Sex workers in HIV prevention: From Social Change Agents to Peer Educators.

We utilised a comparative ethnographic approach to study the implementation of a community mobilisation intervention addressing HIV risk among female sex workers (FSWs) in India, as implemented first by an non-governmental organisation and after oversight of the intervention was transitioned to the government. We demonstrate that the work of peer outreach workers changed from Social Change Agents within a community-led structural intervention (CLSI) to Peer Educators within a targeted intervention (TI). In the CLSI approach, built on the assumption that FSW risk for HIV is rooted in power inequality and structural vulnerability, peer outreach workers mobilised their peers through community-based organisations to address underlying conditions of inequality and vulnerability. In contrast, the TI approach, which views FSW risk as a function of limited knowledge and barriers to services, addressed peers' access to information and health services. Analysis of changes in the function of peer outreach workers reveals critical differences of which we discuss four: assumptions about conditions that produce HIV risk; degree of emphasis placed on collective mobilising and building collective power; extent to which community mobilisation and HIV prevention goals are linked; and the intervention's use of peer input. We discuss the implications of these findings for HIV prevention programming.

Background

Peer workers are used extensively in HIV prevention activities targeting female sex workers (FSWs) (Shahmanesh, Patel, Mabey, & Cowan, 2008). As FSWs who themselves fall in a high-risk group, interventions assume peer workers to be most capable of delivering desired HIV services. Using a case study of the implementation of a community mobilisation intervention to address HIV risk among FSWs in India, we demonstrate that peer workers do not simply educate FSWs about behaviours that can put them at risk of HIV. They also act as community organisers who mobilise their peers to change their risk environment. We develop this argument by focusing on the change in the function of peer workers in non-governmental organisations (NGOs) implementing an HIV prevention intervention. The intervention’s approach transitioned from an implicit model of community-led structural intervention (CLSI) to a more traditional targeted intervention (TI), while continuing to use the language of ‘community mobilization’. We discuss how these two approaches are reflected in the work of peer workers.

The literature on community mobilisation interventions underscores the importance of collectivization and empowerment strategies for HIV prevention programmes targeting FSWs (Blankenship, West, Kershaw, & Biradavolu, 2008; Evans & Lambert, 2008; Gaikwad, Bhende, Nighi, Saggurti, & Ranebennur, 2012; Gurnani et al., 2011). In these programmes, community mobilisation takes on such varied forms as running a 24-hour crisis response (Beattie et al., 2010; Reza-Paul et al., 2012); campaigning for social entitlements (Gurnani et al., 2011); building community-based organisations (CBOs) (Chakravarthy, Joseph, Pelto, & Kovvali, 2012); and stakeholder advocacy (Gaikwad et al., 2012; Gurnani et al., 2011). Interventions using peer workers and community mobilisation approaches are positively associated with outcomes related to HIV prevention such as improved knowledge about condoms, reported increased condom use with clients and reduced incidence of sexually transmitted infections (STIs) (Basu et al., 2004; Blankenship et al., 2008; Reza-Paul et al., 2008; Swendeman, Basu, Sankari Das, Jana, & Rotheram-Borus, 2009). Barring exceptions (Biradavolu, Burris, George, Jena, & Blankenship, 2009; Cornish & Campbell, 2009), the evidence reveals little about the context in which peer workers mobilise FSWs and how they overcome challenges to initiate and sustain community mobilisation interventions. We contend that understanding implementation is important because it reveals conditions that make community mobilisation a distinct approach for HIV prevention among FSWs.

The case study for this paper, an HIV prevention programme among FSWs implemented as part of the Bill & Melinda Gates Foundation (BMGF)-supported Avahan India AIDS Initiative (Wheeler et al., 2012), is an example of an approach characterised by peer education and community mobilisation within a CLSI framework. The Avahan package of prevention interventions used peer workers – self-identified women who exchanged sex for money, who had leadership qualities and strong links to the FSW community – to reach out to members of their social networks who were at risk of HIV. They provided support and information to negotiate condom use, distributed free condoms and encouraged FSW attendance at STI clinics and other community programmes. They mobilised FSWs to strengthen their collective agency and advocate with stakeholders to secure an enabling environment (Avahan, 2009). In 2008, Avahan initiated a planned transition, primarily involving the transfer of funding and management of its HIV prevention interventions to the Indian Government (Rao, 2010; Sgaier et al., 2013). The expected outcome was that the Government’s National AIDS control organisation would finance and monitor the HIV prevention intervention to be implemented by CBOs, organisations whose membership comprised ‘community members’ or FSWs in this case. CBOs would be supported by NGOs, organisations whose workers are not necessarily ‘community members’, but who are able to provide the CBOs with the technical expertise to run an intervention, albeit with fewer financial resources than were previously available when funded by the BMGF.

Instead of realising this vision, the transition essentially separated the intervention into two distinct phases. The intervention that preceded transition represented a CLSI approach to HIV prevention (Campbell & Cornish, 2010) with detailed attention to local context (Evans, Jana, & Lambert, 2010), where peers worked as community organisers to change conditions in the social environment that shaped the HIV risk of FSWs. What followed the transition was a TI approach characterised by peer outreach to FSWs, with greater attention to giving condom use and STI clinic information, and less attention on community mobilisation, collectivisation and forming of CBOs (NACO, 2007).

The purpose of this paper is to provide a comparative ethnographic account (Parker & Erhardt, 2001) of the work of peer workers in an HIV prevention programme for FSWs that started first as a CLSI and transitioned into a TI. We demonstrate the ways in which the changes following transition influenced not only the functionality of peer workers but also the meanings and avenues for community participation. We discuss the implications of these interventions for effective community mobilisation around HIV prevention. Our analysis is focused on peer workers as their work reveals the underlying assumptions about the sources of HIV risk among FSWs and the types of interventions necessary to address them.

Study site, methods and data sources

The data for this paper are drawn from work done by Project Parivartan, a multimethod study supported by the BMGF to analyse the implementation and impact of community mobilisation interventions to reduce vulnerability and HIV risk among FSWs. From the end of 2004 through mid-2007, four Rajahmundry-based ethnographers spent time at an NGO that was responsible for implementing the intervention in and around Rajahmundry city in the East Godavari District of Andhra Pradesh state in southern India. For the TI phase, from mid-2007 through the end of 2012, in addition to the East Godavari sites, we expanded our research to NGOs implementing the intervention in and around Nellore and Guntur towns in the eponymously named districts in Andhra Pradesh. We draw on multiple data sources over an eight-year period covering the CLSI and TI phases of the intervention.

CLSI phase

We used ethnographic data in the CLSI phase, including informal conversations with peer workers, 22 observations of peer worker monthly meetings and 40 observations of the CBO leadership group meetings. At the side lines of these meetings, we had informal conversations with FSWs, peer workers and NGO staff about different aspects of implementing the intervention. In this phase, we held two rounds of formal interviews at our research site office with purposively selected FSWs, peer workers and CBO leaders about their work and about the CBO’s involvement in implementing the intervention. Participants provided verbal informed consent before the start of the interviews and were compensated a nominal amount to cover transportation costs. Study procedures were conducted in the local language. The interviews were audio-recorded and summary reports of the audio-recordings were used in this analysis.

TI phase

For the TI phase, data are drawn from 28 observations of peer worker and intervention staff meetings. We used informal conversations with peer workers, intervention managers, supervisors of peer workers and the staff of district STI clinics, government hospitals and HIV testing centres. Additionally, we draw on 20 informal conversations with CBO leaders. Informal conversations, guided by broad questions about intervention implementation, were held at the side lines of meetings we observed, during time spent at NGO/CBO offices and government health facilities, and when we met people in public places. These conversations were not audio-taped, and participants were not compensated monetarily; ethnographers wrote detailed field notes of these conversations at the end of each day. We also draw on eight records of research team meetings. Two authors supervised data collection in both phases. Ethical approval was obtained at the Institutional Review Boards of Yale University, Duke University, American University and through YRG Care, a research institution in India.

Following the principles of content analysis (Graneheim & Lundman, 2004), one author read all the TI data, identified emerging themes and coded the data using NVivo. All authors periodically discussed emerging interconnections between codes and themes. Next, one author read the CLSI data, viewing the information from the codes and themes that emerged from the TI data. Similarities and differences within and across intervention phases were analysed. All authors discussed these themes at several meetings to reach a consensus on the emerging findings of this comparative analysis. The strength of our methodology is that we analysed what is unique about the CLSI intervention by focusing on the differences that were revealed when the intervention transitioned to a more traditional TI approach, specifically focusing on differences in the function of peer workers.

Findings

An analysis of the work of peer workers, as the intervention shifted from CLSI to TI, reveals critical differences between a CLSI and a TI approach to HIV prevention. We focus on four distinct areas of difference: assumptions about the conditions that produce HIV risk; degree of emphasis placed on the work of collective mobilising and building collective power; extent to which community mobilisation and HIV prevention goals are linked; and the intervention’s use of peer input.

Assumptions about the conditions that produce HIV risk

In the CLSI phase, peer workers organised FSWs to address a range of conditions that produced HIV risk, whereas in the TI phase they provided HIV prevention information to FSWs and motivated them to adopt safe practices. These differences signal the differing strategies adopted by the CLSI and TI approaches, as well as the different roles of peer workers within each approach. In the CLSI approach, peers were designated as Social Change Agents and conceptualised as:Three years into the intervention, peer workers said they not only distributed condoms, brought sex workers to the clinic and provided FSWs information about HIV/AIDS and STDs, but also helped sex workers deal with police interactions, spoke to district officials to obtain ration cards and housing sites for FSWs, advised FSWs about saving money and represented FSW interests at meetings of specialised committees set-up by the NGO-run intervention. FSWs also said that peer workers were ‘… working hard for our welfare … trying to decrease humiliation about sex work and create respect for us in public’.

(A) nucleus of strength in the community for bringing social change started by the community … related to human rights, legal rights … empowered to act as a torch bearer in the struggle for emancipation of the community … and advocate for them and fight against all oppressions and atrocities. (Care India, 2006)

In the TI approach, the designation ‘Social Change Agent’ was replaced with ‘Peer Educator’, which was used across all intervention sites. The function of peer workers was defined as:Using this frame, the range of activities undertaken by the peer workers narrowed. Community mobilisation continued to be featured in discourse as a central component of the TI approach (NACO, 2007), yet, our data indicate that the intervention focused most of its attention on achieving predetermined targets as set by the government. Observations revealed that the better part of peer worker meeting times was spent holding peers accountable for the achievement of such predetermined targets, including numbers of FSWs contacted, taken for clinical services and registered as CBO members, in addition to the number of free and subsidised condoms distributed through social marketing. Field notes indicated that when weekly meetings listed other agenda items such as peer worker response to violence or collection of CBO membership fees, there was typically scant discussion of these issues.

providing information on HIV, STIs and harm reduction, and promoting condom use among colleagues and peers, which ultimately resulted in building peer pressure for behaviour change. They can also distribute condoms, lubricants … and provide basic data for monitoring the project. (NACO, 2007, p. 62)

Degree of emphasis on collective mobilising and building collective power

In line with their mandate for social change, peer workers in the CLSI approach mobilised FSWs to work collectively to address shared problems. This attention to rework power relations was not emphasised in the TI approach. For instance, the CLSI intervention formed a crisis intervention team comprising intervention staff and peer workers using various strategies to leverage the CBO’s collective power to lead sustained action against police violence and harassment (Biradavolu et al., 2009). Similarly, at large public events, peer workers used their collective power to address the stigma attached to sex workers (Blankenship, Biradavolu, Jena, & George, 2010). Field notes from the CLSI phase indicate that peers networked with a wide range of people in positions of power and authority. For example, they met the District Superintendent of Police to request that police stop arbitrary arrests and the District Magistrate to avail social entitlements for FSWs. Peer workers facilitated meetings for the FSWs with the district legal aid authority to learn about their rights. They engaged regularly with district health officials to avail free condoms and facilitate access to HIV testing for CBO members and met the elected local government representative several times to register their opposition to the local government’s proposed relocation of brothels to an outlying area. These networking efforts revealed the intervention’s assumptions of sources of HIV risk and positioned peer workers and their CBO as a force for authority figures to engage with. This point is illustrated through the following report of a peer worker informing her peers about her interaction with the state’s Home Minister and the Director General of Police:This example of networking illustrates the power of a peer worker, supported by the CBO, to initiate structural change within the government.

When the Minister was arguing with me that the government will make a provision for rehabilitation, we told him that sex work is continuing from many years and many people are doing this business. How many people can the police arrest or rehabilitate? I also said that if a police officer beats a sex worker instead of simply arresting her how can the sex worker later continue sex work? … Police does not have any right to beat sex workers. While listening to the arguments, the Home Minister ordered the police officer to stop arresting women who were found carrying condoms and also to stop beating sex workers at the time of arrest.

The TI intervention approach still monitored ‘crisis’ indicators; however, FSWs reported no instances of police violence. The 24-hour-a-day crisis response system supported by the intervention was replaced by a cadre of CBO members who were trained and paid to work in crisis resolution. Furthermore, the TI approach advocated use of public resources, and FSWs were encouraged to approach the district legal aid authority for crisis resolution. One outcome of these new arrangements was that instead of collectively working on crisis resolution, the data indicate that FSWs settled crises individually. The goal of the transition to have all HIV prevention interventions for FSWs managed by CBOs and not NGOs led to various conflicts. Power struggles ensued between NGO staff and CBO leaders that managed HIV prevention interventions. A study researcher recorded her discussion with an intervention manager who said that the staff did not respect or obey the CBO President because they said, ‘“ She is from that (sex worker) community” and they don’t pay much attention to her word’. This and other instances of the eroding power and authority of CBO presidents suggest that in the CLSI approach, although there were instances of power struggles among CBO members, peer workers and intervention staff, the focus of all was to reduce stigma and increase the power of sex workers vis-à-vis external stakeholders. In contrast in the TI approach, the key power struggle reported at all sites with varying intensity was between the FSW community on the one hand and intervention staff on the other hand.

Links between community mobilising and HIV prevention goals

Based on an understanding that FSWs needed to continually engage in collective actions to reduce their risk of HIV, the CLSI approach institutionalised forums where peer workers linked the intervention’s community mobilising activities with those of the CBO. We contend that this bridging by peers reinforced the community mobilisation goals of both intervention and CBOs. In contrast, TI and CBO activities were delinked and the TI offered peer workers few social spaces where they could meet FSWs. Furthermore, in the CLSI phase, the intervention’s HIV prevention and FSW mobilisation activities were functionally separated yet linked through the activities of peer workers. Besides providing HIV prevention services, peer workers also worked to form CBOs, which the intervention described as an operational strategy for social mobilisation of sex workers. Peer workers regularly met FSWs in CBOs to discuss and strategise about how to collectively address their concerns.

Interviews with peer workers three years into the CLSI showed that they understood their work as related to both HIV prevention and community mobilisation. The following excerpt from an interview summary illustrates a peer worker’s links with the intervention and the CBO:Integration of HIV prevention and community mobilisation was also reflected in the intervention’s monthly staff meetings which routinely reviewed CBO activities, such as management of safe spaces for FSWs and formation of a clinic management committee under the oversight of the CBO.

As peer worker, the respondent gives condoms to sex workers, brings them to the clinic, tells them about HIV/AIDS … and cautions them to always use condoms. She also attends meetings of [the CBO] leadership group to discuss problems and [conveys] decisions taken in these meetings … [to] the local CBO. Also, if any sex worker faces any emergency situation, it is her responsibility to inform the [crisis response] team.

The CLSI had various platforms where peer workers could meet FSWs and plan their work for social change. Peers met FSWs in their local areas at weekly solicitation-site group meetings and monthly CBO meetings. They mobilised FSWs to participate in events organised jointly by the intervention and the CBO to mark International Women’s Day, International Sex Workers’ Day, World AIDS Day and International Workers Day. Peer workers had a dedicated monthly meeting. Alongside intervention staff, peer workers were members of various specialised committees that met regularly to work on tasks such as procuring condoms from government, running static and mobile clinics and managing the 24-hour crisis response system. These forums were designed purposefully to engender a sense of community among peer workers.

In the TI phase, the functional and organisational separation of the CBO and HIV prevention was pronounced, but not absolute. The intervention’s achievement indicators focused on what staff referred to as the 4Cs: Condoms, Clinics, Crisis and CBO. However, because peer workers were hired, monitored by and accountable to only the intervention, which now had a decidedly disease-prevention focus, their work on crises and building CBOs was mainly for the instrumental potential to affect increased uptake of clinic services and condoms. Field notes indicated that in the TI phase, peer workers from all sites reported very few crisis cases, all of which were related to intimate partner violence. Peer workers told researchers that the few FSWs who spoke of police harassment said they solved the matter privately; they did not seek peer or CBO mediation, as was the case in the CLSI phase. At peer worker and intervention staff meetings, the primary information sought was the numbers of FSWs registered at the CBOs, which then became the benchmark against which clinic utilisation and condom distribution were gauged. During this phase, there were fewer structured opportunities for peer workers to meet amongst themselves and with FSWs. In part this was a reflection of the separation of the intervention and CBO, with reduced and interrupted funds flowing to the intervention. Apart from Annual General Body meetings of the CBO, there were fewer reports of large gatherings of FSWs. Although on paper, specialised committees to coordinate intervention activities offered a forum where CBO leaders interfaced with peer workers and intervention staff, the evidence indicated that not all committees had peer workers as members and these committees met infrequently. Fewer occasions to meet translated into fewer occasions for peer workers to reinforce a sense of community among FSWs. Speaking to a researcher about this situation, one peer worker observed ‘After the takeover (transition), FSWs [who are not CBO members] lost the linkages with the CBO … CBO members were not getting any information from CBO for any meeting or event’.

In the TI approach, CBOs were tasked with mobilising FSWs for activities other than HIV prevention. CBOs created and filled new positions at the CBO to do the work previously done by peer workers. An informal norm of one person–one position, which precluded peer workers from taking on additional positions in CBOs, was enforced to develop a pool of leaders within the FSW community. Consequently, some peer workers decreased their participation in CBO activities. For example, peer workers at one intervention site did not attend a local CBO’s Annual General Body meeting because they reported it was not a part of their work. Without large gatherings and frequent local group meetings – occasions to recruit new FSWs to the intervention and reinforce a sense of community among FSWs – peer workers reported difficulty in mobilising FSWs to use services at government clinics and persuade FSWs to attend CBO meetings. For instance, when a supervisor admonished peer workers for not bringing the targeted numbers of FSWs to a CBO meeting, they replied:These words highlight that not only did FSWs refuse to associate with CBOs they perceived to not be working for their concerns, but also that peer workers too were at a loss to explain the relevance of CBOs to potential members. It must be noted that it was possibly the success of the CLSI that had brought FSWs to the point where they confidently stated that they could independently access clinic services and condoms. However, these data highlight that peer workers of a TI-focused intervention could not mobilise FSWs as easily as their peers in a CLSI-focused intervention.

So far we brought many community members to various meetings conducted by the CBO. In those meetings, except fighting and unnecessary arguments, nothing was done. If any of them [FSWs] ask us about the need and purpose of CBO then what must we say?

They [FSWs] are also asking us what CBO does for us. They say “We don’t want condoms as customers will bring those and we can go alone and get tested [for STIs]. It’s all waste of time and money [to attend CBO meetings]”.

Intervention’s use of peer worker inputs

The CLSI approach remained relevant to FSWs through a continuous feedback loop among the FSWs, peer workers and intervention managers; indeed, beyond serving as educators, peer workers were responsible for presenting FSW concerns to intervention staff. One example of this was the way peer workers routed FSW feedback to the intervention, which then revised clinic timings and procedures so that FSWs’ use of services would not be compromised. The project manager emphasised continually that it was the peer workers’ task to identify barriers that prevented FSWs from accessing treatment, and she attributed the high numbers of women accessing clinic services to the community mobilising efforts of the peer workers and CBO formation. She said, ‘Clinic service would be effective if [peer workers are] totally involved in the programme, addressing these issues and the community mobilization would be high. We find collectivization is a motivator to bring more [FSWs] to the clinic’. Later, using peer worker inputs that identified outlying areas where FSWs needed clinical services, the intervention introduced mobile clinics to those areas. Similarly, when peer workers reported at a CBO leadership group meeting that not all peer workers distributed the condoms they procured, the intervention and CBO leaders discussed solutions. Although these solutions were not always successful, the attention given to peer worker inputs by intervention staff indicates their value in providing relevant information for FSW programming.

In the TI, field notes indicated that peer workers reported they could not meet their targets, sometimes for several months, because of non-availability of testing kits or personnel at government hospitals or because FSWs in their area had migrated out. In these cases, intervention staff merely encouraged peer workers to meet targets by motivating other FSWs to attend. Field notes from the three sites indicated that peer workers repeatedly informed intervention managers and CBO leaders that their work was made tougher because FSWs were not interested in clinic services only without other concerns being addressed. A peer worker reported at a weekly meeting, ‘More than 10 times I do the visits for mobilizing community to attend meetings, medical camps but all in vain. None of the members show interest to get clinic services or attend the meetings’. However, neither the intervention nor the CBOs acted on this information to affect positive changes in their services or activities.

Discussion

The ethnographic material comparing the function of peer workers between a CLSI-focused approach and more traditional TI-focused approach to HIV prevention intervention reveals clear differences between the two strategies. We have shown that in the CLSI approach, the everyday work of peer workers went beyond just outreach and providing HIV information and prevention services, to also helping FSWs develop a sense of collective identity and power, address police violence and economic vulnerabilities and access social entitlements. The integration of the peer workers’ HIV prevention work with that of the CBO’s mandate of FSW collectivization mutually strengthened both goals. In contrast, peer workers in the TI approach had a narrow remit such that their work was effectively downgraded to outreach for STI prevention, increasing FSW uptake of clinic services and condoms and enrolling new members in the CBO. With a sharper separation between the CBO and intervention in the TI approach and limited spaces for peer workers to operate outside the clinic and condom distribution framework, peer workers remained instruments of programme implementation, efficiency and reaching pre-set targets rather than agents of social change.

Differences in the work of peer workers in a CLSI- and a TI-focused approach reflect different assumptions of the sources of HIV risk and the role of peer workers to address them. The CLSI approach recognises that the sources of HIV risk lie in structural conditions (Blankenship, Friedman, Dworkin, & Mantell, 2006; Gupta, Parkhurst, Ogden, & Mahal, 2008; Parker, Easton, & Klein, 2000) and prevention efforts need to attend to fundamental socio, political and economic complexities operating at the local sites (Evans et al., 2010). Peer workers’ work as Social Change Agents in the CLSI approach at Rajahmundry was based on the recognition that ‘… powerlessness at the community or group level, and the economic and social conditions inherent to the lack of power, are major risk factors for poor health’ (Bohidar, Jani, & Schroeder, 2008, p. 2). This perspective offered a wide scope for FSW participation; individually, peer workers became agents, and collectively CBOs became partners in the intervention’s social change agenda (Cornwall, 2003). In contrast, the TI approach viewed the sources of HIV risk primarily in individual behaviours; accordingly, it used the strategy of peer education based on the assumption that with appropriate information provided by trusted leaders and accessible preventive services, the strong influence of peer pressure to adopt safe sex norms would be an effective mode of HIV prevention (Campbell & Cornish, 2010). The radical dimensions of participatory empowerment and action found in the CLSI approach were absent in the TI approach. Although both approaches leveraged peer workers’ insider status to reach FSWs and address their risks, only in the TI approach peer workers were involved functionally and instrumentally to enlist FSWs into the intervention programmes, secure their compliance and to make projects run more efficiently by involving them in delegated responsibilities.

Community mobilising continued to exist as an important stated aspect of peer workers’ work in the CLSI and TI phases; yet in practice, the change in its form, intent and the extent to which peer workers engaged in community mobilisation reflected the importance of this tactic in each approach. Jana (2013) argues that community mobilisation is a political process which raises different types of challenges within and outside an intervention’s structure, and which takes skill and commitment to implement. The CLSI and TI approaches used the language of community mobilisation in the expectation that the process would build CBOs that would own, operate and sustain HIV programming (Jana, 2013). Yet, each approach differed on its commitment to engender community mobilisation. In the CLSI approach, mobilisation was part of a Freirian type of empowerment education (Wallerstein, 1992) process through which FSWs were made aware of their shared problems and potential power to collectively address them. The CLSI approach supported peer workers’ community mobilisation efforts by creating organisational spaces where peer workers participated in planning, decision-making, implementing and monitoring intervention activities (Avahan, 2009). In the TI approach, with a sharper separation of CBO and intervention and a narrower remit of peer workers’ activities, the extent of peer workers’ use of community mobilisation decreased to be used only instrumentally in implementing programmes and reaching targets.

Limitations

This paper has several limitations. The research strategies used for the CLSI phase and for the TI phase are different. CLSI phase data are only from one site, while the TI phase data come from three sites. While this may not pose a problem for comparative ethnographic analysis, caution must be exercised in generalising these findings to other cross-site or cross-model comparisons.

Conclusions

Implications for HIV prevention

We have shown that the scope of work of the peer workers changed when the intervention transitioned from Avahan to the government. This suggests that, despite maintaining the language of community mobilisation, the TI no longer reflects the same underlying assumptions about the sources of HIV risk and hence no longer represents a community mobilisation intervention. These findings have implications for HIV prevention programming. Having peer workers and a community mobilisation strategy in and of itself may not be enough for FSWs to access HIV prevention services; two other significant conditions appear to matter as well.

One condition is the underlying assumptions about the sources of HIV risk and the programmatic approach to address these risks, where both peer workers and community mobilisation are key elements. HIV prevention interventions may be less effective if structural sources of HIV risk are not identified and addressed and targets are focused solely on individual-level changes in HIV risk behaviours. The CLSI model reflected the view that the risk of HIV was a result of power inequalities that constrained FSWs’ ability to engage in safer sex practices. Peer workers’ mobilisation of FSWs to demand changes in contexts that promote risk was a means of challenging these power inequalities. Interventions that addressed FSWs’ vulnerabilities such as policing (Biradavolu et al., 2009), violence (Beattie et al., 2010) and lack of social entitlements (Blanchard et al., 2013) and collective solidarity (Bhattacharjee et al., 2013; Halli, Ramesh, O’Neil, Moses, & Blanchard, 2006) may raise awareness of the intervention among FSWs who are vulnerable and powerless; it is expected that this awareness may then translate into improved utilisation of the intervention’s HIV prevention services. We found that interventions that did not appear to address FSWs’ stated concerns did not appear to influence their utilisation of HIV prevention services. Moreover, as FSWs’ powerlessness is related to economic vulnerability and limited control over work conditions in sex work, interventions that collectively empower FSWs to address these concerns may reduce the HIV-related risk behaviours of FSWs (Blankenship et al., 2008).

A second condition that is important to consider is the modes of participation available to peer workers, and the communities they are tasked to mobilise in the intervention (Cornwall, 2003). The CLSI approach showed that peer workers’ community mobilisation and collectivization work hold the potential for redressing power imbalances. In the TI approach, community mobilisation was diluted of its transformative dimensions and used instrumentally only to lend legitimacy and bring efficiency to the work of interventions and its peer workers. When the two approaches are compared, evidence indicates that a CLSI approach offers better outcomes for HIV prevention than does a TI approach (Basu et al., 2004; Cornish & Campbell, 2009). This difference is attributed not only to changing the social and environmental context in which an intervention is situated (Evans & Lambert, 2008), but also to strengthening FSWs’ perceived sense of power (Blankenship et al., 2008) and building FSWs’ capacity to bring about changes (Tori, 2013). Addressing power and powerlessness in HIV prevention interventions can best be done by peer workers whose work goes beyond education and clinic referrals to mobilising their peers through collective action and addressing the social and economic complexities that put them at risk of disease.