The development of a Community Mental Health Education and Detection (CMED) tool in South Africa

Abstract

Background:

Low demand for mental health services in sub-Saharan Africa is driven by poor mental health literacy, stigma, and poor service availability.

Objective:

To develop a Community Mental Health Education and Detection (CMED) Tool for adults for use by community health teams in South Africa aligned with their roles of health promotion, screening and linkage to care.

Methods:

Formative evaluation methods involving four processes: (1) Ongoing engagement with the KwaZulu-Natal Department of Health (KZN DoH) to ensure co-creation of the CMED tool and alignment with routine community health team activities; (2) Adaptation of the CMED tool from the Community Informant Detection Tool (CIDT), used to promote help-seeking of people with mental health problems in Nepal; (3) Review of the CMED vignettes and illustrations by a panel of local and international mental health care experts to establish accuracy and contextual and cultural relevance; (4) Process mapping and focus group discussions (FGDs) with community health teams in one district to establish cultural and contextual appropriateness as well as coherence and compatibility with existing community-based services.

Results:

The resulting CMED tool consists of five case vignettes and related illustrations to facilitate psychoeducation and the detection of possible depression, anxiety, psychosis, harmful alcohol use, and drug use by community health teams. Based on prototype matching, it includes two structured questions to guide the community health teams in the detection and referral process. The tool was acceptable, culturally and contextually appropriate, and helpful for the services provided by community health teams. Challenges of working in households and the importance of self-care were highlighted as important considerations when developing training content and piloting the tool.

Conclusion:

Extensive consultation with the KZN DoH, community health teams, and the expert mental health panel resulted in developing a tool that was perceived to be culturally sensitive and relevant to the community package of services.

1. Introduction

Sub-Saharan Africa faces an increasing burden of mental disorders, with projections that the disability burden from these conditions will increase by 130% in the next 40 years (Charlson, Diminic, Lund, Degenhardt, & Whiteford, 2014). There is thus an urgent need to close the treatment gap that is estimated to be between 76.3% and 85.4% in less-developed countries (Demyttenaere et al., 2004). South Africa has a 75% treatment gap for common mental disorders [CMDs] (Herman et al., 2009), with only 25% of South Africans with CMDs receiving mental health treatment of any kind (Seedat et al., 2009). In order to address the treatment gap in South Africa, an increased supply of mental health services has been encouraged through integration into the existing primary health care (PHC) platforms using a task-sharing approach (South African National Department of Health, 2013). There is ample evidence of the effectiveness and cost efficiency of this approach in low- and middle-income countries [LMICs] (Patel et al., 2016, 2018). However, there is also a need to strengthen the demand for services to address the treatment gap (Jordans, Kohrt, Luitel, Komproe, & Lund, 2015; Shidhaye et al., 2017). Many people with mental disorders do not seek care due to poor mental health literacy, which includes a lack of information and knowledge about the signs and symptoms of mental health problems, a lack of awareness of service availability, stigma, and misinformation about treatment (Egbe et al., 2014; Ganasen et al., 2008; Patel, 2007; Saraceno et al., 2007; Shidhaye et al., 2017). A study on stigma and discrimination experienced by service users in South Africa found that stigma is perpetuated by family, friends, community members and health workers and is often caused by misconceptions about mental illness leading to delays in accessing care, worsening symptoms and delays in recovery (Egbe et al., 2014). Developing interventions targeted at raising mental health awareness in communities was recommended in addressing stigma (Egbe et al., 2014).

While community case detection by lay workers has been found to improve help-seeking for mental health problems in other developing contexts (Jordans et al., 2015; Jordans, Luitel, Lund, & Kohrt, 2020; Shidhaye et al., 2017), the need for interventions to increase mental health literacy in LMICs are also indicated (Dang, Lam, Dao, & Weiss, 2021; Ganasen et al., 2008; Shidhaye et al., 2017). These interventions do, however, need to consider the context (including family and cultural belief systems/understandings) and build on existing knowledge (Ganasen et al., 2008). Evidence from LMICs indicates that such interventions at a community level contribute to increased demand for mental health services at a health facility level (Shidhaye et al., 2017).

Community Health Workers (CHWs) have been shown to contribute significantly to increasing coverage of key interventions in communities, particularly those in LMICs with human resource constraints (Bhutta, Lassi, Pariyo, & Huicho, 2010; Shidhaye et al., 2017; Zulu, Kinsman, Michelo, & Hurtig, 2014). South Africa has embarked on a PHC re-engineering strategy that includes community health teams formally known as ward-based primary health care outreach teams linked to PHC facilities (South African National Department of Health, 2017) that function in households and community venues in designated municipal wards (Assegaai & Schneider, 2019). Outreach Team Leaders (OTLs), who are higher-level (professional) or mid-level (enrolled) nurses appointed in local PHC facilities, supervise the CHWs who are responsible for day-to-day household visits for a defined number of households in their wards (Schneider, Besada, Sanders, Daviaud, & Rhode, 2018, pp. 59–65). CHWs minimum required level of formal education is a school level grade 10 supplemented with basic training (10 day courses followed by practicums of up to a year of study with a focus on the health system, priority health areas, and social support (Schneider et al., 2018, pp. 59–65). Although mental health is a part of the CHW health promotion manual a comprehensive focus on mental health was not included in their training at the time of the study. CHWs typically live in the communities they serve, and their roles include health promotion and education in communities, screening and identifying individuals and families at risk of ill-health, tracing defaulters and facilitating onward referral for relevant care (Assegaai & Schneider, 2019). Their focus, to date, has been primarily on physical health conditions with key areas including maternal and child health, HIV and TB.

This study is part of the Southern African Research Consortium for Mental health INTegration (SMhINT) project that has been evaluating the scale-up of a collaborative care package for the integration of mental health into the care provided for chronic care patients at a PHC level in collaboration with the KwaZulu-Natal Department of Health (KZN DoH) using a learning health system approach (Petersen et al., 2021). The collaborative care package – known as the Mental health INTegration (MhINT) package uses training, support tools/materials, and continuous quality improvement [CQI] (Institute for Healthcare Improvement, 2003; O’Neill et al., 2011) as strategies to implement and scale-up the collaborative care package. Through this learning health system approach, the need to increase identification of CMDs, including depression, anxiety, and substance misuse, at a PHC level was identified. This led, inter alia, to the development and validation of the Brief Mental Health Screening Tool (BMH) (Bhana et al., 2019) designed to be used by enrolled nurses at a PHC facility level.

In the context where the South African DoH was undergoing a PHC reengineering process in line with community oriented PHC, the lack of a standardised screening tool for mental disorders at a community level was identified as a gap in the battery of screening tools used by CHWs. The possibility of using the BMH at the community level was deliberated on by the South African and KZN DoH, but excluded because of its symptom checklist approach. It was felt that in the context of poor mental health literacy, it could potentially lead to labelling and stigmatisation of people who screened positive at a community level. These risks were mitigated at a PHC facility level where the BMH is administered, as confidentiality and the rights of patients were protected by professional ethical codes of conduct. The South African DoH and the KZN DoH thus requested the SMhINT team to develop and validate a mental health tool that could be used at the community level by community health teams to identify community members with possible mental health conditions during their routine household visits. Given the low levels of mental health literacy in South African communities as well as the screening and health promotion role of community health teams (Andersson et al., 2013; Ganasen et al., 2008; Hugo, Boshoff, Traut, Zungu-Dirwayi, & Stein, 2003), it was agreed that psychoeducation is included as part of this tool to raise mental health awareness.

The Community Mental Health Education and Detection (CMED) tool was thus developed to provide psychoeducation on mental health conditions and identify people with potential mental health problems at a household level who may benefit from available mental health interventions. This paper aims to report on the formative research processes informing the development of the CMED tool.

2. Methods

2.1. Theory

A mixed-methods approach informed by Normalisation Process Theory [NPT] (Murray et al., 2010) was adopted as a broad framework to elucidate the factors needed to ensure a culturally and contextually relevant tool that would ease integration into routine CHW services (normalised). Regardless of the success of an intervention, its long-term impact depends on its effectiveness in the real-world context and how widely it is implemented. Implementation and sustainability of interventions need to be considered from the outset and can be evaluated using NPT. NPT is relevant in the early implementation stages to when an intervention becomes a part of routine services (normalised). The four components of NPT include i) coherence (meaning and sense-making by users), ii) cognitive participation (engagement), iii) collective action (work needed to ensure adoption, compatibility with the existing system), and iv) reflexive monitoring (benefits/costs of the intervention).

2.2. Setting

The study was conducted in the Amajuba District of KwaZulu-Natal province of South Africa, where the larger SMhINT study was being conducted. It is made up of three sub-districts – Newcastle (urban), Emadlangeni (rural), and Dannhauser (semi-urban). There are 53 community wards and, at the time of the study, a total of 15 community health teams. The study was conducted in the Newcastle sub-district [population of 389 117] (Statistics South Africa, 2016). It comprises both urban and rural areas and is serviced by a district and provincial hospital with 14 PHC facilities and five fully formed (OTL and CHWs) community health teams.

2.3. Procedure

The formative research involved four processes: (1) Ongoing engagement and collaboration with the KZN DoH to ensure co-creation of the CMED tool; (2) Development of the CMED tool materials; (3) Review by an expert panel; and (4) Process mapping and focus group discussions with community health teams in the Newcastle sub-district to inform a standard operating procedure for use in routine household visits.

2.3.1. Process one: engagement with the Department of Health

To ensure that the development of the CMED was aligned with the KZN DoH strategic vision, their needs, and priority areas with respect to the community health team roles, functions, and services provided for other conditions, we had ongoing engagements with the KZN DoH. To this end, a total of 10 joint meetings were held with the KZN DoH, as well as email communication. In tandem, we reviewed the current community health team guidelines (South African National Department of Health, 2017), the national mental health policy (South African National Department of Health, 2013), the existing community health worker curriculum and the community package of services.

2.3.2. Process two: adaption and development of vignettes and illustrations for the CMED tool

The CMED tool was adapted from the Community Informant Detection Tool (CIDT) that was used to promote help-seeking of people with mental health problems in Nepal as part of the Programme for Improving Mental Health Care (PRIME) (Jordans et al., 2015; Subba, Luitel, Kohrt, & Jordans, 2017). As with the CIDT (Subba et al., 2017), the CMED tool is based on the prototype matching approach. Prototype matching is where a diagnosis is made by matching a patient’s presenting symptoms with a paragraph-length description of the disorder (Westen, 2012). It has been found to have several advantages relating to clinical utility over the traditional method of counting symptoms to diagnose patients (Westen, 2012). The format of the CMED and Nepalese CIDT consists of vignettes (prototype paragraph) and illustrations for five mental health conditions. Each vignette (in the CMED and CIDT) is followed by three structured questions (Table 1) that aid the health worker in matching symptoms with the prototype vignette and determines if the family member requires a referral for further care. The mental disorders included in the CMED tool were depression, anxiety, harmful alcohol and drug use, and psychosis. These disorders were chosen on the basis of being the most common and high burden mental and substance use disorders in South Africa (Herman et al., 2009). As with the CIDT, although the symptoms of these mental disorders were drawn from the WHO mhGAP intervention guide (World Health Organisation, 2016), local idioms of mental health problems were incorporated into the development of the vignettes drawing on the literature (Campbell et al., 2017; Davies, Schneider, Nyatsanza, & Lund, 2016; Den Hertog, de Jong, van der Ham, Hinton, & Reis, 2016; Den Hertog, Maassen, de Jong, & Reis, 2020; Petersen, Hanass-Hancock, Bhana, & Govender, 2013; Sibeko, 2016; Swartz, 1998) and the everyday rhetoric of the local context. Local idioms of distress refer to ways of expressing distress that may not involve specific symptoms or syndromes, but provide collective ways of experiencing and talking about distress in local contexts (Nichter, 2010).

Table 1.

Comparison of structured questions in the CIDT and CMED.

CIDT Structured Questions	CMED Tool Structured Questions
1. Does the narrative apply to the person you are talking to now? (four-point scale: no match, moderate match, good match, very good match)	1. Does this story remind you of anyone in the household? Yes/No
2. Do the problems have an impact on daily functioning?	2. Do the problems have a negative impact on daily activities? Yes/No
3. Does the person want support in dealing with these problems?	3. Refer family member and provide healthy lifestyle information

The development of the vignettes was undertaken by a team of clinical, counselling and research psychologists from the Centre for Rural Health, University of KwaZulu-Natal, and psychologists/mental health practitioners from the South Africa HIV- Addiction Technology Transfer Centre (ATTC) at the University of Cape Town who have developed a mental health training for CHWs (Sibeko, 2016). A graphic artist was employed to illustrate the key symptoms of the protagonist with the disorder in each story.

The process described above followed Flaherty’s et al. (1988) translational research methods. These include ensuring content equivalence, that involved making sure that content was relevant to the phenomena being studied, in this case screening for mental health conditions in a South African context. Semantic and conceptual equivalence was also ensured. In this regard, although vignettes were different to those used in the CIDT, the overall meaning of the tool was retained and the same theoretical construct has been adopted (Flaherty et al., 1988).

Standard translation and back-translation procedures were used to provide an isiZulu version of the CMED tool. The translation process was conducted by two bilingual clinical psychologists and a bilingual research psychologist who had an in-depth knowledge of local terms and idioms used by people when discussing and defining mental health.

2.3.3. Process three: review by expert panel (establishing criterion-related validity of the CMED tool)

The resulting CMED tool was then reviewed by an expert mental health panel to assess prioritisation of key symptoms and the appropriateness of the vignettes and illustrations to the South African context. This approach was modelled on the process used to inform the development of the CIDT (Subba et al., 2017), ensuring conceptual, criterion, and semantic equivalence (Flaherty et al., 1988).

The expert panel was selectively targeted because of their particular expertise in working with local populations and we wanted the panel to reflect the range of mental health workers in South Africa. The panel included a mix of health professionals working in the public health sector including psychiatrists, psychologists, psychiatric nurses, social workers and registered counsellors. The panel also included government representatives from the KZN DoH including the KZN Mental Health Directorate and the KZN District Services Community Task Team, as well as academic mental health specialists. The vast majority of the expert panel was local. The international expert included was part of the team that developed the CIDT. These insights were valuable given that we were adapting the CIDT and wanted to ensure that Flaherty’s conceptual, criterion and semantic equivalence was retained through the adaptations. We also wanted insights from the CIDT team as they had already tested their tool in a LMIC.

The question about relevance of symptoms was modelled on the three-point scale (high relevance, low relevance, no relevance) used in the CIDT. We adapted the questions (3-point scale) used in the CIDT and posed these as three different questions to the panel. The expert panel in the CMED were asked to review each vignette using the following guiding questions:

1. “The vignette includes symptoms commonly found among people with this condition.” With the response options being a) Yes or b) If No, what are the problems and what changes would you recommend?

2. The second question to the panel was “Are the graphics used appropriate in depicting some of the symptoms of the disorder? With the response options being a) Yes or b) If No, what changes would you recommend?

3. Additionally, experts were asked an open-ended question regarding additional comments or suggestions including the cultural appropriateness of each vignette to the South African context. Local experts were asked to review the relevance of the description of symptoms in the vignettes and to comment on the relevance of the characters, the stories, graphics, and language used to the South African context.

All responses were collated in an excel document. “No” responses and associated problems/suggestions were discussed by the research team and amendments were then made to the vignettes and graphics. The open-ended nature of the panel questions allowed the experts to provide detail on how the vignettes should be improved.

Additionally, the understanding of the vignettes was also assessed in Process Four with CHWs as they were asked what their understanding of each vignette was. Through this process we wanted to ensure that the mental health conditions included in the vignettes were easily understood by the CHWs, who are lay workers with a school leaving certificate and minimal mental health knowledge.

2.3.4. Process Four: process mapping and focus group discussions with community health teams

2.3.4.1. Process mapping of community health team day-to-day roles and activities.

Process mapping is part of the Continuous Quality Improvement (CQI) toolkit (Institute for Healthcare Improvement, 2003; O’Neill et al., 2011) that enables an in-depth understanding of complex systems and adoption of improvement interventions to local contexts (Antonacci, Reed, Lennox, & Barlow, 2018). Process maps are particularly useful when designing new interventions or programs to help understand the current processes and thus think through the most effective and efficient ways of embedding new innovations within existing systems; in this case, the CMED tool. Information from different sources, including both interviews and direct observations, are helpful in validating process maps given the multiple methods used (Antonacci et al., 2018).

A CQI process mapping workshop was conducted with 54 community health team members (44 CHWs, 10 OTLs/supervisors) from five PHC facilities (all CHWs belonged to formal community health teams) in the Newcastle sub-district at a central location to gain an in-depth understanding of the roles and activities of community health teams including planning, home visits, linking patients for referral, follow-up, and administration. This understanding was necessary to ensure coherence and compatibility of the CMED with the existing activities of the community health teams. For the process mapping exercise, the 54 participants were divided into six groups: two OTL/supervisor groups and four CHW groups to ensure that diverse perspectives were captured from people with different roles within the community health teams in the generation of the process maps (Antonacci et al., 2018).

The community health teams were asked to create a visual map of their daily activities from the start to the end of the week (Box 1).

Box 1. Process Mapping Instructions.

Community Health Team Process Map (Instructions).

We would like to understand your work.

Please describe a typical week as a community health team member (CHW or OTL)?

What happens first? (Write down first step)

What happens next? (Continue for all steps and draw arrows connecting steps)

Note: How long does each step take?

Include who is responsible for each step and if/how it is documented

Remember to include what is actually happening rather than what is supposed to happen

In addition to the process mapping workshop, observations of household visits were also conducted in three community wards to directly observe community health team activities, to better understand their daily work and how the CMED tool could be integrated into the current community package of services. The same steps were employed as described in Box 1, where CQI mentors shadowed community health teams during their household visits.

2.3.4.2. Focus group discussions.

FGDs were conducted with 54 community health team members divided into six groups (the same groups that were involved in the process mapping workshop) to explore if the purpose of the proposed CMED tool was understandable, relevant to their context (including if vignettes and illustrations were culturally appropriate/realistic, and if the language used was acceptable) as well as their perceptions of the feasibility of administering the tool at a household level. We also asked the groups what their understanding of each vignette was to ensure that the meaning of each condition was clearly conveyed and understood in the way we intended. Five groups were conducted in isiZulu and one group was conducted in English. Hard copies of the CMED tool were given to the groups in both English and isiZulu. Each group was given one vignette from the CMED tool to discuss so that each vignette could be explored in depth. Recommended changes to the tool were also discussed.

2.4. Data analysis

All twelve audio-recordings of the process mapping and FGDs were transcribed and isiZulu transcripts were then translated into English. Framework analysis (Gale, Heath, Cameron, Rashid, & Redwood, 2013; Ritchie & Spencer, 1994) was used to analyse the data using qualitative software (NVivo version 12). After reading five transcripts (familiarisation) a thematic framework was developed based on a priori interview topics and new themes emerging from the data. The remaining transcripts were coded and analysed using the framework developed.

3. Ethics

Ethical approval was obtained from the University of KwaZulu-Natal Biomedical Research Ethics Committee (BREC) (BREC Ref No: BF190/17) and the KwaZulu-Natal Department of Health. All participants provided written informed consent.

4. Results

4.1. Process one: outcomes from collaboration with and review by Department of Health

The KZN DoH reviewed and commented on the CMED tool in terms of the local relevance of the tool and congruence with routine DoH processes and referral pathways within the system. The KZN DoH team flagged low mental health literacy in their communities as a challenge in accessing care and raised the need to educate communities in understanding mental health conditions and treatment options (including self-help) if help is required. The team emphasised that the CMED tool should help families to self-identify leading to a referral rather than a process of labelling.

The KZN DoH team recommended an algorithm be added to the tool to help guide the CHWs in choosing which vignette to read at the household visit, and this was added accordingly (see Fig. 1). Changes were also made to the structured questions found in each vignette. In the CIDT and the first draft of the CMED tool, the family member was asked if they would like a referral (Question 3). On advice from the KZN District Services Community Task Team, this was modified to providing a referral as is the practice with existing tools that CHWs use for other health conditions.

Fig. 1.

Mental health question flowchart.

4.2. Process two: developing vignettes and illustrations

The revised CMED tool consists of five case vignettes and related illustrations to facilitate the detection of possible depression, anxiety, psychosis, harmful alcohol and drug use by CHWs. The CMED tool, unlike the CIDT, has an initial flowchart (Fig. 1) which helps direct CHWs as to which vignette to read in a household. The vignettes are identified by character e.g. Nontobeko, and not by the associated mental health condition e.g. depression to obviate labelling or stigmatisation. Graphics of four key symptoms of each condition is shown to family members while the CHW reads the vignette. A calendar book format was used where family members are shown the illustrations and associated captions (Fig. 2), while on the reverse side the CHW is guided by the vignette (Fig. 3). The family is asked what has happened in the character’s life, the main symptoms of the condition are summarised and further information about the condition is given by the CHW. The CMED differs from the CIDT in that after each vignette, an interactive psychoeducation discussion is introduced to promote an understanding of the symptoms of the different conditions.

Fig. 2.

Vignette of Nontobeko (depression) for viewing by family members.

Fig. 3.

Vignette of Nontobeko (depression) to be read by the CHW.

The interactive psychoeducation component draws on our past work of training lay counsellors in mental health and includes the head, heart and feet model in each vignette to guide the CHW in encouraging the family to reflect on what the protagonist in each story is thinking including thoughts or cognitive disturbances that may be evident (head), what they are feeling/emotional impact of the condition (heart) and how they are behaving (feet). This model draws on Cognitive Behavioural Therapy as a tool to develop healthy thinking skills and has been used in the training of the CHWs to aid reflection and internalisation of learning following adult education techniques (South Africa HIV Addiction Technology Transfer Centre, 2020). The CHW then goes onto provide further information on each condition (health promotion). Following the psychoeducation discussion, the family is then asked by the CHW whether someone in the household reminds them of the character in the vignette as contained in the pictures and the story. They are then asked the extent to which the family member matches the prototype vignette. In place of four-point scale used in the CIDT and where the person administering the tool would assess the match in which subtle differences are explained, the CMED tool asks the family members to assess the match using a binary format of “yes” and “no”, that obviates the need for explanation. A positive match together with a positive response to whether the symptoms impact on the person’s daily functioning leads to a referral to existing PHC mental health services (See Fig. 3). It must be noted that the CMED is a first point of detection, referral leads to screening for a mental health condition at the primary health care facility by an enrolled nurse using the brief mental health (BMH) screening tool, following a positive screen they are then referred on for assessment, diagnosis and care by a professional nurse.

In the event that no family member is identified as matching any vignette, the CMED tool prompts the CHWs to use the healthy lifestyle page with related illustrations to provide families with information to support mental wellbeing at home, e.g. diet, adequate sleep, exercise, social support (Fig. 4).

Fig. 4.

Healthy lifestyle advice.

The method of assessment in the CMED tool (technical equivalence) as compared to the CIDT has been retained in the design and administration of the structured questions and interpretation of the CMED tool as a positive screen for a mental health problem and in need of referral or negative screen where no action is needed (criterion equivalence) (Flaherty et al., 1988).

4.3. Process three: review by expert panel

Anxiety and harmful substance (drug) use required the most extensive changes. In relation to anxiety, the panel recommended that interpersonal trauma leading to anxiety was more prevalent in a South African context than the example of a taxi accident that was used in the first draft. The vignette was subsequently changed to a mugging incident. For substance use (drugs), the initial vignette included the character Brian, spending time at the local tavern drinking alcohol and smoking marijuana. The panel recommended that the focus should be taken away from alcohol with a clear focus on harmful drug use. The location where Brian spends time with his friends to smoke drugs was thus changed from a shebeen/tavern to a street corner. The drugs he uses in the vignette was also expanded to include “whoonga” (a heroin-based drug used in South Africa) as recommended by the panel. On recommendation from the panel a theme of money or household items going missing was also included. The illustrations were amended to match the changes in the text.

4.4. Process Four: process mapping and FGDs with community health teams

Both the process mapping workshop and the observation of household visits provided important contextual information about the routine care offered by community health teams at a household level and how the CMED tool could be most easily absorbed into their routine activities. The process map unpacked key activities, the relationship between the household and the clinic facility as well as gaps in processes, with the observations confirming the information generated from the process maps.

The CHWs’ process maps indicated that when visiting a household for the first time, an in-depth registration is conducted, and this household profile is updated at each subsequent visit and recorded in the CHW diary. This information was identified as being helpful for identifying households with family members with potential mental health problems for targeted administration of the CMED tool. Routine services provided by CHWs within households included health promotion and screening for physical health conditions such as TB, hypertension, and diabetes. People who screened positive were routinely referred to the PHC facilities for further screening, diagnosis and care. An exception was patients with a medical emergency who were referred to the district hospital, with an ambulance being requested where indicated. Understanding the CHW referral pathway was important in developing the CMED tool as it guided the researchers in developing the algorithm for referral for care, including emergencies in each vignette.

The OTLs indicated in their process mapping exercise that they routinely provide support and supervision to CHWs by providing regular in-service training and accompanying CHWs on bimonthly visits to households to assess where support and training is required. OTLs visit households flagged by CHWs as having difficult cases. This information was useful for informing the training and mentorship model that the CMED tool would need to follow as per routine care.

The following common themes emerged from the FGDs.

4.4.1. Challenges of working in households

All groups indicated that they experienced threats to personal safety, violence in households and being chased by dogs on their daily visits. They also spoke of the desperate situation of many of their households with layers of stressors, including poverty, unemployment, substance use, violence, domestic violence, sexual abuse and health challenges.

P2: The mother knows because she says “who is going to feed me if I say that her father is raping the child? Who will come home carrying plastics [shopping bags of food] in this household? … who will buy us electricity in this household?” Its best I keep quiet with this thing- but you [CHW] got a report at school that the child is misbehaving … she is frightened, crying and she sleeps. If you follow-up at home that she isn’t coping at school … So now they saying … [to me]as a CHW “there is a child from the Mabaso household that isn’t coping and has changed, all along she was fine but now she isn’t in a good situation”. You know that you don’t go to the Mabaso’s [the child’s home] … how are you going to go there, what will you say? (CHW, FGD 5)

All groups including the outreach team leaders spoke of the emotional labour that they experience when visiting households and felt they needed more training on managing their own mental health when visiting households. The need for structured debriefing sessions was also highlighted by participants to process challenges experienced, focus on lessons learned and to develop a way forward for community health teams and households.

P 4: … I feel bad because I may cry but I have to avoid that and not cry (mm). I must give them hope in life that this shall pass [background noise] I shouldn’t cry, I shouldn’t cry in front of a patient but let me give them hope and talk to them although it’s difficult. It not easy. (CHW, FGD 4)

P2: Let us say we have encountered some challenges. Where do we do our debriefing? As an OTL we encounter such problems that a visit to one household feels like a visit to 10 households. When we come back from that we do not know where we will de-brief. … However, this thing is causing you sleepless nights. You keep thinking about what you discovered in that particular household. Maybe you were told that the man in that household rapes his own child in front of the child’s mother. Where do you get a debriefing about such a thing?.. So, we wish that we could also have debriefing sessions. (OTL, FGD 6)

More than half of the groups spoke of sharing similar challenges as their households making it difficult for them to contain their own emotions and be emotionally supportive for household members:

P4: What I can say is that you don’t cry but you just see the-the-the tears come out because the patient has a problem to the extent that they explain this problem to you maybe I have also experienced it. Maybe I am crying because of this situation that there is someone with a problem like mine. You shouldn’t cry but it just happens then you cry. (CHW, FGD 2)

4.4.2. Confidentiality

The importance of confidentiality in the work of the community health teams by households was raised as a key factor in being granted access to homes and in fostering trust in more than half of the group discussions. The challenge of being both a neighbour and a health worker was raised by the CHWs as households feared that their personal information would be shared through gossip in the community. This dynamic was often frustrating for CHWs as they were often aware that households needed help but were unable to provide support due to household reticence.

P 6: They say that we must work close to our homes because with the stipend we are getting you can’t travel with it … but it’s very challenging … Sometimes you find that you don’t get information because they are hiding it because you know them and they know you. They don’t trust that you could be confidential with their information. (CHW, FGD 3)

Additionally, one group of CHWs spoke of households mistrusting the information they provided as they were seen as community members rather than health workers. A dominant theme of this group discussion was the challenge that CHWs had in proving their worth as legitimate health workers.

P7: So sometimes they doubt us but when we start giving them vitamins and talking about danger zones they get a bit confused because they see me as a normal member of their community. I would have received training by then so people should not be scared. (CHW, FGD 1)

4.4.3. Relevance of the CMED tool

All the groups indicated that the tool would be relevant to the work they were doing in households. Community health teams reported that the CMED tool would be useful for both health promotion in households and in the community and for screening purposes. The groups suggested the CMED tool include lifestyle advice on what families could do at home to improve their mental health which was then included (see Fig. 2).

All the groups felt that a mental health screening tool was necessary and that it would fit into existing community health team roles as screening was an existing part of the package of care delivered to households:

P 10: From my side I think it can work because as we are used to using some screening tools, I think we won’t have a problem because we are used to using screening tools. Even the family, I don’t think they will mind because they are used to, after you are reading something or if you want to screen something you have to go through some questions. (CHW, FGD 2)

All groups indicated that stories in the CMED tool were things that were happening in the community:

P4: … I think stories that are similar to Thandeka (character in the CMED) are there in the community (yah) or other challenges that we experience (mm). eh … (CHW, FGD 4)

More than half of the groups reflected on how the CMED tool would enable them to link patients to care and spoke of how the tool provides them with something that is tangible that they can now use to help them to screen for mental health problems:

P2: I was going to say it (CMED) is better since we have something that tangible now (to use for referral to care) but we have to know not every household will be smooth. There are households that will be difficult (mm) we have to be strong but at least we have something that’s tangible. (CHW, FGD 4)

Two CHW groups were able to relate to the symptoms of traumatic anxiety described in the story of Thandeka and recognised that patient life stories would not necessarily be the same as the vignettes but that symptoms will be common:

P6: Another thing that I am saying is that it won’t always be stories of knifes (mm) and Thandeka. You will find that a person say “my mother passed away” or “my child was raped” (FGD 4)

4.4.4. Potential challenges in administering the tool

It was suggested by the majority of groups that the administration of the CMED tool may require great sensitivity on the part of CHWs, especially with families that are experiencing harmful alcohol use and domestic violence. In the extract below a community health worker describes the complexity of discussing alcohol use and domestic violence with a family where these issues are present and the need to contain emotions when discussing sensitive issues was raised.

Facilitator: … In this family there is Sifiso [character in the CMED relating to harmful alcohol use] and if you get a hold of him it’s difficult (mm) to talk to Sifiso. So as a CCG what do you do?

P3: Uhm … It a fragile issue because now you have to talk in a way that … won’t make Sifiso [character in the CMED relating to alcohol misuse] hit Nontobeko [character in the CMED relating to depression] when you leave. (CHW, FGD 3)

Suggestions were made to improve isiZulu translations, and these were amended accordingly.

The development of the CMED tool has been iterative in nature and based on feedback from the research in Processes One to Four the tool has undergone several amendments reflected in the 27 versions of the tool.

5. Discussion

The CMED tool was developed to improve demand for mental health services in South Africa through providing a psychoeducational and detection tool to strengthen mental health literacy as well as help CHWs identify possible cases requiring referral for further screening, assessment and diagnosis. The coherence of the CMED tool in terms of its meaning and relevance for CHWs and families in the local context (cognitive participation) was an important factor in the design of the CMED tool.

The prototype matching approach using vignettes was adopted given coherence with the oral storytelling culture within Africa society (Scheub, 1985), reducing the chances of labelling and providing a less threatening way of talking about sensitive topics, thereby assisting CHWs to initiate difficult conversations about mental health. The latter is of relevance given that mental illness is highly stigmatised in African contexts (Egbe et al., 2014). The prototype matching approach assists by providing distance from the topic at hand by allowing people to talk about their problems and those of their family members through the characters in the story, as well as enabling participants to define the situation in their own terms (Barter & Reynold, 1999; Petersen, Mason, Bhana, Bell, & McKay, 2006). It provides an elegant alternative to screening tools that use mental health checklists of symptoms. The latter have mostly been developed in high income contexts and lack cultural sensitivity which is especially important when discussing mental health (Bass, Bolton, & Murray, 2007; Subba et al., 2017). Checklist based screening tools also have the potential to promote labelling, which is a particular concern in contexts where there is low mental health literacy.

Through the four processes adopted in this formative study, the CMED tool was developed to optimise contextual and cultural sensitivity by using local idioms and illustrations in the vignettes and psychoeducation components and introducing new knowledge within a culturally relevant framework of everyday rhetoric. Further, the addition of the psychoeducation component in the CMED tool allows for the opportunity to raise awareness about mental health and in doing so reduce stigmatising beliefs about mental illness and increase demand for services.

Constituting an expert panel helped to ensure the clinical validity of the vignettes portrayal of mental health conditions while simultaneously being relatable through depicting experiences that commonly occur in the community (culturally relevant). In addition, focus group discussions were conducted with community health teams to ascertain their views on the whether the scenarios portrayed in the vignettes were relevant and commonly experienced in the community. These two processes assisted in establishing construct validity (Evans et al., 2015).

Process mapping, observations of community health team activities as well as FGDs were used to ensure congruence of the CMED tool with the functioning and routine services provided by community health teams. Overall the process mapping exercise and observations gave an in-depth understanding of the routine community health team services and functioning. The need for a tool such as the CMED to assist in promoting mental health literacy that would fit with their routine health promotion activities, as well as screening for mental health conditions that would fit with their other screening activities was affirmed as a meaningful activity. A number of lessons emerged from these three methods that helped inform the development of the tool, together with the accompanying standard operating procedures and training and support that needed to accompany the implementation of the tool. Firstly, in addition to the CMED being useful as a health promotion tool at any visit, CHWs could also use their in-depth local knowledge and community profiles of their households to specifically target households where they feel a family member may have a mental health condition. Secondly, clarification of referral pathways for other conditions, including emergency referrals, helped ensure that the referral pathways for people identified as having a potential mental health problem was aligned with existing referral pathways for other conditions. Thirdly, an understanding of the OTL’s role in supervision, mentorship and CHW in-service training helped provide insights into how the CMED tool could be introduced into the system using OTLs to provide on-site training and mentorship in the use of the tool; as well as support CHWs when dealing with challenging home circumstances that the CMED tool may unearth. The need for supportive supervision and mentorship for community health teams cannot be understated, given the strong link to CHW programme performance (Assegaai & Schneider, 2019). Fourthly, the need for support of self-care and containing emotional experiences that they may encounter in households emerged as also being important. The CHW consultation in a home presents a different dynamic to a consultation in a clinic as the CHW has to directly contend with the layers of challenges that families face. CHWs typically live and work in the communities that they serve and so families are known to them and they are known to families. Hardships are thus not only relayed verbally but the community health teams get to know families and personally witnesses their hardships. This is part of the reason why CHWs are effective but also brings an emotional burden as CHWs navigate the dual role of health worker and neighbour (Grant et al., 2017).

Normalisation Process Theory was used as a broad framework for sensitising the researchers in this study to important concepts that need to be considered when designing the CMED tool. Defining the context is a key component when using NPT in guiding the design of an intervention (Murray et al., 2010) and speaks to the coherence (meaning and benefits to users) of an intervention as well as cognitive participation/engagement from users.

The extensive consultation process with KZN DoH, the CQI process mapping, and FGDs with community health teams was undertaken to ensure the co-creation of a tool that would be relevant to the routine community package of care offered in households. This process also provided insights into the needs at a DoH management level as well as the needs of CHWs on the ground – which has been invaluable in understanding the context and informing the design of the CMED tool. Although the tool differs from other tools used by CHWs because of its prototype matching vignette format and interactive nature, the researchers took into account current community health team guidelines (South African National Department of Health, 2017) and ensured that the CMED tool fell within the CHW scope of practice of screening and health promotion and careful consideration was given to designing a tool that would be compatible with the existing services provided. This process speaks to NPTs collective action where the impact of the CMED tool on CHWs work and the compatibility of the tool within the existing system is brought to the fore. The NPT process is also iterative in nature, and this is reflected in the continuous amendments that were made to the CMED tool based on feedback from the research process.

6. Strengths and limitations

While the CMED was modelled on the CIDT and followed the steps used in the formative evaluation of the CIDT, it was developed, in part, in response to the needs expressed by the KZN DOH. This demanded a more participatory co-production approach in the South African context, resulting in a lengthy iterative process that included DoH management and the community health teams in addition to academics and mental health experts in the co-production of the tool. While deviating slightly from the steps used in the development of the CIDT, this strengthened the appropriateness of the tool to the cultural systemic context where the tool was to be used.

The formative research was conducted in one local district in the KwaZulu-Natal Province in South Africa. There is a need for further studies to assess the applicability of the CMED tool in different contexts. The CMED includes vignettes for various mental health conditions. Although symptoms may be common, life stories may differ and it is important that CHWs are trained to acknowledge this with their patients.

7. Recommendations

The next steps include testing the CMED tool for accuracy and piloting the tool for feasibility with community health teams and family members. Once accuracy and feasibility are established it is important that the CMED tool be tested as part of routine care and integrated with the other routine community screening tools and health promotion activities of CHWs, as well as integrated in the training of CHWs inclusive of other conditions.