Between March 2018 and September 2019, we used CBPR to iteratively co-create and contextually adapt a biosocial intervention to be delivered by community-based peer-navigators. We used a logic model to show how and why the components of the intervention were theorized to improve demand, uptake and retention along the HIV prevention cascade (Theory of Change), in an area of rural KwaZulu-Natal, South Africa with high HIV incidence rates.
Setting
This study was conducted in the Africa Health Research Institute (AHRI) population-based intervention area in the uMkhanyakude district of KwaZulu-Natal, South Africa. The study area of ~ 500 km2 with a population of ~ 17000 16–29 year-olds was the site of the scale-up of structural and behavioural interventions through the DREAMS partnership between 2016 and 2018 [7]. The study area is mostly rural and poorer than most other parts of South Africa, with high levels of unemployment and a high burden of HIV [1, 9, 23].
Ethics
Ethics approval was received by the Biomedical Research Ethics Committee (BREC BFC515/18) of the University of KwaZulu-Natal, South Africa and the REC (Rec numbers (5672/002) of University College London, United Kingdom.
Participatory intervention development
The process followed six steps in which through listening and responding to young people a biosocial intervention iteratively emerged (Fig. 1).
Step 1: Community engagement to identify and train youth to co-create and deliver the intervention: With the support of the AHRI community advisory board and public engagement unit (PEU), local traditional and municipal leadership from the 21 administrative areas (izigodi) were asked to identify 4–5 young men and women living in their area who were aged 18–30 years to undergo 20 weeks of training and then work 24 hours per week to support youth in their areas. The criteria for selection that were developed with the AHRI PEU were that the youth had to: be aged 18–30; have completed high school and matriculated; and be actively engaged in their communities and considered opinion leaders [24]. We strongly encouraged leadership to suggest both young men and women. Participants underwent training which covered, youth development, HIV and sexual health information, accredited HIV counselling and testing course, confidentiality, ethics, and research methods. The training had originally been planned for 16 weeks. Progress was evaluated using written and oral assessments to select ~ 55 area-based peer-navigators to co-create and implement the intervention.
Step 2: Synthesise the existing evidence of structural, behavioural and biological drivers of HIV: We used quantitative and qualitative data from the evaluation of the local scale-up of structural and behavioural interventions through the DREAMS partnership between 2016 and 2018 to summarise the structural, behavioural and biological drivers of HIV, and of engagement with the HIV prevention cascade. The evaluation data arose from: (i) surveys with a representative sample of 13-35-year-old-males and females (n = 4918); (ii) rapid ethnographic community mapping (n = 4); (iii) provider and user interviews (n = 22 and n = 58 respectively); and (iv) group discussions (n = 29). Detailed methods and findings of the contextual factors that drive risk and create barriers to effective HIV testing, care and prevention have been previously reported [1, 6, 9, 15, 25–28]. For the purposes of the intervention development the findings were summarised and converted into vignettes, case-studies and simple infographics by a team of social scientists, statisticians and clinicians who had been engaged in the data collection and peer-navigator training.
Step 3: Participatory workshop to develop the logic model for the theory of change (ToC). During a full day participatory workshop, participants were divided into mixed gender small groups of 6–8 individuals. Each group was moderated by social scientists. Participants had clear instructions to emerge from the groupwork with interventions they could implement. They used the vignettes as a vehicle to brainstorm practical approaches to mitigate the particular structural, behavioural and biological drivers of HIV and poor engagement with HIV care that the vignette signified. The candidate interventions were then plotted to a Theory of Change (ToC) (Fig. 2).
Step 4 Community entry and mapping the delivery of intervention components to each area (izigodi): The peer navigators were divided into three groups who each worked closely with one social scientist and a support team of clinicians (professional nurses) and social scientists. Each group: i) physically mapped the health, education and social services within their own communities (including introduction to the primary healthcare staff and community care givers in their areas); ii) mapped the places that young people gather; iii) identified potential adult youth champions to support them; iv) piloted the data collection tools; v) implemented the health promotion intervention under the supervision and observation of their supervisors; and vi) identified places for youth-friendly mobile healthcare services (Isisekelo Sempilo (‘Foundation of Life” in isiZulu). They then came together in a second workshop to harmonize approaches in each izigodi.
Step 5: Pilot of the intervention and process evaluation. The intervention was piloted across the 21 izigodi with some of the larger or more densely populated areas having more than one pair of peer navigators. We conducted a process evaluation of the pilot using a combination of routine data, data collected from the training and supervision and semi-structured interviews with a purposive sample of peer-navigators. We collated data from anonymised programme data records from the peer navigators’ daily reporting of their outreach activities. This included date and time, age, gender, area of recruitment, the peer navigator ID and the service provided by the peer navigator. This was supplemented with notes taken by the social science supervisors from bi-weekly training sessions, and weekly supervisory debriefings to assess the feasibility and acceptability of the intervention. Semi-structured interviews were conducted between April and August 2019 with 34 Thetha Nami peer navigators aged 20–30 years (24 female and 10 male). The interviews were conducted in IsiZulu by a different team of social scientists, audio-recorded, transcribed and later translated to English. Participants’ views were explored about the acceptability and feasibility of Thetha Nami intervention and interviews lasted between 30 to 60 minutes. Interview transcripts were managed and coded using NVIVO software and data were analysed thematically following an interpretivist approach. Initial findings reported here were derived from reflexive notes from the field to iteratively feed into step 6.
Step 6: Participatory intervention development workshops to refine the intervention and ToC. We conducted a third participatory workshop to discuss and rank the challenges to implementation and refine the intervention and ToC (Fig. 3) – in preparation for an effectiveness trial. Peer navigators were presented with the process evaluation findings and asked to reflect on their own experiences. They were divided into three moderated groups and asked to identify the challenges and rank their top three. They then deconstructed each challenge, described what its signified and brainstormed potential solutions. These were then presented pictorially and presented back to the whole group.