Quantifying intersectional stigma in HIV prevention contexts: adaptation and validation of a scale among female sex workers in Namibia Authors: Kaitlyn Atkins, H T Saleem, A L Stangl, S D Baral, R Eakle, D Mali, L Bikinesi, L Ashipala, J B Reed, S M Murray Published: 2025 ABSTRACT AbstractBackgroundKey populations disproportionately affected by HIV face stigma and discrimination related to their multiple, converging marginalized identities. Engaging these populations in HIV prevention services requires understanding intersectional stigma and its effects and addressing it through targeted programs. New approaches are needed to quantify intersectional stigma as a determinant of HIV and HIV prevention engagement among key populations, particularly in sub-Saharan Africa, where this is understudied. We applied an intersectional approach in adapting existing stigma measures with female sex workers (FSW) in Namibia, with the goal of refining measures to more comprehensively assesses how stigmas related to FSW’s converging identities shape HIV-related outcomes.MethodsUsing survey data collected from 214 FSW in Namibia, we measured sex work-related stigma, PrEP-related stigma, and non-attributed stigma (i.e., stigma that was not attributed to any one condition, behavior, or identity). We evaluated the psychometric properties of these scales and combined them, using exploratory factor analysis to refine an intersectional stigma measure. We assessed construct validity by testing for correlations with healthcare provider trust, racial discrimination, depressive symptoms, and social cohesion.ResultsWe identified three dimensions of intersectional stigma among this sample of FSW: experienced intersectional stigma (α = 0.88), anticipated intersectional stigma (α = 0.84), and PrEP-related stigma (α = 0.45). Subscales were positively correlated with racial discrimination, depressive symptoms, and healthcare provider distrust.ConclusionsOur adapted measure is novel in relying on pre-existing measures, providing a proof of concept that researchers and programs can use existing validated tools to study stigma and its impacts through an intersectional lens. Our approach can be replicated to inform HIV prevention programs with key populations in other contexts. Further, our adapted tool demonstrated preliminary evidence of validity and reliability, and can be used by programs with FSW to consider how different dimensions of stigma may impact engagement with HIV prevention providers and programs.Keywords: Pre-exposure prophylaxis, Stigma, Sex work, Intersectionality, Scale development Background The HIV pandemic continues to disproportionately affect marginalized populations including female sex workers (FSW), gay, bisexual, and other men who have sex with men (MSM), people who inject drugs, transgender people, prisoners and incarcerated people, and their sexual partners [1]. Members of these groups face complex stigmas which can negatively impact HIV outcomes [2]. These can be conceptualized as “intersectional stigma,” which refers to stigma at the intersection of multiple interlocking systems of power [3, 4] and may reflect stigma related to health conditions (e.g., HIV, mental illness), sociodemographic characteristics (e.g., ethnicity, gender), behavior (e.g., sex work), or other characteristics [5]. Coined by Black feminist scholar Kimberlé Crenshaw to describe race and gender inequalities in the U.S. legal system [6, 7], intersectionality draws on decades of Black feminist scholarship [8, 9] and has been increasingly incorporated into HIV-related research and intervention [5, 10], but there are gaps in availability and use of validated intersectional measures to quantify multiple marginalization in the context of HIV services [11]. While approaches to quantifying intersectional stigma and its impacts have expanded [12–14], there is no consensus on which methods are most appropriate [13]. One challenge regards the presence or absence of stigma attribution. Attributional approaches focus on the specific experiences of stigma among a given population (e.g., Black men; sex workers; people living with HIV); however, these approaches can pose challenges for multiply marginalized individuals related to the cognitive burden of identifying the basis of experiences of stigma and discrimination [15] and the response burden of using repeated, parallel measures to assess multiple forms of stigma [16]. As an alternative, some have called for use of non-attributional scales, which ask about stigma experiences based on one’s identity or social position as a whole [15]. These measures can apply to a wide range of populations; however, their broad nature may limit their specificity. The present study was conducted among FSW in Namibia, where evidence suggests FSW in the country face vulnerability related to criminalization [17], poverty [18], violence [19], and negative social norms around sex work. FSW may experience intersectional stigma and structural vulnerabilities related to ethnicity, sexuality, poverty, HIV, sex work, and more [18]. These stigmas not only increase FSW’s vulnerability to HIV [20] but influence how FSW engage with HIV prevention services such as oral HIV pre-exposure prophylaxis (PrEP) [21]. To date, no studies have examined how to best characterize intersectional stigma in the context of HIV prevention among this population. To fill this gap, we evaluated attribution-specific and non-attributional measures of stigma adapted for use among FSW accessing PrEP services in Namibia. Methods From January to April 2021, we conducted a cross-sectional study among FSW, MSM, and young women accessing PrEP in three regions of Namibia [22]. At study facilities, PrEP services tailored to key populations were provided by the Ministry of Health and Social Services and supported by U.S. Agency for International Development (USAID) and local partners. We recruited from non-governmental and government facilities in Erongo, Khomas, and Zambezi regions, selecting sites in partnership with USAID and PrEP partners based on client volume. Procedures We used convenience sampling to recruit FSW during PrEP visits. Providers referred interested clients to on-site data collectors for eligibility screening. FSW were included if they were 18 or older, resided in one of the study regions, received oral PrEP services from a study site in the past six months (including on the day of the survey, for those who were initiating PrEP), and were cisgender women who exchanged sex for money, goods, or something else of monetary value in the last year. Following enrollment, trained study staff administered a questionnaire in the language of participants’ choice (Afrikaans, English, Silozi, Nama/Damara, Oshiwambo, Otjiherero) to assess their characteristics, health behaviors, and stigma experiences. Questionnaires were administered face-to-face in private rooms at PrEP facilities, and responses were recorded using tablets. All staff received training in intersectionality, stigma, and questionnaire administration with historically marginalized communities, including how to properly ask sensitive questions. FSW received NAD 100 (~ USD 6) as reimbursement for transportation and associated costs. Measures Participant characteristics included age, marital status, education level, formal employment, distance lived from PrEP site, primary language, region, self-reported condom use with clients in last month, and PrEP disclosure. Sex work stigma was assessed using 12 items from the WHO Biobehavioural Survey Guidelines, developed for administration with populations at risk of HIV [23]. This scale included items related to internalized, anticipated, and experienced stigma [24] on account of sex work. We adapted the phrasing of the attribution from “because you sell sex” to “because you exchange sex for money or goods,” to clarify FSW in Namibia receive varied payments in exchange for sex. Two items were positively framed (e.g., “I am not ashamed to say I am a sex worker in a gathering with other people who are sex workers”); these were reverse coded for analyses. PrEP-related stigma was measured using 8 items from the PrEP Anticipated Stigma Scale, developed among PrEP-inexperienced women in the U.S [25]. While not previously used with sex workers, this was the only published and validated PrEP-related stigma scale tested among women at the time of study enrollment. In consultation with the Ministry of Health and Social Services and local organizations offering PrEP programs with FSW, we dropped “People would assume I was gay if they knew I took PrEP” and added “I feel empowered because I take PrEP,” which was reverse coded for analyses [26]. Non-attributional stigma was assessed using two of three subscales of the Intersectional Discrimination Index [15] asking about anticipation or experiences of discrimination on account of “who you are.” To minimize burden and given redundancy with other items in our survey, we excluded 13 items measuring major discrimination. We included 7 items measuring anticipated stigma and 5 items measuring experienced stigma. Healthcare provider trust was measured using five items about clients’ trust in their providers’ knowledge and protection of information (a = 0.84). These items were validated with FSW in Togo and Burkina Faso [27] and had Likert-style responses (strongly agree, agree, disagree, strongly disagree). Racial discrimination was measured using the Everyday Discrimination Scale-Racism [28, 29], which measures frequency of day-to-day discrimination because of individual’s race or shade of skin color (a = 0.90). This scale has been found suitable for use among racially diverse adults in South Africa [28]. In consultation with local partners about potential harm introduced due to Namibia’s history of apartheid, we inserted a screening question: “Have you ever been treated differently because of your race or shade of skin color?” Participants who responded “no” (n = 158) skipped the scale. Due to low response rates, these items were only used during construct validation. Depressive symptoms were assessed using the 9-item Patient Health Questionnaire (PHQ-9; a = 0.83) [30], which has been found reliable among FSW globally [31, 32]. Social cohesion was measured using 12 items from the WHO Biobehavioural Survey Guidelines (a = 0.88) [23, 33]. The scale has been adapted and validated with sex workers in low- and middle-income countries [34–36]. Resilience was measured using the 6-item Brief Resilience Scale (a = 0.91) [37], which assesses the ability to bounce back from stress. Data analysis Because individual stigma scales had not been validated among the study population, we first examined the properties of each stigma subscale independently using the same methods we describe below. Based on these findings, we dropped “Has someone ever blackmailed you because you exchange sex for money or goods?” and “Because of who I am, a healthcare provider might treat me poorly.” No items were dropped from the PrEP-related stigma scale. To examine the conjoint latent structure of stigma conceptualized intersectionally, we included all 30 items from the three revised scales in a combined exploratory factor analysis (EFA). Before EFA, we examined sampling adequacy using the Kaiser-Meyer-Olkin (KMO) test and sphericity using Bartlett’s test [38]. We considered data suitable for EFA based on KMO values > 0.7 and rejection of the Bartlett’s test null hypothesis that variables were orthogonal (i.e., p < 0.001). To determine the number of factors to extract during EFA, we performed a principal component analysis (PCA) and considered the number of eigenvalues > 1.0 and percentage of variance explained per eigenvalue. We also examined the scree plot and ran a parallel analysis. We used iterative principal factor estimation with oblique rotation in the EFA, retaining items with factor loadings ≥ 0.40, uniqueness below 0.8, and which did not cross-load on multiple factors. We iteratively performed PCAs and EFAs for each item set, dropping one item at a time, until we reached a suitable solution. Factor interpretation was guided by findings from an ongoing qualitative sub-study with the population. We considered the items loading onto each extracted factor to comprise a subscale measuring the given construct and calculated mean scores per subscale, where higher scores indicate higher levels of stigma. For each subscale, we calculated Cronbach’s alpha to measure internal consistency [39]. We considered dropping items that would improve the overall alpha of the subscale by at least 0.05. We assessed convergent construct validity of each subscale using correlations with racial discrimination, healthcare provider trust, depressive symptoms, and social cohesion. We hypothesized intersectional stigma would correlate positively with racial discrimination and depressive symptoms, and negatively with healthcare provider trust, social cohesion, and resilience [40–44]. All analyses were conducted in Stata Version 15. Results Sample characteristics The final sample of 214 FSW included 114 individuals initiating PrEP (i.e., new PrEP clients) and 100 ongoing PrEP clients (Table 1). Median age was 29 (interquartile range [IQR]: 24–33), and the majority (92.1%) of participants were not married. Around half (53.7%) of FSW had at least a secondary education, and most (91.1%) were not formally employed outside of sex work. Table 1Sociodemographic and behavioral characteristics of sample (n = 214)CharacteristicN (%)Type of PrEP client Initiating PrEP114 (53.3%) Ongoing PrEP user100 (46.7%) Age, median (IQR)29 (24–33)Marital status Unmarried†197 (92.1%) Married or domestic partnership17 (7.9%)Education level Less than secondary school‡99 (46.3%) Secondary school or higher§115 (53.7%)Formal employment status No formal employment195 (91.1%) Formal employment19 (8.9%)Recruitment location Erongo (Walvis Bay sites)52 (24.3%) Khomas (Windhoek sites)131 (61.2%) Zambezi (Katima Mulilo sites)31 (14.5%) Distance from site (km), median (IQR)3 (1.5-4) Erongo (Walvis Bay sites)1 (1–3) Khomas (Windhoek sites)3 (2–5) Zambezi (Katima Mulilo sites)3 (1.5–15)Primary language Nama/Damara104 (48.6%) Afrikaans40 (18.7%) Oshiwambo19 (8.9%) Silozi14 (6.5%) Otjiherero12 (5.6%) Subia11 (5.1%) Other¶14 (6.5%)Percent condom use with clients, last 30 days # Less than 50% of encounters36 (16.8%) 50–75% of encounters29 (13.6%) > 75% of encounters149 (69.6%)Disclosed current or planned PrEP use to at least one other person No108 (50.5%) Yes106 (49.5%)†Includes divorced (n = 6) and separated (n = 4) FSW‡Includes one FSW who reported no formal education§ Includes three FSW who reported university-level education¶Includes English (n = 4), Rukwangali (n = 2), Sifwe (n = 2), Mbalangwe (n = 2), Ndebele (n = 2), Shona (n = 1), and Yeyi (n = 1)#FSW were asked how many times they had exchanged sex for mony or goods in the last 30 days. Of those times, they indicated how many times they used a condom. We used these self-reported frequencies to calculate a percentage Sociodemographic and behavioral characteristics of sample (n = 214) †Includes divorced (n = 6) and separated (n = 4) FSW ‡Includes one FSW who reported no formal education § Includes three FSW who reported university-level education ¶Includes English (n = 4), Rukwangali (n = 2), Sifwe (n = 2), Mbalangwe (n = 2), Ndebele (n = 2), Shona (n = 1), and Yeyi (n = 1) #FSW were asked how many times they had exchanged sex for mony or goods in the last 30 days. Of those times, they indicated how many times they used a condom. We used these self-reported frequencies to calculate a percentage The majority of FSW (61.2%) were recruited in Khomas. FSW lived an average of 3 km from the PrEP site at which they were recruited; FSW in Erongo lived closer to facilities (within 1 km, on average) than FSW in Khomas or Zambezi. Thirteen languages were represented in the sample, with most FSW speaking Nama or Damara (48.6%) or Afrikaans (18.7%) at home. With clients in the last 30 days, 69.6% of FSW reported using condoms more than 75% of the time. Around half said they had disclosed current or planned PrEP use to someone; this was more common among ongoing PrEP users (69.0%) than those initiating PrEP (32.5%). Factor structure Following PCA of the 30 items, there were 7 eigenvalues over 1 explaining 78% of the variance. In the final PCA (with 18 items), there were 4 eigenvalues over 1 explaining 71% of the variance. However, scree plots and parallel analyses consistently favored a three-factor model. Across iterations, KMO values were > 0.7 and we rejected the Bartlett’s test null hypothesis that variables were orthogonal (p < 0.001). Based on review of factor loadings, uniqueness, and Cronbach’s alpha values, we dropped 12 items: six from the sex work stigma scale, four from the PrEP-related stigma scale, and two from the non-attributional stigma scale. All dropped sex work and PrEP-related stigma items failed to load on any factors at or above 0.4. The two dropped non-attributional items (“Because of who I am, I worry about being harassed or stopped by police or security” and “Because of who I am, people might try to harm me physically”) consistently cross-loaded at 0.4 or greater on more than one factor. The final factor structure, which included 18 items, is summarized in Table 2. Factor loadings ranged from 0.42 to 0.93. We named the three factors experienced intersectional stigma (10 items from sex work and non-attributional stigma scales, α = 0.87), anticipated intersectional stigma (four items from non-attributional stigma scale, α = 0.84), and PrEP-related stigma (five items from PrEP-related stigma scale, α = 0.53). Scores for experienced intersectional stigma ranged from 0 to 2, with a mean score of 0.82. Scores for anticipated stigma and PrEP-related stigma each ranged from 0 to 3, with mean scores of 1.6 and 1.8, respectively. Experienced stigma and anticipated stigma had a moderate, negative correlation (r=−0.55; Table 3). Anticipated and PrEP-related stigma were weakly positively correlated (r = 0.28), while experienced and PrEP-related stigma were weakly negatively correlated (r=−0.32). Table 2Intersectional stigma item factor loadings and correlations (n = 163)ItemFactor 1: Experienced StigmaFactor 2: Anticipated, Non-Attributional StigmaFactor 3: PrEP-related StigmaUnique-nessHave you ever felt excluded from family activities because you exchange sex for money/goods?0.73−0.15−0.100.26Have you ever felt that family members gossiped about you because you exchange sex for money/goods?0.850.010.050.31Has someone ever scolded you because you exchange sex for money/goods?0.86−0.050.070.25Has someone ever physically harassed or hurt you because you exchange sex for money/goods?0.53−0.16−0.030.56Has someone ever forced you to have any type of sex when you did not want to, because you exchange sex for money/goods?0.42−0.300.100.63People would assume I slept around if they knew I took PrEP.−0.23−0.100.510.65People would think I am a bad person if they knew I took PrEP.0.100.210.560.62People would assume I was HIV-positive if they knew I took PrEP.−0.30−0.260.550.60I feel ashamed to tell other people that I take PrEP.0.290.380.500.63Because of who I am, I might have trouble getting/keeping a job.−0.120.800.070.20Because of who I am, I might have trouble finding housing.0.040.86−0.010.30Because of who I am, I worry about being treated unfairly at work/school.−0.290.61−0.010.36Because of who I am, I fear that I will have a hard time finding relationships.−0.260.50−0.090.57Because of who you are, have you heard or seen others laughing about you?0.77−0.09−0.100.25Because of who you are, have you been called names?0.930.040.010.18Because of who you are, have you been told that you should be more like others?0.850.13−0.060.35Because of who you are, have you heard that you or people like you don’t belong?0.70−0.120.010.41Because of who you are, have you been treated as if you are less capable?0.63−0.20−0.050.39Factor CorrelationsFactor 1Factor 2Factor 3Factor 1: Experienced Stigma1Factor 2: Anticipated Intersectional Stigma−0.551Factor 3: PrEP-related Stigma−0.320.281 Intersectional stigma item factor loadings and correlations (n = 163) Table 3Correlation matrix of intersectional stigma (IS), healthcare provider (HCP) distrust, depressive symptoms, and social cohesion (n = 214)Antic. StigmaExper. StigmaPrEP-related StigmaHCP TrustRacial Discrim.Depress. Sympt.Soc. Cohes.Anticipated Stigma1.0Experienced Stigma0.60***1.0PrEP-Related Stigma0.28***0.27***1.0Health Care Provider Trust−0.25***−0.17*−0.0011.0Racial Discrimination0.30*0.64***0.33**−0.031.0Depressive Symptoms0.28***0.42***0.19**−0.090.31*1.0Social Cohesion−0.130.12−0.150.080.03−0.051.0Resilience−0.19**−0.18**−0.170.16**−0.13−0.27**0.13*p < 0.05, **p < 0.01, ***p < 0.001 Correlation matrix of intersectional stigma (IS), healthcare provider (HCP) distrust, depressive symptoms, and social cohesion (n = 214) *p < 0.05, **p < 0.01, ***p < 0.001 Construct validation Correlations between intersectional stigma and other constructs identified as part of the nomological network of stigma are presented in Table 3. Overall, subscales demonstrated the expected positive correlations with racial discrimination and depressive symptoms. No subscales were correlated with social cohesion. The strongest correlation was between experienced stigma and racial discrimination (r = 0.64). Experienced stigma correlated weakly and negatively with provider trust and resilience, and weakly and positively with depressive symptom severity. Anticipated and PrEP-related stigma scales had weak positive correlations with racial discrimination and depressive symptoms, and weak negative correlations with provider trust and resilience. Discussion We adapted and validated three subscales to assess stigma among FSW in Namibia. The measures integrate attribution-specific (PrEP and sex work) and non-attributional measures to capture diverse stigma experiences. We identified two intersectional stigma subscales, one measuring anticipated stigma and one measuring experienced stigma. A third subcale, measuring PrEP-related stigma, included items loading distinctly from the two other scales. We found experienced and anticipated stigma were distinct, regardless of attribution, supporting a growing body of literature that suggests the detrimental impacts of stigma are not dependent on the identities, social positions, conditions, or behaviors to which it is attributed [14, 45]. Future studies should consider measuring experienced, anticipated, perceived, and internalized stigma as separate constructs, regardless of attribution, to capture stigma manifestations more fully. Indeed, given the diversity of FSW in our sample and in Namibia [46], non-attributional questions may have allowed participants in our study to consider how sex work along with other, more salient identities (e.g., ethnicity, religion) influenced their stigma experiences. At the same time, retaining attribution-specific items allowed us to examine specific stigma related to sex work and PrEP use as documented in the literature [47–50]. Our 10-item experienced and 4-item anticipated stigma scales were reliable measures of stigma among FSW that demonstrated preliminary evidence of validity and fill an important gap in the literature. To refine these final scales, we dropped three items about FSW shame, two about FSW stigma in healthcare settings. It may be that, when combined with items from the non-attributional and PrEP-related stigma scales, these shame and healthcare stigma items remained too distinct to jointly load with other items. Further, two of the items about shame were positively framed and may have been confusing to participants [51]. We also dropped two items about policing and arrest. The item about arrest (“Have you ever been arrested because you exchange sex for money or goods?”) was infrequently endorsed in our sample. However, the item about policing (“Because of who I am, I worry about being harassed or stopped by the police”) was commonly endorsed, and others have documented policing as a concern for FSW in Namibia [18, 46]. It may be that this item failed to load with others because it was the sole item assessing this domain. The five-item PrEP-related stigma subscale was less psychometrically sound, with poorer internal consistency and weaker correlations with other constructs. All items had relatively high uniqueness (above 0.60), suggesting failure to fully capture stigma experiences in this domain. Notably, all dropped PrEP-related stigma items were positively framed (“I feel empowered because I take or am planning to start PrEP” and “My family/friends/partner would approve of me taking PrEP”); may have represented empowerment or resilience, rather than stigma. Further, items around family, partner, and friend approval may have been inappropriate given mixed levels of PrEP disclosure. Our findings suggest that PrEP-related stigma may represent a separate construct from other, intersecting stigmas faced by FSW. The four retained items assessed PrEP stereotypes and shame rather than other stigma experiences, so it may also be that they reflected different manifestations of stigma (i.e., shame and stereotypes vs. experienced and anticipated stigma), rather than specific PrEP-related experiences. This should be explored further, given literature on PrEP-related stigma’s close relationship with HIV and sex work stigma [26, 52–54]. As additional measures of PrEP-related stigma are developed, it will be important to examine their relationship to intersectional stigma experiences among PrEP users, who often represent multiply marginalized communities [52]. Each of the three subscales was correlated with at least one related construct (healthcare provider trust, racial discrimination, depressive symptoms, or resilience). These constructs are consistently associated with HIV-related outcomes, including prevention-effective use of PrEP [55], making our subscales useful for understanding social determinants of HIV. It will be important to further examine how intersectional stigma influences HIV outcomes, and via what pathways. It is noteworthy that all three scales had significant correlations with racial discrimination, which others have shown plays an important role in healthcare experiences in Namibia [56, 57]. Further research is needed to more fully validate the subscales, including in new populations and with non-latent indicators of stigma. We found weak or non-significant correlations between intersectional stigma and resilience and social cohesion, respectively. However, much of the literature suggests that these attributes serve as protective factors moderating the relationship between stigma and adverse health outcomes among FSW and other marginalized groups [43, 58, 59]. If resilience and social cohesion indeed act as moderators rather than outcomes of stigma, that may help to explain why their pairwise correlations were weaker. In other words, FSW’s levels of resilience may not be strongly correlated with experiences of stigma, but those with higher levels of resilience may be better enabled to cope with stigma and mitigate its impacts. This should be explored further. Limitations The intersectional stigma measures used in this study had not been trialed in the study languages prior to survey administration. While translations were validated by members of the study team fluent in each language, limited pre-testing may have resulted in varied interpretations of the questions, particularly the phrase “because of who you are.” Due to missingness, we were unable to include racial discrimination items in our EFA. When endorsed, most racial discrimination items were often reported as occurring rarely, a few times a year, or a few times a month; reports of daily or weekly discrimination were uncommon. Given high levels of racial segregation in Namibia, it could be that day-to-day experiences of interpersonal racial discrimination are less common, whereas internalized or anticipated racial discrimination (unmeasured in our study) may be more relevant. Our measure of PrEP-related stigma was also limited. We were unable to develop a measure of PrEP-related stigma, and at the time of enrollment, there was only one published, validated measure of PrEP-related stigma. This measure was developed among PrEP-inexperienced women in the U.S., whose experiences likely diverge from the FSW in our sample. For example, our sample may have been more comfortable with or motivated toward PrEP than the sample with which the scale was developed. Additionally, to reduce participant burden, we did not include any measures of internalized stigma, which limited our ability to assess this domain of stigma from an intersectional perspective. This is an important area for future psychometric work. The context of survey administration may have caused FSW to emphasize their sex work identities when responding to stigma items, including non-attributional items. Although we included introductory language explaining the items were being asked in relation to participants’ identities as a whole, consent forms specified that recruitment was on the basis of being a FSW. Further, surveys were administered following PrEP visits, where clients often answer questions about their sexual behavior that reinforce their positions as FSW. Though it cannot be assumed, it is therefore possible that for some FSW the phrase “because of who you are” may have been understood as redundant with “because you exchange sex for money/goods.” Finally, recruitment of FSW from healthcare contexts limits the generalizability of our findings. Conclusions This was one of the first studies to pilot and validate an intersectional stigma measure combining attribution-specific and non-specific items among FSW. Our findings suggest manifestations of experienced and anticipated stigma cut across attribution, while PrEP-related stigma may be distinct. As approaches to measuring intersectional stigma evolve, this study underscores the utility of combining attributional and non-attributional measures, providing a new path forward for measuring and addressing this critical construct that influences HIV prevention, care, and treatment services [10]. Acknowledgements Not applicable. Abbreviations Human immunodeficiency virus exploratory factor analysis Female sex workers principal components analysis HIV oral pre-exposure prophylaxis United States Agency for International Development Authors’ contributions KA, JBR, LA, LB, DM, and RE designed the research study, protocols, and instruments. HTS, ALS, SDB, and SMM contributed to instrument development. KA analyzed the data. KA wrote the first draft of the manuscript. HTS, ALS, SDB, RE, DM, LB, LA, JBR, and SMM provided critical feedback and revision. All authors have reviewed the manuscript and approved the final version. Funding KA acknowledges support from the National Institute of Mental Health (F31MH124583, R25MH083620). This work was made possible with support from the U.S. President’s Emergency Plan for AIDS Relief, through the United States Agency for International Development funded RISE program, under the terms of the cooperative agreement 7200AA19CA00003, and the Johns Hopkins University Center for AIDS Research, a National Institutes of Health-funded program under grant P30AI094189. The results presented in this manuscript do not necessarily reflect the views of the National Institutes of Health, USAID, or the United States Government. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability The datasets analyzed for this study are not publicly available due to the highly marginalized nature of participants and potential identifiability. Data may be made available on reasonable request. Declarations Ethics approval and consent to participate This study was approved by Johns Hopkins School of Public Health Institutional Review Board and the Research Ethics Committee of the Namibia Ministry of Health and Social Services. All participants provided written informed consent. With partners, we developed referral protocols for participants with mental health concerns (i.e., acute depression or suicidality). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets analyzed for this study are not publicly available due to the highly marginalized nature of participants and potential identifiability. Data may be made available on reasonable request.