. Frequently, EDs care for socioeconomically disadvantaged patients, who lack regular access to medical services because they are uninsured, and who have multiple comorbidities, including several behavioral risk factors associated with IPV. For instance, binge drinking, higher scores in the Alcohol Use Disorder Identification Test (AUDIT), post-traumatic stress disorder (PTSD), and alcohol and or illicit drug use are all present among ED patients and have all been positively associated with IPV.^4,8–11 These risk factors also have an additive effect; the more risk factors are present the higher is the risk for IPV.¹²

The present paper examines interaction effects between gender and some of the risk factors identified above, namely, depression, PTSD, impulsivity, drug use, adverse childhood experiences, at-risk drinking, and a hazardous drinker partner. There are many well-established risk factors for IPV, such as having a history of childhood maltreatment ^13,14 and others identified above. Less is known about how the associations between these risk factors and IPV may vary by gender. ^4,15 Some evidence supports gender specific associations between anxiety disorders, intermittent explosive disorder, alcohol abuse/dependence, and common mental disorders combined (e.g., depression, panic disorder, phobia) and IPV only among women.^16–18 Identification of interactions between risk factors and gender will provide knowledge of importance for screening and more precise identification of IPV origin in dyads reporting IPV.

With this background, the general hypothesis to be tested in this paper is that each one of the seven interactions tested (gender and depression, gender and PTSD, gender and impulsivity, gender and drug use, gender and adverse childhood experiences, gender and at-risk drinking, and gender and a hazardous drinker partner), will have a statistically significant and positive effect modification of the association between gender and IPV. This proposed interaction effect will strengthen the positive association between female gender and both perpetration only and mutual IPV (i.e., both perpetration and victimization; also referred to as bidirectional IPV) for each of the risk factors to be tested. This is because previous analyses by Cunradi et al. ⁷ and others showed that female partner violence (FMPV) is higher than male partner violence (MFPV) and that there are specific associations between certain risk factors and female gender.

METHODS

Sample and data collection

Trained, bilingual (English and Spanish) research assistants (RAs) recruited non-emergent patients in the ED of an urban Level I trauma center and county safety-net hospital, which serves mostly a low-income patient population, many on Medicaid, without medical insurance, and no ability to pay. The initial sample size estimate called for the enrollment of 800 married, cohabiting, or dating adults aged 18-50. This was based on calculations that using linear regression analyses, power would be 80% to detect a small overall effect (R² = .02) with 20 predictors, α = .05, and n=800. Power would be 85% to detect small incremental changes of adding single variables to the regression equations (ΔR² = .01) with 19 prior predictors, a prior R² of .10, and α = .05.

Participant eligibility criteria included: 18-50 years old; English or Spanish speaker, residence in the county where the study was conducted, and married, cohabiting, or in a romantic (dating) relationship for the past 12 months. The upper age limit was set based on consistent research evidence showing that most IPV occurs in younger age groups.¹⁹ Patients who were intoxicated, experiencing acute psychosis or suicidal or homicidal ideation, were cognitively and/or psychologically impaired and unable to provide informed consent, in custody by law enforcement, or in need of immediate medical attention were excluded. No data were collected on reasons for visiting the ED and on whether patients were undergoing behavioral or mental health treatment.

Figure 1 shows the recruitment sequence, with the final response rate of 87%, and the final sample size N=1037. Two interviewers per shift staffed the ED during weekday peak volume hours (9 am– 9 pm) to recruit patients to the study. Interviewers located and conducted face-to-face screening with patients in the ED waiting room or in a treatment cubicle. Eligible participants were offered the opportunity to participate in a confidential, face-to-face survey interview for which they would receive a $30 grocery store gift card incentive. Data were collected from February through December 2017. Patients could opt to be interviewed in English or Spanish. A Spanish version of the questionnaire, which had been validated through translation into Spanish and re-translation into English followed by verification, was used.

Figure 1:

Study sample recruitment

Once informed consent was obtained, patient survey data were collected by the RAs using computer assisted personal interview (CAPI) with computer tablets running the Qualtrics platform. The average survey interview completion time was 37 minutes (SD = 20.7). The project was approved by the Institutional Review Board of the hospital where the study was conducted.

Measurements

Unidirectional and Bidirectional Intimate Partner Violence:

Past-12 month physical IPV was measured with the Revised Conflict Tactics Scale (CTS2),²⁰ which has been used in prior ED-based IPV studies.^21–23 Patients were asked about violent behaviors they may have perpetrated against their spouse/partner, and that their spouse/partner may have perpetrated against them. This allows for identification of the patient as a perpetrator of violence, a victim of violence, or in a relationship with bidirectional violence. The CTS2 has 39 items divided into 6 subscales, physical assault (12 items), psychological aggression (8 items), negotiation (6 items), injury (6 items), sexual coercion (7 items). Respondents are asked to identify which items they experienced and report their frequency in an 8 points scale ranging from “never happened” to “more than 20 times in the past year”. Items are scored with a value that represents the mid-point for the response categories. For example, if the behavior happened 6 to 10 times in the past year, the score is 8. The study being reported used only the physical assault scale. Cronbach’s α for the scale in the dataset under analysis was .85.

Partner hazardous drinking:

The 3-item AUDIT-C was used to measure the respondent’s assessment of his/her spouse/partner’s drinking.^24,25 Male partners with a score above 4, and female partners with a score above 3 in the test 0-12 scale were considered hazardous drinkers. Internal consistency reliability for this scale in the data set under analysis as measured by Cronbach’s alpha was 0.81.

Cannabis use:

Participants were asked how many times during the past 12 months or 365 days they had used cannabis or hashish (weed, pot, hash) without a doctor’s instruction. Recreational cannabis use became legal in California in November 2016.

Illicit drug use:

This measure covered illicit drug use in the 12 months preceding the interview. Respondents were asked how many days they used the following drugs: amphetamines, cocaine, heroin, and prescription pain relievers “not prescribed for you.” Drug use was operationalized as any or no drug use.

At-risk drinking:

Respondents who drank alcohol in the past 4 weeks were asked: “What was the greatest number of drinks you had on any day in the past 4 weeks?”. A “drink” was defined as a 12-ounce can of beer, a five-ounce glass of wine, or a one-ounce shot of liquor. Respondents who did not use alcohol in the past 4 weeks were asked the same question over the past year. Women/men were considered at-risk drinkers if they had had four/five or more drinks on any one day in the past 4 weeks (past 12 months for past year drinkers).

Impulsivity:

This was measured with three items assessing respondents’ agreement with the following statements: I often act on the spur-of-the-moment without stopping to think; You might say I act impulsively; Many of my actions seem to be hasty. ^26,27 Four response categories ranged from “not at all” to “quite a lot,” with scores ranging from one to four per item. For the present analysis scores were divided into tertiles, and the scale was dichotomized with the two bottom tertiles coded as “zero” and the top tertial coded as “one.” Alpha reliability for the scale in the data set under analysis was 0.79.

Post-traumatic stress disorder (PTSD):

This measure is from the Primary Care Screener for PTSD, ²⁸ and it too has been successfully used in ED studies (see ^29,30). It asks subjects about past-month symptoms resulting from a “frightening, horrible or upsetting” experience. Answers were coded yes or no, and a score of three or more is considered positive. The internal consistency of this scale in the data set under analysis was α = 0.83.

Adverse childhood experiences:

This modified ACE measures exposure to six adverse experiences respondents may have had “during their first 18 years of life:” 1) exposure to a mentally ill person in the home, 2) parent/caregiver alcoholism, 3) sexual abuse, 4) physical abuse, 5) psychological abuse; and 6) violence directed against the respondent’s mother.³¹ The first 2 exposures were measured with one dichotomous question “yes” or “no”. The remaining 4 exposures were measured on a scale of 1 to 5 (never – very often). These six exposures were then summed to create the ACE variable (score range = 0–6). Internal consistency reliability (Cronbach’s alpha) in the data set under analysis was 0.74.

Other sociodemographic variables:

Gender. A dichotomous variable based on participant self-report coded as male and female (reference). Age. Coded as a categorical variable: 18-29, 30-39, and 40-50 (reference). Level of education. Respondents were categorized into four education categories: a) less than high school (reference); b) completed high school or GED; c) some college or technical or vocational school; d) completed 4-year college or higher. Importance of Religion. This variable had 4 categories: very important (reference), somewhat important, not very important, not important at all. Marital status: This is a 3-category variable: a) married (reference); b) cohabiting; c) single, separated or divorced. Food insufficiency: Respondents were asked their level of agreement with the statement, “In the past 12 months, the food we bought ran out and we didn’t have money to get more.” Response categories were never (reference), sometimes true, often true.³² Ethnicity: Based on self-identification. Respondents were asked: What racial or ethnic group(s) best describes you? Response categories were Asian; Black, African American; Latino, Hispanic; White, Caucasian; Native American/Alaskan Native; Native Hawaiian/Other Pacific Islander; Some Other Race (specify). For the analysis in this paper, this variable was recoded in five categories: White, Hispanic, Black, Multiethnic, Other. The multiethnic category included all respondents who identified with more than one race or ethnic group.

Statistical analyses

All analyses were conducted with Stata 17.0.³³ Associations in bivariate analyses (Tables 2 and 3) were tested with chi-square. Multivariable analyses were conducted in two steps. First, a multinomial logistic regression of IPV type with an outcome variable with 4 categories (No IPV, Perpetration only, Victimization only, Mutual IPV) was conducted with Stata’s “logit” procedure. The reference category was “No IPV.” Independent variables were entered in the model in one step. Covariates selection (age, religion, education, marital status, race/ethnicity, food insufficiency) plus risk factors shown in Table 3 and 4 was based on previous analyses of the data set and previous results in the literature.^{4,12,21,34,35}

Table 2:

Prevalence (percent) of Selected Risk Factors for Intimate Partner Violence by Gender

	Men (484)	Women (550)	Total Sample (1034)
Adverse Childhood Experiences2	30.4	39.4	35.2
At Risk Drinking3	34.4	20.9	28.0
Impulsivity ns	29.2	26.9	28.0
Marijuana Use1	30.5	23.9	27.0
PTSD1	22.2	27.6	25.1
Partner Hazardous Drinker ns	19.4	22.5	21.0
Illicit Drug Use 3	17.1	8.1	12.3

Test of gender difference in prevalence:

^nsnot significant;

¹p<.05;

²p<.01;

³p<.001

Table 3:

Intimate Partner Violence among Men and Women Positive for Selected Risk Factors: Percentages

	Adverse Childhood Experiences		Illicit Drug Use		At Risk Drinking		Impulsivity		Marijuana Use		PTSD		Partner Hazardous Drinker
	Men (146)	Women (215)	Men (82)	Women (44)	Men (171)	Women (115)	Men (141)	Women (148)	Men (146)	Women (130)	Men (106)	Women (152)	Men (94)	Women (124)
IPV Type
Perpetration	2.02	8.8	4.82	11.4	1.73	13.0	2.82	11.5	2.72	9.2	2.82	9.2	2.11	7.2
Victimization	10.3	4.6	19.5	0.0	8.8	3.5	12.0	4.7	12.3	3.8	16.0	7.2	17.0	8.0
Mutual	23.3	18.2	26.8	38.6	19.3	26.9	24.1	26.3	26.0	36.1	31.1	21.7	31.9	22.6
No IPV	64.4	68.4	59.7	53.2	70.2	56.5	60.9	57.4	58.9	50.8	50.0	61.8	48.9	62.1

Chi square test of differences in the distribution of IPV types within each factor of risk by gender:

¹p<.05;

²p<.01;

³p<.001

Table 4:

Logistic Regression of Mutual Intimate Partner Violence (IPV) on Sociodemographic, Psychological and Drinking-Related Variables.

	Mutual IPV vs. No IPV
	OR (95%CI)
Men (Ref: Women)	.64 (.36-1.15)
PTSD Positive (Ref: Negative)	1.46 (.77-2.77)
Male x PTSD Positive (Ref: Women/No PTSD)	3.06 (1.22-7.69)1
Days Marijuana Use Past 12 Months	1.003 (1.002-1.005)3
At-Risk Drinker (Ref: Not at Risk)	1.73 (1.07-2.77)1
Partner Hazardous Drinker (Ref: Negative)	2.19 (1.35-3.55)2
Any illicit Drug Past 12 Months (Ref: No Drug)	2.09 (1.18-3.71)2
Adverse Childhood Experiences	1.23 (1.06-1.42)2
Impulsivity (Ref: Negative)	2.04 (1.29-3.22)2

¹p<05;

²p<.01;

³p<.001.

Analyses also controlled for the effects of age, importance of religion, education, marital status, race/ethnicity, and food insufficiency.

The main source for the selection of interactions to be tested in the analyses was a previous paper ¹² that examined IPV severity and its association with a multiple risk index with the following factors: depression; adverse childhood experiences; illicit drug use (including cannabis); impulsivity; PTSD; at-risk drinking; and partner’s hazardous drinking score on the Alcohol Use Disorders Identification Test (AUDIT-C). Among these variables, preparatory analysis for the present paper identified 7 factors (see Tables 2 and 3) that had statistically significant associations with IPV type: illicit drug use, impulsivity; cannabis use, PTSD; at-risk drinking; adverse childhood experiences; and partner’s hazardous drinking. Seven interactions were tested, those between each of these factors and gender, which was seen as an effect modifier.

The process for testing the 7 interactions was the following: Once a multinomial logistic model with main effects only was fitted, each of the interactions to be tested was entered in the model one at a time. The strength of the statistical association between the variable in the interaction to be tested and gender, the chi square “p” value in Table 2, determined the order of testing. When two or more variables had the same “p” value, the order of testing was determined by the prevalence of the variable in the sample; variables with higher prevalence were tested first. If the interaction was not statistically significant, it was dropped from the model, and a new interaction was added and tested. This process showed that none of the interactions had a statistically significant association with IPV perpetration only or victimization only. Because of these null findings and because results from a multinomial analyses with main effects for perpetration only and victimization only had already been published,¹⁰ the data on perpetration and victimization were dropped from the multivariable analysis (N=103). A logistic model contrasting mutual IPV with No IPV (Table 4) was fitted (N=926) using Stata’s “logistic” procedure to test the effect of the interaction between gender and PTSD on Mutual IPV.

RESULTS

Missing data were negligible; none of the variables analyzed in this paper had more than 2.6% information missing. Therefore, no imputation was conducted to address missing data, which were left as missing.

Sample sociodemographic indicators and IPV rates

Men had a mean age higher than women and this difference was statistically different. Marital status was not different between men and women, but education was. There was a higher percentage of women than men with some college education and college graduation. Racial and ethnic distribution between genders was not different. The majority of men and women in the sample was Hispanic. Unemployment affected almost a third of the men and about a quarter of the women. This difference was statistically significant. Food insufficiency was high both among men and women, but the rate among women was higher than among men. IPV type distribution across gender was different. IPV perpetration was high among women, and consequently IPV victimization was higher among men.

Prevalence of Intimate Partner Violence Risk Factors

The most prevalent IPV risk factor in the sample was adverse childhood experiences (Table 2). This was also the most prevalent risk factor among women. At-risk drinking was significantly higher among men, being reported by a little over a third of them, than among women (reported by a fifth). The prevalence of impulsivity was not significantly different between men and women. Cannabis use was present in about a quarter of the sample. Its prevalence was significantly higher among men than among women. PTSD was the only risk factor with a prevalence that was significantly higher among women than among men. The proportion of men and women who had a partner who was a hazardous drinker (about a fifth) was similar and not statistically different. Finally, illicit drug use was present in about a little more than a tenth of the sample, and it was 2 times higher among men than among women.

Prevalence of Intimate Partner Violence by Gender for Selected Risk Factors

Chi square analysis in Table 3 tests the statistical significance of differences between men and women in the distribution of IPV types within each factor of risk. Results show significant differences in the distributions of IPV type for each factor of risk. IPV perpetration only was higher among women than among men for all risk factors. Rates were 7.5 (at-risk drinking) to 3.2 times (PTSD) higher among women than among men across the different risk factors. IPV victimization only was higher among men than among women. Rates of victimization were, in general, 2 to 3 times higher among men than among women. Mutual IPV associated with illicit drug use, at-risk drinking, cannabis use, and impulsivity was higher among women than among men. The reverse was true for mutual IPV associated with PTSD and having a partner hazardous drinker; it was significantly higher among men than women.

Sociodemographic, Psychological, and Drinking-Related Correlates of Mutual Intimate Partner Violence

The association between PTSD and mutual IPV varied by gender. First, there is no difference in the odds of involvement in mutual IPV between men and women (Table 4). Second, there also is no difference in the odds of involvement in mutual IPV between PTDS positive and PTSD negative participants. However, men who are positive for PTSD are about three times more likely than women with no PTSD to be involved in mutual IPV. Black and Multiethnic respondents compared to Hispanics (reference) were more likely to report mutual partner violence. Higher frequency of cannabis, at-risk drinkers compared to non-at risk drinkers, those who reported illicit drug use in the past 12 months compared with those who did not report this behavior, those with a partner who was a hazardous drinker, those with higher ACE scores, and those in the upper third of the impulsivity scale were also more likely to report mutual violence.

DISCUSSION

The main goal of this paper was to test interaction effects between selected risk factors and gender on IPV. First, the crosstabulations showed that the prevalence of most risk factors vary by gender (Table 1), with exception of impulsivity and having a partner who is a hazardous drinker, and that IPV type varies by gender depending on which risk factor is present (Table 2). For instance, adverse childhood experiences, illicit drug use, at risk drinking, cannabis use, and impulsivity are all associated with a higher percentage of women than men reporting IPV perpetration only, which is in accordance with previous research.^7,36

Table 1.

Sample characteristics (N=1034)

	Men	Women	Total Sample
	(484)	(550)	(1034)
	% or Mean	% or Mean	% or Mean
Age Mean (range 18-50) ***	36.5	34.0	35.2
Marital statusNS
Married	44.6	37.5	40.2
Cohabiting	30.1	32.8	31.5
Single, separated, divorced	25.3	29.7	27.6
Education **
Less than high school	37.7	28.8	32.9
High school graduate/GED	33.5	31.9	35.7
Some college	21.7	25.2	22.4
College graduate+	6.9	11.1	9.1
Race/ethnicityNS
Hispanic	52.9	47.6	50.1
African American	26.2	31.3	28.9
Multiethnic	4.3	6.4	5.4
Other	10.1	8.2	9.2
White	6.4	6.5	6.5
Employment ***
Unemployed	29.9	26.7	28.3
Food Insufficiency *
Sometimes/often	44.4	54.1	49.6
Never	54.9	45.3	49.8
Past-year intimate partner violence ***
None	76.6	76.6	76.6
Perpetration only	1.26	6.0	3.8
Victimization only	8.8	4.0	6.2
Mutual	13.4	13.3	13.3

NS: not significant;

^*p<.05;

^**p<.01;

^***p<.001.

The multivariable analyses showed that no interactions between selected factors of risk were present for perpetration and victimization only, but an interaction between gender and PTSD was significantly associated with mutual IPV. Past research has showed a positive association of PTSD with psychological and physical IPV both among men and women.^37–40 This association has been linked to the dissociative dimensions of PTSD.⁴¹ Dissociation may be associated with a breakdown in information processing, leading to an increase in general aggression and IPV.³⁹ PTSD may also disrupt anger regulation, which together with hypervigilance leads to threat misperception.³⁷

This effect modification changed the main effect of male gender on mutual IPV from a non-significantly protective association to a factor of risk. These results are of importance to the role of emergency medicine in addressing IPV. EDs are mandated by the Joint Commission on Accreditation of Health Care Organizations to implement universal IPV screening. ⁴ This is in part because ED patient populations have many factors of risk associated with IPV, including PTSD. The 12-month and lifetime prevalence of PTSD in the US general population is 4.7 and 6.1%.⁴² But these percentages are higher among ED patients as seen in Table 2 herein, which shows a past month rate of 25.1%. To our knowledge, this interaction effect between gender and PTSD has not been reported in the literature but given previous findings discussed above it is not difficult to understand. However, previous research has also seen PTSD as a consequence of IPV.^43,44 Longitudinal studies that could elucidate the direction of the association between PTSD and IPV have not been conducted yet. It is also important to remember that the results in Table 4 show that various other factors of risk have positive and statistically significant associations with mutual IPV, but their effect is just not modified by gender. This indicates that while the assessment of PTSD is important, IPV identification in EDs should be multifaceted and comprehensive, including other factors of risk. This is so especially because as previously reported by Caetano et al.,¹² the risk associated with factors is additive, increasing as the number of factors present increases.

The analysis and the results described have limitations. PTSD was identified with a primary care setting screener, the PC-PTSD. This screener has 5 questions, which do not completely cover all the criteria required for a diagnosis of PTSD in the 5^th Edition of the Diagnostic and Statistical Manual of the American Psychiatric Association (DSM-5). Further, the PC-PTSD is based on self-assessment, and it relies on a cut-off point to identify PTSD, which may not be optimal given that PTSD may not be an all-or-none phenomenon.⁴⁵ However, the PC-PTSD has strong psychometric properties. Its sensitivity was 95% with a cut score of 3 in a Veteran primary care sample.⁴⁶ Other research has reported similar results.^46–49

Other limitations are the following: The subjects enrolled were a convenience sample and may not be representative of the population. Results are from analyses of data from a single urban ED; findings may not generalize to other EDs and other health settings. In addition, recall bias may have affected subjects’ information about events that reached back over 12 months, and patient self-reporting of sensitive facts as IPV may lead to underreporting. Another limitation is that data were obtained from only one member of the couple. This may result in an underestimation of IPV, and partner substance use patterns. Finally, self-reports of IPV can be affected by social desirability bias, leading to under-reporting of these events. To minimize the impact of this factor, interviews were conducted in private settings, and respondents were assured of the interview confidentiality, i.e., the research team would have access to deidentified data without any information that could link responses to the interviewee.

In conclusion, the results suggest that gender has a strong association with IPV that is not modified by many factors. The identification of potential interaction effects helps directing clinical actions to target couples that are at higher risk for IPV, and who need specific types of interventions to address, for example, the dyad where one of the members has been diagnosed as presenting PTSD. Individual patients in EDs are frequently asked as part of intake procedures whether they “feel safe” a home or have experienced emotional or physical aggression from a partner. This screen is important, but it should be extended to include specific questions about presence of PTSD especially among male partners. Positive findings should then lead to more detailed assessments of IPV. Public health interventions can be focused on providing resources that can minimize stress between partners and support couples such as affordable childcare, pediatric care, affordable housing, community safety and other support. IPV risk assessment in EDs must be trauma oriented with attention not only to patients’ gender but also to the presence of other risk factors that can modify the risk of IPV seen among men and women.