To decrease TB stigma, action is required at multiple levels. Group discussions with people with TB, caregivers and health workers revealed how stigma is experienced across the socio-ecological model, and guided by HCD principles, participants identified potential intervention components at each level. Intervention components across different levels shared common strategies, namely education, empowerment, engagement, and innovation, which align with intervention functions in the behaviour change wheel framework,39 and can facilitate the efficient design of effective interventions. At the individual level, proposed improvements to existing counselling included involving TB survivors, clarifying that not everyone with TB has HIV, emphasising TB is curable and can affect anyone, and providing examples of people who have overcome TB. At the interpersonal level, participants suggested TB support groups and family-centred counselling could dispel myths around TB and foster support. At the institutional level, participants recommended health worker stigma training informed by TB survivors and restructured service delivery models including home-based care. At the community level, approaches to improve knowledge and reduce misinformation included awareness-raising events led by TB survivors, incorporation of TB into school curricula, and increased outreach by CHWs. At the policy level, solutions focused on reducing the visibility generated by a TB diagnosis and resultant stigma in facilities and shifting tasks to CHWs.
These data raise questions about the optimal approach to stigma intervention design that includes intervention components across socio-ecological levels. In a recent systematic review we considered how stigma reduction interventions were implemented at each level of the socio-ecological model.21 Existing interventions – all in the pilot stage – typically targeted one or possibly two levels; for example, TB support groups or clubs (individual/interpersonal),40–42 home visits (individual/interpersonal),42,43 education for patients/caregivers (individual/interpersonal),44 health worker training (institutional),45–47 and community-volunteer led education (community)48. This highlights a tension whereby, given limited resources, trade-offs are made between interventions at different levels of the socio-ecological model. Our participants also identified this friction, with strategies targeting individuals (individuals experience stigma, therefore interventions should seek to foster resilience) often pitted against those targeting the wider family, clinic or, community (interventions should either change attitudes or reduce visibility to prevent stigma from occurring at all). For example, when trying to address hesitation about TB disclosure, some participants argued for structural changes to reduce visibility (removing identifying features in clinics such as cards/uniforms, facilitating home-based care), and others advocated that people with TB should be empowered to freely disclose their TB status (counselling and education to foster resilience).
While single-level interventions may seem more feasible from a funding or policy perspective, a synergistic approach that seeks to foster individual resilience amongst people with TB while strengthening efforts to address community-based stigma is important. This should include providing destigmatized, compassionate health care and dispelling misconceptions about TB that drive community-based stigma. This is consistent with another systematic review of health-related stigma reduction interventions, which commented that effective interventions would be oriented at multiple levels.49 Our findings align with other data suggesting peer navigators could improve patient-centred TB education and counselling.50 Further, TB survivors can play a critical role in stigma reduction efforts across multiple levels, from providing individual and interpersonal support to informing health worker training and raising community awareness, which may increase the feasibility of implementing multi-level interventions.
The impact of intersectional TB-HIV stigma across all levels of the socio-ecological model, also identified in other studies,17,51 was often emphasised by participants. The move towards greater TB-HIV care integration is aligned with the aims of Differentiated Service Delivery (DSD) for HIV treatment.52,53 While studies indicate the effectiveness of DSD for TB-HIV when considering indicators such as treatment success,54 our findings – in line with other studies55 – suggest this could exacerbate intersectional TB-HIV stigma. This has implications for calls to integrate TB with other care services such as diabetes56,57 or mental health.58–60 Other tensions related to optimal service delivery provision included the challenges of one-size-fits all programmatic approaches. While some patients may appreciate support such as home medication delivery and ongoing counselling throughout treatment, others may find this unhelpful and even condescending. There is a desire for individualised care, tailored to the needs of each person with TB and fostered through empathetic relationships with health workers. Yet resource constraints common in high TB incidence settings were apparent in our data, with health workers reporting how high workloads impact their ability to deliver high quality care. Most literature on individualised care focuses on treatment of chronic conditions in high income countries,61 with proposals for tailored TB care largely limited to individualised treatment duration.62,63 Implementation research can identify how best to target and deliver individualised care for TB, for example through risk factor screening to identify vulnerable individuals.64
Key strengths of our study include the use of HCD and CBPR. Using HCD enabled us to translate stigma experiences and service provision gaps into actionable intervention components. By talking to survivors, caregivers, and health workers in separate and combined groups, we were able to identify differing views and points of synergy. While discussions were conducted in a single clinic in Khayelitsha, Cape Town, which limits generalisability of the findings,65 this in-depth focus on a single community is the most appropriate for co-design projects, to form a long-term collaboration with affected groups in a specific community and design an intervention tailored to their specific needs.66 Although the use of peer research associates raises the possibility that their lived experience may influence discussions, their involvement improved rapport with participants, understanding of local references, and contextualisation of their narratives.36 We thus sought to analyse findings reflexively, bearing in mind the positionality of facilitators and researchers.67