Siemieniuk, Reed A.C. MD*,†; Krentz, Hartmut B. PhD*,‡; Miller, Patricia MSW*,§; Woodman, Kate PhD‖; Ko, Karen MA*; Gill, M. John MB, ChB*,¶
Intimate partner violence (IPV; also known as domestic violence, violence against women, or gender-based violence) is a common and damaging experience that is attracting increasing attention because of its well-documented impact on both personal and public health. The Institute of Medicine now recommends screening and counseling all women for IPV.1 Violence against women is closely linked to HIV infection worldwide. In fact, the United Nations General Assembly's Declaration of Commitment on HIV/AIDS explicitly calls for national strategies to eliminate violence against women as a means to reduce vulnerability to HIV/AIDS.2 IPV is an established risk factor for incident HIV infection as reported in both prospective studies in Africa3 and in large cross-sectional studies from the United States4 and India.5 The burden of IPV among those living with HIV is therefore also likely to be high. To our knowledge, comprehensive population-based studies have yet to be reported among HIV-positive women; however, pilot clinic data suggests past or present IPV in up to 43%–68% of HIV-positive women.6–9
IPV may directly contribute to an increased risk for acquiring HIV infection in women through sexual assault or more frequently from long-term behavioral implications such as a reduced ability to negotiate safe sexual practices with partners.4,10 In 1 study, perpetrators of IPV were additionally noted to have a high prevalence of HIV.5 The effects of IPV may also lead indirectly to HIV acquisition through increased distress and its association with riskier sexual practices and drug use (including intravenous drug use).3,11 Importantly, the interaction between IPV and HIV infection is bidirectional, as HIV disclosure subsequently puts women at risk of new or increased violence.12
Despite the high prevalence of IPV now being reported in HIV-positive women and the strong link between IPV and HIV infection, little is known about the risk factors for and the clinical outcomes of IPV once HIV infection is established. We hypothesized that IPV is associated with conditions such as poor mental health, illegal drug use, riskier sexual practices, and a history of incarceration. Furthermore, it may not only simply predispose individuals to HIV infection but also then negatively impact HIV-related health outcomes. We examine and describe the clinical implications of IPV in HIV-positive women living in well-defined geographic areas of Canada (southern Alberta).
The Southern Alberta HIV Program provides exclusive outpatient multidisciplinary care to all adults living with HIV in a large Canadian region. Patients accessing care sign a voluntary informed consent form approved by the University of Calgary Ethics Committee allowing blinded administrative data to be used in research.
Routine screening for domestic violence in general (IPV and child abuse) was implemented in May 2009 for all HIV-positive patients receiving care in southern Alberta. All accrued data up to January 1, 2012 were analyzed. IPV was defined as experiencing abuse within a current or previous intimate adult relationship (older than 16 years). A practical standardized screening instrument was used and previously described.6 Briefly, after a short preamble defining domestic violence, a registered nurse or research coordinator asked the patient, as part of a routine clinic visit, if he or she had ever been exposed to domestic violence in a current or past relationship. The interviewer also asked if there was a history of childhood abuse. Further details, including if the abuse occurred within a current and/or a past intimate relationship and the type of violence (physical abuse, sexual abuse, emotional abuse, financial abuse, isolation, neglect, or intimidation), were recorded.
Patients disclosing IPV were offered consultation with a social worker with expertise in IPV and HIV. All patients were screened unless there were overriding clinical or social factors. If the patient did not disclose IPV, the opportunity was left to discuss the issue at a future visit should the need arise. Results from the first visit disclosing IPV were used. All patients who self-reported as female were included in the study.
Sociodemographic and clinical care markers are recorded at the initial visit to the clinic and updated prospectively. Sociodemographic data collected included age at the IPV screening visit, years in care, self-reported ethnicity [white, Aboriginal (First Nations, Métis, and Inuit), black (>90% immigrants from Sub-Saharan Africa), or other], location of original HIV diagnosis (within Southern Alberta vs. elsewhere), living arrangement (alone vs. cohabitating), housing situation (homeless or supported/temporary vs. stable), and history of incarceration. Living arrangement and housing situation was recorded as the most recent response to a standardized screen performed at each visit. Unstable housing was defined as not having a fixed address and/or living in homeless or women's shelters. Substance use since the last visit was also assessed: smoking (current and former vs. never), alcohol excess (>9 drinks per week or any binge), and any illicit drug use. The reason for the initial HIV test was categorized to include screening (pregnancy, immigration, insurance, and tissue donation), high-risk exposure (intravenous drug use, sexual contact, and blood of unknown serostatus), HIV-associated symptoms, patient or physician request, or unspecified. Clinically relevant psychiatric disease before HIV diagnosis (major depression, anxiety disorder, and self-reported attempted suicide) was recorded at initial visit. All subsequent consultations with the HIV program's psychiatrists were documented. Patients were routinely asked about recent safe or unsafe sexual activity. Health-related quality of life (HRQoL) was assessed with the question “how would you describe your overall health today” and reported as poor, fair, good, very good, or excellent.13
To optimize data quality, we restricted analyses with variables obtained at the initial HIV diagnosis a priori to patients diagnosed locally in southern Alberta by excluding those who had moved into the region after an HIV diagnosis outside the region. These analyses included mental health parameters before HIV diagnosis and CD4 count at diagnosis.
Standard of care in the region entails physician visits at a minimum of every 4 months. Lost to follow-up (LFTU) was defined as an interruption in care for ≥365 days. LFTU with clinically important implications was defined as moving from an undetectable viremia to a detectable viremia (≥500 copies/mL) after being LFTU for ≥365 days. Levels of viremia ≥500 copies/mL were felt to represent an interruption in adherence or development of resistance rather than a clinically insignificant viral load “blip.”
Clinical variables, including CD4 count at HIV diagnosis (per cubic millimeter), nadir CD4 count, and CD4 count and viral load at time of screening (closest to and within 6 weeks), use of antiretroviral therapy (ART), and history of AIDS were also reviewed. Effective viral load suppression (<500 copies/mL) was analyzed within the context of current ART use (ART vs. no ART). A predefined secondary analysis of AIDS incidence was stratified by CD4 count at presentation to care (<200 vs. ≥200 cells/mm3).
All hospitalizations between January 1, 1996 and December 31, 2010 were verified via a centralized database. Hospitalization rates, defined as the number of admissions per thousand years-at-risk, were evaluated. Years-at-risk was defined from HIV diagnosis for local patients or date of transfer to our region to the first of death or December 31, 2010. Hospitalizations were coded as HIV-related or HIV-unrelated based on the most responsible diagnoses for admission, using previously described methodology.14 Hospitalizations concurrent with initial HIV diagnosis were excluded in a predefined secondary analysis to evaluate hospitalizations felt to be mostly avoidable with adequate outpatient care.
Multivariable analyses were conducted with Poisson regression, adjusted a priori for age, months living with HIV infection, log of initial CD4, location of HIV diagnosis (locally or not locally), country of birth, and when the specific variable was determined (if appropriate, same visit vs. prior visit). Statistical significance was defined as P < 0.05. The χ2 test was used for categorical data, and Student's t test was used for continuous data. Statistical analyses were performed with SPSS v20.0 (IBM, Armonk, NY).
Three hundred thirty-nine (79.2%) of the 428 women who received HIV care in southern Alberta between May 2009 and January 2012 were screened for IPV. Those screened had similar clinical characteristics to those not screened: median nadir CD4 counts were 230 versus 218 per cubic millimeter (P > 0.05), and median number of days in HIV care were 1263 versus 1286 (P > 0.05) for those screened and not screened, respectively.
One hundred thirty-seven (40.4%) of the 339 patients reported a history of IPV. Of those, 20.4% (n = 28) reported abuse in a current relationship and 87.6% (n = 120) reporting abuse in a previous relationship (Table 1). Eleven women (8.0%) reported both past and current IPV. Seventy-three women (21.5%) had a history of childhood abuse, of whom 60 (82.2%) subsequently experienced IPV as an adult (Table 2). The adverse effects described were found similarly among women currently experiencing IPV as found among women with a history of IPV.
Emotional and physical abuses were the most commonly reported forms of IPV, followed by sexual abuse. Most victims, however, experienced multiple types of abuse (n = 98, 71.5%), on average disclosing 2.6 forms of abuse.
There were no significant differences in age, the time elapsed since engaging into HIV care, or living arrangement between victims of IPV and nonvictims after adjusting for cofounders. There were, however, significant ethnic disparities in the prevalence of IPV (P < 0001). Aboriginal and white women reported the highest prevalence of IPV at 65.0% and 61.4%, respectively, whereas black women (92.0% of whom had migrated from Sub-Saharan Africa) reported IPV less frequently at 22.1%.
Overall, most women reported cohabitating with at least one other person (81.0%) and renting or owning their residence (95.7%). IPV was, however, independently associated with unstable housing [adjusted prevalence ratio (APR) = 4.49, CI: 1.14 to 17.74].
A history of IPV significantly predicted recent use of illicit substances (APR = 7.58, CI: 2.45 to 23.26; Table 3). Women who disclosed IPV were also more likely to be current smokers (APR = 5.07, CI: 1.19 to 9.68) or have a history of previous smoking (APR = 3.39, CI: 1.19 to 9.68). IPV was not a predictor of recent excessive alcohol use. Of note, 19 of the 25 women with a history of incarceration reported IPV (APR = 4.84, CI: 1.85 to 12.68).
Before HIV diagnosis, a higher rate of depression requiring medical therapy was noted in women disclosing IPV (APR = 1.87, CI: 1.10 to 3.16). The prevalence of a preexisting anxiety disorder was also much higher among those with IPV (APR = 5.75, CI: 2.10 to 15.63). Furthermore, although numbers were small, IPV was a strong predictor of a suicide attempt before HIV diagnosis (APR = 53.86, CI: 5.21 to 556.48). We did not find a difference noted in usage of specialized psychiatric resources, however, use of general psychiatric services not specializing in HIV could not be evaluated in this study.
Health-Related Quality of Life
Women in HIV care reporting past or present IPV had a significantly worse HRQoL (P = 0.006 for trend), reporting poor or fair health 2.94 times (1.40 to 6.16) more commonly than very good or excellent health.
Women who reported IPV were more likely to have been diagnosed with HIV earlier in the course of infection, defined by CD4 count (CD4 >500 vs. CD4 <200; APR = 4.33, CI: 1.11 to 16.95). There were no detectable differences in reasons for testing after adjusting for cofounders.
Interruptions in care (ie, LFTU for ≥365 days) were more common among those reporting IPV compared with those not reporting IPV in the adjusted (APR = 1.90, CI: 1.07 to 3.39) and unadjusted analyses: 20.4% versus 11.9% (P = 0.006).
Despite having similar CD4 counts at the time of IPV screening (P = 0.95), patients disclosing IPV were 45% less likely to be using ART (APR = 0.55, CI: 0.34 to 0.91) and 2.10 (1.23 to 3.58) times more likely to have uncontrolled viral replication ≥500 copies/mL.
Hospitalization rates were higher for women disclosing IPV at 256 (223 to 292) per 1000 patient-years compared with 166 (143 to 193) among those not disclosing IPV (P < 0.001; Table 4). Most of these hospitalizations were unrelated to HIV (relative risk = 1.42, CI: 1.16 to 1.73); however, HIV-related hospitalizations were also borderline significantly higher than in women with no experience of IPV (P = 0.070). After the patient was engaged into routine HIV care, however, HIV-related hospitalizations were much higher in women who had experienced IPV (relative risk = 2.19, CI: 1.01 to 4.85) suggesting less successful uptake of HIV care.
IPV is common in our population of HIV-positive women with more than 40% disclosing past or present IPV. The prevalence may, in reality, be even higher, as others have reported that many women who experience IPV do not interpret their experiences as IPV, often minimizing their situation.15 Population-based studies in general North American populations report a lifetime prevalence of IPV of 19%–35% among women.16–18
Particularly, high rates of IPV among Aboriginals, both HIV-positive and HIV-negative, have been identified previously,6,16,19 and may be related to their disadvantaged socioeconomic status.20 However, the equally high prevalence of victimization among white women has not been previously reported and may reflect the effect of the overlap of risk factors for HIV infection and IPV among white women in Canada, such as risky sexual behavior and illicit drug use. Relatively, lower rates of IPV have been reported among recent immigrants and may explain the low rates of IPV that we observed among black patients, most of whom recently migrated to Canada from Sub-Saharan Africa.6,21 However, this was not corroborated by a recent study of HIV-positive women in the United Kingdom.7 This discrepancy in prevalence may be explained by real differences or perhaps differences in recognizing and reporting IPV among different populations.
The negative health implications of IPV observed among HIV-positive women are, in many ways, similar to those observed among HIV-positive gay and bisexual men in our region.19 However, as there are very different societal and relationship dynamics involved, a distinct consideration of the outcomes is required.
IPV was associated with high rates of depression and anxiety in our study, similar to studies of HIV-negative populations.22 Those reporting IPV were also much more likely to have attempted suicide, which is the strongest predictor of future attempts.23 Psychiatric illness unrelated to IPV strongly predicts poor HIV outcomes.24
There was a high degree of illicit substance abuse and smoking among women who have experienced IPV. The relationship between substance abuse and IPV may be complex and bidirectional; however, helping patients address IPV in a constructive manner may also mitigate morbidity and mortality from illicit substance use25 and smoking.26 Illicit drug use is also predictive of poor retention in HIV care and may play a role in the poor retention associated with IPV in our study.27
Patients who experienced IPV were diagnosed with HIV earlier in the course of their disease (with a higher CD4 count), which is likely explained by increased contact with the medical system for related and unrelated complaints.6,28 While early diagnosis of HIV is a very strong predictor of superior HIV-related outcomes,29 those with IPV still had a worse prognosis despite their earlier diagnosis. Women disclosing IPV experienced twice as many hospitalizations for HIV-related diseases despite the benefit from an earlier diagnosis and early connection to HIV outpatient care. Furthermore, IPV was strongly associated with lower levels of HRQoL,13 even after adjusting for confounders, strengthening the above evidence that HIV-positive women who have experienced IPV also experience worse health.
There is a survival benefit associated with good engagement in care, not only because of increased adherence to and monitoring of ART but also because patients can receive treatment for medical and psychiatric comorbidities. The poor HIV-related outcomes seen in our study among women experiencing IPV are likely related directly to interruptions in care, which is a well-described predictor of HIV-related mortality.30–33 By directly addressing IPV, we may support adherence to care and thereby improve outcomes. HIV-positive patients experiencing IPV face multiple barriers to adequate engagement in care, such as the immediate threats to the safety of themselves and their dependents, which in turn requires less imminent threats, such as HIV infection, to fall in priority. Many may also experience IPV as a consequence of HIV infection and its disclosure. As such, these women may be risking violent repercussions by engaging in care.12 Our study supports this hypothesis, as women who experienced IPV found it more difficult to remain engaged in HIV care.
The idea that HIV-positive women experiencing IPV have worse outcomes because of poor engagement in care is further supported by the finding that women disclosing IPV were less likely to be receiving ART and more likely to have uncontrolled viral replication, both of which negatively affect medical outcomes.34 Similar barriers that exist for regular clinic follow-up described above also exist for ART adherence and maintaining controlled viral replication: victims of IPV may be prioritizing immediate personal safety over regular adherence to ART and the erratic nature of their day-to-day lives in itself may preclude their ability to follow regular ART dosing regimens. This lower level of ART use may partially account for the increased number of HIV-related hospitalizations seen among women experiencing IPV after engaging into care. Given that the increased risk of HIV-related hospitalizations occurs after an initial contact with HIV outpatient care, IPV may be amenable to focused intervention by primary HIV caregivers. Specifically, a primary objective should be to increase retention in care by developing safe and trustworthy relationships with HIV-positive women and increase ART adherence among patients disclosing IPV, which will, in turn, improve outcomes.
Our study has some limitations. It was conducted in a diverse population of HIV-positive women in Canada but may not entirely reflect populations elsewhere. The study also used a cross-sectional and retrospective approach, and thus, causality should be interpreted with caution. The use of a single question IPV screen, while being practical and specific, may not be as sensitive for different aspects of IPV as longer more detailed screening questionnaires. In a large systematic review, however, multiple screening questions were reported to be no more sensitive for general screening than a single question.15 In contrast to in-depth surveys, the single question was easily adapted to routine clinical practice and was well received by patients and staff.
Health care practitioners and victims of IPV alike may not be fully aware of the direct and indirect impacts of IPV on health and health-seeking behaviors. It is therefore necessary to address this important social determinant of health systematically in HIV-positive women. HIV clinics provide the optimal environment for addressing this sensitive issue because the long-term health care practitioner–patient relationship builds trust and allows for ongoing follow-up. Identification of IPV in the clinical HIV setting can assist the patient in addressing the issue (through appropriate referrals to specialized agencies or allied health care workers) and in managing the underlying barriers to optimal follow-up. Ultimately, this may improve the patient's safety (and that of any children in her care), reduce current IPV, while increasing the efficacy of clinical interventions and improve health outcomes.
We have identified a 40% prevalence of IPV among a population of HIV-positive women. A history of childhood abuse, incarceration, illicit substance abuse, smoking, and psychiatric disease all independently predicted past or present IPV. Women with a history of IPV were more often hospitalized for both HIV-related and HIV-unrelated causes, which markedly increased after initiating outpatient care. Victims also had decreased HRQoL. This worse prognosis for HIV-positive victims of IPV was mediated through poor retention in care and may therefore be amenable to meaningful interventions.
The identification of IPV as a common and negative social determinant of health in HIV-positive women opens up the opportunity to develop targeted interventions that address the effects of this destructive social behavior. Such interventions might help these women engage in HIV care, improve quality of life, and improve health outcomes.
The authors thank the HIV care staff in Southern Alberta for working so passionately to ensure a successful domestic violence screening program.
1. National Research Council. Clinical Preventive Services for Women: Closing the Gaps. Washington, DC: National Academies Press; 2011.
3. Jewkes RK, Dunkle K, Nduna M, et al.. Intimate partner violence, relationship power inequity, and incidence of HIV infection in young women in South Africa: a cohort study. Lancet. 2010;376:41–48.
4. Campbell JC, Baty ML, Ghandour RM, et al.. The intersection of intimate partner violence against women and HIV/AIDS: a review. Int J Inj Contr Saf Promot. 2008;15:221–231.
5. Decker MR, Seage GR III, Hemenway D, et al.. Intimate partner violence functions as both a risk marker and risk factor for women's HIV infection: findings from Indian husband-wife dyads. J Acquir Immune Defic Syndr. 2009;51:593–600.
6. Siemieniuk RA, Krentz HB, Gish JA, et al.. Domestic violence screening: prevalence and outcomes in a Canadian HIV population. AIDS Patient Care STDs. 2010;24:763–770.
7. Dhairyawan R, Tariq S, Scourse R, et al.. Intimate partner violence in women living with HIV attending an inner city clinic in the UK: prevalence and associated factors. HIV Med. 2013;14:303–310.
8. Cohen M, Deamant C, Barkan S, et al.. Domestic violence and childhood sexual abuse in HIV-infected women and women at risk for HIV. Am J Public Health. 2000;90:560–565.
9. Gielen AC, McDonnell KA, Burke JG, et al.. Women's lives after an HIV-positive diagnosis: disclosure and violence. Matern Child Health J. 2000;4:111–120.
10. Johnson MP. Conflict and control: gender symmetry and asymmetry in domestic violence. Violence Against Women. 2006;12:1003–1018.
11. Frye V, Latka MH, Wu Y, et al.; INSPIRE Study Team. Intimate partner violence perpetration against main female partners among HIV-positive male injection drug users. J Acquir Immune Defic Syndr. 2007;46:S101–S109.
12. Zierler S, Cunningham WE, Andersen R, et al.. Violence victimization after HIV infection in a US probability sample of adult patients in primary care. Am J Public Health. 2000;90:208–215.
13. Crane HM, Van Rompaey SE, Dillingham PW, et al.. A single-item measure of health-related quality-of-life for HIV-infected patients in routine clinical care. AIDS Patient Care STDs. 2006;20:161–174.
14. Krentz HB, Gill MJ. Cost of medical care for HIV-infected patients within a regional population from 1997 to 2006. HIV Med. 2008;9:721–730.
15. Ramsay J, Richardson J, Carter YH, et al.. Should health professionals screen women for domestic violence? Systematic review. BMJ. 2002;325:314.
16. Tjaden P, Thoennes N. Full Report of the Prevalence, Incidence, and Consequences of Violence Against Women: Findings from the National Violence Against Women Survey. Washington, DC: United States Department of Justice; 2000:71.
18. Black MC, Basile KC, Breiding MJ, et al.. The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 summary report. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention; 2011. 37–50.
19. Siemieniuk R, Miller P, Woodman K, et al.. Prevalence, clinical associations, and impact of intimate partner violence among HIV-infected gay and bisexual men: a population-based study. HIV Med. 2013;14:293–302.
20. Reading CL, Wien F. Health Inequalities and Social Determinants of Aboriginal People's Health. British Columbia, Canada: National Collaborating Centre for Aboriginal Health; 2009.
21. Du Mont J, Hyman I, O'Brien K, et al.. Factors associated with intimate partner violence by a former partner by immigration status and length of residence in Canada. Ann Epidemiol. 2012;22:772–777.
22. Trevillion K, Oram S, Feder G, et al.. Experiences of domestic violence and mental disorders: a systematic review and meta-analysis. PLoS One. 2012;7:e51740.
23. Beghi M, Rosenbaum JF. Risk factors for fatal and nonfatal repetition of suicide attempt: a critical appraisal. Curr Opin Psychiatry. 2010;23:349–355.
24. Carrico AW, Riley ED, Johnson MO, et al.. Psychiatric risk factors for HIV disease progression: the role of inconsistent patterns of antiretroviral therapy utilization. J Acquir Immune Defic Syndr. 2011;56:146–150.
25. Egger M, May M, Chêne G, et al.. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet. 2002;360:119–129.
26. Pines H, Koutsy L, Buskin S. Cigarette smoking and mortality among HIV-infected individuals in Seattle, Washington (1996-2008). AIDS Behav. 2011;15:243–251.
27. Giordano TP, Hartman C, Gifford AL, et al.. Predictors of retention in HIV care among a national cohort of US veterans. HIV Clin Trials. 2009;10:299–305.
28. Mugavero MJ, Pence BW, Whetten K, et al.. Childhood abuse and initial presentation for HIV care: an opportunity for early intervention. AIDS Care. 2007;19:1083–1087.
29. Castilla J, Sobrino P, de la Fuente L, et al.. Late diagnosis of HIV infection in the era of highly active antiretroviral therapy: consequences for AIDS incidence. AIDS. 2002;16:1945–1951.
30. Giordano TP, Gifford AL, White AC Jr, et al.. Retention in care: a challenge to survival with HIV infection. Clin Infect Dis. 2007;44:1493–1499.
31. Krentz HB, Siemieniuk RA, Gill MJ. Similar challenges with retention in care. Clin Infect Dis. 2007;45:1527.
32. Gill MJ, Krentz HB. Unappreciated epidemiology: the churn effect in a regional HIV care programme. Int J STD AIDS. 2009;20:540–544.
33. Krentz HB, Worthington H, Gill MJ. Adverse health effects for individuals who move between HIV care centers. J Acquir Immune Defic Syndr. 2011;57:51–54.
34. Bangsberg DR, Perry S, Charlebois ED, et al.. Non-adherence to highly active antiretroviral therapy predicts progression to AIDS. AIDS. 2001;15:1181–1183.