The HIV epidemic in India, the world's second most populous country, is of global importance.1 Interventions targeting sex workers are pivotal to HIV prevention in India.1,2 Community mobilization is considered by the National AIDS Control Program to be an integral component of this strategy.1,2 Yet, with the exception of exemplary programs in South India's high-prevalence states, HIV prevention in sex workers is suboptimal.1
Despite mounting evidence for targeting female sex workers (FSWs) as part of a comprehensive HIV prevention strategy,3-5 only 10 countries provide HIV prevention services to the majority of their FSW.6 This failure reflects the tension between 2 opposing philosophies: harm reduction versus prohibition. Policy makers and the public often prefer the “quick fix” allure of the latter.
Baina beach, a well-demarcated slum of 0.09 km2, situated in Goa's largest port, was a renowned red-light area. Having emerged in response to docking ships, it expanded and adapted to the needs of tourism, and its army of migrant workers. HIV prevention interventions, consisting of behavioral change, condom promotion, and syndromic treatment of sexually transmitted infections (STIs), had been provided for the 2000-3000 FSWs living and working in Baina red-light area since the early 1990s. Routine surveillance recorded a predemolition HIV prevalence of 30%-50% in Baina FSWs and high levels of HIV awareness and condom use.7
The Immoral Trafficking Prevention Act regulates sex work in India. Although the act does not prohibit prostitution per se, prostitution carried out in a brothel or within 200 m of certain public places is a criminal offense. On June 14, 2004, the Government of Goa implemented in part of a Mumbai High Court ruling and demolished Baina red-light area.8 No rehabilitation or relief was provided and a decade of HIV prevention was brought to an abrupt close.8 Since the demolition of Baina, several high-profile closures of red-light areas and dance bars have taken place in India.9 There is an urgent need to systematically document the public health consequences of prohibitive approaches to sex work. We hypothesized that after the demolition, area-based sex work would be replaced by more dispersed and clandestine types of sex work, less amenable to HIV prevention. In this cross-sectional study of FSWs, conducted in the year after the demolition, we compare the sex work concentrated in Baina with the dispersed sex work that materialized.
The study was set in Goa, a small coastal state with a population of 1.37 million.10 After the demolition, the relatively homogeneous population of red-light district sex work rapidly evolved into a more heterogeneous dispersed and clandestine types.11 This study was conducted throughout Goa in collaboration with Positive People, the largest and most experienced HIV organization in Goa. Recruitment took place from December 2004 to 2005.
FSWs were defined as women who were currently providing sexual services in exchange for goods or money.
The FSWs in the survey were recruited using respondent-driven sampling (RDS).12 This variant of chain sampling delves deeper into the hidden networks by rationing the number of recruits per respondent, increasing the number of waves of recruitment, and providing financial incentives to the “recruiter.” The initial recruiters (seeds) were defined as FSWs or community members close to sex work. Based on detailed ethnographic mapping of Goa, seeds were purposively selected from various ethnicities, ages, areas of Goa, and sex work typologies.13
Data Collection and Management
This has previously been described in detail.13 Data were collected through face-to-face interviews with a female interviewer in 1 of 4 Indian languages. The questionnaire was a composite of questions derived from several sources on demographics, psychosocial factors, sex work and sexual risk, knowledge and exposure to HIV prevention interventions, reproductive health, and health-seeking behavior. An informal confidential voting interview collected responses to sensitive questions from each respondent. Two independent reviewers checked all questionnaires and inconsistencies were referred back to the field. Data were double entered into a Microsoft access database (Microsoft, Redmond, WA) and underwent range and consistency checks.
Biological samples were collected using self-administered vaginal swabs, previously validated in Goa, and dried blood spots. For women who refused to take a vaginal swab, first void urine samples were collected to test for Neisseria gonorrhoeae (NG) and Chlamydia trachomatis (CT). One vaginal swab was inserted into a sterile universal container and the other was inserted into an InPouch TV culture kit (Biomed Diagnostics, San Jose, CA). Five blood spots were taken according to protocol. All samples reached the laboratory within 24 hours of collection. The InPouch TV culture was incubated at 37°C for up to 5 days and underwent daily microscopy for Trichomonas vaginalis (TV). The other samples were stored in a −70°C freezer until processing. Polymerase chain reaction using the Roche Amplicor system (Roche Molecular Systems, Alameda, CA) was used to diagnose chlamydial and gonococcal infection. Dried blood spots were tested for antibody to HIV using 2 enzyme-linked immunosorbent assay (ELISA) tests, Vironostika Uni-Form II plus O (Organon Teknika, Boxtel, The Netherlands), HIV enzyme immunoassay (Ani Labsystems, Ltd, Oy, Vantaa, Finland). Discordant tests were confirmed by Murex HIV 12O (Abbott Laboratories, Abbott Park, IL). The laboratory participated in quality control for molecular (United Kingdom) diagnostics annual QC and National AIDS Research Institute external quality control of the dried blood spots.
To adjust for potential biases in recruitment, data were weighted by the inverse of the approximate probabilities of recruitment.12 The weights were calculated, based on network size, age, ethnicity and area, according to how these factors were related to recruitment, using RDS Analysis Tool 5.4.0 (Cornell University, Ithaca, NY). All analyses were performed using Stata8 (Stata Corporation, College Station, TX), incorporating the weights through the standard survey analysis functions. The exposure was defined as ever having worked in Baina red-light area. Characteristics of FSWs who had ever worked in Baina (ex-RLD-SWs) were compared with FSWs who had never worked in Baina (non-RLD-SWs). The association between the exposure, that is, being a ex-RLD-SW and the following outcomes, curable STI, HIV, self-reported condom use, and ever having attended HIV prevention sessions, were measured using multivariate analysis and adjusting for potential confounding factors. Confounders were defined for each outcome as factors associated with both exposure and outcome (P < 0.2) and not on the explanatory pathway. Age, religion, ethnicity, marital status, socioeconomic status, number of regular and nonregular paying customers, and duration of sex work were included as a priori confounders based on published literature. Finally, factors potentially on the explanatory pathway were added into the model to explore their effect on the relationship between STIs and having worked in Baina.
The main reason for missing values was that samples were not received or inhibition of the sample occurred during polymerase chain reaction. Seventeen (5%) of the STI samples were missing (4 CT, NG, TV not received, 6 CT, NG inhibitory, and 7 TV only not received). Only 1 (0.3%) HIV sample was not received. Missing cases were excluded from the analysis.
Ethical approval was obtained from the Independent Ethics Commission, Mumbai, and University College London's ethics committee. A community advisory board mediated community engagement. We campaigned against the demolition and provided material support in the immediate aftermath. All participants and their partners were offered presumptive treatment for curable STIs and treatment based on laboratory tests. HIV results were anonymous; however, voluntary counseling and testing for HIV, treatment for STIs, and HIV risk reduction counseling were made available to participants and nonparticipants alike.
Three hundred twenty-six sex workers were recruited from 35 of the 59 seeds approached. Following up to 6 recruitment waves, each seed gave rise to between 2 and 30 participants. Based on our extensive mapping, we became aware of networks that we were unable to recruit; these comprised mainly of women who did not self-identify as FSWs. All areas of Goa and types of sex work identified through mapping were represented in the sample.
Of the 326 FSWs recruited, 125 [44.8% (95% confidence interval, CI: 39.1 to 50.5)] had ever worked in Baina red-light district (ex-RLD-SWs) and 201 [55.3% (95% CI: 49.5 to 60.9)] had never worked in Baina red-light district (non-RLD-SWs). Ninety-six [47.9% (95% CI: 40.9 to 55.0)] of the non-RLD-SWs had started sex work since the demolition.
Table 1 summarizes the differences between ex-RLD-SWs and non-RLD-SWs. Although, the median age of non-RLD-SWs and ex-RLD-SWs was similar, 27 (interquartile ratio: 23-35) and 28 (interquartile ratio: 24-35) respectively; the proportion aged younger than 20 years was higher in the non-RLD-SWs. The non-RLD-SWs were more likely to be Goan, part time, street, and home based. They were more likely to work in more than 1 place and entertained fewer customers, more of whom were regular clients. They were more likely to experience sex-based violence and report suicide attempts, but less likely to be economically disadvantaged than their Baina counterparts.
After adjustment for confounding (Table 2), ex-RLD-SWs had 3 times lower odds of having a curable STIs, 20 times greater odds of exposure to HIV prevention, and a 100 times greater odds of reporting consistent condoms use with clients compared with non-RLD-SWs. The greater odds of HIV in univariate analysis did not remain statistically significant after adjustment.
After further adjustment for behavioral factors potentially on the explanatory pathway, that is, condom use with paying and nonpaying male partners, HIV knowledge, and exposure to HIV prevention interventions, ex-RLD-SWs were still less likely to have curable STIs, adjusted odds ratio 0.30 (95% CI: 0.19 to 0.94).
We believe that this is the first published description of sex work in the aftermath of dismantling a red-light district. Our study suggests that the homogeneous ex-RLD-SWs had lower risk sexual behavior and better access to HIV prevention services compared with the heterogeneous dispersed sex workers who dominated after the demolition.
Part time, dispersed, home, and street-based FSWs filled the void left by the demolition. The non-RLD-SWs were more likely to have an STI, a biological marker of recent sexual risk and/or poor access to STI treatment. This finding was in keeping with their lower likelihood of reporting consistent condom use with clients14 and supports the hypothesis that the non-RLD-SWs were more likely to engage in high-risk behavior. Although this may partly be explained by lack of exposure to HIV prevention interventions, the persistence of higher odds of STIs even after adjusting for behavioral and knowledge indicators suggests that nonpaying sexual networks and more proximal determinants, such as their greater experience of sexual and intimate partner violence, the more socially isolated types of sex work, and lack of collective identity, may also play a part.13,15
Ex-RLD-SWs were more likely to have HIV, although this difference was not statistically significant after adjustment for confounding. Ex-RLD-SWs reported a longer duration in sex work, more clients, and were more likely to have migrated from the higher prevalence areas of northern Karnataka, which was an independent risk factor for HIV.13
We have compared ex-RLD-SWs with non-RLD-SWs. The speed with which events unfolded meant that quantitative predemolition data were unavailable for comparison. As recruitment started soon after the demolition, it was likely that the ex-RLD-SWs would approximate to the sex workers working in Baina before the demolition, whereas the non-RLD-SWs would represent the types of dispersed sex work that dominate the sex trade after the demolition. However, the possibility that ex-RDL-SWs who continued to practice in Goa are different to those who left is an important source of selection bias. Moreover, recruitment took place over a period of a year, and the prevalence of curable STIs can rapidly change over a shorter period, potentially diluting the difference between ex-RLD-SWs and non-RLD-SWs.
A dedicated team, familiar with sex work, conducted this study. We actively engaged in advocacy and provided material support post demolition. This lack of “distance” may have resulted in interviewer bias. Similarly, our association with HIV prevention may have resulted in social desirability bias. Behavioral data collected through the survey was, however, consistent with data collected through qualitative methods, informal confidential voting interview, and biological markers.13
To reduce selection bias, we used chain sampling where the probability of recruitment can be calculated; however, this is not a probability sample survey. Although we are able to incorporate variability in the probability of recruitment through weighting to provide unbiased analysis, bias may arise in our analysis if the selection of network members for recruitment is based on factors related to our outcome measures. Furthermore, the full complexity of the RDS sample is not reflected in the standard errors, and so the CIs should be viewed as approximate.
Ex-RLD-SWs had better access to HIV prevention and were more likely to consistently use condoms with their clients. Although the methodological limitations make the difference in STIs difficult to interpret, it is notable that the increased likelihood of having an STI among dispersed non-RLD-SWs is of a similar magnitude to the reductions recorded in successful sex worker interventions.5,16-19
In conclusion, sociopolitical events that destabilize the context and force sex workers into unorganized, fragmented, and clandestine working conditions create barriers to effective HIV prevention and community mobilization.20-22
We are grateful to the Wellcome Trust for supporting this work through a Fellowship grant to M.S. We thank Anil Pandey and Beethoven Fonesca for administrative support to the research project and Sushila Mendoza for cleaning the data. We thank the board members and staff of Positive People and Sangath for supporting us in our work and, in particular, the research team for all their tireless work under grueling conditions. We thank the laboratory staff in Sangath for processing the samples and rapidly adapting to the changing requirements of field-based sampling. We thank the sex workers of Baina and Goa for receiving us with open arms despite the harsh circumstances, for participating in this study, for implementing the findings and providing us with constant and stimulating critical feed back through the community advisory board and peer educators. We thank Beryl West and Rosanna Peeling for their training and supervision of the laboratory staff and Dr. Risbud of the National AIDS Research Institute, Pune, for quality control of our samples. Authors contributions: M. Shahmanesh declares that she designed and implemented, reviewed, analyzed, and interpreted the data. She wrote the first and subsequent drafts of the articles and has final approval of the published article. S. Wayal declares that she participated in the implementation of the study, collection and analysis of the qualitative data, and critical appraisal of all the drafts of the articles and final approval of the published article. A. Copas declares that he supported the statistical analysis of the quantitative data and was involved in the critical appraisal of all the drafts of the articles and final approval of the published article. V. Patel, D. Mabey, and F. Cowan declare that they participated in the design of the study, interpretation of the data, and critical appraisal of all the drafts of the articles and final approval of the published article.
1. Chandrasekaran P, Dallabetta G, Loo V, et al. Containing HIV/AIDS in India: the Unfinished Agenda. Lancet Infect Dis
3. Rekart ML. Sex-work harm reduction. Lancet
4. Steen R, Dallabetta G. Sexually transmitted infection control with sex workers: regular screening and presumptive treatment augment efforts to reduce risk and vulnerability. Reprod Health Matters
5. Shahmanesh M, Patel V, Mabey D, et al. Effectiveness of interventions for the prevention of HIV and other sexually transmitted infections in female sex workers in resource poor setting: a systematic review. Trop Med Int Health
7. Goa State AIDS Control. HIV/AIDS in Goa: Situation and Response 2005-2006
. Goa, India: Navaditya Packaging; 2007.
9. Research Centre for Women's Studies. Working Women in Mumbai Bars: Truths Behind the Controversy
. Mumbai, India: SNDT College; 2005.
10. Government of India. Census of India 2001
. New Delhi, India: Government of India; 2001.
11. Shahmanesh M, Wayal S. Targeting commercial sex-workers in Goa, India: time for a strategic rethink? Lancet
12. Magnani R, Sabin K, Saidel T, et al. Review of sampling hard-to-reach and hidden populations for HIV surveillance. AIDS
13. Shahmanesh M, Cowan F, Wayal S, et al. The burden and determinants of HIV and sexually transmitted infections in a population based sample of female sex workers in Goa, India. Sex Transm Infect
14. Peterman TA, Lin LS, Newman DR, et al. Does measured behavior reflect STD risk? An analysis of data from a randomized controlled behavioral intervention study. Project RESPECT Study Group. Sex Transm Dis
15. Blanchard JF, O'neil J, Ramesh BM, et al. Understanding the social and cultural contexts of female sex workers in Karnataka, India: implications for prevention of HIV infection. J Infect Dis
. 2005;191(Suppl 1):S139-S146.
16. Ghys PD, Diallo MO, Ettiegne-Traore V, et al. Effect of interventions to control sexually transmitted disease on the incidence of HIV infection in female sex workers. AIDS
17. Alary M, Mukenge-Tshibaka L, Bernier F, et al. Decline in the prevalence of HIV and sexually transmitted diseases among female sex workers in Cotonou, Benin, 1993-1999. AIDS
18. Kaul R, Kimani J, Nagelkerke NJ, et al. Reduced HIV risk-taking and low HIV incidence after enrolment and risk-reduction counselling in a sexually transmitted disease prevention trial in Nairobi, Kenya. J Acquir Immune Defic Syndr
19. Reza-Paul S, Beattie T, Syed HU, et al. Declines in risk behaviour and sexually transmitted infection prevalence following a community-led HIV preventive intervention among female sex workers in Mysore, India. AIDS
. 2008;22(Suppl 5):S91-S100.
20. Blankenship KM, West BS, Kershaw TS, et al. Power, community mobilization, and condom use practices among female sex workers in Andhra Pradesh, India. AIDS
. 2008;22(Suppl 5):S109-S116.
21. Jana S, Bandyopadhyay N, Mukherjee S, et al. STD/HIV intervention with sex workers in West Bengal, India. AIDS
. 1998;12(Suppl B):S101-S108.
22. Moses S, Ramesh BM, Nagelkerke NJ, et al. Impact of an intensive HIV prevention program for female sex workers on HIV prevalence among antenatal clinic attendees in Karnataka state, south India: an ecological analysis. AIDS
. 2008;22(Suppl 5):S101-S108.