Stigma and intersectionality: a systematic review of systematic reviews across HIV/AIDS, mental illness, and physical disability

Background Stigma across HIV/AIDS, mental illness, and physical disability can be co-occurring and may interact with other forms of stigma related to social identities like race, gender, and sexuality. Stigma is especially problematic for people living with these conditions because it can create barriers to accessing necessary social and structural supports, which can intensify their experiences with stigma. This review aims to contribute to the knowledge on stigma by advancing a cross-analysis of HIV/AIDS, mental illness, and physical disability stigma, and exploring whether and how intersectionality frameworks have been used in the systematic reviews of stigma. Methods A search of the literature was conducted to identify systematic reviews which investigated stigma for HIV/AIDS, mental illness and/or physical disability. The electronic databases MEDLINE, CINAHL, EMBASE, COCHRANE, and PsycINFO were searched for reviews published between 2005 and 2017. Data were extracted from eligible reviews on: type of systematic review and number of primary studies included in the review, study design study population(s), type(s) of stigma addressed, and destigmatizing interventions used. A keyword search was also done using the terms “intersectionality”, “intersectional”, and “intersection”; related definitions and descriptions were extracted. Matrices were used to compare the characteristics of reviews and their application of intersectional approaches across the three health conditions. Results Ninety-eight reviews met the inclusion criteria. The majority (99%) of reviews examined only one of the health conditions. Just three reviews focused on physical disability. Most reviews (94%) reported a predominance of behavioural rather than structural interventions targeting stigma in the primary studies. Only 17% of reviews used the concept and/or approach of intersectionality; all but one of these reviews examined HIV/AIDS. Conclusions The lack of systematic reviews comparing stigma across mental illness, HIV/AIDS, and physical disability indicates the need for more cross-comparative analyses among these conditions. The integration of intersectional approaches would deepen interrogations of co-occurring social identities and stigma.


Background
Stigma is a dynamic process enacted through structures and individuals, and mediated by relationships of power, control, and domination that are continuously produced and reproduced by actors [1]. At its foundation, stigma is about social inequality and social control, which create a hierarchy that devalues stigmatized people [1].
Stigma is especially problematic for people living with HIV/AIDS (Human Immunodeficiency Virus Infection and Acquired Immune Deficiency Syndrome), mental illness, and physical disabilities because it can create barriers to accessing health care, education, employment, and affordable housing, which in turn, may exacerbate the experience of marginalization [2,3]. Furthermore, people often live with more than one of these health conditions and may simultaneously experience different kinds of health-related stigma. For example, research indicates that people living with HIV/AIDS have higher rates of depression and anxiety in comparison to the general population [4], and people with physical disabilities are at an elevated risk for depressive symptoms and major depressive disorder [5].
The overlap of different kinds of disease stigma and the rooted-ness of stigma in larger systems of inequality and webs of power have pushed researchers to consider different ways to investigate and analyze it. Given stigma's links to historical and contemporary manifestations of inequality, power, and systems of domination; intersectionality offers a promising theoretical approach to examine research on stigma. Black feminists, who coined and produced theory on the concept of intersectionality, highlighted how multiple oppression and structural inequalities exist in matrices of domination, which in turn, reinforce unequal relationships of power amongst people; and between people and social institutions such as healthcare, housing, and the law [6][7][8]. Originally used in feminist theory to describe Black women's positions within webs of power, intersectionality has been taken up by health sciences researchers to help deepen their analyses of structural and systemic issues in health, and the inequalities and inequities they create [9,10].
This review of reviews seeks to contribute to the knowledge on stigma by advancing a cross-analysis of HIV/ AIDS, mental illness, and physical disability stigma, and exploring whether and how intersectionality frameworks have been used in the systematic reviews of stigma.

Methods
We adapted Arskey and O'Malley's [11] scoping review framework to guide the methodology of our review. We used the same subheadings as the authors for the "identifying the research questions" [11] and "identifying relevant studies" [11] stages. However, 'relevance review' was used instead of "study selection" [11], and we collapsed "charting the data" and "collating, summarizing and reporting the results" [11] into a single subheading called 'data extraction, collation and analysis' to reflect our methodology.

Identifying the research question(s)
The research questions guiding our review were: 1. What are the characteristics of systematic reviews examining sources of and influences on stigma among those living with HIV/AIDS, mental illness, and/or physical disability? 2. Has intersectionality been used and how has it been used in systematic reviews of stigma and stigma reduction interventions for those living with HIV/ AIDS, mental illness, and/or physical disability?

Identifying relevant studies
The search strategy and electronic database searches were developed and conducted with the assistance of a librarian at the University of Ottawa Health Sciences Library. Electronic databases that publish health-related research and information were accessed. Five databases were searched: MEDLINE, Cumulative Index to Nursing and Allied Health Literature (CINAHL), EMBASE, COCHRANE (Database of Systematic Reviews, EBM Reviews-ACP Journal Club, and EBM Reviews-Database of Abstracts of Reviews of Effects), and PsycINFO. Several databases indicated that prior to the year 2005, stigma was not a mesh heading, and terms such as "discrimination", "stereotyping", and "prejudice" were commonly used. Although we were searching for publications from 2005 and later, we used both the newer and older search terms to ensure that we captured all relevant reviews. The search headings used in all five electronic databases included: "stigma", OR "prejudice", OR "social discrimination", OR "social stigma", OR "stereotyping", OR "stereotyped attitudes", OR "shame". Keywords used in the search of the databases also included truncated versions of the following terms: discrimination stigma, and prejudice.
Search filters (hedges) were used in the MEDLINE, EMBASE, and PsycINFO databases to limit retrievals to systematic reviews. These filters were not necessary for the COCHRANE database because it only publishes systematic reviews. A filter to restrict retrieved papers to systematic reviews was also applied to the CINAHL database search.
Inclusion criteria for the database searches were: 1. Reviews written in English language 2. Reviews published between January 2005 and November 2017 (inclusive).
Exclusion criteria for the database searches were: Zotero, a software reference package, was used to manage the citations.

Relevance review
Following the first database search, both authors independently reviewed a sample (n = 15) of retrieved titles and abstracts for relevance. They then met to discuss discrepancies in their assessments, and refine the final inclusion criteria for reviews, which were: Systematic reviews using qualitative, quantitative, or mixed methods; Focus on health-related stigma experienced by study populations with HIV/AIDS, mental illness, and/or physical disability. The definition of mental illness used in the review is consistent with the following definition: "a spectrum of cognitions, emotions and behaviours that interfere with interpersonal relationships as well as functions required for work, at home and in school" [2]. The definition of physical disability used in the review is consistent with the following definition: "any infirmity, malformation or disfigurement that is caused by bodily injury, birth defect or illness and, without limiting the generality of the foregoing, includes diabetes mellitus, epilepsy, a brain injury, any degree of paralysis, amputation, lack of physical co-ordination, blindness or visual impediment, deafness or hearing impediment, muteness or speech impediment, or physical reliance on a guide dog or other animal or on a wheelchair or other remedial appliance or device" [12] Stigma is an outcome, result and/or theme of the review and discussed in the research findings; and, Review includes measurement tools; conceptual frameworks and theoretical frameworks such as, but not limited to intersectionality; guidance documents; and/or methodological approaches for exploring stigma.
The titles and abstracts of all citations were then screened for relevance by the authors. When relevance could not be ascertained, the full paper was retrieved and reviewed to make a relevance decision.

Data extraction, collation and analysis
Data were extracted from the reviews using the following categories: aim/objective, specific health issue addressed (i.e. type of mental illness or disability), type of systematic review and number of primary studies included in the review, their geographic location, study design (qualitative, quantitative or mixed methods), study population, type(s) of stigma addressed (interpersonal stigma, intrapersonal stigma, and structural/institutional stigma), and destigmatizing interventions used. We also extracted key findings and recommendations from each review. Data were entered into a table in Microsoft Excel. To ensure that we captured all descriptors of intersectionality, we then did a keyword search of each eligible review paper using the terms "intersectionality", "intersectional", and "intersection". We extracted all definitions and descriptions of these terms from these papers as well as any related findings. We used matrices to compare the characteristics of reviews and their application of intersectional approaches across the three health conditions.

Results
The electronic database search yielded 2405 citations. In the first exclusion phase, 691 duplicates were eliminated leaving 1714 citations for relevance review. In the second exclusion phase, 1487 papers were eliminated because they were ineligible or found to be additional duplicates. In total, 227 papers were identified for a full text review. One hundred twenty-nine papers were found to be ineligible during the third exclusion phase. In total, 98 retrieved systematic reviews of stigma were included in our review (See Fig. 1 for an overview of the search results, and Table 1 for an overview of the reviews).

Disease/condition focus, publication date, and geographic location in the primary studies
Primary studies were reported from over 60 countries; all continents were represented except Antarctica. The majority of the reviews were disease specific, with the largest proportion found for mental illnesses (61%, n = 60), followed by HIV/AIDS (34%, n = 34); a smaller number of reviews were found for physical disability stigma (3%, n = 3). A single review (1%, n = 1) looked at stigma across all three health conditions [13] and included other stigmatized health conditions including leprosy, tuberculosis, and epilepsy.
Most reviews of HIV/AIDS and mental illness stigma had been published within the last 5 years (64%, n = 63). With only three reviews for physical disability stigma, no publication pattern was discernable.

Sample characteristics of the systematic reviews
Across the reviews, the study populations were mostly comprised of people living with one of the three health conditions. For example, people living with HIV/AIDS (PLWHA) were most commonly included in primary research studies on HIV/AIDS stigma (37%, n = 22). In the reviews on mental illness stigma, the predominant study populations in the primary studies were people living with mental illness and mental healthcare consumers/ users (50%, n = 17). There were three reviews in the physical disability stigma category: one examined people living with a physical disability (33%, n = 1), the second focused on those who interacted with people living with disabilities (33%, n = 1), and the third explored how disability has been considered in nursing and healthcare literature (33%, n = 1) [14]. Table 3 shows the number and percentages of each type of stigma investigated in the systematic reviews. Nearly half (47%) of the reviews discussed more than one stigma type, even when the type of stigma described was not an eligibility criterion. Across the health conditions, various stigma types were examined: 78.5% examined intrapersonal forms of stigma (i.e. self-stigma, internalized stigma, perceived stigma, affiliate stigma), 48% of the reviews looked at interpersonal forms of stigma (i.e. social stigma, public stigma, enacted stigma, cultural Search terms: Search Headings: "Stigma", OR "prejudice", OR "social discrimination", OR "social stigma", OR "stereotyping", OR "stereotyped attitude", OR "shame", OR "psychosocial" Keyword search (truncated): discriminat, stigma, prejudic Protocols and filters: hedges, "review", "systematic review", "meta-analysis"         Interventions to manage, reduce, and prevent stigma were included in 36% (n = 35) of the reviews. Among these, interventions for mental illness stigma were the most common (74%, n = 26), followed by interventions for HIV/AIDS stigma (23%, n = 8), and a smaller number of interventions for physical disability stigma (3%, n = 1) (See Table 4 for a complete list of interventions and their characteristics).

Stigma type and interventions included in the primary research
Behavioural interventions such as psychoeducation, informational approaches, and/or social contact were most commonly reported in the reviews of interventions (94%, n = 33). In comparison, only one review of HIV/AIDS stigma described structural interventions. In this latter review, a structural intervention (universal precaution supplies to healthcare workers) was combined with a behavioural intervention focusing on information and skill-building to combat HIV/AIDS stigma [15]. Additionally, only 3 reviews described community-based interventions, which included a community-empowerment approach to respond to HIV/AIDS among sex workers [16], community mobilization and involvement to address HIV/AIDS stigma in India [17], and advocacy-based approaches for mental illness stigma in Canada [18].

Intersectionality
Our text search yielded 13 reviews (17%) whose authors had used an intersectional lens to analyze primary research studies. The majority of these reviews were found in the work on HIV/AIDS stigma (92%, n = 12); just one review of mental illness stigma (8%) used intersectionality. While all 13 of these reviews mentioned intersectionality when describing how a mental illness or HIV/AIDS diagnosis intersected with culture, power and/or other differences to reinforce social conditions for stigmatization [19,20], only three of these 13 reviews (23%) [21][22][23] provided a definition for intersectionality. Loutfy et al. [21] and Monteiro et al.'s [22] used Crenshaw's [6,7] concept of intersectionality in discussing how health inequities are impacted by categories of difference like HIV status, race, gender, and sexuality. Both reviews used this concept of intersectionality to analyze the primary studies and highlighted its usefulness in improving knowledge about how processes of marginalization overlap and are impacted by HIV/AIDS stigma [21,22]. Sandelowski's [23] review used a more recent concept of intersectionality, which characterized it as a research paradigm and methodological intervention [24,25]. For instance, Sandelowski [23] explained that intersectionality acknowledges intra-category diversity and can help to investigate relationships between and among dynamic categories of difference like gender in health research. However, Sandelowski [23] found that non-intersectional, unitary analytical approaches that isolated overlapping categories of difference like race, gender, and HIV stigma from one another were overwhelmingly used in the primary research studies.

Intersectional stigma and interventions
Five of the reviews that included intersectionality used it as a framework to discuss the occurrence of intersectional  [21] review as the "multiple, simultaneous and dynamic interchanges among categories of social difference as it interlinks with power and privilege, and systemic oppression and its operation at the micro, mesa, and macro levels" (p. 2). The HIV/AIDS reviews tended to focus on stigma amongst groups that have been socially and historically marginalized such as Black women [21,26], Black men who have sex with men [27], and sex workers [16]. In these reviews, focusing on intersectional stigma allowed other forms of social inequality experienced by people living with HIV/AIDS like racism, sexism, and homophobia to be included and highlighted the existence of overlapping forms of oppression and marginalization.
Although interventions for stigma were described in 36% of the reviews, none of the authors indicated whether or how intersectional interventions were used in the primary studies. However, three of the 35 reviews (8%) highlighted the lack of intersectional interventions designed to address intersectional stigma in the primary studies, and advocated for their use in primary research studies [15,21]. Loutfy et al.'s [21] review found there was an absence of stigma-reducing interventions that addressed co-occurring stigmas experienced by Black women who are HIV positive, and that most focused on interpersonal and intrapersonal stigma. Kerrigan et al.'s [16] review of community-empowerment interventions to counter HIV stigma amongst mainly female sex workers found that one of the greatest structural barriers to the implementation and scale-up of these interventions was the presence of intersectional stigma.

Discussion
To our knowledge, this is the first review to provide a cross-analysis of systematic reviews of HIV/AIDS, mental illness, and physical disability stigma. It is also the first review of reviews to examine whether and how intersectionality has been used as an analytic approach on stigma.
The nearly complete lack of reviews that examined stigma across these three health conditions may be partly attributed to trends in stigma research [28], since systematic review topics are constrained by primary studies on a topic. Researchers conducting effectiveness studies of interventions concentrate on specific target Table 2 Types of systematic reviews, definitions, and references Type of systematic review Definition and reference(s) Critical synthesis A review which aims to demonstrate that the writer has extensively researched the literature and critically evaluated its quality. This review technique incorporates analysis and conceptual innovation [112].
Integrative review A technique that integrates review, critique, and synthesis of representative literature on a topic to create new frameworks and perspectives on the topic [113]. It also includes experimental and non-experimental research studies and combines theoretical and empirical data to gain a more comprehensive understandings of a phenomenon [114] .
Meta-analysis A review technique that "systematically combines the results of quantitative studies to provide a more precise effect of the results" [112].
Meta-ethnography An interpretive and inductive approach that combines and sometimes compare the findings of ethnographic research or qualitative research to provide a higher level of analysis, generate new research questions, and reduce duplicate research [115,116].
Meta-study A research approach that involves the analysis of theory, methods, and findings of qualitative research and synthesizes these insights into new ways of thinking about some phenomena [117].
Mixed methods review A review technique that combines qualitative and quantitative approaches [112].
Scoping review A research approach that provides a "preliminary assessment of the potential size and scope" of the research on a particular subject. The aim of this review is to identify the nature and size of research [112].
Qualitative or quantitative systematic review Engages in a systematized search, appraisal, and synthesis of research evidence which adheres to a set of guidelines [112]. Some qualitative and quantitative reviews include an assessment of methodological quality, while others may not.

Interpersonal Stigma
The process in which members of the general public direct stigma towards individuals with a specific health condition 47 48 Institutional/Structural Stigma Practices initiated at the institutional level that work to disadvantage a stigmatized group or person 3 3 Note: a Percentages do not add up to 100 because some reviews included more than one type of stigma HIV/AIDS Educational interventions groups and conditions, in part because the aim of effectiveness research is to determine whether a specific outcome can be attributed to a particular intervention. Thus, researchers doing effectiveness studies are more likely to focus on homogenous (e.g. populations with a primary diagnosis such as HIV/AIDS) rather than heterogeneous populations. There is also a predominance of disease-specific funding, which may preclude and/or discourage cross-analytical work. Funding calls that require stigma research on heterogeneous samples would foster this kind of work, enabling cross-analyses of stigma for these health conditions. We noticed that in reviews of intervention studies, there is a dominant focus on behavioural rather than structural interventions for stigma, and that reviews typically focus on interpersonal and intrapersonal stigma rather than structural and institutional stigma. This gap may be due to a dearth of primary studies with a structural focus, reflecting a persistent person-centric orientation towards stigma reduction [29]. In addition to the need for more reviews that compare stigma and destigmatizing interventions across disease conditions, there is a need for reviews that compare behavioural and structural interventions or their combination. This would provide an important basis for comparing stigma reduction approaches with either (or both) orientations.
Our review indicates that research on stigma has begun to move in the direction of acknowledging the intersectionality of these experiences and grappling with how stigmatization overlaps with other forms of oppression. Although just one of the mental illness stigma reviews used intersectionality, it highlighted the potential value of intersectionality as an analytical framework, noting that it captured the interlocking effects of various kinds of oppression as it overlaps with mental illness stigma [30]. A small proportion of systematic reviews of HIV/AIDS stigma were categorized as using intersectionality frameworks. We think the integration of intersectionality within some HIV/AIDS reviews could be a result of the expansive body of diverse and cross-cultural, cross-racial, and cross-geographical research on HIV/AIDS stigma, which may have stimulated the application of analytical frameworks that acknowledge the convergence of multiple kinds of stigma and structural inequality.
The limited number of reviews addressing physical disability stigma is noteworthy. We do not know if this gap reflects a lack of primary research studies on this topic. Nevertheless, Boyles et al. [14] stated that Mental Illness

Contact-based interventions and education interventions
If the cell is empty, no descriptions were provided historically, most research on disability has been designed, conducted, and managed by people who do not live with disabilities, which limits knowledge creation about these health conditions. This is in sharp contrast with research on HIV/AIDS, which has a long history of involving persons living with HIV/AIDS in many facets of research studies. If more opportunities were created to meaningfully include and facilitate leadership roles for people living with disabilities in health research, we might see an increase in both primary research and systematic reviews on this topic. Purposefully engaging those living with physical disability in studies on HIV/ AIDS and mental health would help to build the evidence-base on co-occurring stigma.

Recommendations for future research
This review of reviews indicates there is a need for more work that focuses on structural interventions to reduce stigma in both primary studies and systematic reviews. While several reviews highlighted the lack of structural interventions in their findings, we believe it is also imperative to identify concrete examples of these kinds of interventions when reviewers present recommendations for future work on this topic. At the organizational level, examples might include research that examines institutional interventions implemented to reduce stigma in healthcare settings like culturally-specific mental healthcare programs, clinical assessments that omit problematic or pathologizing questions about gender and sexuality, and anti-stigma training for healthcare professionals. At the state level, policies such as legalizing/decriminalizing homosexuality or enacting legislation that protects the rights of people living with disabilities and mental illness are examples of structural interventions that might be expected to have an impact on stigma.
To strengthen the reviews of stigma across these health conditions and in cross-comparative work we call for more explicit integration of intersectionality frameworks in the methodology of systematic reviews. It is not sufficient for reviews to use the language of intersectionality as an afterthought in the conclusion or discussion sections or to hint at co-occurring inequalities or people's multiple social identities without context, clear definitions, and critical reflections on how the term has evolved. Researchers need to work towards more accurate and meaningful inclusions of intersectional approaches, which use the concept to deepen analyses of stigma, particularly as it applies to understanding the presence of stigma from more than one health condition and from other co-occurring sources of stigma emanating from social identities like race, gender, and sexuality.

Limitations
We searched five reputable databases that publish medical and health research. However, we did not search databases that are exclusively in the social sciences and may have missed some pertinent reviews as a result. We did not use intersectionality as a search term. This may have led to the omission of some eligible reviews although that seems unlikely since our broader search terms likely captured any reviews that included intersectional approaches. Nevertheless, the patterns we observed are overwhelming and it seems unlikely that a different pattern would have emerged even if some additional reviews had been found.
We were cautious in our categorization of reviews as using or not using intersectional approaches, and reviews that included a discussion or analysis of other social categories like race or gender were not categorized as intersectional on that basis alone. Our conservative approach is consistent with a literature that warns against misappropriation of the concept, describing how it is sometimes used to gloss over identity politics in research or to treat categories like race and gender as independent variables rather than as reflections of social practices that are linked to larger processes of inequality [9,30,31]. We acknowledge that more conventional systematic review methods may not be congruent with reporting on this deeper and more nuanced approach to the intersectional analyses of stigma.
Our approach to identifying and selecting articles focused on reviews rather than primary studies. While it seems reasonable to conclude that the gaps we identified from systematic reviews mirror gaps among primary studies, we are not able to confirm this.

Conclusions
The nearly total lack of systematic reviews examining stigma across mental illness, HIV/AIDS, and physical disability indicates there are ripe opportunities for further primary research and systematic reviews that undertake a cross-comparative analyses among these health conditions. Approaches such as intersectionality that deepen our interrogation of intersecting stigma and that acknowledge and address larger processes of inequality and inequity that occur alongside health stigma are needed. Such approaches may inform intervention design as well as research methods; these are needed to avoid reproducing and exacerbating inequalities and inequities among population that experience marginalization due to their health condition(s).