Agency as a mediator in the pathway from transactional sex to HIV among pregnant women in Swaziland: a multigroup path analysis

• Type of Material: Article
• Authors: Rebecca Fielding-Miller, University of California San Diego; Kristin L. Dunkle, South African Medical Research Council; Craig Hadley, Emory University; Hannah Cooper, Emory University; Michael Windle, Emory University
• Language: English
• Date: 2017-07-18
• Publisher: International AIDS Society
• Publication Version: Final Published Version
• Copyright Statement: © 2017 Fielding-Miller R et al; licensee International AIDS Society.
• License: Creative Commons Attribution 3.0 Unported
• Final Published Version (URL): http://dx.doi.org/10.7448/IAS.20.1.21554
• Title of Journal or Parent Work: Journal of the International AIDS Society
• ISSN: 1758-2652
• Volume: 20
• Issue: 1
• Start Page: 1
• End Page: 11
• Grant/Funding Information: Financial and material support was provided by a Fulbright US Student Grant, the Health Community Capacity Collaborative, Swaziland with support from the Presidents Emergency Plan for AIDS Relief (AID-OAA-A-12-00058), and the Center for AIDS Research at Emory University [P30AI050409].; This study was supported by the Swaziland National Emergency Response Council on HIV/AIDS.; Support in preparing this manuscript was provided by NIDA grants T32DA023356 and R21DA039782 as well as NIMH grant K01MH112436.
• Keywords: CONDOM USE; SOUTH-AFRICA; MADAGASCAR; FOOD INSECURITY; Infectious Diseases; RISK-FACTORS; Swaziland; WESTERN CAPE; cultural consensus modeling; RURAL MALAWI; structural equation modeling; Science & Technology; social status; transactional sex; BEHAVIOR; GENDER; HIV/AIDS; agency; Life Sciences & Biomedicine; SUBJECTIVE SOCIAL-STATUS; Immunology

Abstract

INTRODUCTION: Transactional sex is a structural driver of HIV for women and girls in sub-Saharan Africa. In transactional relationships, sexual and economic obligations intertwine and may have positive and negative effects on women's financial standing and social status. We conducted a clinic-based survey with pregnant women in Swaziland using a locally validated transactional sex scale to measure the association between subjective social status, transactional sex, and HIV status, and to assess whether this association differed according to a woman's agency within her relationship. METHODS: We recruited a convenience sample of 406 pregnant women at one rural and one urban public antenatal clinic in Swaziland and administered a behavioural survey that was linked to participant HIV status using clinic records. We then conducted a multigroup path analysis to test three hypotheses: (1) that more engagement in transactional sex is associated with decreased condom use and increased subjective social status; (2) that subjective social status mediates the relationship between transactional sex and HIV status; and (3) that these relationships are different across groups according to whether or not a woman reported any indicator of constrained agency within her relationship. RESULTS: The amount and value of material goods received from a sexual partner was significantly and positively associated with higher subjective social status among all participants. As the amount of material goods received from a partner increased, women who reported no indicators of constrained agency were less likely to use condoms. Conversely, there was no relationship between transactional sex and condom use among women who reported any indicator of constrained relationship agency. Among women who reported any indicator of constrained agency, HIV was significantly associated with lower subjective social status. CONCLUSIONS: Relationship agency likely plays a key role in determining which mechanisms create HIV risk for women in transactional relationships. Interventions to mitigate these risks must address social forces that penalize women who engage in sexual relationships as well as structural drivers of gendered economic disparity that reduce women's agency within their sexual and romantic relationships.