Structural factors, including social, cultural, economic, political, legal and environmental factors, are known to affect individual risk and vulnerability to HIV, and to operate at several different societal levels (individual, interpersonal, community, culture and policy) [1]. HIV prevention in the context of female sex work continues to be a high priority globally, particularly in settings where heterosexual contact is the main transmission mode, but the epidemic has yet to become generalised [2–4]. Early interventions with female sex workers (FSWs) typically relied on individual behaviour change involving peer education, condom promotion and provision of sexual health services [5], and many were limited in scale [6]. Exceptions to this included the 100% condom programme in Thailand which was a government-led structural intervention, targeting the mainly brothel-based sex worker population and their clients [7]; and the Sonagachi programme in Kolkata, India, which provided one of the first examples of a rights-based FSW HIV prevention programme, focusing on the community mobilisation and empowerment of sex workers alongside engagement with power structures [8–10]. Although these examples differed in approach, both recognised the need to address broader structural factors to create an enabling environment within which individual behaviour change could occur. More recently, a comparison of two structural intervention models with FSWs in the Dominican Republic found that an intervention which combined community mobilisation with government policy was more effective, and more cost-effective, than community mobilisation alone in reducing HIV and STI risk among FSWs [11, 12].

Karnataka state in South India has a population of approximately 60 million and, with an adult HIV prevalence of approximately 1% in several districts, ranks in the top four states in India with regard to epidemic severity [13, 14]. Heterosexual contact is the main route of transmission, with commercial sex work a key area of concern. The FSW population has been estimated to number over 100,000, and HIV rates have reached more than 25% among FSWs in some districts [15, 16]. In contrast to Thailand, Kolkata and the Dominican Republic, where brothel-based sex work largely features, the majority of FSWs in Karnataka are street- and home-based, working in rural as well as urban settings [17, 18].

Prior to 2003, condom availability throughout Karnataka was low [19], and HIV prevention programmes targeting female sex workers were operating on a small scale in only a few of the 27 districts. As part of Avahan, the India AIDS Initiative, funded by the Bill & Melinda Gates Foundation, which works with core and bridging populations in 83 districts in six Indian states [20], the Karnataka Health Promotion Trust (KHPT) was established in 2003 with the aim of rapidly scaling up HIV prevention efforts through targeted interventions in 18 of Karnataka's then 27 districts. Programme implementation at the community level was through non-governmental organizations (NGOs) or community-based organizations (CBOs), and included a rights-based approach [21]. The gradual handover of ownership to the FSW community and the Indian government has now begun, with 30% of the districts transitioned in April 2011, and plans for the remainder to be transitioned by April 2012. This process has been managed by KHPT, which will continue to provide post-transition support for another one to two years to help ensure the maintenance of programme quality.

In the early stages of the programme, needs assessment workshops were conducted with FSWs in each district, where women were asked about the problems and challenges they were facing in their lives. Key themes to emerge from these workshops were that issues relating to poverty (housing, food and schooling insecurities), stigma and discrimination, and violence and harassment, were major and common concerns of FSWs and were contributing to their vulnerability to HIV. For FSWs to prioritise their health needs, it became clear that, alongside the provision of standard programme initiatives including peer education, condom promotion and sex-worker specific sexual health services, these context-specific broader structural barriers contributing to FSW vulnerability would also need to be addressed. An enabling environment needed to be created in which HIV prevention programmes and services could be embedded [22–24].

We have previously reported on the impact of the HIV prevention programme in terms of reductions in violence, increases in condom use, and reductions in HIV and STI prevalence among the female sex worker population [16, 21, 25, 26]. In this paper we provide details of the community components of the programme, how we have worked to bring the FSW community in Karnataka together through processes of collectivisation, community mobilisation and empowerment and how, in partnership with the FSW community, we have engaged with policy makers (government), secondary stakeholders (the police, the media and human rights lawyers), and primary stakeholders (FSWs themselves) to address the broader structural barriers which contribute to FSW vulnerability to HIV and other sexually transmitted infections (STIs) (Figure 1).

To assess the impact of these structural interventions on FSW collectivisation, mobilisation and empowerment, as well as on FSW experience of stigma, discrimination, violence, harassment and social equity over time, we examined data from routine programme management information systems and from daily monitoring of newspapers in Karnataka.

HIV preventive interventions focused on vulnerable groups such as sex workers have traditionally targeted individual behaviour only, with impediments to success frequently including structural factors beyond the scope of these programmes [5, 6]. In this paper, we have described a structural approach to HIV prevention which sought to change those context-specific social, economic, political and environmental factors determining HIV risk and vulnerability, as part of a large-scale HIV prevention programme in Karnataka state. We focused on addressing structural factors identified by the FSW community as contributing to their HIV vulnerability and, following collectivisation and community mobilisation of FSWs, worked in partnership with policy makers, secondary stakeholders (police, journalists and human rights lawyers), and primary stakeholders (FSWs themselves) to address issues of stigma and discrimination, violence and harassment, and social inequity facing the FSW community.

Previously we have reported significant increases in condom use, reductions in HIV and STI rates and reductions in violence against FSWs in Karnataka [16, 21, 25, 26]. In this study we have reported that the HIV prevention programmes have supported FSWs to re-dress over 90% of reported incidents of violence and harassment against them. In addition, we have presented evidence of widespread FSW mobilisation and collectivisation, substantial increases in the numbers of FSWs gaining access to government social benefits, and substantial changes in the reporting of HIV/AIDS and sex work in district and state level newspapers. The impact of the various components of a structural intervention on reducing HIV vulnerability can be difficult to assess [1]. This is particularly true of structural barriers which are more distal to HIV risk, such as stigma and discrimination of FSWs by the general population, compared with more proximal factors such as sexual violence or micro-environmental factors affecting the conditions and resources of individuals (e.g. living conditions, food availability and education). Elsewhere, studies have demonstrated strong evidence of the detrimental impact on HIV risk of structural factors such as violence [32, 36, 43, 44], poverty [45–47], and stigma and discrimination [48–51]. In addition, some studies have sought to address structural barriers contributing to HIV/STI risk [52, 53], although there have been few studies demonstrating the direct impact of structural interventions on HIV risk among vulnerable populations [11, 12, 54]. Indeed, although the 100% condom programme in Thailand and the Sonagachi programme in Kolkata have both reported reduced or low rates of HIV among their target populations, without a randomised control trial design, it is difficult to directly assess the impact of the structural intervention components of these programmes on HIV prevalence [7, 9, 10, 37, 55].

Although we have presented evidence here suggesting that we are managing to successfully address some of the structural barriers identified as important by the FSW community, it is not possible for us to tease out from the data available, which components of our structural intervention have been most important in contributing to the reductions in HIV/STI rates. We are planning to conduct qualitative and process evaluation studies with FSWs to help in this respect. Evidence from studies elsewhere suggests that increased social capital at both the individual and community level could be important in reducing vulnerability to HIV [56–58]. In addition, evidence from elsewhere suggests that had we focused solely on individual behaviour change and not addressed the context-specific structural barriers contributing to the HIV vulnerability of FSWs, it is unlikely our HIV prevention programming would have been as successful [1, 10, 11].

There were several challenges to implementing this structural intervention, and lessons learned. One lesson is that it is important to develop the vulnerability reduction strategies from programme inception, as risk and vulnerability go hand in hand. In addition, vulnerabilities faced by FSWs are very contextual, and thus the involvement of sex workers in defining the structural drivers or factors causing vulnerability is crucial. For example, we found that while addressing violence was more important for certain FSW typologies and in certain geographic areas, for other FSWs improving social equity was more important. Addressing structural drivers is a long term intervention and requires patience from the donor and the implementer. Furthermore, any structural intervention work comes into direct conflict with power structures, which can sometimes lead to short-term increases in violence, discrimination and inequity. It is important that there is solidarity among the FSW population to be prepared for this, and that the supporting agencies provide maximum support to FSWs during this time. Finally, measuring the success or impact of structural interventions is complex, and the results of these interventions are not always consistent (for example, levels of reported violence can initially increase following a violence intervention due to violent reactions from established power structures, and increasing capability in violence reporting by FSWs) [25].

There are several limitations to a study such as this one. The absence of a precise baseline profile is a clear limitation, but it was not possible for us to conduct studies with the FSW community until trust had been established and the HIV prevention programme had begun. From our discussions with FSWs it is clear that prior to the programme commencing, no support systems (e.g. crisis management teams) or social entitlements were available to them, and condom availability was low [19]. The lack of cohort data makes it difficult for us to examine changes to individual behaviour over time, or to unpack the impact of the various structural interventions on changes to individual behaviours or STI/HIV rates. In addition, as most of the newspapers in Karnataka are not available online, the tracking of news reports is currently conducted manually, and may be subject to reporting error. Finally, it was beyond the scope of this paper to examine the cost and cost-effectiveness of this structural intervention, but such evaluations of the programme overall have been reported elsewhere, with results suggesting that this HIV prevention programme is not costly compared with other HIV prevention programmes, and that it is cost-effective [59, 60].

Taken together, we maintain that stigma and discrimination, violence and harassment, and social equity issues are all critical concerns of this FSW population, and all needed to be addressed, to enable vulnerable women to become sufficiently empowered to adopt the safer sexual behaviours which are required to respond effectively to the HIV epidemic. Although measuring the impact of a structural intervention is complex, addressing broader structural barriers contributing to HIV vulnerability as part of HIV prevention programming is not only possible, but is key to reducing the vulnerability of women.