Violence Against Women in Selected Areas of the United States

Abstract

Violence against women is increasingly recognized as a critical national public health concern in the United States, as evidenced by the recent signing of the Violence Against Women Act by President Obama.¹ Based on nationally representative samples, it is estimated that in their lifetime, nearly 1 in 3 US women has survived physical violence, and 1 in 10 has survived rape.² Women who experience emotional, physical, and sexual violence not only experience the injury of the initial trauma, but also have higher rates of depression, posttraumatic stress disorder (PTSD), substance abuse, re-victimization, and high-risk sexual behaviors.^3–7

There is strong evidence that supports the relationship between experiencing intimate partner violence (IPV) and HIV risk, as well as acquiring HIV.^8–20 In the United States, the relationship between IPV, especially sexual violence, and HIV came to light almost 20 years ago.⁵ Several studies^{6,10,17,18,20,21} during the past decade reported high co-occurring rates of violence, HIV risk, mental illness, and substance use in the United States among incarcerated women,²² female substance users,¹⁹ women in shelters,^23,24 women living in impoverished areas,^25–28 and women engaging in prostitution.²⁹ Not only do these conditions frequently co-occur, but the presence of one may magnify the effects of the others, thus demonstrating the notion of a “syndemic,” which is a term used to refer to a set of synergistic or intertwined and mutually enhancing health and social problems facing vulnerable populations such as women living in poverty.³⁰ However, interpretation and generalizability across studies has been difficult because of small sample sizes, convenience samples (e.g., women in methadone treatment, shelters, or clinics), narrowly defined study populations, the inclusion of both women living with HIV and women living without HIV, and poorly standardized study variables that typically did not include emotional abuse (e.g., combining different types of violence or combining childhood and adult violent experiences).^5,6,10,31

Overall, most published US data have suggested that women living with HIV experience IPV at the same rate as women not living with HIV from the same population, but that women living with HIV experience such violence more frequently and with increased severity.^9,31 In addition, substance abuse, poverty, and other HIV risk factors were associated with experiencing violence and therefore, also contribute to the HIV and IPV relationship.^9,32 However, many of these studies focused on “intimate” partner violence specifically and not on violence overall. For example, using a large representative sample of US women (n = 13 928), Sareen et al.⁸ found that women who experienced any IPV in the past year were more than 3 times as likely to report an HIV/AIDS diagnosis by a health professional as women who had not experienced IPV. They postulated that nearly 12% of HIV/AIDS infections among US women in intimate relationships was caused by IPV. Despite its novel contribution to the examination of the relationship between HIV infection and IPV among US women, this study was limited because it only examined physical and sexual IPV in the past 12 months. Another large-scale domestic study by Stockman et al. assessed specific types of sexual coercion in a nationally representative sample of 5857 US women and found positive associations among coerced sex, using drugs and alcohol, and having multiple sexual partners.²⁰ However, this study was limited in the way it assessed HIV risk by not examining perceived partner risk factors. Other recent studies have further elucidated this complex clustering of risk factors, but sample sizes have remained small, and measures of sexual HIV risk behaviors have varied widely.^9,33,34

Because of the concentration of HIV/AIDS in key areas of the United States, the HIV Prevention Trials Network (HPTN) study 064 was uniquely designed to recruit a representative sample of women living in US areas with high poverty and a high prevalence of HIV.^26,27 Unlike the more diffuse HIV epidemics seen in other countries, HIV in the United States has striking socioeconomic and racial disparities that are concentrated in key “hot spots” of the Northeast, South, and West. However, HIV among women in the United States is still not fully appreciated in the current HIV prevention research.^26,27 HPTN 064 made a significant public health contribution by assessing key social and behavioral factors that contribute to HIV acquisition among women in the United States by exploring the risk of HIV infection among certain populations of US women and providing information about their risk behaviors.³⁵ We used this large data set to assess the following: (1) the prevalence of emotional abuse, physical violence, forced sex, and experiencing 2 or more types of violence; and (2) sociodemographic characteristics, personal HIV-risk behaviors, and perceived sex partner behaviors associated with each type of violence.

METHODS

We used baseline data from the HPTN 064, the Women’s HIV SeroIncidence Study for our analyses. HPTN 064 was a multisite, prospective observational cohort study that has been described in detail elsewhere.³⁵ Briefly, we used ethnographic mapping and venue-based sampling to enroll 2099 women from 10 urban and peri-urban communities in 6 geographic areas in the United States (i.e., Atlanta, GA; Baltimore, MD; New York City, NY; Newark, NJ; Raleigh-Durham, NC; Washington, DC) with high rates of poverty and HIV.

As part of the eligibility criteria, participants had to reside in census tracts that were in the top 30th percentile of HIV prevalence for that area, which also had at least 25% of residents living below the US federal poverty threshold, as defined by the 2008 US Census Bureau.^36,37 Other eligibility criteria included being between the ages of 18 and 44 years, self-identifying as a woman, having at least 1 episode of unprotected vaginal or anal sex with a man in the 6 months before enrollment, and willingness to undergo HIV rapid testing and receive HIV test results. In addition, potential participants had to report at least 1 additional personal or partner HIV risk behavior in the past 6 months (e.g., drug use, sexually transmitted infection [STI] diagnosis, binge drinking, or exchanging sex) or incarceration in the past 5 years. Women were ineligible if they reported a history of a positive HIV test, current enrollment in an HIV prevention trial, current or past participation in an HIV vaccine trial, or anticipated absence from the community for more than 2 consecutive months during the follow-up period.

We recruited for and enrolled women in the study between May 2009 and July 2010. We obtained informed consent before the initiation of study procedures. Participants received monetary reimbursements for their time and travel to all study visits. The reimbursement amount varied by study site and was approved by the site institutional review board. Audio Computer-Assisted Self-Interviewing (ACASI) was used to collect data at behavioral baseline, and at 6 and 12 months.

RESULTS

The demographic characteristics profile of our sample of 2099 women was described in detail elsewhere.³⁵ In brief, 86% of participants were African American, with a median age of 29 years. Most reported having a high school diploma or less (74%), being nonpartnered (67%), and being unemployed (64.7%). At baseline, 44% had an annual household income of less than $10 000, and 46.3% reported food insecurity. Nearly 35% and 29% reported symptoms of depression or PTSD, respectively. Forty-five percent reported childhood abuse (Table 1).

The overall baseline prevalence of emotional abuse, physical violence, forced sex, and experiencing 2 or more types of violence in past 6 months were 31%, 19%, 7%, and 17%, respectively (Table 1). Table 2 includes results from bivariate analyses between covariates (sociodemographic characteristics and HIV risk) and violence outcomes. Many bivariate relationships were significant; however, race, ethnicity, education, poor health status, and unprotected vaginal sex were noteworthy exceptions.

Independent associations between types of violence and dependent variables from a multivariable model are shown in Table 3. Relative to participants who did not report emotional abuse, the odds of being food insecure were 1.8 times higher, surviving childhood abuse were 2.0 times higher, having symptoms of depression were 1.8 times higher, and having symptoms of PTSD were 2.0 times higher for those who experienced emotional abuse (all P < .001). Living with parents, partner, or friends was significantly associated with emotional abuse (adjusted odds ratio [AOR] = 1.40; 95% CI = 1.09, 1.80; P < .01). Several HIV risk factors were also significantly associated with emotional abuse, including exchanging sex for commodities (AOR = 1.40; 95% CI = 1.06, 1.85; P < .05), having a previous self-reported STI diagnosis (AOR = 1.52; 95% CI = 1.08, 2.16; P < .05), having a binge-drinking partner (AOR = 1.58; 95% CI = 1.25, 2.01; P < .01), and reporting indirect concurrency (AOR = 1.53; 95% CI = 1.18, 1.97; P < .01).

Physical violence in the past 6 months was significantly associated with younger age (P < .05) and not being married, but living with a partner (P < .05; Table 3). The odds of being food insecure or experiencing symptoms of depression were nearly 2 times higher among participants who reported physical violence in the past 6 months compared with participants who did not report physical abuse (P < .001). Physical violence was also associated with childhood abuse (AOR = 1.71; 95% CI = 1.28, 2.27; P < .01). Engaging in unprotected anal sex (AOR = 1.47; 95% CI = 1.11, 1.96; P < .01), having a drug-using sexual partner (AOR = 1.41; 95% CI = 1.02, 1.96; P < .05), and experiencing indirect concurrency (AOR = 1.40; 95% CI = 1.03, 1.89; P < .05) were significantly associated with experiencing physical violence.

Participants who reported being forced to have sex in the past 6 months had 2.9 times higher odds of being nonpartnered (P < .05), 2.6 times higher odds of identifying as a commercial sex worker (P < .05), and 2.4 times higher odds of experiencing PTSD symptoms (P < .01). In addition, forced sex was associated with unprotected anal sex (AOR = 1.68; 95% CI = 1.10, 2.57; P < .05) and having a drug-using sexual partner (AOR = 1.91; 95% CI = 1.18, 3.07; P < .01).

Experiencing 2 or more types of violence in the past 6 months was significantly associated with not being married (P < .05) and living with a sexual partner, friend, or parent (P < .01) (Table 3). Similar to other violence variables, participants who reported multiple forms of violence in the past 6 months had nearly twice the odds of reporting food insecurity (P < .001), childhood abuse (P < .01), and symptoms of depression (P < .001) compared with participants who did not report multiple forms of violence. Multiple types of violent experiences were also associated with PTSD symptoms (AOR = 1.69; 95% CI = 1.21, 2.35; P < .01), engaging in unprotected anal sex (AOR = 1.61; 95% CI = 1.19, 2.19; P < .01), having a drug-using sexual partner (AOR = 1.46; 95% CI = 1.04, 2.05; P < .05), having a binge drinking sexual partner (AOR = 1.40; 95% CI = 1.02, 1.92; P < .05), and experiencing indirect concurrency (AOR = 1.56; 95% CI = 1.13, 2.16; P < .01) (Table 3).

Pairwise odds ratios between types of violence were also calculated. All forms of violence were highly correlated (all P < .001). Specifically, the odds ratio between emotional abuse and physical violence was 14.3 (95% CI = 11.00, 18.75); between emotional abuse and sexual violence, the odds ratio was 7.8 (95% CI = 5.31, 11.56); and between physical violence and sexual violence, the odds ratio was 9.8 (95% CI = 6.85, 14.04).

DISCUSSION

In our sample, nearly 1 in 3 women reported experiencing some form of violence in the past 6 months. Emotional abuse, which is an understudied form of violence against women that has been associated with serious mental and physical health outcomes, was the most commonly reported type of violence.^44–46 Violent or abusive experiences had multiple associations, including marital status, food insecurity, housing stability, symptoms of depression and PTSD, childhood abuse, unprotected anal sex, sex work, and having a partner believed to be engaged in drug use, binge drinking, or a concurrent sexual relationship. Although differences in research design and assessment made it difficult to compare our findings across studies,^47–49 our study added to the existing literature by systematically sampling young women living in impoverished areas of the United States with a high prevalence of HIV. Our results demonstrated that violence is a common experience within this population of women, and that this experience is associated with several other factors that increase the risk of acquiring HIV.

However, there were key variations across demographic variables that should be highlighted. In bivariate analysis, race/ethnicity was not associated with violence, but poverty represented by measuring food insecurity was strongly linked to 3 types of violence in both bivariate and multivariable analyses. Although our results could not determine causality, several posited pathways might explain the relationship between violence and food insecurity, which has been found among other groups of impoverished women.⁵⁰ One potential pathway was that women who were being abused might be denied access by their abuser to financial resources and employment opportunities necessary to purchase food, and subsequently, these women became food insecure.^32,51,52 Another reason was that women who left abusive situations might be at greater risk of becoming food insecure because of the drastic socioeconomic changes that result after leaving a financially supportive abuser.⁵¹ Lastly, both violence and food insecurity are associated with poverty, and therefore, both were more likely to occur among residents of impoverished areas.⁵¹ Impoverished women tend to lack the social resources, instrumental support, and educational and economic opportunities available to more financially stable women. Consequently, these women might be less likely to leave financially supportive abusers.³² Structural interventions designed to reduce community-level economic disadvantage and to improve the financial stability (e.g., microfinance and financial empowerment) of US women of various racial/ethnic groups might be promising approaches to prevent violence overall and in specific US communities.^32,53

The poverty experienced by our sample might also contribute to violence rates through its association with marital instability.^54,55 Consistent with previous research,⁵⁶ marriage had a protective influence against violence, which might be caused by more stable relationships, more protective resources, or having partners who were more invested in sustaining the relationship. However, only 7.5% of our sample was married. Low marriage rates are common in impoverished communities, especially in predominantly African American communities, because of fewer men who are financially able to support a family and the low gender ratios between men and women.^54,55,57

All forms of violence were associated with experiencing symptoms of depression or PTSD. This relationship was strongest among survivors of forced sex and survivors of multiple types of violence. Numerous other studies have found similar results.^58–61 There was empirical evidence for both a bidirectional and a causal relationship between experiencing violence and symptoms of depression.⁶² It was suggested that poverty played a role in this relationship because of its association with both mental health outcomes and violence against women.⁶¹ Our findings reiterated the critical need for trauma-informed mental health care for survivors of violence, especially in impoverished communities.

Consistent with the current literature that examines the relationship between sexual risk behaviors and violence,^4–6,63 we found that experiencing any form of violence was associated with participation in sexual risk behaviors that increased the risk of acquiring HIV, either directly (i.e., exchanging sex, identifying as a commercial sex worker, and having unprotected anal sex) or indirectly (i.e., having a sexual partner who is believed to engage in HIV risk behaviors). Interestingly, we did not find a significant relationship between unprotected vaginal sex and any type of violence. However, we did find significant associations between experiencing physical violence, forced sex, and multiple types of violence, and unprotected anal sex. Although research has shown that experiencing violence impairs a woman’s ability to successfully negotiate condom use with future male sexual partners, there were mixed results in studies that examined the relationship between unprotected sex and violence.^6,64 These results might be because of a failure to separate anal and vaginal sex. Future research that examines condom use among survivors of violence should specify type of sex, especially because anal sex is associated with greater risk of HIV transmission.⁶⁵

Although we did not limit our measurement of violence to IPV, perceived partner HIV risk behaviors were still important to consider, because participants who experienced some form of violence were more likely to have male sexual partners who they perceived as exhibiting HIV risk behaviors, including drug use, binge drinking, or indirect concurrency. Previous research substantiated a strong relationship between substance use and perpetrating violence, as well as experiencing violence.⁶⁶ Our findings were similar to the findings by Coker et al., who found that partner drug use and binge drinking, as reported by the participant, were associated with experiencing different types of violence even after controlling for the participant’s substance use.⁶⁷

Numerous studies have examined the relationship between violence and HIV among men and women.^8–11,32 However, our study contributed to the existing literature by using a large data set to examine the association between 4 specific forms of violence against women, including emotional abuse and several well-defined personal and perceived partner HIV risk behaviors. As a result, our findings might elucidate details not previously explored about the posited factors involved in the syndemic relationship among poverty, mental illness, violence, and HIV risk in impoverished US women.^{6,23,29,31,34,68,69} The role of emotional abuse was of particular interest, because it is rarely examined as a standalone form of violence. The similarity of the constellation of associated factors for emotional abuse and physical violence compared with forced sex suggest the need for additional examination of the details surrounding the violence experienced rather than just if violence occurred.

The presence of a syndemic would indicate that HIV prevention programs among similar populations should be comprehensive, should address these co-occurring issues, and should broadly focus on contextual factors that influence HIV risk behavior. More research, particularly longitudinal research, is needed to improve our understanding of syndemic relationships, and thereby, to develop effective interventions for vulnerable women living in the United States. Although several behavioral HIV prevention studies included violence prevention content, only a handful of interventions addressed these complex relationships by including changes in violence measures as a primary endpoint or by recruiting survivors of violence as the study population.^{12,53,70–77}

However, to date, most of these large-scale interventions were conducted outside of the United States.⁷⁶ One potential reason for this gap is the lower HIV incidence rates in the United States; studies such as these would require much larger sample sizes than those in locations with higher HIV rates. Consequently, limiting studies to the use of HIV endpoints, which are preferred in HIV prevention research, results in limited data regarding US women. However, the unique structural, cultural, socioeconomic, and political factors that influence violence against women and HIV risk among US women limit generalizability from international research and necessitate further research among US women. Therefore, the use of HIV-related proxy measures and HIV incidence modeling may need to be considered so that this critical research can be conducted among US women. To narrowly focus prevention efforts in the United States on interventions that decrease HIV risk behaviors without considering the greater context of the at-risk population may result in brief, short-lived changes in HIV risk behaviors without substantial long-lasting impact on HIV incidence. The literature is replete with examples of HIV interventions (both behavioral and biomedical) that had limited or no efficacy to prevent HIV infection, perhaps because of the somewhat narrow understanding of the broader context of the study populations.^78–80

Acknowledgments

Human Participant Protection

References