Community mental health in South Africa needs participatory and liberatory practice, not only task-sharing

Introduction

When people in South African health services say “community” mental health, they usually mean care that has been moved closer to where people live. Clinics, generalist nurses, lay counsellors and the occasional visiting specialist. That is a real achievement in a country where psychiatric care was concentrated in tertiary hospitals, and it is also a fairly thin version of what community psychology set out to be. The service-delivery picture is bleak in places. In a study of five primary clinics, a day-care facility and three hospitals in a rural part of the Eastern Cape, frontline staff described heavy workloads, no on-site psychologists or psychiatrists, and a heavy reliance on sedating medication (Rall & Swartz, 2025). Reviewing the wider literature, Truter (2023) found barriers that recur across the South African studies, among them staff shortages, informal referral and weak feedback between levels of care.

My argument is that task-sharing and integration, useful as they are, cannot stand in for the participatory and liberatory tradition that South African community psychologists have built, and that the two need to be read together. Task-sharing asks how existing clinical work can be spread across more hands. The liberatory tradition asks who defines the problem, who benefits from the answer, and what structures keep producing distress in the first place (Ratele & Malherbe, 2022; Malherbe & Canham, 2024). Those are different questions. If only the first one gets asked, community mental health risks becoming individual-focused treatment delivered more cheaply, with community members as recipients rather than participants. The evidence is uneven, and I say where it is thin. Still, taken together (Davies et al., 2022; Kathree et al., 2023) it points one way. The paper begins with what the liberatory tradition claims, turns to how community-based care is actually working, then looks at lay and community workers, and finally at the few participatory efforts that have been documented.

What the liberatory tradition claims

Community psychology in South Africa has never been only a method for taking therapy to more people. Ratele and Malherbe (2022) ask what a liberatory, Africa-centred community psychology would look like, and they build the answer partly from Neville Alexander, a political thinker and activist who had no formal link to the discipline. Their point, as I read it, is that psychological work in communities should be drawn from Africa’s own histories, politics and cultures and should sit inside wider efforts at social change, rather than being imported whole and applied. That is a claim about where knowledge comes from (Malherbe & Canham, 2024), which is not a point individual-focused models usually bother with.

Malherbe and Canham (2024) show what this looks like in practice by reading the archive of Mohamed Seedat, whose community work stretched across almost forty years and was rooted in the working-class south of Johannesburg. The archive documents community praxis under apartheid and after it, and the authors treat it as evidence of how a liberation psychology actually got done. One thing they stress is that the field is loosely bounded and thinly resourced, and that many practitioners never used the label. That matters for anyone teaching or studying the field, because it means the tradition has to be reconstructed from practice, not looked up in a single textbook definition. It also hints at a weakness. A tradition that lives in archives and scattered projects is hard to measure, and thinly resourced work can struggle to sustain itself, as the advocacy evaluation discussed later suggests (Malherbe & Canham, 2024; Davies et al., 2022).

A third strand concerns structure. Malherbe et al. (2025) argue that community violence is badly understood when it is treated as a set of observable incidents. Drawing on critical realism, they propose looking at how visible violence intersects with epistemic violence, meaning harmful and inaccurate representation of people, and at the structural mechanisms behind both. Their worked example comes from a low-income South African community whose members made and screened a documentary about community violence and collective resistance, and used it to engage xenophobic violence among other things. Whatever one makes of the framework, it models the kind of move that individual-focused psychology rarely makes, which is to ask what social arrangements generate the thing being counted and who gets to describe it (Malherbe et al., 2025; Ratele & Malherbe, 2022).

Community-based care as it is actually delivered

Set beside this, the daily reality of community-based services looks rather different. Rall and Swartz (2025) interviewed 14 frontline health workers in the Dr Beyers Naudé Local Municipality in the Eastern Cape between March and July 2022. Staff told them that mental health had no clearly responsible person once it was folded into general care, that patients waited long and sometimes dropped out, and that clinics got little or no feedback after referring someone to hospital. Specialists were often reached by phone and never saw the patient. Sedating medication was basically the main tool available. The sample is small and comes from one district, and the first author was also the only clinical psychologist for the provincial health department in the area, which the authors acknowledge may have shaped the study. Even so, the picture is hard to wave away because it matches what other work reports (Truter, 2023).

Truter (2023) reviewed 25 South African studies published between 2002 and October 2021 and sorted collaboration into four levels: between professionals, within communities, across sectors, and between Western and traditional or faith-based healing. Most of the research sat at the first two levels. Very little examined intersectoral work or collaboration between healing paradigms, which seems a gap given that Truter lists it as one of the four levels. The barriers she lists include unequal power among professionals, unclear roles, competition for resources that encourages working in silos, and stigma. That list is worth pausing on, because it is mostly about power and organisation, not about the technical content of treatment. Her recommendation that conditions for collaboration be checked before a model is rolled out is sensible, and it is also, in its own right, a participatory instinct (Truter, 2023; Rall & Swartz, 2025).

There is a political reason these patterns persist. Van Rensburg (2021) examines the post-apartheid reforms that shaped public, community-based services for people with severe mental illness and argues that those services have been moulded by the relations among the state, private for-profit health care and civil society. His treatment of the Life Esidimeni tragedy reads it as the product of corruption, administrative mismanagement and a power imbalance in which vulnerable people were treated as commodities, and he links that commodification to neoliberal ways of governing poverty and severe mental illness. This is one author’s reading of the policy record, so it works better as a framing than as proof. What it does suggest is that the difficulties in Eastern Cape clinics are not simply local failures of management, and that a community psychology which stays at the level of the clinic will miss part of the explanation (van Rensburg, 2021; Truter, 2023).

Task-sharing and the people who carry it

Task-sharing is the most credible response to the shortage of specialists, and the best evidence comes from studies that look at it closely. Kathree et al. (2023) followed 627 chronic care patients who screened positive for depression in 10 primary care clinics in the Amajuba district of KwaZulu-Natal, where nurses assessed and diagnosed, and existing HIV counsellors delivered a manualised counselling intervention. Nurses did not diagnose depression in 42% of patients and diagnosed without referring in another 23%. Patients who were diagnosed and referred had lower depression scores at three months, with roughly twice the odds of a symptom reduction of more than half, but the difference had gone by nine months. The authors suggest that spontaneous remission among mostly mild to moderate cases is part of the reason (Kathree et al., 2023).

Read carefully, the Amajuba study supports task-sharing and complicates it at the same time, a pattern that Truter (2023) also reports. It supports it because a model built from nurses and existing counsellors could be run in ordinary clinics. It complicates it because the design was observational, follow-up was incomplete at around two thirds of participants, and the counsellors had competing duties that undermined both referral and counselling. Nobody was referred to a doctor for antidepressants. So the bottleneck turned out to be the organisation of work around the clinic, which no amount of counsellor training addresses by itself. That fits the reports of unclear roles and weak feedback in other work (Kathree et al., 2023; Truter, 2023; Rall & Swartz, 2025).

There is also a question about the workers themselves. Marindi and Batisai (2025) studied 20 women doing community home-based care in Soweto and describe a workforce that is mostly Black, working-class and female, in precarious jobs where unfair labour practices and poor organisation have led to strikes and unionisation. The women were nonetheless hopeful, and the authors read their optimism as a form of self-empowerment and agency. The authors also call for the gendered hierarchies of care work to be dismantled and not merely described. If community mental health expands by leaning harder on workers of this kind, then it is drawing on the very structures of race, class and gender that liberatory psychology wants to examine. Calling that expansion empowerment, when it may be cheap labour in a different guise, would be a stretch (Marindi & Batisai, 2025; van Rensburg, 2021).

What participation has and has not delivered

It would be unfair to claim that participatory work has solved what service delivery cannot. The documented cases are few and mostly small. Peters et al. (2022) used photovoice, as a form of participatory action research and decolonial praxis, with young men who identify as coloured in Bishop Lavis, Cape Town. Working through interviews, photo-narratives and a focus group, they set out to ask how the men construct their community as raced, classed and gendered, given that existing stereotypes link such communities to violence, drugs, gangsterism and alcohol. The chapter reports that the men’s experiences were shaped by race, age, class, gender and place together, which complicates the single-story images that individual-level research can reproduce. The chapter gives little basis for judging how far anything changed in Bishop Lavis afterwards (Peters et al., 2022; Malherbe et al., 2025).

The national advocacy programme for mental health care users gives a more sobering test of empowerment. Davies et al. (2022) interviewed 18 care users, 9 service providers, 6 support persons, 4 NGO directors and 4 departmental managers about a programme run by the national Department of Health and the South African Federation for Mental Health. Participants valued the information in the empowerment sessions, yet very few advocacy groups were set up afterwards. Reasons given were a lack of follow-up support, stigma from communities and from health workers, low self-confidence and a lack of money. Facilitators included having mental health champions at clinics and funding NGOs. In other words, a single round of empowerment sessions did very little without sustained resources, which is hardly surprising, but the finding is a useful corrective to talk of empowerment as if it were an attitude people could simply acquire (Davies et al., 2022; Marindi & Batisai, 2025).

Put the two cases next to each other and an uncomfortable asymmetry shows up. The photovoice and documentary projects (Peters et al., 2022; Malherbe et al., 2025) are rich on meaning and weak on demonstrable outcomes, and the advocacy programme is the reverse, with structure and institutional backing but little lasting participation. Neither has the evidence base that a trial of counselling can show. That asymmetry is probably one reason participatory work struggles for funding, and it is a reason to be careful about the claims being made for it. The honest position, I think, is that participation has shown it can change how problems are described and who describes them, and has not yet shown it can reliably change services (Davies et al., 2022; Malherbe & Canham, 2024).

What the South African context changes

Three features of the context make this argument sharper here than it would be in many other countries. The first is history. A psychology that grew up partly inside apartheid has a particular obligation to ask whose interests its models serve, and Malherbe and Canham (2024) show that a body of community practice answering that question did exist, even if it was patchy and under-resourced. The second is inequality. Services for people with severe mental illness have been shaped by the relationship between the state, private providers and civil society, and the Life Esidimeni tragedy shows what can happen to people when that relationship goes wrong (van Rensburg, 2021). The third is the density of violence, including xenophobic violence, which Malherbe et al. (2025) treat as something communities themselves analyse and resist, not only suffer.

What follows for practice is modest but specific. A community-based service that takes the liberatory tradition seriously would involve users and community workers in deciding what counts as the problem, treat the conditions of lay workers as part of its remit, and keep the structural questions about funding and governance in view even when the immediate task is a referral pathway. Truter’s (2023) suggestion of pre-intervention research could be done participatively, so that communities help establish whether the conditions for collaboration exist (Truter, 2023; Kathree et al., 2023). None of this needs new buildings, though it does take time and a willingness to hear things that complicate the plan.

Conclusion

Community mental health in South Africa has two traditions that do not always talk to each other. One is a health-systems tradition of integration, task-sharing and collaborative care. The other is the participatory and liberatory tradition of community psychology. The evidence reviewed here suggests that the first is necessary and incomplete: it shows that nurses and counsellors can extend care (Kathree et al., 2023), yet it keeps running into referral gaps, thin supervision, and precarious labour that it did not create and cannot fix alone (Rall & Swartz, 2025; Marindi & Batisai, 2025). The second supplies the questions the first leaves out about power, representation and structure (Ratele & Malherbe, 2022; Malherbe et al., 2025), but it has a thin record when it comes to measurable change (Davies et al., 2022).

My position, then, is that the two should be judged by each other’s standards (Truter, 2023; Malherbe et al., 2025). Participatory work needs to say what it changed, and task-sharing needs to say whose problem it is solving and who is carrying the load. A lot of this rests on small studies, a few qualitative accounts and edited-book chapters, so the conclusion is a cautious one. Even so, it seems hard to defend a version of community mental health in which communities are served but not consulted, and the South African literature gives good reason to expect that version to disappoint (Truter, 2023; Peters et al., 2022; van Rensburg, 2021).

Reference list

Davies, T., Roomaney, R., Lund, C., & Sorsdahl, K. (2022). Evaluation of an advocacy programme for mental health care users in South Africa: A mixed methods study. Community Mental Health Journal, 58(4), 720-728. https://doi.org/10.1007/s10597-021-00877-8

Kathree, T., Bachmann, M., Bhana, A., Grant, M., Mntambo, N., Gigaba, S., Kemp, C. G., Rao, D., & Petersen, I. (2023). Management of depression in chronic care patients using a task-sharing approach in a real-world primary health care setting in South Africa: Outcomes of a cohort study. Community Mental Health Journal, 59(7), 1261-1274. https://doi.org/10.1007/s10597-023-01108-y

Malherbe, N., & Canham, H. (2024). Reading a liberation psychology archive in South Africa. South African Journal of Psychology, 54(4), 475-487. https://doi.org/10.1177/00812463241267295

Malherbe, N., Seedat, M., & Suffla, S. (2025). Understanding community violence: A critical realist framework for community psychology. Journal of Community Psychology, 53(1), Article e22660. https://doi.org/10.1002/jcop.22660

Marindi, I., & Batisai, K. (2025). Optimism and resilience among the precariat: A gendered analysis of community home-based care work in South Africa. Home Health Care Services Quarterly, 44(1), 1-25. https://doi.org/10.1080/01621424.2024.2421537

Peters, S. M., Kessi, S., & Boonzaier, F. (2022). Constructing race and place in South Africa: A photovoice study with ‘coloured’ men in Bishop Lavis. In S. Kessi, S. Suffla, & M. Seedat (Eds.), Decolonial enactments in community psychology (pp. 139-166). Springer. https://doi.org/10.1007/978-3-030-75201-9_8

Rall, D., & Swartz, L. (2025). ‘The community lives on sleeping medication and antidepressant[s]….’: Health care workers’ experiences of mental health service provision in rural South Africa. PLOS Mental Health, 2(11), Article e0000350. https://doi.org/10.1371/journal.pmen.0000350

Ratele, K., & Malherbe, N. (2022). Liberatory Africa(n)-centred community psychology of psychosocial change. In S. Kessi, S. Suffla, & M. Seedat (Eds.), Decolonial enactments in community psychology (pp. 13-27). Springer. https://doi.org/10.1007/978-3-030-75201-9_2

Truter, Z. M. (2023). Collaborative care for mental health in South Africa: A qualitative systematic review. South African Journal of Psychology, 53(1), 18-31. https://doi.org/10.1177/00812463221093525

van Rensburg, A. J. (2021). The political economy of mental illness in South Africa. Routledge. https://doi.org/10.4324/9780429201455