Risk and Resilience Among Female Sex Workers in the Dominican Republic: Using a Person-Centered Approach to Examine the Relationships Between Syndemic Health Outcomes and with Theorized Protective Assets
Open AccessBackground: Female sex workers (FSWs) live and work at the intersection of multiple marginalized identities that place them at high risk for various negative health outcomes. Violence is both an outcome of concern unto itself as well as a risk factor for substance use, poor mental health, and poor HIV care and treatment outcomes. Resilience theory asserts that an individual or community needs assets from which they can draw to survive adversity, such as the chronic discrimination and oppression FSWs face. This dissertation employs a person-centered approach to investigate and compare patterns of both violence exposure and resilience assets and their respective relationships with syndemic health outcomes for cisgender and transgender FSWs living with HIV in the Dominican Republic. Methods: Two cross-sectional survey datasets of cisgender and transgender FSWs living with HIV (N=211 and 100, respectively) are used from companion studies. Latent class analysis (LCA) is a person-centered approach that allows for the identification of unmeasured class membership among a group. LCA is employed in each manuscript - first to define patterns of violence and then patterns of assets. For each manuscript, predictors of class membership and associations between class membership and syndemic health outcomes (substance use, mental health, and HIV care and treatment outcomes) are identified using forms of logistic regression. Additionally, a comparison of the cisgender and transgender samples is made in order to illuminate similarities and differences between the two groups of FSWs. Results: From Manuscript 1, two classes were found among the cisgender sample: Low Reported Violence Exposure (Class 1, 86.30% of participants) and Sex Work-related Police Harassment (Class 2, 13.70%). Being drunk or high while exchanging sex for money increased the likelihood of being in Class 2, compared to Class 1, by roughly 10 times (aOR=10.11; 95%CI 1.93, 52.89; p<0.01). The average number of clients per week was marginally significant as a predictor of Class 2 membership, compared to Class 1 (aOR=1.08; 95%CI 0.97, 1.21; p=0.16). Class 2 participants had greater odds of ART disruption (OR=3.40; 95%CI 1.38, 8.41; p<0.01), having borderline abnormal or abnormal anxiety (OR=4.81; 95%CI 1.35, 17.13; p<0.01), and having moderate-to-severe depression (OR=5.41; 95%CI 1.82, 16.04; p<0.01), compared to Class 1. The transgender sample produced three classes: Low Reported Violence Exposure (Class 1, 48.65% of participants); Sex Work-related Police Harassment (Class 2, 31.89%); and Sex Work-related Violence and Harassment (Class 3, 19.46%). Educational attainment reduced the likelihood of being in Class 2 (aOR=0.49; 95%CI 0.24, 0.98; p<0.05). Average monthly income reduced the likelihood of being in the Class 3 (aOR=0.93; 95%CI 0.88, 0.98; p<0.05), while sex work-related travel increased the likelihood of being in Class 3 (aOR=5.32; 95%CI 1.18, 23.86; p<0.05), compared to Class 1. Class 3 participants had greater odds of having borderline abnormal or abnormal anxiety (OR=6.83; 95%CI 1.94, 24.03; p<0.01) and moderate-to-severe depression (OR=6.83; 95%CI 1.94, 24.03; p<0.01), while Class 2 participants had greater odds of ART adherence (OR=3.09; 95%CI 1.04, 9.18; p<0.05), compared to Class 1. From Manuscript 2, the transgender sample produced two classes: Internal and External Multilevel Assets (Class 1, 42.62% of participants) and External Institutional Assets (Class 2, 57.38%). Being born in Santo Domingo (OR=3.02; 95%CI 1.14, 8.01; p<0.05), being self-employed (OR=4.93; 95%CI 1.03, 23.67; p<0.05), and the number of years as a sex worker (OR=1.04; 95%CI 0.99, 1.08; p<0.10) all marginally increased the odds of transgender participants being in Class 1 compared to Class 2. In contrast, average monthly income marginally decreased those odds (OR=0.98; 95%CI 0.95, 1.00; p<0.10). However, none of these predictors were statistically significant in the multivariate model. Being in Class 1 was associated with ART adherence (OR=9.67; 95%CI 1.19, 78.69; p<0.01), current ART use (OR=2.60; 95%CI 0.78, 8.72; p<0.10), and reporting physical or sexual violence in the past 6 months (OR=2.23; 95%CI 0.97, 5.12; p<0.10), compared to Class 2. The cisgender sample produced three classes: Internal and External Multilevel Assets (Class 1, 36.64% of participants); External Institutional Assets (Class 2, 54.15%); and Low Reported Assets (Class 3, 9.21%). Only age (aOR=1.11; 95%CI 1.01, 1.23; p<0.05 for Class 1; aOR=1.12; 95%CI 1.02, 1.23; p<0.05 for Class 2) and educational attainment (aOR=4.74; 95%CI 1.71, 13.13; p<0.01 for Class 1; aOR= 3.31; 95%CI 1.15, 9.55; p<0.05 for Class 2) were statistically significant as predictors of membership compared to Class 3; average monthly income was marginally significant (aOR=1.16; 95%CI 0.98, 1.37; p<0.10 for Class 1; aOR= 1.15; 95%CI 0.98, 1.36; p<0.10 for Class 2). Being in Class 1 was significantly associated with being ART adherent over the past four days (OR=2.90; 95%CI 0.92, 9.14; p<0.05) and marginally associated with being virally suppressed (OR=3.63; 95%CI 1.11, 11.90; p<0.10), compared to Class 3. Being in Class 2 was marginally associated with currently taking ART (OR=7.25, 95%CI 1.51, 34.82, p<0.10), compared to Class 3. Discussion: Findings from both papers underscore the role of gender identity as a key factor that helps define a FSW’s violence exposure and the assets she holds. Transgender FSWs were found to experience more diverse and extreme forms of violence compared to their cisgender counterparts, while the cisgender sample was found to have more diverse patterns of assets. Further, we found that different patterns of violence have different impacts on cisgender compared to transgender FSWs, but a consistent finding points to the toll violence takes on their mental health. We also see some consistent patterns of assets across the two groups that point to the resilience of these women, particularly how having a combination of interpersonal, community, and institutional assets improves HIV care and treatment outcomes when compared to having low assets for the cisgender participants or having only external institutional assets for the transgender participants. These results from both papers are supported by existing literature, but the use of LCA, rather than a data-centered approach, allows for a more nuanced understanding of what violence looks like for cisgender and transgender FSWs, respectively, the impact of these patterns of violence, and patterns of assets that may protect against the weight of intersecting systems of marginalization. This nuance can offer greater precision when designing and implementing interventions to ensure that those most at risk for violence and those with the fewest assets to draw from are supported.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.