Engage actively and collaboratively with local stakeholders

When designing and planning local services, it is important to do so using participatory methods such as ToC workshops. These workshops proved to be a useful approach to developing a logical structure for mental health plans in PRIME, through accessing contextual details for implementing the plans in district sites and obtaining stakeholder buy-in (Breuer et al. Reference Breuer, De Silva, Fekadu, Luitel, Murhar, Nakku, Petersen and Lund2014). The participatory nature of the workshops allows for an exploration of the suitability of locally based services, and involves all stakeholders working together to map out the ToC for the district, while keeping the intended outcomes or end goals in mind. To ensure inclusion and maximum buy-in, we recommend including a range of stakeholders, such as ‘mental health specialists, researchers, policy makers, district level health planners and management, and service providers’ (Breuer et al. Reference Breuer, De Silva, Fekadu, Luitel, Murhar, Nakku, Petersen and Lund2014, p. 5). Also, when conducting ToC workshops, it is important to keep the hierarchical nature of health systems in view. We found that some contexts may require separate workshops to involve different levels of participants, in order to facilitate maximum participation, and high-level stakeholders may potentially require separate interviews. Buy-in is only achieved with consultation with all stakeholders, through both ‘a top down and bottom up approach’ (Breuer et al. Reference Breuer, De Silva, Fekadu, Luitel, Murhar, Nakku, Petersen and Lund2014, p. 10).

Use primary care systems to access vulnerable populations

Improving access to mental health care for vulnerable populations can often be achieved through primary health care centres. For example, in South Africa, there is a very good uptake of antenatal service attendance at primary care clinics: 93% of pregnant women accessed these services in 2014 (StatisticsSA., 2015). In addition, previous epidemiological studies have identified high prevalence of antenatal and postnatal depression and anxiety in this population, in Khayelitsha (Cooper et al. Reference Cooper, Tomlinson, Swartz, Landman, Molteno, Stein, McPherson and Murray2009). The AFFIRM South Africa trial took this opportunity to access depressed pregnant women in antenatal clinics. This carries the potential to improve maternal and child health outcomes, in keeping with the priorities of the sustainable development goals. We found that while working in primary care systems, it is essential to achieve buy-in from local providers such as nurses. Meetings, individual discussions and networking with health care providers before implementation are vital for an intervention to be effective.

Use cultural concepts of distress and narrative-based vignettes to identify persons with potential mental health problems

The PRIME site in Nepal developed a Community Informant Detection Tool (CIDT) for detecting mental health problems by NSWs. This tool incorporates locally identified cultural concepts of distress and manifestations of mental health problems that are commonly understandable, and uses vignettes rather than a checklist of symptoms, with good accuracy (Jordans et al. Reference Jordans, Kohrt, Luitel, Komproe and Lund2015). The Nepali PHQ-9 also incorporated cultural concepts of distress to improve accuracy and reduce the screening burden on NSWs using a stepped algorithm approach. In AFFIRM, local concepts of maternal depression were identified through in-depth interviews with participants and these were incorporated into the development of the AFFIRM counselling manual to ensure its local relevance and acceptability (Davies et al. Reference Davies, Schneider, Nyatsanza and Lund2016; Nyatsanza et al. Reference Nyatsanza, Schneider, Davies and Lund2016).

Use manual-based approaches to deliver care through non-specialist health workers

There is now a reasonably well-established evidence base for using manuals to guide non-specialist counsellors in the delivery of psychological interventions (Rahman et al. Reference Rahman, Malik, Sikander, Roberts and Creed2008; Chibanda et al. Reference Chibanda, Mesu, Kajawu, Cowan, Araya and Abas2011; Balaji et al. Reference Balaji, Chatterjee, Koschorke, Rangaswamy, Chavan, Dabholkar, Dakshin, Kumar, John, Thornicroft and Patel2012). Both PRIME and AFFIRM are using manualised counselling interventions for perinatal depression, general adult depression and alcohol use disorders, which provide a systematic and structured method of working with these illnesses for NSWs. Piloting the manual-based interventions, and evaluating them for understanding and comprehension, is essential, before rolling them out with NSWs.

Set up systems of ongoing supervision and support

Experiences from PRIME and AFFIRM have shown that in order for programmes to be sustainable and effective it is critical for NSWs to have ongoing, structured and supportive supervision. We found that supervision needs to be conducted by specialist mental health workers, and not by non-specialists in existing supervisory structures, unless they have been specifically trained. This however requires the availability of specialists who can provide the necessary support and referral platforms, in order for NSWs to be able to implement the required mental health services (Mendenhall et al. Reference Mendenhall, De Silva, Hanlon, Petersen, Shidhaye, Jordans, Luitel, Ssebunnya, Fekadu, Patel, Tomlinson and Lund2014). Training in itself is not enough to provide task-shared services by NSWs if there are not sufficient specialists available to provide supervision and referral platforms for severe cases.

In the AFFIRM-SA trial, we found weekly group and monthly individual supervision by a mental health specialist were essential for NSWs to competently treat women with maternal depression with manual-based counselling sessions. We found that 5 days of training was insufficient to obtain competence in counselling for community health workers. With hindsight, more extensive training, and more in-depth supervision would have further improved counsellors’ competence. We thus recommend supervision to occur weekly if in groups, or bi-monthly or monthly if individualised. PRIME has developed a rating scale which was used in Nepal to monitor training and supervision, which can be administered across programme types [The ENhancing Assessment of Common Therapeutic factors (ENACT) rating scale] (Kohrt et al. Reference Kohrt, Jordans, Rai, Shrestha, Luitel, Ramaiya, Singla and Patel2015), as well as a self-report competency evaluation form to monitor NSWs competency post training. These and other tools such as counselling manuals are available on the PRIME website (www.prime.uct.ac.za).

Along with supervision, it is necessary to monitor progress against the programme aims, throughout implementation, to (a) ensure implementation is being carried out according to protocol; and (b) ensure the intervention is continuously appropriate, beneficial and relevant to the local setting. This can be conducted by having clearly stated monitoring methods and goals written in project protocols, which can guide those monitoring progress and data. AFFIRM's mental health specialist was responsible for monitoring number and quality of sessions provided by NSWs, while also listening to recorded sessions in order to monitor quality. PRIME monitors progress and data through ongoing multi-method evaluation strategies across different levels, which examine changes in contact rates, detection of disorders, patient outcomes and the process of implementation (De Silva et al. Reference De Silva, Rathod, Hanlon, Breuer, Chisholm, Fekadu, Jordans, Kigozi, Petersen, Shidhaye, Medhin, Ssebunnya, Prince, Thornicroft, Tomlinson, Lund and Patel2016).

Adequately compensate NSWs for their services

In PRIME, the NSWs across sites stated the need for adequate compensation in order to add mental health care to their work load within the primary care systems. Lack of adequate compensation can be a barrier to continuation of services and commitment by health workers because of their already heavy workloads. PRIME involves varying already-established NSWs, who receive compensation as part of their employment in existing routine Ministry of Health or non-governmental (NGO) services. What is slightly more difficult to sustain is compensation of higher level supervisors or specialists, some of whom have had to be funded by PRIME in order to demonstrate how the model could work in routine settings. We are currently working with Ministries of Health and local NGOs to take over these additional functionalities in a sustainable manner, beyond the life of the programme.

Respond to crises by ‘building back better’

When primary care mental health systems are strengthened, more resources are available to respond to local crises. The WHO report on mental health in humanitarian disasters recommends using these opportunities to ‘build back better’ out of a disaster (WHO, 2013a ). The PRIME site in Nepal provided an example of this when it was able to respond to the crisis of the earthquake in 2015 by making PRIME Mental health care plans available to assist with strengthening mental health services in earthquake affected districts. These included psychosocial skills modules, counselling manuals, detection tools, and mhGAP implementation plans.

Make use of policy windows

Lastly, we recommend making use of ‘policy windows’, by providing solutions or recommendations for problems that have been identified and acknowledged by government at the time when they are open to finding solutions; in other words when there is political will or a desire for change. For example, PRIME took the opportunity to implement their mental health care plan in the Dr. Kenneth Kaunda district in South Africa, which was a national Department of Health (DoH) pilot site for the planned National Health Insurance reforms. In the case of the AFFIRM trial in South Africa, the Western Cape province DoH has launched a ‘First 1000 days initiative’, which draws attention to the importance of the first 1000 days of a child's life (DoH, 2016). This initiative focuses largely on maternal and child health and education, and provides an optimal policy window to use the AFFIRM trial findings to promote and initiate maternal mental health interventions in routine maternal health services. Similarly, in Ethiopia, the Ministry of Health developed a strategy to implement and scale up the mhGAP, which provided the opportunity for PRIME to present and use their tools in Ethiopia. In India, the PRIME mental health care plans were linked to a governmental initiative to expand access to mental health care in the state of Madhya Pradesh: the Strategy for Healthy and Active Minds (SOHAM) initiative.

Crucial to integrating mental health services into primary care is financing. AFFIRM was an externally funded RCT, and while it funded independent counsellors not employed by government, the purpose of the research was to demonstrate the cost-effectiveness of a task sharing model to policy makers, in the hope that they will invest in and scale up the interventions. PRIME on the other hand, worked closely with Ministry of Health partners, and relied largely on Ministry of Health financing to implement the interventions with training, technical support and supervision provided by PRIME. However, in some settings, where services were not in place, PRIME provided additional funding to initiate these services, in order to demonstrate the feasibility and impact of new models. PRIME is now working with Ministries of Health and NGOs to ensure sustainability and scale up of these interventions by local funding sources, including Ministries of Health and NGOs.