‘Leaving no one behind’: reflections on the design of community-based HIV prevention for migrants in Johannesburg’s inner-city hostels and informal settlements

A relatively young city, Johannesburg is now the economic hub of sub-Saharan Africa, with an estimated 4.4 million residents [21]. Since the end of Apartheid, migration patterns have shifted as national borders have opened up and restrictions removed on the movement of South African nationals into and within urban spaces [22, 23]. Roughly three quarters of the city’s population are black African, and English and isiZulu are the two most widely spoken languages [21]. Johannesburg is now an increasingly cosmopolitan city: in 2008, around 14% of its population were estimated to be foreign-born, mainly Zimbabwean, Mozambican and Nigerian [24]. The inner-city, in particular, is home to a heterogeneous population of internal and cross-border migrants, as well as refugees and asylum seekers from across the African continent and beyond [25, 26]. Single-sex hostels in the inner-city provide an entry point to urban life, and are an important source of cheap housing for men (and, more recently, women and children) who have moved to Johannesburg to seek work [27]. Originally created by mining companies to accommodate men from rural areas at minimal cost, hostels were part of an Apartheid strategy to limit the urbanisation of black people from rural Bantustans. They have been described as “institutions of incarceration” [28], for their barrack-style environments that were – and remain – typically overcrowded and lacking in privacy [27].

Our study focused on hostels in the Benrose area, in ‘Region F’, south-eastern Johannesburg. These hostels were initially located on the outskirts of the city, but increased urban growth has made them much more centrally located [29]. Following mine closures in the early 1990s, the departure of former employers and the subsequent decay of management structures, the hostels are now occupied by unemployed migrants rather than miners [9]. Over time they have become increasingly unstable, unsafe and unsanitary, characterised by high levels of crime, violence and a lack of functioning amenities. Upgrading programmes devised by the Provincial government, including the transformation of hostels into family units, have had mixed success [30].

Prior to developing the intervention, formative research activities were undertaken to better understand the demand- and supply-side barriers to delivering HIV prevention activities within this community. This included community mapping, a baseline survey, client-simulation exercise in local public sector clinics, and interviews with traditional healers operating near the hostels.

Community mapping was carried out in 2003–4 to understand the layout of the study setting and the types of settlements occupied by residents. This was supplemented by participant observation by a trained fieldworker who spent several nights in one of the hostels. Together, these activities provided valuable information on the physical structure and layout of each hostel, as well as the gender dynamics and nature of everyday life within the hostels. It also brought to light the scale of informal settlements located on the periphery of hostel borders, and the complex networks and relationships forged between hostel residents and their neighbours in these settlements.

Supply-side issues were elucidated through simulated client visits in 2004 to public sector services located in the vicinity of the hostels. Staff at selected health facilities were informed in advance that simulated clients would be visiting them over the course of the subsequent three months, presenting with the following scenarios: (1) symptoms of an STI; (2) a request for voluntary counselling and testing of HIV (VCT), and (3) enquiries about antiretroviral therapy (ART) for a sick relative. Immediately after their clinic visits, these clients were interviewed in isiZulu by a trained fieldworker about the encounter.

Traditional healers previously identified during community mapping activities were included in the assessment of local health services if they were recognised by community members as legitimate traditional health practitioners, provided services to residents of the hostels and informal settlements, and had experience with providing care for STIs and/or HIV-related complaints. Semi-structured, in-depth interviews with these healers were undertaken in isiZulu by a trained field worker, and were audio-recorded, translated and transcribed.

Our initial intention was to design and implement a peer-education intervention to improve knowledge of STIs and HIV, and increase safer sexual practices among male hostel residents. As results from the formative research made clearer the demographic and health status of the communities in this setting, we began to ask what priorities community members would themselves identify as the basis for an intervention, if given the chance to articulate them. A brief participatory exercise was undertaken to elicit and capture these priorities. Community members from both the hostels and surrounding shack settlements were recruited with the assistance of a community NGO partner working in the hostels, to participate in a pair-wise ranking exercise [31]. In a series of six sessions, participants were placed in groups and asked to record problems faced by their community on blank cards, one problem per card. The list of problems generated by this activity was then read out by a representative from each group. All participants present proceeded to group these problems under ten main categories. Each card was placed next to an appropriate symbol for each category, and comparisons between the categories were made. Participants then generated a final priority list by ranking each of the ten categories according to the number of needs recorded, and by expanding, clarifying or consolidating some of the categories. Results from this exercise were fed directly into the process of intervention design, alongside findings from the formative research phase.

Once implementation of the intervention was complete, a process evaluation was carried out to assess fidelity to project planning, and to identify major strengths and weaknesses. This involved an evaluation of actual outputs against a matrix of planned activities and key project objectives for selected components of the project. The ‘Quality Living Clubs’ (QLC) – the only component described here in any detail – were evaluated using a variety of methods, including analysis of attendance registers; use of an attendance tracking tool; assessment of module content, and exit interviews with QLC facilitators and 123 of the 148 ‘graduates’ of the QLCs (those who completed at least one module, i.e. two out of three learning sessions and one out of three reflect circles). Interviews were recorded and transcribed before being entered into Epi-Info, coded, and imported into an Excel spreadsheet for analysis of basic statistics. Qualitative portions of the interview data were reviewed by the research team and thematically analysed to triangulate quantitative findings, particularly on components more difficult to evaluate, namely, fidelity and implementation. Illustrative quotes were also identified from the interview data to highlight key findings.

Mapping activities confirmed the location and size of six hostels in the intervention area (at the time of the study, four were administered by the municipality and two privately), and identified specific localities within this area associated with high levels of substance abuse. Importantly, the mapping led to a revision of population estimates from an initial figure of 7000 residents to around 30,000. At the time of the intervention, Denver, Jeppe and George Goch Hostels each had between 3200 and 3700 registered residents, but were estimated to house as much as three times as many men (see Fig. 1). Demand for housing adjacent to the hostels was high: abandoned buildings and warehouses located nearby had been taken over by slumlords and subdivided for letting, while people were living in an abandoned sports stadium in the area and in shacks built on former factory plots. These shacks and other informal forms of housing had emerged partly as a result of hostel traditions enforced by the gatekeepers – local indunas (IsiZulu: ‘headmen’) – which did not allow women and children within the strongly male-dominated environments of the hostels. These prohibitions were strictly enforced and, according to local informants, women could be raped or assaulted if they were found inside the hostels. Consensual sexual activity with women – i.e. sex within the context of a relationship, or sex with a sex worker – was also reportedly prohibited within the hostels, leading many male hostel residents to construct informal housing nearby for their wives and partners. Our survey showed that around two thirds of men had family members either living in the hostel with them or in neighbouring informal settlements and that despite several years of residence in Johannesburg, both populations still overwhelmingly regarded the city as a temporary location, and saw their ‘home’ as being outside the province. Overwhelmingly, most of the men (94%) were from KwaZulu-Natal province, and described themselves as being of Zulu ethnicity. Half reported travelling to their rural homes at least once a year. Around three quarters of women (76%) were also internal migrants from KwaZulu-Natal (76%).

Hostel life was largely defined by overcrowded living conditions: 70% of the men surveyed shared a room with at least five other men, and as much as 44% shared with ten or more. Water and sanitation facilities were often lacking, with toilets reportedly unusable at least once in a month. Informal settlements surrounding the hostels were also described as overcrowded, and lacking in adequate sanitation and electricity (see Fig. 2). The latter problem in turn created a demand for alternative cooking fuels, such as paraffin, which posed a potential fire hazard. On average women had not lived in the informal settlements as long as men had been in the hostels: only a third of women had lived there for more than 4 years. Around half of the men (51%) had at least one child, but both men and women reported having to provide for a number of financial dependants on what was, for most, a meagre and erratic income. Some small businesses, such as car maintenance and shoe repairs, had sprung up in these informal settlements, alongside spaza shops (small informal grocery shops), shebeens (taverns), restaurants and butcheries. Seventy-six per cent of women interviewed and 56% of men stated that they were not working.

Concurrent sexual partnerships were common: at least 47% of men claimed to have more than one sexual partner at the time of the survey, and a notable portion of these men said they had exchanged money (5%) or gifts (10%) for sex. Only 2% of women from the informal settlements reported having been paid for their most recent episode of sex and none identified themselves as “sex workers”. Eighty nine percent of women had boyfriends, who were generally two or more years older than themselves. As measured by biological samples taken during the survey, more than twice as many women (56%) were found to be living with HIV than men (24%). These figures were considerably higher than national HIV prevalence estimates for 2004, the year of the survey (16.6% of men and 20.2% of women aged 15–49) [32], and, indeed, higher than more recent surveillance data from 2012 (14.5% and 23.2% respectively) [7].

Consuming alcohol and marijuana, playing cards and gambling, illicit activities and crime were described as part of the fabric of life in the hostels, with “fighting” and “shooting” reportedly commonplace. The gender dynamics of local interactions and women’s comparatively greater vulnerability to violence were clearly evident: a third to a half of all men considered their community to be unsafe (47%), compared to 70% of women. Around a quarter of women reported ever being beaten by their partner or husband, 11% had ever been threatened into having sex, and 8% had ever been raped. In addition, 6% of men acknowledged that they had ever used physical force against a woman to have sex. Despite these perceptions of a lack of personal safety, residents expressed a strong sense of belonging and community cohesiveness both within the hostels and informal settlements. They described feeling “alienated” from the broader city of Johannesburg, and men said they felt closer to other hostel residents – most of whom were also migrants from the same province – than to people in other parts of the city.

In terms of accessing healthcare, public-sector clinics, health centres and hospitals were most commonly consulted by women and men alike, although men were twice as likely to seek private healthcare as women (14% vs. 7%). Women generally had greater familiarity with available services for STI treatment and HIV testing; four times as many women reported having had an HIV test, compared to men (41% vs. 10%). More than half of women used no contraception, and consistent condom use was also low, with less than a quarter of men and women reporting use of a condom at last sex. Sexual decision-making was perceived by more than three-quarters of women to be the prerogative of their male partners.

During the assessment of available health facilities, ten government health units were identified within a 2 kilometre radius of the hostels in the study area. In addition, three pharmacists, nine private general practitioners and ten traditional healers were practising in the area. A total of 44 simulated client visits were completed by 12 trained hostel and informal settlement residents (six male, six female) who visited nine public sector health facilities. These visits showed that the greatest challenge to the provision of quality health services in the inner-city of Johannesburg was human resource capacity, in which a lack of qualified staff in clinics led to prolonged waiting times. Attitudes of health care providers to clients were found to vary with the reason for the consultation. Clients enquiring about HIV testing or ART received better care than those presenting with the STI scenario, suggesting that healthcare providers may have looked more favourably upon those proactively seeking health information, and that patients with STI symptoms were likely to be stigmatized. There was a general failure to follow national guidelines on STIs, including the physical examination of patients reporting STI symptoms. Further problems included inappropriate questioning of clients by untrained staff at clinic reception areas and the absence of particular services on some days of the week.

Interviews with traditional healers confirmed that the majority of clients visiting these healers came from the neighbouring hostels and surrounding informal settlements. The healers reported treating male clients primarily for complaints of impotence, backache, “doropo”¹ (gonorrhoea), swollen feet and “idliso” (bewitchment, or magical poisoning). In most cases, patients who were believed to be living with HIV were allegedly referred to mainstream public sector clinics for care.

Prior to the collection of formative data, our intention had been to devise an intervention to improve knowledge and awareness of STIs and HIV, and increase safer sexual practices (e.g. partner reduction, increased condom use, HIV testing and seeking symptomatic STI treatment) among men in urban hostel settings. In addition, it was envisaged that the programme would strive to increase residents’ utilisation of primary healthcare (PHC) services for Sexual and Reproductive Health (SRH) needs. Collectively, however, the formative research findings necessitated a radical revision of these objectives.

The mapping exercise, for example, led to an important adjustment of the study parameters. It identified a large population of mostly migrant women living in informal settlements surrounding the hostels, and largely invisible to city authorities. This meant that the intervention could no longer engage male hostel residents alone. Survey findings indicated that local gender dynamics and women’s precarious living arrangements rendered them particularly vulnerable to violence, HIV infection and poor health – suggesting that the intervention needed to approach HIV as part of this nexus of vulnerability and recognise its spatial dimensions. The client-simulation exercise revealed that both women and men experienced poor quality SRH services, thanks largely to inadequacies in staffing and training of health-care workers. There was apparently low demand for these services, as shown by low uptake of contraception among women and low HIV testing rates by all.

While we anticipated that the built environment and social context of the hostel setting would shape intervention design, what was unexpected was the extent to which active involvement of local communities would impact on project scope and activities. Through the participatory pair-wise ranking [31] exercise undertaken with community members at the close of our formative research phase, we became aware of the contrasting priorities of men and women living in this setting – and the relatively low priority afforded to HIV (see Fig. 3).

These divergent priorities were driven to some extent by their different living conditions. For men in the hostels, water, sanitation and housing were not as much of a concern as they were for informal settlement residents, and where men prioritised the need for jobs, women identified poverty alleviation more strongly as a priority. To some extent, this may reflect local gender roles, in which men are traditionally the (employed) providers, while women are expected to care for the home and family – roles that are directly impacted on by inadequacies in housing, sanitation and access to water. Women also identified a broad range of everyday issues (included in the “other” category) as concerning, including alcohol abuse, absent or unreliable electricity services and shack fires. In spite of the relatively high HIV prevalence in this community (particularly amongst women), both women and men identified HIV as less of a priority than general health concerns. Instead, most wanted action to be taken on what they saw as more pressing issues associated with daily survival: principally, employment and financial stability.

Incorporating this local input meant that – to be relevant – the programme had to shift from its core concern with HIV to a broader focus on community development and health. During project planning, the study team had thought to use peer-education to promote safer sex practices, but these community-driven insights highlighted the inappropriateness of such an activity in this setting and context. The objectives of the intervention were therefore re-conceptualised to focus more broadly on increasing livelihood security and wellbeing, rather than on promoting safer sex practices and improving SRH services alone.

Intervention design was partly influenced by the concept of the poor collectively acting to break down structural inequalities and build a ‘capacity to aspire’, as formulated by Appadurai [33], together with Freire’s notion of learning through the creation of a ‘critical consciousness’ [34]. In practical terms, we also drew heavily on a Sustainable Livelihoods Approach (SLA), which primarily aims to expand the level of control that participants have over their lives. The SLA presupposes that individuals occupy a position on a livelihood continuum ranging from surviving (where choice and negotiation are absent) through coping (where short-term solutions to newly identified challenges are possible) to aspiring (which involves planning for the future) [35, 36]. As individuals move along this continuum from surviving to aspiring – although in theory they could also move in the opposite direction, depending on circumstances – they expand their ‘capacity to aspire’ [33]. Accordingly, the level of control over their own lives increases, which in turn builds assets and enables a greater ability to cope with social challenges (see Fig. 4). It was hypothesised that participation in activities to promote health and long term well-being would enhance community members’ resilience to the shocks and stresses associated with migration and urban living, and ultimately enable them to actively engage with HIV and other SRH interventions. In other words, the route towards promoting and entrenching safer sex practices was neither as direct nor as explicit as initially envisaged.

Ultimately, the intervention comprised three related components and was implemented in 2005–6. These focused on multi-sector, health sector and community sector activities respectively: (1) developing a coordinated response to the community’s diverse priorities through strengthening working relationships between the City of Johannesburg, the Gauteng Provincial government and local NGOs; (2) training healthcare workers to deliver quality, relevant SRH services – in particular, STI treatment; and (3) formation of community-based ‘health clubs’ (known as Quality Life Clubs, or ‘QLCs’) to serve as a central vehicle for driving change in the intervention communities. Only the third component is detailed here, as developing these health clubs drew most heavily on participatory approaches, and embodied most clearly the shift in project focus.

Based on a fusion of several innovative methodologies used elsewhere, including “study circles” [37], community health clubs [38, 39], and the “Reflect” methodology [40], the QLCs developed within Mpilonhle-Mpilonde explicitly acknowledged the local contexts of poverty, social marginalisation and the particular vulnerability experienced by urban migrants. Similar clubs have been implemented in rural settings in both Zimbabwe [38] and Sierra Leone [38], as a long-term strategy to enhance people’s control over the determinants of health. In our intervention setting, QLCs were established as voluntary, community-led structures to encourage change through a process of participatory learning and reflecting [41]. In a series of carefully structured “learning sessions” facilitated by a trained, expert facilitator from partner organisations (Idasa, Soul City and The Greenhouse Project), community members identified challenging and problematic aspects of their environment, and collectively formulated responses to bring about change.

Ideally, QLCs would meet weekly over a six-month period for group learning sessions, and then gather later in the week in smaller “reflect circles” [37, 40] to discuss and reflect on what they had learnt that week. To facilitate positive group dynamics, discussions of community needs began with practical, non-sensitive topics, such as environment and sanitation, or food and nutrition, and only moved on to more sensitive topics – such as gender roles, violence or HIV – when club members had acquired sufficient confidence and group management skills to actively engage with these more complex issues. While some reflect circles focused largely on individual capacity-building, like developing curriculum vitae writing skills, others were devoted to broader community-focused activities, such as neighbourhood clean-up campaigns, community gardening, support for raising issues with local authorities, and community mobilisation for HIV ‘Know Your Status’ campaigns. Yet other reflect circles made income-generating activities the central focus. These included building skills to make household goods and crafts for re-sale, and support for initiating small-scale catering businesses. In a community made up largely of unemployed migrants, such initiatives became especially important ways of fostering a sense of security, positive community identity, and hope for the future.

Between February 2005 and July 2006, three rounds of this intervention were facilitated by research staff and partners, involving 479 community members. By December 2006, a fourth round was implemented by community members who had been mentored by the research team in earlier rounds. Interestingly, modules focusing on HIV had the highest attendance over the course of the three rounds monitored (652 participants attended at least one learning session or reflect circle, indicating that some participants actually attended these modules more than once), closely followed by those discussing gender and violence (553 attendants; some attended more than once). Women formed the majority at both learning sessions and reflect circles regardless of topic, possibly reflecting local gender norms that assume men will take ownership of projects as leaders, while women will make up the bulk of ordinary participants and volunteers in grassroots community development [20]. Gender also featured strongly in the breakdown of those attending sufficient sessions to qualify for ‘graduation’: 83% of graduates were women.

A key motivation for joining the QLCs, cited by most of the QLC graduates interviewed for the process evaluation, was a desire for knowledge acquisition, particularly knowledge on health issues. One fifth of the members interviewed said that the most important thing they had gained was a sense of empowerment, while 86% of graduates had noted a change in their lives as a result of participation in the programme. One young woman reflected that “I see things differently now, and my behaviour and lifestyle has changed” (DIS096, aged 25), while a male participant said, “[A] lot of things has changed, my behaviour, the way I think and talk” (JO29, aged 37). Testimony from facilitators and participants alike suggested that women in particular had grown in confidence as a result of their participation in the QLCs. One facilitator of the reflect circles at George Goch Hostel described the transformation he had witnessed in a woman attending these circles. Initially, “…she was very shy, with no sign she can engage. It is amazing how her confidence has increased”. Other female participants commented, “In Reflects, I have learnt to speak out my mind” (DIS021), and “I am able to do some of the things that I was not able to do before I joined Quality Life Clubs” (JO23, aged 36).

Finally, an important question posed of the QLCs in the process evaluation related to their sustainability. While six QLCs had been initiated in 2005, and peer-to-peer recruitment campaigns had even sprung up to encourage other residents to join the QLCs, by 2006 these had been consolidated and only three clubs remained operational. Membership decreased between rounds one and two for all of the QLCs, and some clubs ceased to function altogether, mainly owing to participants leaving to pursue employment opportunities or volunteer in government programmes. There were also challenges in securing a regular meeting venue. Countering these apparent signs of non-sustainability, to some extent, was the fact that at least seven graduates were known to have started their own QLCs after the programme ended. In one instance, there was even a ‘ripple effect’ felt beyond the study setting, as one male migrant exported the concept of QLCs back to his hometown in KwaZulu-Natal shortly after completing the Mpilonhle-Mpilonde programme. In the intervention site itself, QLC members took over the programme after the third round, and launched their own community-based organization during the graduation ceremony. This organisation, named ‘Sipho Esihle’ (‘Precious Gift’) was formed to provide networking support for community projects trying to create sustainable, healthy livelihoods. Overall, as community driven structures, the QLCs and the derivations they later spawned, demonstrated a clear potential to create a sense of community and to build demand for health and civic engagement more broadly.