Introduction

HIV stigma has widely been recognized as a substantial barrier to accessing and using HIV prevention, treatment, care and support services [1, 2]. Systematic reviews suggest that HIV stigma and discrimination are associated with reduced access to HIV treatment/care [3] and ART adherence [4]. HIV stigma and discrimination has also been linked to factors that have been identified as antecedents to HIV treatment and care outcomes (e.g., ART adherence), including social support, self-rated health [5, 6], physical health, mental health [6, 7] and quality of life [8]. Growing research has contributed toward understanding the negative health consequences of HIV stigma among WLWH [6, 9,10,11,12]. In response to this evidence, interventions incorporating stigma reduction strategies, tailored for people living with HIV (PLWH) or more broadly (e.g., health care workers; church groups) have been developed. Many studies suggest some reduction in stigma [13, 14], but less evidence demonstrates an impact on behavioural or biomedical HIV outcomes (e.g., HIV testing; use of antiretroviral therapy (ART) [14]. These results demonstrate the complexity of measuring and addressing stigma, and further suggest that addressing research gaps in understanding the correlates of HIV stigma could contribute to improved intervention outcomes [1, 15].

Several types of HIV stigma have been described [7, 12]. These include anticipated stigma (i.e., fear of disclosure of HIV status and the consequences of disclosure); enacted stigma (i.e., experiencing stigmatizing behaviours by others, which can include violence and exclusion); perceived stigma (i.e., beliefs about how others perceive individuals living with HIV negatively); and internalized stigma (i.e., negative beliefs about oneself and other people living with HIV) [16, 17]. There is limited quantitative research on the correlates of different types of HIV stigma. Existing research suggests that lower educational level and higher anxiety are associated with heightened perceived HIV stigma among Canadian women born in Canada. For Canadian women born outside of Canada, negative or judgemental attitudes by reproductive and other health professionals (e.g., regarding pregnancy) have been associated with higher perceived HIV stigma [18]. Gender and racial differences in HIV-related stigma were identified in a large cohort of men and women living with HIV (WLWH) in Ontario, Canada, with black women reporting significantly higher HIV stigma (versus white men) [19].

Goffman’s (1963) work on stigma, which originated well before the first cases of HIV were recognized [20], alongside Parker and Aggleton (2003)’s work, underscores the importance of structural factors and conditions that enable or inhibit the development of HIV stigma [1, 21,22,23,24]. HIV stigma is produced and experienced uniquely across societies, cultures and settings, with important sex/gender differences [6, 25, 26]. Gendered social inequities continue to exist, and result in unique challenges for cisgender (cis) and transgender (trans) WLWH that can shape the intensity and frequency of how women experience HIV stigma [2]. Restrictive gendered social norms about sexuality, combined with deep-rooted social-cultural associations between women’s sexuality and HIV, including sex work have been shown to intensify women’s experiences with HIV stigma [19, 27,28,29]. Intersecting or intersectional stigma refers to the convergence of marginalized social identities (e.g., drug use, sex work, ethno-racial minority) and corresponding social marginalization processes (e.g., drug use stigma, sex work stigma, racism) [6, 27]. Intersectional stigmas further shape experiences of HIV stigma for WLWH [6, 27]. Women may require more access to health services than men, including sexual health (e.g., Papanicolaou [Pap] tests) and reproductive health services (e.g., prenatal testing). These health care contacts provide more opportunities for learning about their HIV serostatus as well as potentially stigmatizing interactions in health care.

In Canada, the rate of new HIV infections among women has showed a slightly increasing trend between 2013 (2.5/100,000) to 2017 (3.2/100,000), comparable to the increase in men (9.2 in 2013 to 9.9/100,000 in 2017) [30]. In British Columbia for the year ending 2019, it was estimated that there were approximately 1867 WLWH (18% of the total population living with HIV), with 88% of women linked to care and 69% of women on ART [31]. Women continue to experience reduced ART adherence and viral suppression compared to men [32]. Research gaps continue to exist in understanding the correlates of HIV stigma among WLWH, with even fewer studies investigating multiple stigma types. Given these gaps, and the persistence of HIV stigma experienced by WLWH and impact on HIV treatment and care outcomes, it is important to investigate correlates of heightened HIV stigma among WLWH. These investigations will help understand how different types of HIV stigma are produced, as well as guide recommendations for the scale-up of existing programs and design for new programs and interventions to reduce and address the effects of different types of HIV stigma. Our study therefore examined correlates of each of four stigma types as outcomes among WLWH in Metro Vancouver, Canada.

Methods

Study Design and Sampling

Data were drawn from 2 years (September 1/15 to August 31/17) of a longitudinal community-based participatory open cohort of cis or trans WLWH who lived and/or accessed care in Metro Vancouver, Canada (2014-present) (Sexual Health and HIV/AIDS: Women’s Longitudinal Needs Assessment “SHAWNA”). SHAWNA is a partnership of more than 20 women’s HIV and community service providers, and is guided by a Community Stakeholder Advisory Board that meets semi-annually, and a Positive Women’s Advisory Board of WLWH that meets every 2–3 months. SHAWNA was developed from over 6 months of community consultations with WLWH, HIV care providers and policy experts.

Eligibility for SHAWNA includes being 14 years or older, identifying as a cis or trans woman and living and/or accessing HIV/AIDS services in Metro Vancouver. Various recruitment methods based on community-based research principles were used to recruit participants. These included self-referrals and referrals from HIV care providers, peer navigators, HIV/AIDS service organizations and clinical outreach, including BC’s primary referral centre for WLWH (Oak Tree Clinic, based at BC Women’s Hospital). Participants who have provided informed consent complete a semi-annual questionnaire at baseline and 6-monthly follow-ups, administered by trained peer researchers/community interviewers.

In a visit with the study nurse, participants are also offered HIV viral load/CD4 measurements as well as education and referrals to other needed health and social services. All tests, referrals and other services are voluntary and offered regardless of enrolment in the study. At each visit, participants receive CAN$50 (Canadian dollars) remuneration for their time, travel and expertise. SHAWNA holds ethical approval through Providence Health/University of British Columbia Research Ethics Board and BC Women’s Hospital.

Measures

HIV stigma was measured using nine questions from Wright’s shortened 10-item version [33] of Berger’s HIV-stigma scale [34]. The question “Most people think that a person with HIV is disgusting” was perceived by the study participants and team as re-traumatizing and was removed. Cronbach’s alpha, based on responses at participants’ first study visit, was 0.79 for the anticipated stigma sub-scale, 0.88 for the enacted stigma sub-scale and 0.87 for the internalized sub-scale (0.84 for the overall scale).

Explanatory variables (correlates) were chosen based on a literature review and ongoing discussions with clinicians, community members, peer researchers and members of our Community Stakeholder Advisory Board and Positive Women’s Advisory Board. We describe here definitions for some key social-demographic variables. Indigenous people represent diverse and heterogenous cultures and languages in the geographic area that is now referred to as Canada through settler colonialism practices and policies; “Indigenous” in this study, reflecting the context and setting, refers to and includes those identified as any of First Nations, Métis, or Inuit. We use the term “im/migrants” to include the diversity of refugee, immigrant, and migrant women born in other countries who entered Canada, inclusive of long-term and recent arrivals, refugees, asylum seekers, economic, and undocumented im/migrants. We use the term “gender minority” to refer to all participants who identify as: being a member of gender minority communities, including transgender, transsexual, two spirit, genderqueer or any other gender vs. cisgender; and the term “sexual minority” to refer to all participants who identify as a member of sexual minority communities, including gay, lesbian, bisexual, two spirit, asexual, queer or other sexuality vs. straight.

Social demographic factors included: age; identifying as a member of gender minority communities (last 6 months); identifying as a member of sexual minority communities (last 6 months); race (defined as Indigenous [First Nations, Metis, or Inuit], other racialized persons [African/Caribbean/Black/other visible minority including East/South/Southeast Asian and Middle Eastern], vs. white); im/migrant to Canada; and having a physical disability that limits mobility (last 6 months). Structural vulnerability factors included homelessness (i.e., sleeping on the street/having no place to sleep one night or longer) (last 6 months); incarceration (i.e., staying overnight or longer in jail) (last 6 months); sex work (last 6 months); injection drug use (last 6 months); and non-injection drug use (last 6 months). Interpersonal factors included: gender-based physical and/or sexual violence (by any perpetrator, last 6 months); HIV disclosure without consent (last 6 months); and HIV-related violence (verbal and/or physical violence due to known/perceived HIV-positive status) (last 6 months). HIV-related individual factors included duration since HIV diagnosis (measured in years); disclosure of HIV status to anyone at diagnosis; HIV having negative physical effects/symptoms (e.g., ‘weight loss’; ‘night sweats’) (last 6 months); and viral load suppression (< 50 copies/ml HIV-1 RNA viral load) (last 6 months). Healthcare-related factors included ever being advised not to become pregnant by a healthcare professional and healthcare barriers in the last 6 months related to: empowerment and choice (e.g., ‘couldn’t get doctor of preferred gender’; ‘lack or programs/services tailored for me’) or to safety and respect (e.g., ‘lack of sensitivity to privacy’; ‘poor treatment by health professionals’); the list of survey responses to healthcare barriers were developed with extensive participant and community consultation in our own setting.

Analysis

Descriptive statistics were calculated for all HIV stigma measures (agree/strongly agree vs. neutral/disagree/strongly disagree) and sample characteristics. Descriptive statistics included frequencies and proportions for categorical variables, and measures of central tendencies (i.e., mean, medians, and interquartile ranges [IQRs]) for continuous data. Sample characteristics were compared according to the overall HIV stigma measure. Pearson’s chi-squared tests (or Fisher’s exact test for small cell counts) were used for categorical and Wilcoxon rank-sum test were used for continuous variables (since all of the continuous variables were not normally distributed).

We considered the sub-scale outcomes ordinal measures, and used bivariate and multivariable cumulative logistic regression using generalized estimating equations (GEE) and an independent correlation structure to screen for independent associations between potential correlates and each of the four types of HIV stigma outcomes. GEE were used in order to adjust standard error estimates to account for correlations arising from repeated measurements on the same participants over time, as our sample included time-varying variables over multiple follow-up surveys. Unadjusted/adjusted odds ratios (OR/AOR) and 95% confidence intervals (CI) were used to measure the strength of association between correlates and each of the four types of HIV stigma outcomes.

We used a p-value cut-off of < 0.10 to consider variables for inclusion in the multivariable models. A backward stepwise approach was used to construct the final multivariable model, beginning with a full model, followed by the removal of variables with the highest p-value and an assessment of the quasi-likelihood under the independence model criterion (QIC) after the removal of each subsequent variable. The QIC value provides an indication of the combination of variables that best explain the variability in our outcomes, the HIV stigma sub-scales. We use a manual selection approach assessing both significance and fit of the model at each step in order to strive for model parsimony while retaining important explanatory variables. Backwards elimination allows us to examine the effect of including all hypothesized explanatory variables in the model (i.e., the full model) [35]. The final model presented is the model with the lowest QIC value and was tested for collinearity. All p-values are two-sided, and SAS software version 9.4 was used for all statistical analyses (SAS Institute Inc., Cary, NC).

Results

Sample Characteristics

Overall, 215 women responded to ≥ 1 follow-up survey, of four possible surveys included in this study, with 509 total observations over 2 years of follow-up (September 1/15 to August 31/17).

Table 1 presents social and structural factors representing sample characteristics reported at participants’ baseline interviews, and bivariate comparisons at baseline between these factors and experiencing any HIV stigma (agree/strongly agree to any of the nine HIV Stigma measures). At baseline, the median age of participants was 46 years (IQR: 39–53 years), with a median of 15 years (IQR: 9–21) since first diagnosed with HIV (Table 1). Indigenous women were over-represented in the sample (56.7%) relative to the BC population (5% in 2011), with 35.4% of participants reporting being white and 7.9% as African, Caribbean or black, or other racialized persons. Overall, 39.1% of participants reported identifying as a member of sexual minority communities and 11.2% reported identifying as a member of gender minority communities. In the last 6 months, 10.7% reported being homeless, 23.3% reported sex work and 47.0% and 48.4% reported injection and non-injection drug use, respectively (Table 1).

Table 1 Social and structural sample characteristics at first follow-up of our sample of women living with HIV in Metro Vancouver
Full size table

Table 2 shows the prevalence of reporting agree/strongly agree vs. neutral/disagree/strongly disagree at participants’ baseline interview, overall and for the four types of HIV stigma. In the last 6 months (ever), 68.4% (ever = 73.5%) of participants reported anticipated HIV stigma, 38.1% (ever = 62.8%) reported enacted stigma, 36.3% (ever = 45.6%) reported perceived stigma and 28.4% (ever = 30.7%) reported internalized stigma, with 80.0% (ever = 87.9%) reporting agree/strongly agree to any HIV stigma measure (Table 2).

Table 2 Definitions of four HIV stigma types based on the HIV Stigma Scale and proportion (n) reporting strongly agree or agree to HIV stigma measure at first follow-up
Full size table

Multivariable Analysis

Unadjusted and adjusted odds ratios from logistic regression models using GEE and 95% CIs are presented in Table 3. In multivariable analysis, HIV disclosure without consent was significantly associated with anticipated (AOR: 2.33 [1.43–3.81]); enacted (AOR: 4.35 [2.65–7.12]); and perceived HIV stigma (AOR: 2.69 [1.58–4.58]) stigma. Verbal or physical violence related to HIV status was significantly associated with heightened enacted (AOR: 2.86 [1.65–4.95]), internalized (AOR: 2.38 [1.26–4.48]) and perceived HIV stigma (AOR: 2.66 [1.45–4.86]). Experiencing negative physical effects/symptoms of HIV was significantly associated with anticipated (AOR: 1.68 [1.16–2.42]), enacted (AOR: 1.68 [1.16–2.44]), internalized (AOR: 2.13 [1.48–3.05]) and perceived HIV stigma (AOR: 1.62 [1.10–2.37]). Physical/sexual violence by anyone was significantly associated with perceived HIV stigma (AOR: 1.67 [1.00–2.77]). Healthcare access barriers related to safety and respect (AOR: 1.74 [1.03–2.93]) and ever being advised not to become pregnant (AOR: 2.09 [1.25–3.50]) were associated with enacted HIV stigma. Injection drug use (AOR: 1.56 [1.04–2.34]) and duration since first diagnosed with HIV (AOR: 0.97 [0.94–1.00] for each 1-year increase in duration) were significantly associated with internalized HIV stigma (Table 3).

Table 3 Bivariate and multivariable odds ratios for the relationship between social and structural correlates and four types of stigma
Full size table

Discussion

Evidence from our study suggests that WLWH experience high levels of HIV stigma, both recently as well as historically throughout their lives. There were consistent patterns as well as important differences that emerged in the relationships between factors investigated and the four different types of stigma, with key implications for public health policy.

Our study provides quantitative evidence showing that HIV status disclosure without consent is associated with anticipated, enacted and perceived HIV stigma. Evidence from studies suggests that disclosure of HIV status without consent (e.g., through assertive HIV outreach to WLWH’s place of residence inadvertently identifying women as HIV-positive) can be a devastating violation of privacy with serious negative consequences for WLWH, including rejection, discrimination and violence [29, 36, 37]. Decisions to disclose are highly personal and it is important for women to retain autonomy of when and to whom to disclose [38]. Safe disclosure of sensitive health information in general is thought to be beneficial, and for WLWH can be important in terms of facilitating connections and reducing barriers to health services including HIV treatment and care [39, 40]. Not disclosing HIV status can act as a barrier to health access (e.g., appropriate referrals) [41]. Disclosure of one’s HIV status has been shown to be positively correlated with social support, including in a review and meta-analysis [42] and quality of life [40] and can reduce psychological stress and build community, highlighting the importance of supporting safe disclosure practices for WLWH.

Our findings suggest it is important to address both structural and institutional factors to reduce situations where women’s HIV status would be disclosed involuntarily and instead support safe disclosure on women’s terms. In Canada, where HIV non-disclosure is criminalized, concerns and confusion regarding legal repercussions of non-disclosure laws complicate disclosure practices. This is especially true for women, who may be in relationships where gendered power dynamics in favour of male partners and fear of gendered violence make it difficult to safely disclose HIV status [43, 44]. Disclosure of HIV status without consent may be used as a threat by partners to enact or reinforce emotional or financial control within the relationship [44].

Criminalized approaches have been associated with heightened HIV stigma and forced disclosure of HIV status [43]. Moreover, evidence suggests these approaches do not reduce HIV prevalence or increase ‘public safety’ [45]. Reforming such laws could potentially have a substantial impact on supporting safe disclosure and reducing HIV stigma. Safe disclosure and confidentiality related to HIV status is also paramount for broader health providers, who have a responsibility to avoid inadvertently disclosing patients’ status. Health providers can facilitate access to guidelines and counselling to support people living with HIV through a safe disclosure process and minimize negative impacts of disclosure [46].

Several measures reflecting negative interactions with health providers, including being advised not to become pregnant and health barriers related to safety and respect (e.g., ‘lack of sensitivity to privacy’; ‘poor treatment by health professionals’), were associated with enacted HIV stigma. These results indicate a correlation between health provider approaches or attitudes towards participants and outward acts of discrimination against participants based on their HIV status. Existing research suggests that broad health providers, including dental health professionals, physicians and nurses can be sources of enacted HIV stigma in ours and other settings [47,48,49], and that stigmatizing experiences with health professionals can act as a barrier to health care access [47]. In line with some existing services in Metro Vancouver [50, 51] and elsewhere that provide such services, policies and programs that emphasize the importance of safe, non-judgemental care for PLWH at all levels of health education and practice is essential to shifting attitudes and approaches toward reducing HIV stigma.

Our results also suggest a relationship between physical effects/symptoms of HIV and HIV stigma. These results are in line with previous studies have suggested that individuals with visible symptoms of HIV find it more difficult to conceal their HIV status, and experience higher levels of perceived HIV stigma [52, 53]. Health care providers can help ameliorate this relationship by providing safe, respectful environments to support retention in care and facilitate positive connections with peers and referrals to community supports.

Physical and verbal violence related to HIV status was linked to heightened enacted, internalized and perceived HIV stigma; physical/sexual violence by any perpetrator was associated with perceived stigma. Physical and verbal violence and enacted stigma could occur in related or the same events, or be produced by the same driver (e.g., involuntary HIV disclosure; HIV-related disability [54]), helping to explain the correlation. The mechanisms through which verbal and physical violence related to HIV status, a form of enacted HIV stigma, may influence the development of internalized and perceived HIV stigma, are complex and require further study [54, 55]. Historically, intrapersonal stigma types like internalized and perceived HIV stigma have been thought to develop in parallel or iteratively through individual-level processes that lead to absorption of negative beliefs/actions/attitudes related to HIV status. These processes include those related to psychological morbidity (e.g., depression symptoms) or social-cognitive factors (e.g., shame and self-blame, reduced coping, self-esteem, self-worth and self-efficacy) [54, 55]. More recent research has proposed that social and structural factors also play an important role alongside or outside of experiences of enacted HIV stigma and that internalized HIV stigma “operates within mutually reinforcing relationships with other marginalized social statuses based on sex, age, gender identity and expression, racialisation, sexual orientation and behaviors, illicit drug or alcohol use, sex work, criminalization and incarceration” [55].

Much more work needs to be done to understand the mediating factors between different types of violence and HIV stigma to identity optimal areas in which to intervene. Strategies to reduce gender-based violence against women should regardless be incorporated more broadly into HIV prevention and care programs. High prevalence of violence and trauma among WLWH has led to calls for trauma-informed and women-centred approaches in ours [50] and other settings [56]. Trauma-informed models of HIV care guided by WLWH should continue to be expanded on to address drivers of HIV stigma and HIV stigma itself. Strong peer networks and increased community connections may help WLWH counter negative self-image and ameliorate the effects of violence [57, 58]. Other studies have suggested that over time since their initial HIV diagnosis, women have resisted the effects of internalized and enacted stigma through various tactics. These include identifying and seeking out supportive allies; redefining stigma as ignorance; and becoming advocates themselves for people living with HIV [38]. Our study results suggesting that longer duration of HIV and reduced internalized stigma are correlated are in line with these observations [53, 55]. Participatory programs and interventions that include WLWH as leaders could thus make strong contributions toward addressing negative effects of violence, disclosure and HIV stigma.

Our study had a number of limitations as well as strengths. Our study may not be representative of all WLWH in Metro Vancouver nor be generalizable to Metro Vancouver or other settings, especially as HIV stigma is highly contextual. However, recruitment of study participants for SHAWNA is achieved by broad coverage of diverse AIDS Service Organizations and clinical settings by peer research associates, community interviewers and clinical outreach. It is important to note that such recruitment approaches may result in a sample that is over-represented by participants who are connected with HIV services and care. However, we feel that our sample reflects the diversity of WLWH and we regularly come into contact with WLWH who are not connected to HIV services or care through outreach. We have been able to successfully re-connect or connect these participants to services and care. We did not examine differences in patterns of association among sub-populations, including by gender/sexual identity or race; however, given the intersecting stigmas that WLWH identifying as gender/sexual diverse, and ethno-racial minorities, and Indigenous peoples experience [27], future studies should investigate these differences. For instance, a recent study of sexual orientation differences among women with HIV in Canada reported increased racial and gender discrimination, violence, sex work and drug use among sexually diverse women with HIV vs. heterosexual [59].

To conclude, study results suggest the need to develop strategies to address social and structural violence against WLWH. These strategies include amending Canada’s restrictive HIV disclosure laws as a structural intervention to reduce HIV stigma and promote safe disclosure for WLWH. Paramount to all discussions of how to understand and reduce HIV stigma for women is the meaningful inclusion of the expertise and perspectives of WLWH.