Early identification and management of mental illness in childhood and adolescence is important for averting debilitating mental illness in adulthood [1]. Child and Adolescent Mental Health (CAMH) refers to a range of mental, neurological and substance use (MNS) disorders that occur in childhood and adolescence. Mental health has been a neglected global health area and the attention given to CAMH has until recently been disproportionately low, compared to mental disorders in adults and the elderly [2]. As a result, a huge treatment gap for CAMH conditions persists, one that could be addressed by improving access to CAMH care in LMIC [3].

The importance of access to mental health services for all is accentuated by the UN Sustainable Development Goals (SDGs), which acknowledge mental health as a development priority for which service coverage indicators are imperative [4]. The inclusion of mental health in the SDGs came in the wake of global discussions in the last decade on eliminating barriers to equitable care [5–7], a position adopted by the World Health Organization (WHO) in 2001 [2].

Access to mental health services is important for Primary Health Care (PHC) [2]. This opinion is reinforced by the 2008 World Health Report which revisited the Alma Ata goals and emphasised the need for engaging culturally competent providers who respect patient beliefs [8]. By providing care in line with indigenous knowledge and belief systems, traditional healers fit this description [9, 10]. Two WHO strategies currently endorse the involvement of traditional healers in care: the WHO’s traditional and complementary medicine strategy 2014–2023 which highlights traditional healers as a potential solution to achieving universal health coverage [11]; and the WHO mental health action plan 2013–2020 which encourages greater collaboration with traditional healers to promote mental wellbeing [12]. However, there are no clear examples of collaboration between traditional healers and the biomedical health system for mental health care.

Collaboration between traditional healers and clinicians in alleviating mental suffering among children and adolescents is particularly important in settings where access to CAMH services is poor, such as Uganda, which has fewer psychiatric facilities than the global average [13]. However, hindrances to successful collaboration between traditional healers and biomedical mental health systems are cited in qualitative studies. For instance, traditional healers express a preference for referring clients to another healer rather than to a clinician, maintaining an ‘internal’ referral network. Other traditional healers express skepticism regarding the value of biomedical psychiatric treatments because of their perceptions of the underlying spiritual cause of mental disorders [14]. Conversely, biomedical service providers think that the difficulty around establishing the scientific validity of traditional and faith healers’ practices makes referral to traditional healers very difficult [15]. These attitudes reflect a lack of trust between traditional healers and biomedical providers. Yet trust drives cooperation between agents of the health system; without it, collaboration breaks down, corroding the interactive nature of health systems [16].

The interaction between the different agents of health care is embodied by the classical Explanatory Models of Illness theory proposed by Kleinman and others. Under this model, patient satisfaction with care is more likely with providers who explore and address patients’ explanatory models of illness [17]. Thus, health care is composed of different sectors each with its unique institutions and each relying on different patient explanatory models (Fig. 1). The model states that while the professional / biomedical sector is one sought by patients who ascribe illness to biological causes, the folk and popular sectors tend to attract clients who attach a social or a cultural explanation to illness. Each of the sectors offers unique remedies, related both to the expertise of providers within the sector and to the perceived cause of illness. Clients may utilise a combination of two or more sectors for a single episode of illness and therefore, an interaction between the sectors is required to achieve access to and satisfaction with care [18].

Estimates indicate that more than 80% of African populations attend traditional healers for health reasons and that 40–60% of these have some kind of mental illness [11, 19]. In fact, traditional healers are the first recourse for approximately half of individuals seeking care for mental disorders in Africa [20]. In Uganda, available research indicates that up to 60% of Ugandan patients attending traditional healers’ shrines have moderate to severe mental illness [21]. Several studies have documented positive outcomes among adults of mental health care from traditional healers in sub Saharan Africa [21–23], a factor which could explain traditional healer popularity. The widespread use of traditional healers could also be attributed to the accessibility of traditional healers relative to medical professionals; across sub Saharan Africa, there is a traditional healer for every 500 people, compared to a doctor: population ratio of 1: 40,000 [24]. While it has not been possible to determine the total number of traditional healers in Uganda, estimates indicate that compared to a doctor: population ratio of 1:8547, there is one traditional healer for 700 Ugandans [25].

Traditional healers are widely consulted worldwide for treatment of various ailments and practice traditional medicine, which is defined as knowledge and remedies based on beliefs indigenous to different cultures used in the prevention, diagnosis or management of physical and mental illness [11]. For purposes of this paper, traditional medicine refers to such practices applied with the aim of diagnosing, treating or otherwise alleviating mental suffering among children and adolescents. In this paper, the term ‘bio medical personnel’ is used interchangeably with ‘clinicians’. The aim of this study was to explore traditional healers’ views on their collaboration with biomedical health systems so as to inform the implementation of strategies to improve access to CAMH services in Uganda. Specifically, we aimed to examine the experiences of traditional healers with CAMH; their perception of and attitude towards collaboration with biomedical mental health systems.

Two main themes arise from the data: 1) Treating mental illness is cultural; and 2) Mistrust hampers collaboration. The theme “Treating mental illness is cultural” presents data on epistemologies of mental illness in children and adolescents; and on interactions between traditional healers and biomedical providers. “Mistrust hampers collaboration” is categorized into two sub-themes: willingness to collaborate and barriers to collaboration.

The main findings of the study were: 1) Epistemologies of mental illness in children and adolescents are shared by traditional healers; 2) traditional healers have limited interactions with the biomedical health system for mental illness; and 3) traditional healers’ willingness to collaborate with biomedical providers is hampered by mistrust. Whereas previous studies [21, 27] have addressed traditional healers’ views on help-seeking behaviour and have characterised mental health conditions and outcomes, this is the first study in Uganda that has explored the views of CAMH-experienced traditional healers on their integration with formal health systems.

In line with the classical Explanatory Models theory proposed by Kleinman, the epistemological view of mental illness among traditional healers was tinged by their belief systems. Specifically, their reliance on spiritual and supernatural explanations of mental disorders is in line with their identity as traditional healers. This has been found in other studies as well [28]. Nevertheless, some causes of CAMH illness they identified, namely substance abuse and fevers, are shared by biomedical practitioners, illustrating the intersection between the traditional and professional sectors of CAMH care.

That traditional healers are engaged in managing CAMH cases is unsurprising and has been documented by other studies in Uganda and elsewhere [19, 29, 30]. What hasn’t been widely reported, however, is the cross-over between biomedical and traditional treatments, exemplified in the present study by the traditional healer who routinely uses largactil®, a brand name for chlorpromazine. The use of biomedical remedies by traditional healers reinforces the notion that epistemologies are not clear cut – one sector of health care may borrow technologies from another. It also further highlights Kleinman’s conceptual overlap between the professional and folk sectors of care. As suggested by Kleinman therefore, this overlap is one that can be explored in enhancing collaboration between the two sectors.

As in this study, Van Niekerk et al., (2014) reported a low level of referral from clinical practitioners to traditional healers in South Africa. Likewise, Musyimi et al., (2016) found that traditional healers in Kenya had no experience of referral from clinical practitioners [15, 31]. In keeping with Kleinman’s theory [17], the most prevalent referrals found are self- referrals. Patients appear to use the traditional and biomedical sectors interchangeably, possibly depending on factors that may include access and illness interpretation.

The low levels of referral between traditional healers and clinicians could be related to the lack of mutual trust between traditional healers and clinical practitioners found in this study. As noted, trust is a prerequisite for successful collaboration between health care managers and practitioners [16, 32, 33]. In this study, the mistrust is limited to mental health conditions, exemplified by the non-mental health referrals being made. Lack of trust for mental health services has been found elsewhere [15, 34], and is justified by recent findings from sub Saharan Africa that confirm clinicians’ perception of traditional healers as dirty and unprofessional [15, 34, 35]. Suggested ways of addressing this negativity is creating means for dialogue between traditional healers and clinical providers, which has been successful in Kenya (30). Specifically, the health system needs to address the negativity with which patients referred from traditional healers are received at biomedical clinics. Trust of other traditional healers is also poor, and may seem surprising until one considers the high mean age of 53 in this sample; older, more experienced traditional healers whose reputation is based on years of practice could be wary of younger healers who advertise their services and charge exorbitantly for their services, jeopardizing the quality of their occupation. However, this finding suggests too that traditional healers are not homogenous, and potential for competition exists alongside potential for collaboration.

In this study, traditional healers believe themselves to be superior to biomedical providers in treating CAMH disorders, as biomedical providers possess no understanding of spiritual matters. This finding is in sharp contrast to their belief that clinicians regard them as inferior; and suggests that adoption of biomedical explanatory models, treatments and remedies by traditional healers is as unlikely as the adoption of traditional models by clinicians might be. At first look this would appear to raise a conflict that would constrain collaboration. However, viewed in light of the different sub-sectors of health care suggested by Kleinman [17, 18], this finding reinforces the need for a collaborative model, in which the relative competencies of the different sub sectors are recognized and respected.

Traditional healers perceive their devaluation by biomedical providers as a hindrance to collaboration. On the other hand, traditional healers also tend to devalue doctors’ CAMH epistemologies, even as they borrow certain treatments from the biomedical sector. This suggests that traditional healers, whilst professing superior knowledge, find value in biomedical sector therapies. This contradiction is one that appears to mask a willingness to collaborate more with biomedical providers. In fact, willingness to collaborate has been documented elsewhere [15, 34, 35]. The fact that willingness in this study is conditional on traditional healers’ contributions not being overshadowed by the clinical providers suggests strongly that an element of competition exists between the two categories of provider. Competition is counter to the Explanatory Models of Illness theory which is more collaborative than competitive. It is worth noting that whereas the traditional healers’ attitude to biomedical epistemologies is dismissive, clinicians are perceived as less of a threat than ‘neo-Pentecostalism’.

The results of this study have several implications for improving access to CAMH services through traditional healers’ collaboration with clinical mental health systems. First, advocates for collaboration should consider the different explanatory models existent among the two categories of provider; and find mutually acceptable ways to stimulate collaboration, recognising that the two categories have a complementary role to play in stimulating access to CAMH services. This should involve, among others, improving clinicians’ acceptance of traditional healers’ explanatory models for illness; and vice versa. Secondly, trust between the two categories of provider needs to be enhanced to improve interaction between the two sectors, which currently operate in isolation. In particular, the perceived negativity by clinicians of traditional healers needs to be addressed. Thirdly, quality of care by traditional healers needs to be addressed to improve hygiene and eliminate unethical practices like extortion of money from clients. These could form suggestions for inclusion in policy suggested by traditional healers to guide collaboration between the two sectors of care.

Some strengths of this study are recognised. The use of a purposive sample of traditional healers with CAMH experiences provides experience-based views. Thus, we believe that these views are transferable to all traditional healers with CAMH experience in Uganda considering that the health system and policy context is uniform across the country. Also, given that traditional healers believe they are disliked by clinicians, conducting interviews at traditional healers’ place of work and the use of a non-medical interviewer probably improved the traditional healers’ ability to express their views freely. As limitations, we note that clinicians’ views were not sought to validate traditional healers’ impressions; however, the aim of this study was to explore CAMH services from the traditional healers’ perspective. Secondly, we recognize that the reliance on a single coder, and the direct translation of interviews into English during transcription could have introduced a bias resulting from the coder’s and transcriber’s own interpretation of the data; and threatened linguistic equivalence since back-translation of local language transcripts was not done. Thirdly, the absence of the PI at interviews might have limited deeper enquiry into the subject matter as the data were collected.

These findings show that potential for collaboration between traditional healers and biomedical health systems for improving access to CAMH in Uganda exists, based on shared epistemologies and technologies; and patients’ movement across the two sectors. However this potential is undermined by the prevailing mutual mistrust and competition between traditional healers and clinicians. The findings highlight the need for implementation of strategies that harness the complementarity of traditional and biomedical sectors of mental health care as a means of improving access to CAMH services.