Introduction

Despite significant advancements in HIV prevention, early detection, and treatment, 37.7 million people are living with HIV around the world, and there were an estimated 1.5 million new infections in 2020 [1]. The UNAIDS global targets for 2025 focus on primary prevention of HIV as well as ensuring that 95% of people living with HIV become aware of their status; 95% of people diagnosed with HIV receive Antiretroviral Therapy (ART); and 95% of people living with HIV on ART achieve sustained viral suppression [2]. Collectively, the goal is to end new HIV infections by 2030. However, the stated goal for 2020 was to reduce new infections to 500,000, which was not achieved due in part to limited progress in reducing stigma affecting people at risk for and living with HIV [2]. Stigma has been identified as a social determinant of health and a key barrier to improving health outcomes among people living with HIV [3–8]. As such, stigma continues to present barriers to achieving the HIV prevention and treatment targets by interfering with the provision and uptake of prevention, diagnostic, and treatment services. This is particularly true for key populations (sex workers; gay, bisexual, and other men who have sex with men; people who use drugs; and transgender persons) who may experience stigma relating to actual or assumed HIV status in addition to experiencing intersecting stigma related to their actual or assumed behaviors or identities [2]. Stigma’s negative impact on the health and quality of life of people living with HIV and key populations is also well documented [3,9–14].

Stigma is a social process in which an individual or group is linked to a negative stereotype or misconception, often resulting in adverse experiences, loss of social status, and limited opportunities [15,16]. Stigma may occur at the individual, interpersonal, community, and structural levels and can be experienced as anticipated, perceived, internalized, or enacted [17]. Anticipated stigmas refer to the expectation of bias perpetrated by others [17–20]. Perceived stigmas refer to felt stigma and the perception of bias as understood by a person living with a stigmatized identity [21]. Internalized, or self-stigma, is the adoption of negative feelings or devaluing of oneself on account of a stigmatized identity [17,22]. Enacted, or experienced stigma, is the perpetration of mistreatment or discriminatory acts by someone on the basis of a stigmatized identity [23]. Discrimination, part of the social process of stigma [15], includes any distinction, exclusion, or restriction made to human rights and fundamental freedoms, in the political, economic, social, cultural, civil, or any other field [24]. Discrimination can be institutionalized through existing laws, policies, or practices that negatively impact people living with HIV or key populations. Although the conceptualization of discrimination in relation to stigma varies, here, we consider enacted stigma to be inclusive of discrimination [17]. Drivers of HIV and key population stigmas can include individual-level factors such as lack of awareness or education (i.e., misinformation about HIV risk and transmission); however, societal level policies, cultural norms, and religious values can also act as drivers or facilitators of stigma [23,25–27].

In recognition of the central role stigma has in impeding HIV epidemic transition, UNAIDS has established a vision to achieve 3 zeros by 2030: zero new HIV infections, zero AIDS-related deaths, and zero discrimination [2,9]. To support this vision, UNAIDS has also established the 10–10–10 goals that focus on removing societal and legal barriers to HIV services, including reductions in punitive laws and policies, experiences of stigma and discrimination, and experiences of gender inequality and violence by 2025 [28]. Stigma related to HIV and key population statuses have also been recognized as a key factor impeding progress toward the 2030 Agenda for Sustainable Development Goal (SDG 3.3) of achieving the end of the HIV pandemic [24,29]. Despite this consensus on the importance of addressing stigma, measurement of stigma experienced by people living with HIV and key populations continues to be a challenge for public health practitioners and policymakers [18,30]. Stigma measurement has unique challenges given that it is a latent construct, a social process that acts across multiple levels, and a phenomenon that has often been thought of as too complex to measure [31–33]. However, progress has been made over the last decade in developing standardized measures to quantify both HIV stigma and key population stigma [18,34–38]. Specifically, advancements have been made through the Global AIDS Monitoring indicators and WHO Strategic Information guidelines [39,40]. Despite the existence of many validated measures of different forms of stigma related to HIV and key population membership, there is currently no established methodology for bringing these together into more concise but validated summary measures at the country level. A summary measure to quantify progress toward achieving zero discrimination may allow for easy comparisons within and across countries and possibly improved accountability for progress.

The ability to characterize stigma across countries was recommended as a strategy to achieve the end of the HIV pandemic during the 2017 UNAIDS Science Panel meeting on ending the AIDS epidemic by 2030 [41]. In response, the UNAIDS Monitoring Technical Advisory Group (MTAG) created a task team in 2018 comprised of technical experts on HIV and key population stigma from civil society and academia from around the globe. This task team’s mission was to provide a set of consensus recommendations to UNAIDS on the establishment of a country-level summary measure of stigma faced by people living with HIV and key populations at high risk of HIV and the legal and policy environment for the protection of these individuals’ fundamental rights, including their ability to access health and HIV services.

The creation of country-level indices that characterize the degree of stigma existing within a country may facilitate understanding and interpreting the overall burden of stigma and its effects. For the purposes of this work, we consider an indicator to be a single item measuring an aspect of stigma, and an index is a composite of indicators to generate a summary or average value representing a broader stigma construct. Summary index measures may also be used by global- and national-level policymakers, advocates, and program implementers engaged in the AIDS response to track progress on eliminating HIV and key population stigma, potentially increasing the likelihood that this goal can and will be achieved. The question remains, however, whether sufficient data exist to inform country-level indices that characterize levels of stigma for people living with HIV and key populations. This study aims to identify and describe the availability and quality of stigma data across countries through a global data mapping exercise.

Methods

The methods used in this exercise, including preparatory activities for data mapping and the mapping of available data for country-level summary measures of HIV and key population stigma, are outlined in Fig 1. The protocol for this exercise is included as S1 Protocol.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flowchart of methodological approach for data preparation and data mapping.

DHS, Demographic and Health Surveys; IBBS, Integrated Biological and Behavioral Surveys; ILGA, International Lesbian, Gay, Bisexual, Trans and Intersex Association; MICS, Multiple Indicator Cluster Surveys.

https://doi.org/10.1371/journal.pmed.1003914.g001

Results

The summary results of indicator identification, extraction, selection, and assessment for HIV stigma indicators and the key population stigma indicators are presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of stigma indicator mapping identification, extraction, selection, and assessment.

https://doi.org/10.1371/journal.pmed.1003914.t001

Final potential indicators

This mapping exercise identified 24 potential HIV stigma indicators and 10 key population indicators with potential for use in characterizing stigma and creating valid stigma summary measures. The key population indicators include 2 for sexual orientation/behavior stigma related to men who have sex with men; 2 for sex work-related stigma; 2 for drug use stigma; and 4 for gender identity stigma related to transgender people. These are described in detail in Tables 2–6.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Final indicators for HIV stigma.

https://doi.org/10.1371/journal.pmed.1003914.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Final indicators for sexual behavior/orientation stigma related to men who have sex with men.

https://doi.org/10.1371/journal.pmed.1003914.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Final indicators for sex work stigma.

https://doi.org/10.1371/journal.pmed.1003914.t004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Final indicators for drug use stigma.

https://doi.org/10.1371/journal.pmed.1003914.t005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Final indicators for gender identity stigma related to transgender persons.

https://doi.org/10.1371/journal.pmed.1003914.t006

Geographic representation

We found 61 of 193 UN member states with any available data (32%) to have data for a sufficient number of HIV stigma indicators (missing data for less than 40% of indicators). Considering all key population stigma indicators together, 119 of 193 countries (62%) had data on a sufficient number of indicators. Countries that meet the criteria of having sufficient data across the 24 HIV indicators are displayed in Fig 2 and for the 10 key population indicators in Fig 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Countries with sufficient available data on HIV stigma indicators.

Base layer of the map used are from ArcGIS, ESRI: https://arcg.is/8DHLK. *Data were available for more than 40% of indicators.

https://doi.org/10.1371/journal.pmed.1003914.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

Countries with sufficient available data for key population stigma indicators. (A) Sexual behavior/orientation stigma related to men who have sex with men. (B) Sex work stigma. (C) Gender identity stigma related to transgender persons. (D) Drug use stigma. *Data were available for more than 40% of indicators. Base layers of the maps used are from ArcGIS, ESRI: Fig 3A: https://arcg.is/bu8Cu0. Fig 3B: https://arcg.is/Pbzn00. Fig 3C: https://arcg.is/0G0PDG. Fig 3D: https://arcg.is/1HS1zT.

https://doi.org/10.1371/journal.pmed.1003914.g003

Discussion

In 2022, data exist for comprehensively characterizing the nature and level of different forms of stigma affecting people living with HIV and key populations. However, the data mapping exercise described here highlighted a lack of stigma indicators that have ongoing, consistent data collection for assessing change over time, and thus the ability to track progress in mitigating stigma at the country level. In particular, data for indicators outside of the structural stigma domain are much less consistently and rigorously collected and made accessible, particularly for stigma experienced by key populations. There are a few indicators at present that have existing data and are incorporated within specific frameworks for routine data collection mechanisms. Given geographic representation, repeated implementation, and quality, these indicators provide some ability for beginning to explore if and how complex experiences of stigma may be able to be summarized to elucidate the state of stigma within a country and assess change going forward. Despite the current limitations in data availability and quality, participatory approaches involving experts from communities, civil society, academia, and public health practices can be used to help inform the creation of summary measures (i.e., via use of participatory weighting) and to inform data collection priorities moving forward (i.e., via participatory ranking of indicator importance).

Unfortunately, several domains, subdomains, and indicators of stigma were not identified as having sufficient data to use in a summary measure at present despite their importance. No publicly available data sources for appropriately assessing indicators within the domains of social norms and attitudes, or internalized stigma for key populations were identified, and, therefore, these domains had to be removed from key population stigma summary measures at present. Given that social norms and attitudes have been well documented to play a large role in the experiences of stigma among key populations and to be associated with health outcomes, data collection for indicators within this domain should be prioritized in order to include this domain when characterizing stigma at a country level. Indicators within the violence, anticipated stigma, experienced stigma, and stigma and discrimination domains were not retained for consideration as a summary measure for key populations due to limited availability and quality. Therefore, the potential indicators for key population–related stigma only included those from the structural stigma domain. This highlights the importance of expanding data collection for these domains to improve both quality and availability of data that can be used to more comprehensively characterize stigma as it exists in different forms for key populations. Guidance on biobehavioral surveys have been recently developed and can be used by countries to improve consistent data collection of validated measures in the future [49].

Often, only data aggregated at the sample or even country level were able to be located for indicators, which presents a challenge in accounting for differences in sampling strategies when comparing across countries. For instance, individual-level indicators assessing attitudes or experiences of discrimination are often reported at country level, although data may not have been drawn from a nationally representative sample. Statistical methods can be used to better account for sampling approaches to address these concerns of representativeness; however, this is not possible without access to individual-level data, and neither is it possible to recombine or disaggregate samples (such as by age or gender). While there may be consistency in how individual-level surveys are collected, without systems for sharing deidentified data from such surveys, it is not possible to use these data to inform broader comparative analyses aimed at assessing country progress. This limits availability for indicator calculation. The creation of a broader data repository for individual-level deidentified data provides an opportunity to address both of these concerns and keep data for indicators easily accessible for monitoring and analysis. Countries could then better understand the data available and use weighting strategies appropriately to characterize stigma even in the context of limited data. When stigma is measured in surveys, especially when stigmas affecting key populations are assessed, validated measures or scales are often not used, and there is great variety in the measures used across studies [18]. Therefore, an essential step in improving usability of data shared in a repository is increasing the consistency and standardization of stigma measures across countries to facilitate improved comparability of secular trends and health outcomes of stigma.

Stigma mitigation can emerge from both interventions within and outside of the health sector [23]. Measuring secular trends with standardized measures across and within settings can provide insight into the potential impact of interventions aimed at addressing human rights barriers and complementary investments aiming to address stigma [30]. A quantitative measure of stigma, which links directly to diminished uptake of HIV prevention or treatment services, offers an additional indicator against which to measure progress in the AIDS response. HIV (mathematical) modelers have sought such data on barriers to the AIDS response. For example, currently, we do not know if the people who do not know their HIV status do not know their risk or are avoiding testing due to stigma and discrimination. Several indicators identified in this mapping exercise had been revised in recent years, possibly in an effort to improve reporting and measurement. However, the change in definitions over time limits the comparability of indicators over time and complicates the ability to observe change. Among indicators which do currently have sufficient data to compare time periods, most indicators only have data available for a few time points, and many indicators, particularly at the structural level, necessitate longer intervals and follow-up periods to demonstrate meaningful change. Many of the data obtained in this mapping exercise on laws and policies are not available going back more than a few years, although if continued to be prioritized for collection could be used to track changes in stigma into the future.

Summary measures will be meaningful if they provide an accurate characterization of the overall environment of stigma within countries. Combining the indicators and domains identified in this mapping exercise into fewer items as an index may allow for the creation of summary measures. In order for these summary measures to be concrete and valid, indicators will need to be appropriately aggregated and weighted [50]. Based on the findings of this mapping exercise, there are 2 concurrent approaches that would be appropriate for determining indicator weights. The first is empiric, using methods such as exploratory factor analysis, to assess the underlying factor structure of a given set of variables and the strength of the association between a variable and a latent construct when sufficient data exist [51]. As there are insufficient data for conducting this process for key population stigma currently, we also propose a second participatory strategy to determine relative importance and weight indicators. Analytic hierarchy process, a multicriteria decision analysis method that seeks to solve a problem through a hierarchical approach [52], is one example of such a participatory approach. A participatory approach will allow for experts from communities, civil society, academia, and public health practices to fill in gaps where data are currently insufficient. Therefore, despite the limitations in the current data available for indicators identified through this mapping exercise, empiric and participatory approaches can be used alongside the data that do exist to inform the creation of summary measures and set data collection priorities moving forward.

There is heterogeneity in the experiences and burden of stigma within and across populations [4,53]. An essential consideration, true in the creation of any summary measure for a complex construct, is that heterogeneity within countries will be masked in the process of distilling multiple indicators into one or even a few indices. This data mapping exercise demonstrated significantly more data characterizing HIV stigmas than those focused on sexual behavior or orientation, gender identity, sex work, or drug use—challenging our ability to describe heterogeneity within the key population subgroups, such as age, residence, employment, etc. Discussions have emerged in this process to consider the development of a single summary measure of stigma for all key populations in order to clearly guide targeted HIV and human rights policy, advocacy, and practice. Not only may this mask diverse experiences of stigma across these populations, but it may also mask substantial differences in the relative state of stigma and progress toward its elimination. For instance, we found vast differences in the level of criminalization related to different behaviors associated with different key populations. However, creating separate indices may artificially segment experiences that are intersectional for individuals with HIV and membership of a key population. Intersectional stigma is the potentially compounded effect of stigmas among individuals with multiple identities which may be devalued by some in society [54]. Unfortunately, the development of measures for intersectional stigma is still in their nascency and would not be represented in this index [27,55]. Additionally, the emerging field of microaggressions is an important area of ongoing research that aims to characterize and understand stigma processes, and, although there is not yet consensus on how this fits within the stigma framework, there is likely great overlap with individual-level experiences of stigmas. However, this is not represented in the current stigma indicators that were considered and whether and how to characterize such experiences at a broader socioecological level to understand stigma at a country level is an important area of future research. Given that summary measures may not be nuanced enough to appropriately represent intersectionality when combined, in balancing all considerations, we recommend the creation of separate indices for HIV and each key population stigmas with ongoing consideration of how to examine intersectional influences on health. The addition of narratives from advocates and people with lived experience will be essential to consider in conjunction with any summary measure toward this purpose.

The process for creating summary measures to accurately describe and track country-level stigma, while also accounting for heterogeneity and complexity is a challenge, yet not an insurmountable one. As described in conceptual frameworks depicting the processes of stigma, these relationships are complex and often do not occur in isolation [17]. In terms of understanding stigma, there is discussion about the appropriate point of measurement—whether it be the social determinants of stigmas (reflected, for example, in indicators such as acceptability of partner violence, demonstration of controlling behaviors, or having experienced intimate partner violence) or the experiences of stigmas as an outcome or endpoint (reflected, for example, in indicators such as having been socially excluded for living with HIV, having experienced HIV discrimination in healthcare, or having lost a source of income due to living with HIV). Continuing this discussion on the process and methods for stigma measurement at a country level will be an essential step for improving our ability to characterize stigma among and between populations and track country progress in addressing stigma. This should be done in conjunction with, rather than at the expense of, continuing work to measure and assess stigma at other socioecological levels, including the interpersonal and individual levels. Ultimately, measures of stigma at the country level can be used in combination with for instance, individual-level stigma measures, in comparative research to help elucidate the complex interaction between structural stigma and other stigma manifestations.

In addition to the challenges outlined in this discussion, there are several additional limitations to this data mapping exercise that should be considered. The objective of this study was to understand the stigma data that exists for calculating several stigma indicators that have potential use for informing a standardized stigma measure at the country level. As a data mapping exercise, this process was exploratory and therefore may be subject to bias. Although a priori criteria were established and are reported, final decisions on inclusion were based on the state of the data observed during this process. The research and authorship team includes a diverse group of stigma experts from across different academic institutions, multilateral agencies, community networks, and international organizations. This collaborative approach was taken to expand the access to existing stigma data and to minimize selection bias. Our research team used publicly available data, supplemented by data obtained through collaborators, and directly through the research team where possible. This study was therefore unable to assess data that were not publicly available or databases that were not yet available at the time of this data mapping exercise, but could have potentially been shared by institutions outside of this collaborative research team [56,57]. Although systematic reviews of the literature may have informed which data had ever been collected, it would not have yielded the raw data needed to conduct this exercise (i.e., the data would not necessarily be accessible).

Conclusions

Several indicators were determined to have the potential to appropriately characterize the level and nature of stigma affecting people living with HIV and key populations across countries and time based on the data available. Although few available indicators have sufficient data available to use in summary measures for key populations, empiric and participatory approaches for weighting can be used to fill gaps in the current data and inform more comprehensive summary measures. While many countries lack data for the indicators identified in this data mapping exercise and therefore would be underrepresented in any global characterizations, highlighting these gaps can support the direction of funding opportunities, government endorsement, and supportive technical assistance to certain countries/areas to improve the representativeness of measurement of stigma mitigation progress in the future. The creation of a global stigma data repository would also serve to improve the availability and use of stigma measures, and efforts for increased data collection of validated stigma metrics can improve our ability to characterize country-level stigma across various domains in the future. Using the indicators identified as having sufficient data at present for the creation of country-level indices for HIV stigma, sexual minority stigma related to men who have sex with men, sex work stigma, drug use stigma, and gender identity stigma related to transgender persons can better support tracking progress in stigma mitigation for people living with HIV and key populations.

Supporting information

Acknowledgments

We would like to thank all the people living with HIV and members of key populations who agreed to participate in studies across all of the countries where data are available. We would like to thank the community groups and those working to call attention to both measure and address stigmas affecting people living with HIV and key populations. We would like to thank the members of the Task Team developed by UNAIDS Monitoring Technical Advisory Group (MTAG).

The UNAIDS Task Team members not listed as authors in the manuscript were Daniel Reidpath, George Ayala, Erika Castellanos, Ruth Morgan Thomas, Nicholas Niwagaba, Julian Kerboghossian, Sonal Mehta, Joanne Csete, Nazneen Damji, Elena Kudravtseva, Luisa Cabal, Kene Esom, Mianko Ramaroson, and Meaghan Kall.