Setting the global research agenda for community health systems: literature and consultative review

Selection and training of CHWs

There is a lack of evidence on optimal processes for CHW selection and the impacts of different selection policies [13, 14]. For example, selecting CHWs with high educational requirements may prevent representation of certain communities [15], reduce community support for a CHW, and result in higher CHW attrition [16]. Traditionally, in small programs, CHW selection is informal, guided by local support and social norms. As programs grow, selection processes become more formal, with regional administrative selection requirements, which may undermine the community’s role. Leaving the selection process entirely to a local community can, however, be affected by local politics and traditions and lead to selection through sub-optimal criteria. Further research could test appropriate selection strategies under different stages of maturation.

There is also limited research on the optimal design of CHW programs’ accreditation and certification [2]. CHW training typically involves both a theoretical and practical component in the pre-service phase, as well as ongoing in-service training for skills updates and renewals. As a step towards CHWs’ institutionalization, several countries—notably Uganda, Haiti, Mali, and Liberia—are in the process of formalizing their training programs, with the intention of accrediting CHWs [17–21]. A recognized gap in the literature is the need for testing innovative approaches to CHW pre-service training and measuring how CHW baseline characteristic variations, such as gender and education, as well as the length of pre-service training, may affect CHW performance and patient outcomes [4, 22–26]. This all resonated with several policymakers as they expressed a need to understand models of training that will result in high CHW competency [21]. Additionally, there is a need to study the effectiveness of job aids during training [27, 28], including digital tools, compared to more traditional training methodologies [29]. National programs typically require rapid scale up of training infrastructure, using models such as training of trainers (ToT) [2]. More insights are needed around cost of training models as costs vary considerably by the mode of training, as well as the location (e.g., disbursed training close to the community versus bringing CHWs to centralized training facilities), and have significant bearing on how ministries can reasonably proceed [21]. Mediating performance and cost while balancing intensity and frequency of contact are critical considerations in CHW training design and policy [2].

Training content is another consideration, with tension between an emphasis on broad training to address social determinants of health with necessary community mobilization and counseling skills versus focused, competency-based biomedical training [13]. In addition to the type of work to be trained for, there also needs to be guidance on the volume of work that would contribute to CHW efficiency [6]. Some evidence suggests that CHW technical competency declines after training, and therefore follow on, regularly supervised practice, and mentored opportunities are needed [23, 30].

Community embeddedness

A CHW’s level of embeddedness in a community is essential to her or his success. “Community embeddedness” can refer to the level of community CHW acceptance and investment, as well as the level of CHW engagement, including trust and empathetic relationships a CHW has with the community. From a research perspective, community embeddedness is difficult to study, as communities are heterogeneous with complex power dynamics [13]. The utility of community embeddedness comes from the assumption that a CHW with high levels of trust and empathetic relationships within the community is better positioned to enter community members’ homes and counsel them on sensitive health matters. Scott et al. state that acceptance from a community may affect CHW retention, motivation, performance, and accountability, but that there is minimal evidence on how to strengthen a CHW’s connection to the community [2]. What role, if any, does involving local communities in CHW selection and training, and do community structures play, in advancing community embeddedness and subsequent CHW performance and retention? [23, 26, 28, 30–33]. Inversely, how does the work of the CHW empower communities to respond to traditional barriers to care-seeking, harmful practices and stigma? [15, 26, 27, 31, 34]. How do these relationships change, given differing contexts and settings? [29].

Institutionalization of CHW programs

Like CHWs’ horizontal integration with the community, their vertical integration within the health system is integral to their success. Research is needed to identify and test existing mechanisms to integrate and institutionalize CHWs within the formal health system [23, 32] and the extent to which institutionalization enhances performance [28–30]. Although research has not conclusively defined all components of health systems integration, several points emerged as important, including referrals, supervision, and supply chain.

Referrals

There is a need to test whether models of “shared care” involving referral and counter-referral between communities and health facilities can influence CHW performance, especially where communications and transportation systems are weak [16, 23, 27]. Bosch-Capblanch and colleagues suggest the need for more rigorous research comparing models that link peripheral health services with a central managerial unit [35].

Supervision

Numerous programs and research studies emphasize adequate supervision of CHWs for effective performance. Evidence gaps about effective supervisory approaches (e.g., type and frequency) continue, however [24–27, 30, 36]. The need to study supervisory mechanisms also emerged as important needs of the ministries of health in Bangladesh, Haiti, and DRC during the ICH meeting [21]. Evidence suggests supervision quality may be more important than frequency, up to a point [29, 37, 38]. Perceived supervision, integrating the CHW perspective of lived experienced of supervisory quality, is also integral to explore across country contexts [11]. Given this, the comparative efficacy of different models of supervision, such as peer, group, and community supervision, and self-assessment, through checklists and a combination of these approaches, needs to be tested within the broader health system using robust study designs [28, 37]. Regular supervision contributes substantially to costs associated with CHW programs, and the cost-effectiveness of various supervisory approaches should also be evaluated [29]. In addition to supervision, alternate accountability mechanisms should be explored further, from a health systems perspective [33].

Supply chain

Strategies to improve the routine availability of medical commodities emerge in the literature as another research priority [16] and were also emphasized by participants at the ICH meeting [21]. In the recent years, the use of digital tools to report on stock levels with CHWs and the use of data dashboards to improve transparency of district-level stocks have gained support. Further support is needed for studies to understand how these tools can be best integrated with existing inventory management systems and used by CHWs [39].

CHW needs including incentives and remuneration

Historically, scaled CHW programs have suffered from significant attrition over time and low productivity [2], with several reasons cited including inadequate attention to CHW concerns about their pre-service training, supervisory support, financial and non-financial incentives, satisfaction with their role, and professional development opportunities [25, 40]. As with other cadres of workers, CHWs’ incentive satisfaction is closely linked to their motivation, but research on this matter continues to be piecemeal, small in scale, and contextual, with limited generalizability [25, 41, 42]. Existing research suggests that formal salaries for a large cadre of CHWs may be financially unsustainable at a national scale in most low- and some middle-income countries; however, a combination of financial and non-financial incentives such as t-shirts/caps, other social recognition, certifications, resource availability, and positive working relationships may improve CHW motivation and reduce attrition [23, 40, 41, 43]. The ethical dimensions of asking CHWs to volunteer their time need to be considered, and more research is needed on an appropriate combination of these incentives (appropriate training/certification, career opportunities, social recognition, performance-based financing, allowance and salaries) which could be applicable across contexts and commensurate with their job demands [6, 23, 26–28, 34, 40, 42, 44]. Further research is also needed on payment systems that could reward accountability [32]. This information needs to be contextualized to specific country settings to understand what levels, methods, and types of incentives are cost-effective, given a country’s gross domestic product (GDP) and health expenditure trends [16, 43]. These results need to be disaggregated by gender, and more research is needed on meeting the needs of, in some cases, a predominantly female CHW workforce [21, 28]. Incentives’ effects on CHWs’ motivation, as well as their effects on performance and patient outcomes, have scarce empirical evidence [23, 31].

Attention also needs to be paid to the risk of exploiting CHWs, particularly women [2]. The emphasis on achieving Sustainable Development Goals (SDGs) and universal health care (UHC) has resulted in expanded CHW roles beyond health care, across a variety of sectors including education and agriculture; however, limited progress has been made in institutionalizing compensatory mechanisms for their time.

Governance and sustainability of CHW programs

Much remains unknown about how CHW programs, once scaled nationally, can be sustained [21]. Questions remain about what policy and governance structures should ensure their sustainability [21]. Short-term donor funding and different stakeholder agendas can disrupt programs’ sustainability [23, 45]. Research is needed to understand the context and conditions in which it makes sense to implement and integrate CHW programs [16]. Large-scale implementation of programs should be accompanied by research to determine the contextual factors that affect CHW programs’ performance and scale [6, 26, 28, 30, 32, 36].

The effect of decentralization of governance on CHW program implementation has emerged as a critical area for further research [2], especially in contexts like Kenya where the recent devolution has resulted in the simultaneous implementation of different CHW models across its 47 counties [21]. In Bangladesh, with a pluralistic health system, several cadres of CHWs exist; each focused on specific health areas, with some overlapping activities. While the reasons for multiple models of CHW programs are different in Kenya and Bangladesh, the research needs to focus on identifying which of these models are most effective and what types of governance structures can support them [21]. For example, in Bangladesh, it is important to understand the level of coordination required across ministries to harmonize job descriptions and activities of different cadres of health workers [21]. Structural changes in these programs’ governance may also influence CHW motivation and performance [31]. In additional to national and regional level governance, there is also a need to study the effect of civil society players in advocating for CHW programs and improving CHW accountability [21].

Performance and quality of care

How do each of these factors—training, community embeddedness, institutionalization, governance—affect CHWs’ productivity and quality of care, especially as programs scale up? [22, 25, 38]. Traditionally, clients’ satisfaction with CHW services has been used as a proxy measure for CHW quality of care; however, as programs scale, measuring their services’ technical quality becomes critical [15]. An equity perspective in the study of quality of care is integral. Does quality of care provided vary by type of community group, wealth, or gender [6, 15, 16, 21, 32], and how can the services provided by CHWs be channeled to reduce stigma and adapt socio-cultural norms? [36]. What types of drugs can CHWs safely administer, and what might be the implications for antibiotic resistance or a patient’s informed choice of contraceptives? [16]. With the advent of new technologies in the community health space, it becomes important to test innovative technology-based approaches (such as the use of pre-filled injection devices and mobile job-aids) on quality of care [32]. Review studies have pointed out the challenges in assessing the impact of CHW programs on quality services, in part due to low-quality evidence [23, 30] and lack of standardized metrics in quality assessment [28]. To the furthest extent possible, future research should employ validated quality metrics, and where possible use longitudinal research designs, to assess the effects on quality of services over time [27]. CHW performance also needs to be assessed in specific populations such as adolescents [26], refugees and displaced populations [46], and in epidemic settings such the Ebola epidemic in DRC and Liberia [21].

Cost-effectiveness of CHW programs

There is some evidence of cost-effectiveness of CHW programs within the context of delivering services for tuberculosis, HIV, pediatric asthma, and management of malaria [6, 25–27, 47]. However, given the challenges of assessing cost-effectiveness of integrated multi-component interventions with varying typologies, much of the data continues to be of low quality, insufficient, and limited in its generalizability [6, 25–27, 47]. The need for costing data and cost-effectiveness evidence is a common refrain from policymakers, and it emerged as one of the most important questions by government representatives participating in the ICH partners meeting. Some priority areas included the need to assess the cost of the different models and conduct comparative cost-effective work in Bangladesh and Kenya and assess the cost of innovative models such as CHWs using technology to improve quality of care in Uganda and Liberia [21].

Further studies on costs and cost-effectiveness are needed for specific interventions delivered by CHWs such as vaccination programs, malaria prevention and treatment, and mental health programs [24, 32, 34, 48], as well as comparing CHW models that are single-disease focused versus “generalist” CHWs that offer a range of community-based services [26]. Cost and cost-effectiveness studies may be guided by certain principles to be more operationally relevant across contexts—for example, varying the timeframe to explore the impact of the program over short and long terms [23], incorporating detailed costing of personnel (e.g., including supervisory costs) and other resources associated with the intervention, and using multiple comparative scenarios as alternatives [28, 29, 34]. While challenging to implement, costing studies should also consider costs against the quality of service [29]. Vaughan et al. suggest that given the level of embeddedness of CHW programs within the community, cost-effectiveness studies also need to capture societal costs and benefits, including aspects such as social capital, trust, and costs to clients [47].

Prioritization of research areas based on consultations with technical advisory group (TAG)

Discussion with the TAG helped to prioritize and contextualize the research priorities identified through the literature review and consultations with ICH partners. There was consensus on the need to consider the community health system holistically—beyond CHW programs—in that the ability of CHWs to perform effectively often depends on their community and health systems environments. Additional priorities not highlighted in the literature emerged—research on the effectiveness of CHW programs to address non-communicable diseases and the use of digital tools to strengthen CHW programs.

The priority setting survey yielded a range of responses from the 18 experts convened and confirmed many of the priorities that were also identified through the literature.