Stigma and discrimination related to gender identity and vulnerability to HIV / AIDS among transgender women : a systematic review

1 Departamento de Ciências da Vida, Universidade do Estado da Bahia, Salvador, Brasil. 2 Instituto de Saúde Coletiva, Universidade Federal da Bahia, Salvador, Brasil. 3 Instituto de Humanidades, Artes e Ciências, Universidade Federal da Bahia, Salvador, Brasil. 4 Faculdade de Ciências Médicas da Santa Casa de São Paulo, São Paulo, Brasil. 5 Centro de Ciências Biológicas e da Saúde, Universidade Federal do Oeste da Bahia, Barreiras, Brasil. doi: 10.1590/0102-311X00112718


Introduction
HIV prevalence is disproportionally high among transgender women when compared to the general population 1,2,3 .A metanalysis estimated a prevalence of 19.1% in 15 countries, which is 48.8 times higher than in the reproductive-age population in the same countries 2 .
Various studies have explained this disproportionality by a range of complex individual factors: biological (i.e., unprotected anal sex) and behavioral (i.e., lack of condom use, use of psychoactive substances, etc.), together with structural factors such as stigma and discrimination, which also play an important role and can influence behaviors, practices, and attitudes in relation to HIV, limiting access to socioeconomic resources, especially education, work, and prevention services 2,3,4 .Thus, researchers, activists, and health professionals have considered stigma and discrimination two key factors associated with high HIV prevalence rates 5,6,7,8,9 .
Gender performances of transgender women are seen as insubordination to the dynamics established by heteronormative society over bodies and social relations 10,11 .As a consequence, transgender women face intense stigmatization due to the expression of their gender identities in predominantly patriarchal and male chauvinist societies 3,10 .When comparing men who have sex with men (MSM) and transgender women, the latter experience more stigma and discrimination 7 and more stressful psychosocial events, revealing the existence of discrimination even within the LGBT community 12 .They also present higher HIV prevalence rates than MSM 13 .
Stigma and discrimination due to gender identity are frequently related to the unfavorable social, economic, and psychological context for transgender women 14 , which often relates to their involvement in commercial sex, generally as a result of the limited options for accessing the formal labor market 2,3,4,15 .Even so, the current response to the HIV/AIDS epidemic has emphasized biomedical measures to the detriment and less structural issues, which includes the role of activists that are member of the populations most affected by the epidemic 16 .The current article thus intends to conduct a systematic literature review to analyze the relationship between stigma and discrimination related to gender identity of transgender women and their vulnerability to HIV/AIDS.

Methodology
This is a systematic literature review on stigma, discrimination, and vulnerability of transgender women to HIV/AIDS, involving identification, compilation, analysis, and interpretation of the results of selected studies.The review followed the PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), which describe the specific requirements for systematic review studies and metanalyses 17 .

Search strategies and information sources
Independent reviewers (L.M., M. P.-S.) conducted the study search in PubMed, Scopus, Web of Science, Science Direct, and LILACS, using the following combinations of keywords: "discrimination", HIV, "social stigma" or "stigma" "transgender persons" or "transgender" or "transvestite" (Supplementary Material, Table S1: http://cadernos.ensp.fiocruz.br/site/public_site/arquivo/suppl-e00112718ingles_2106.pdf).The review also examined the reference lists from the relevant studies in order to identify other potentially eligible studies.
In Brazil and Latin America in general, the terms "transvestite" and "transsexual woman" are used more frequently by the communities themselves than "transgender woman".These differences can mark political and/or subjective identities and are fluid depending on the context 18 .The terms convey different levels of performances as a woman and demand their identity's legitimacy beyond binary male-female parameters, adequacy of their physical image and bodies via hormone therapy, use of silicone, and other body modifications, and the fact that they wish to be addressed in the feminine and by the name with which they identify.Importantly, there is transit between identities, which are not fixed or isolated categories, but are always in dispute, negotiation, and constant interaction and movement 19,20,21 .This study used the term "transgender women", since most of the literature consulted in Cad.Saúde Pública 2019; 35(J):e00112718 the review was in English, and it is an umbrella term for a wide range of transfeminine identities that blur the sex-gender borders, although the term "transvestite" was also included in the search strategy.
The publications were managed in the Mendeley app (https://www.mendeley.com)to remove duplicates.Data collection lasted from October 2016 to February 2017 and was updated in June 2018.No publication period was determined in advance for the review.

Eligibility criteria
Inclusion criterion: studies that addressed the relationship between stigma and discrimination due to gender identity and vulnerability of transgender women to HIV/AIDS.There was no exclusion of any methodological approach; both qualitative and quantitative articles were included.The review included articles written in English, Portuguese, and Spanish.No articles were excluded on the basis of geographic location or time frame or for the term used to define transgender women (transvestite, transsexual woman, aravanis, hijras, metis, etc.).

Data extraction
Study selection began by reading the titles and abstracts, based on the inclusion criteria.The full texts of the selected articles were read.After the assessment, the studies were selected for inclusion in the review's corpus.An Excel (https://products.office.com)spreadsheet was organized with the following terms: authors, year of publication, study country, study design/methodology, number of persons in study sample, objectives, study population, and main results.

Assessment of risk of bias (quantitative studies) and methodological rigor (qualitative studies)
Next, the methodological quality was assessed according to the study's nature.Qualitative studies were assessed with the Research Triangle Institute Item Bank (RTI-Item Bank) scale, which evaluates risk of bias 22 .RTI-Item Bank contains 29 items to assess studies, 6 of which were applied to the studies included in this review (Supplementary Material, Box S1: http://cadernos.ensp.fiocruz.br/site/public_site/arquivo/suppl-e00112718ingles_2106.pdf):(i) inclusion and exclusion criteria clearly defined; (ii) use of valid and reliable measures to assess inclusion and exclusion criteria; (iii) standardized recruitment strategy for participants in all the groups; (iv) appropriate sample selection; (v) results assessed using valid and reliable measures, implemented consistently for all the study participants; (vi) confounders and effect modifiers considered in the design and/or data analysis 22 .Risk of bias was assessed and classified using the studies' response to the above-mentioned items and classified as follows: high risk of bias -when the study had one or more negative answers to the items; moderate risk of bias -when one or more items were classified as "partially" or "indeterminate"; low risk of bias -when all the items in the scale recorded a positive answer 22 .
Assessment of qualitative studies used the instrument proposed by the Critical Appraisal Skills Programme (CASP), employed in the critical analysis of reports by qualitative studies.This instrument has ten questions that help the assessor think systematically on the study's rigor, credibility, and relevance, considering: (i) clear and justified objective; (ii) appropriate methodological design for the objectives; (iii) methodological procedures presented and discussed; (iv) sample selection; (v) data collection described, instruments and saturation process explained; (vi) explanation of the relationship between researcher and study subject; (vii) ethical precautions; (viii) dense and well-founded analysis; (ix) results presented and discussed, featuring the issue of credibility and use of triangulation; (x) description of the contributions and implications for the knowledge generated by the study, as well as its limitations 23 .Qualitative studies were classified in two categories: A, for studies with high methodological rigor, since they met at least 9 of the 10 items; B, for studies with moderate methodological rigor, meeting at least 5 of the 10 items 23,24 .

Data analysis
The analysis was oriented according to the theoretical references for the concepts of stigma, discrimination, and vulnerability.The study adopted the concept of vulnerability applied to the field of health, specifically to the discussion on the HIV/AIDS epidemic.This concept can be understood by the analysis of three interrelated components: individual vulnerability, aimed at identifying physical, mental, or behavioral factors through risk assessment and/or other approaches; social vulnerability, analyzing the dimensions of culture, religion, morals, politics, economy, and institutional factors, which can determine the means of exposure to diseases and/or injuries; programmatic vulnerability, examining how policies, programs, and services affect persons' social and individual situations 25,26,27,28 .Vulnerability emphasizes the responsibility of government actions and public policies as an integral part of the determinants of the health/disease process 25,26 .In this article, the theoreticalconceptual understanding of this construct expanded the scope of analysis of the articles beyond the behavioral and individual risk issues, including studies that related stigma and discrimination to barriers in accessing health services.
Stigma refers to a person's profoundly depreciative attribute, which is perceived as such through social interaction.The presence of this attribute may confirm or reaffirm the "normality" of specific persons or groups.Stigma highlights a specific trait in individuals, subjecting them to the impossibility of social attention to their other attributes, assigning major discredit to them 29 .Hatzenbuehler & Link 30 recently emphasized the need for progress in the conceptualization and measurement of stigma as a social phenomenon with roots in social structures.The authors define structural stigma as conditions at the broader social and cultural levels and institutional policy norms that construct the opportunities, resources, and well-being of stigmatized individuals.The authors call attention to the intense interaction between the microsocial level, the locus of interpersonal relations, and the macrostructural level.Such structures are not unidirectional and static, but shaped by interpersonal relations and individual factors.
Discrimination can be understood as a practical result of stigma, defined by a conceptual review 31 (p.34): stigma is a profound attribute of discredit, a "mark" or "socially devalued identity"; stigmatization is related to a social process that produces devaluation through labels and stereotypes; a label is an officially sanctioned term applied to conditions, individuals, groups, places, organizations, institutions, or other social entities, since the stereotype is related to negative attitudes and beliefs targeted to the labeled social entities; prejudice is an endorsement of negative beliefs and attitudes related to the stereotype; and discrimination involves the actions targeted to the endorsement and reinforcement of stereotypes to place the labeled persons at a disadvantage.In this article, we thus consider studies on discrimination and stigma related to the gender identity of transgender women.Since there is no consensus in the literature on this issue 32 , we will use "stigma and discrimination" widely speaking throughout the article, but understanding that there are important theoretical and conceptual specificities 33 .
In this analysis, we investigated the methodological issues of the studies analyzed here and established key elements that constituted thematic units 34 .This process identified 65 key elements based on a reading of the articles, which were categorized on an Excel spreadsheet based on the three thematic units in the concept of stigma according to Hatzenbuehler & Link 30 and White-Hughto et al. 14 : individual level (psychological issues such as self-stigma), interpersonal level (person-to-person discrimination), and structural level (state policies that can promote social exclusion).

Characteristics of selected studies
We identified 791 articles in the databases, of which 41 were included in the review.Figure 1 shows the search strategies.The reasons for exclusion of articles were the absence of analysis on stigma, discrimination, vulnerability, and HIV (Supplementary Material, Box S2: http://cadernos.ensp.fiocruz.br/site/public_site/arquivo/suppl-e00112718ingles_2106.pdf).

Measurement of discrimination and stigma in the quantitative studies
To identify how the studies dealt with the construction of the stigma or discrimination variable, we analyzed 12 exclusively quantitative articles and two with mixed methods.Eight studies dealt with the phenomenon as "discrimination" (experience, perception, etc.) 21,35,36,37,38,39,40,41 , three articles analyzed "stigma" (experience, perception, etc.) 6,7,9 , one dealt with the phenomenon of "homophobia" 42 , one with "transphobia" 8 , and one of the mixed-methods articles did not use the quantitative method to assess discrimination and stigma 43 .Many of these studies did not provide a theoretical framework on the distinction between the concepts of stigma and discrimination.
Most of the studies (54%) showed high risk of bias 8,35,37,39,40,41 , and only 31% were classified as low risk of bias 6,7,9,21 .Inadequate sample selection and assessment of the study outcome with valid criteria were the items that most contributed to bias scores in the studies analyzed here.In one study it was not possible to apply the scale of bias, since it did not present quantitative methodological elements for the assessment 43 (Figure 2) (Supplementary Material, Table S2: http://cadernos.ensp.fiocruz.br/site/public_site/arquivo/suppl-e00112718ingles_2106.pdf).
The variables related to stigma and discrimination were built on the basis of an unvalidated scale for the population of transgender women, some inspired by previous scales on racial discrimination 40 , perception of stigma in MSM 6 , and homophobia 8,9 , while others were created on the basis of previous studies with this population, or drawing on a review of the literature 7,21,35,36,41 .A few studies used just one or two questions on perceived discrimination 37,39 and did not provide details 42 .Among the studies that used items to assess discrimination or stigma, the majority used Cronbach's alpha to estimate the questionnaire's reliability 6,8,9,40 , one used the Kuder-Richardson coefficient 41 , one used confirmatory factor analysis 9 , one used exploratory factor analysis 7 , and another employed latent class analysis 21 .Some did not perform any of these analyses 35,36,38 .
Methodological rigor according to the CASP criteria was classified as B (moderate rigor) in four studies 52,54,65,66 .Non-rigorous data analysis, research ethics procedures not specified in the methodology, and lack of specification of interaction between researchers and participants in the field were the items that scored negatively and contributed to the moderate methodological rigor (Table 1).

Stigma, discrimination, and vulnerability to HIV
According to the review, stigma produces discrimination and violence at different levels: structural, interpersonal, and individual, which can play a role in the individual, social, and programmatic vulnerability of transgender women to HIV (Figure 3).

Structural stigma
Structural stigma promotes a totally adverse social context for transgender women through transphobia and discrimination 5,42,44,55,56,63,64 .In some countries, especially those with a strong religious tradition, transsexuality, and homosexuality are still legally criminalized, as exemplified in two studies, one in Malaysia 46 and the other in India 60 .In India, section 377 of the Indian Penal Code, known Cad.Saúde Pública 2019; 35(4):e00112718

Figure 1
Flowchart for the systematic article selection process.
as the "Sodomy Law", which criminalizes persons who have sex with non-vaginal penetration, was reinstated by the Supreme Court in 2013, but repealed in September 2018 60,67 .In India, marriage and procreation, considered key criteria for achieving respect and heterosexual normativity, appear to justify the stigma and violence against groups that do not conform to the hegemonic gender identities 48 .In Lebanon, incarceration on grounds of gender identity or expression has also been reported 38 .
Even in liberal countries (from the legal point of view) such as the United States 41,44,53,54 , Mexico 51 , Japan 68 , and Brazil 42 , transgender women still suffer discrimination in public spaces and experience difficulty in reassigning their name in keeping with their gender identity 12,54,63 .
Family and social stigma was found associated with sex work 7 .It was also reported as an important barrier to access barriers to schooling 43,51,56 and formal employment 5,12,43,46,56,63 , which often leaves them in a situation of socioeconomic marginalization 36,61,69 and entry into the sex work market 43,46,56 .

Table 1
Characteristics of knowledge production in qualitative studies on the relationship between stigma, discrimination, and vulnerability of transgender women to HIV/AIDS, 2004-2018.As for access to health services, various studies have documented that stigma and discrimination can pose serious barriers for transgender women 12,39,43,46,48,53,54,57,58,59,62,63,64,66,69,70 .Many avoid going to health services because they anticipate discrimination 59,65 and others are denied access even in public services 46,52 .Studies that analyze the use of the public health system in some countries indicate that transgender women prefer to avoid this care and pay for private services or self-medicate, due to the stigma 43,48,49 .Lack of access to hormones 12,46,54 and surgical procedures for body modification and gender reassignment 49 has also been identified in the literature as a barrier to a healthy life.

Reference (year)
Stigma and discrimination also pose barriers to access to HIV/AIDS prevention and treatment services, such that many transgender women avoid public healthcare services due to previous experiences of discrimination and mistreatment 46 .From this perspective, many studies report the difficulties of transgender women in access to HIV testing and counseling services 46,59 , lack of access to information on prevention 58,63 , lack of confidentiality of HIV test results in public healthcare services 12,46 , and limited access to condoms 56 .In Brazil, self-perceived discrimination was associated with resistance to HIV testing 37,61 .Even those who have already tested for HIV faced more stigma when accessing HIV testing and care services, when compared to those who had never been tested.Stigmatization can also hinder retention of transgender women in HIV treatment services 9,69 .

Table 2
Characteristics of knowledge production in quantitative studies on the relationship between stigma, discrimination, and vulnerability of transgender women to HIV/AIDS, 2005-2018.Social exclusion due to stigma can cause intense geographic displacement 54 and entry into sex work 5,48,50,51,53,54,56,57,58,62,63 .Sex work in precarious conditions and receiving more money for unprotected sex have been reported in the literature as one of the reasons for unprotected anal sex 56 .

Individual stigma
The combination of interpersonal and structural stigma can cause various negative outcomes in the lives of transgender women, for example, social isolation 48,65 and fear of discrimination 44,52,53,57,59,62,64,66 .The expectation of rejection related to gender was associated with sexual risk behaviors for HIV infection 35 .
Experiences of discrimination are reported as important elements in the internalization of stigma, which can cause a range of psychosocial stress 48 , such as low self-esteem 48,56,68 , and compromise mental health with the occurrence of depression 6,41,46,58,68 , suicidal ideation 48,56 , and attempted suicide 41,46,56 .

Figure 2
Stigma and discrimination based on gender identity and individual, social, and practical vulnerability of transgender women to HIV.

Figure 3
Summary of risk of bias in the selected quantitative studies.
Cad. Saúde Pública 2019; 35(4):e00112718 Alcohol use 9,41,57,58,62,68 and use of other drugs 5,9,41,58,62 are reported in contexts in which transgender women experience high levels of discrimination, besides the use of these substances before sexual relations 9 as a practice that increases the risk of HIV infection, mainly through unprotected anal sex 6,40,46 .
Stigma and discrimination are identified as factors that can directly influence vulnerability to HIV.A study showed that stigma related to transgender identity was more prevalent in transgender women living with HIV than in those without the infection 9 .The relationship between stigma, discrimination, and HIV infection can be explained by transgender women's low capacity to negotiate condom use, resulting in unprotected anal sex 5,6,12,21,48 .In addition, low self-esteem and depression, caused by intense stigmatization of transgender identities, have been reported as important factors for unprotected sex 56 .
Some studies suggest that unprotected anal sex is practiced for validation of female status vis-à-vis the male partner 5,46 , especially with steady partners such as boyfriends or husbands 21 .A qualitative study in Colombia showed that although transgender women say they use condoms in all their relations, unprotected sex means fulfillment and success in the eyes of their stable partners or husbands.In this context, the risk is even greater in stable relationships due to the "active" sexual role (insertive anal sex) played by the partner, often idealized by some transgender women in that country 49 .

Discussion
Analysis of the articles highlighted that stigma due to gender identity, and discrimination, violence, and transphobia, have been identified as structuring elements of the vulnerability to HIV/AIDS among transgender women.Stigma and discrimination were observed wherever the studies were performed, in low, middle, and high-income countries.Nevertheless, some studies documented forms of resistance by transgender women through social activism, participation in support groups, and resilience 48,56,70 .
Research on stigma and discrimination has grown exponentially in the last decade, encompassing various areas and becoming increasingly specific and complex 31 .In relation to the studies' methodologies, we found that the majority took qualitative approaches.A plausible hypothesis for this fact is the complexity of operationalizing the concept of stigma in quantitative studies due to the diversity of definitions for stigma.The quantitative studies reviewed here attempted to solve this problem by using scores for variables related to discrimination based on gender identity (at work, in health services, difficulty in obtaining housing) 8,36 , by factor analysis 7 , or by latent class analysis through the inclusion of specific variables of discrimination (in the family, with friends, with neighbors, at health services, verbal aggression), by adaptation of scales for measuring homophobia 9 , or directly by the self-perception of discrimination 37 .
Quantitative studies were marked by emphasis on the relationship between experiences of stigmatization and risk of HIV infection.It is important to recall that the initial interpretations of the AIDS epidemic (1981-1984) were marked primarily by a biomedical, epidemiological, and behaviorist focus 26 , leading to the identification and stigmatization of population subgroups with the highest likelihood of including persons with the disease when compared to the general population 71 .However, the epidemiological studies reviewed here appear to go beyond a merely behavioral relationship.By reflecting on the concept of stigma, they challenge structural and relational issues that affect analytical dimensions of the concept of vulnerability, producing a shift from exclusively individual issues such as behaviors, attitudes, and risk practices to attention to social factors 26 .
Qualitative studies of a sociocultural nature featured significant contributions to the analysis of stigma and vulnerability to HIV, since they were not limited to the dimension of individual behaviors, but expanded the analytical window to include issues related to labeling, distinction, and exclusion, which sustain stigma as a profoundly depreciative attribute.Based on an analysis of narratives and daily social relations, these studies were able to relate the process of stigmatization to transgender women's social and programmatic vulnerability to HIV.
According to Link & Phelan 72 , stigma exists when a set of interrelated components converge.The first refers to the fact that persons distinguish and label human differences through a substantial Cad.Saúde Pública 2019; 35(4):e00112718 simplification of differences, as if there were no gradation between the various categories.In this sense, dualism between the categories usually prevails: cis/trans, gay/straight, black/white, etc.An important characteristic of this component is that prominent attributes differ drastically according to time and place.The second component involves the association of human differences -which are labeled -with negative characteristics; the connection between these two properties shapes what the authors call stereotype.The third component of stigma occurs when the social labels promote the separation between two categories of persons: "us" and "them".
We thus observe that stigmatization of transgender women produces discrimination, which materializes as social exclusion and various forms of violence.The effects of stigma may take the form of psychiatric outcomes (e.g., suicidal ideation and depression) and substance abuse.Social exclusion may also be related to low schooling and barriers to access to the work market, which in turn can influence entry into the sex trade and the adoption of risky behaviors such as the use of injecting drugs without medical orientation and unprotected anal sex with steady or casual sex partners or clients.
We also found that at the individual level, transgender women face major social isolation, exacerbated by fear of rejection and discomfort or insecurity in public places, producing high rates of depression and suicide, as observed in various studies 4,73,74 .Substance abuse is also closely related to risk behaviors for HIV infection 1,75 .A study in New York produced strong evidence that genderbased discrimination against young transgender women increased the risk of depression and sexual risk behaviors, which in turn increased the likelihood of HIV infection and other sexually transmissible infections 74 .
At the structural level, the studies show that stigma, through discrimination, can affect access by transgender women to health services, including HIV/AIDS testing and treatment services, which is corroborated by other studies that do not focus specifically on the relationship between HIV and stigma 76,77 .A study in Argentina found that 40.7% of transgender women reported avoiding the use of health services because of their gender identity.The study observed that factors related to the stigmatization process were associated with this phenomenon, for example, the report of having experienced discrimination in health services by health professionals or other patients, or having suffered police brutality 78 .
The diverse ways of measuring stigma and discrimination in the quantitative studies may hinder the production of future metanalyses on the impact of stigma on the risk of HIV infection.Another important issue is the diversity of uses of the concept of stigma and discrimination in this field of studies.We thus suggest constructing, standardizing, and validating scales to measure the different facets of stigma (individual, interpersonal, and structural) and discrimination (as the action or effect of stigma) in quantitative studies.We found that qualitative studies were the best methodology for analyses intended to address the relationship between the categories of stigma, discrimination, and vulnerability to HIV.Quantitative studies should also consider the sampling processes, since the choice of non-probabilistic procedures is one of the elements responsible for the high risk of bias in the studies analyzed here.We thus suggest that in future studies on the theme, the sample size and selection of participants should be adequate for comparison of the groups and to control confounding.
In the qualitative studies analyzed here, the depth and analytical rigor were procedures that displayed limitations.In qualitative studies, we suggest greater analytical depth and the adoption of different methods for understanding stigma and vulnerability, such as triangulation of methods.
This review study has some limitations.The first is the lack of a metanalysis with the data from the quantitative studies, considering the heterogeneity of the variables they used.There was also difficulty in synthesizing the results of studies with different methodological approaches, since most guidelines for systematic reviews do not consider the integration of qualitative and quantitative studies in the same review.In addition, the current review did not include all of the grey literature from a relevant body of scientific output published online by international organizations, outside the scope of peer-reviewed scientific journals.These limitations notwithstanding, we adopted consistent methodological procedures performed by independent reviewers and assessed the studies that met the eligibility criteria in order to reduce the possibility of bias.
In this review study, we found that stigma and discrimination are related in various ways to individual, social, and programmatic vulnerability to HIV/AIDS.It is necessary to understand how stigma and discrimination operate in society to produce and reproduce social and health iniquities.Understanding the history of stigma and its consequences for individuals and communities, such as discrimination, can help us develop better measures to fight it or reduce its effects 79 .We thus suggest that health measures and HIV prevention should not be limited to behavioral aspects and risk practices, but should embrace the promotion of a culture of non-discrimination and respect for gender differences.

Contributors
L. Magno participated in the article's conception, systematic literature review, analysis, and writing and final revision.L. A. V. Silva participated in the article's conception and writing and final revision.M. A. Veras participated in the critical revision and approval of the final version.M. Pereira-Santos participated in the data collection and final revision.I. Dourado participated in the study's conception and writing and final revision.

Table 2 (continued)
MSM: men who have sex with men; NA: not applicable; STI: sexually transmited infections.