Background

Community-based health worker (CHW) programs are undergoing a resurgence, as these health workers are envisioned to be culturally adept members of comprehensive and people-centered primary health care teams that will enable universal health care [1]. The last decade has seen both the introduction as well as the re-invigoration of national CHW programs in many low- and middle-income countries (LMICs) [2, 3]. These programs involve the delivery of community-based health services by paid or volunteer health workers with fewer than 2 years training. There is a rapid growth of evidence on the effectiveness of community-based interventions [4, 5], positive experiences with reinvigorated national CHW programs [2], and renewed interest in stronger national CHW programs [6]. Health systems in LMICs and high-income countries (HICs) are expanding their utilization of CHWs in order to meet population health needs, improve access to services, address health inequities, and improve health system performance and efficiency [7]. Policymakers need evidence-based guidance to further develop this cadre of the health workforce. As a first step in developing policy guidance on health policy and systems support to optimize CHW programs, the World Health Organization (WHO) commissioned a systematic review of available reviews related to CHWs.

This systematic review synthesizes existing reviews on CHWs in order to map what is known about these programs. We present evidence on the roles and capacities of CHWs as well as the health system enablers that can support their functionality. We reviewed heterogeneous evidence to identify the types of interventions that CHWs provide, as well as optimal approaches to training, support, supervision, and remuneration, and health system integration (i.e., recognition in national health care planning, regulation, and implementation) [8].

Methods

Search strategy

We searched for articles published between 1 January 2005 and 15 June 2017 in 11 electronic databases: PubMed, Embase, PASCAL Biomed, the Cochrane Library, Ovid’s Global Health, WHO Global Health Regional Libraries, the Database of Abstracts of Reviews of Effects (DARE), Epistemonikos, Health Systems Evidence, PROSPERO, and the National Guideline Clearinghouse of the US Department of Health and Human Services. Searches were developed and conducted by an academic librarian (co-author MG) and peer reviewed by a second librarian prior to implementation.

The systematic literature search used a combination of controlled vocabulary and keywords for two concepts: (1) reviews and (2) community-based health workers (e.g., “community health worker”, “lay health worker”, “close-to-community provider”). We used the validated systematic review filter for PubMed [9] and expanded it to catch 30 key articles. Similarly for Embase, we used the validated Wilcynski and Haynes, “small drop in specificity, substantive gain in sensitivity” systematic review query [10] and expanded it with additional terms (metanalysis; review:ti), to include, for example, all titles with the word “review” in them. In the other nine databases, we did not use pre-developed review filters but instead used simpler search strings for the concept “review.” We did not limit to language. All titles and abstracts relevant to our study were retrieved and searched for full text. See Additional file 1 for the full PubMed search strategy.

Eligibility criteria, screening, and article selection

Articles were included if they were (a) reviews and (b) focused on CHWs. We included systematic reviews as well as non-systematic reviews (such as realist, narrative, scoping, and literature reviews), because many non-systematic reviews provided insight into CHW program design and health system integration. Our inclusive approach brought together reviews on CHWs that used a wide range of synthesis methods to comment on many features of CHW programs, going beyond the effectiveness focus of systematic meta-analysis. We defined CHWs as health workers based in communities (i.e., conducting outreach from their homes and beyond primary health care facilities or based at peripheral health posts that are not staffed by doctors or nurses), who are either paid or volunteer, who are not professionals, and who have fewer than 2 years training but at least some training, if only for a few hours. Adhering closely to this definition led us to include some programs, such as those for peer supporters and traditional birth attendants with some training, that reflect divergent and context-specific understandings of the term “CHW.” We excluded articles that did not directly mention CHWs or mentioned them only in passing without information on their role. Article titles and abstracts were divided and assigned for independent review to two authors from among KS, HBP, SWB, KDR, and GP, with a third author from among the same group selected on a revolving basis to resolve disputes. Full texts of retained articles underwent a final screening for eligibility.

Data extraction and quality assessment

Included articles were divided among KS, HBP, SWB, KDR, and GP for detailed data extraction. Data extractors used a pilot-tested framework (in Excel) that synthesized content on the following topics, adapted from the 2006 World Health Report’s framework on the working life of health care providers [11]: CHW roles and capacities, training, deployment, performance measurement, remuneration and incentives, support and supervision, cost effectiveness, community embeddedness, logistical support and supplies, and integration into health systems. KS spot-checked the data extraction by frequently returning to original articles for verification.

Two authors (SWB and MG) assessed the methodological quality of the systematic reviews using the 11-item validated Assessing the Methodological Quality of Systematic Reviews (AMSTAR) criteria [12]. They began by both rating the same 10 systematic reviews and then compared and discussed their ratings to obtain consensus on how to proceed. They then divided the remaining systematic reviews between them and rated a random sample of 10% in duplicate to check agreement. Disagreements were limited and resolved through discussion. For two AMSTAR items, we assessed the articles according to the original (strict) AMSTAR criteria and also for adapted (relaxed) criteria that we developed to more appropriately assess the quality of included systematic reviews. See Additional file 2 for an explanation of the ratings. The non-systematic reviews used a diverse range of non-systematic approaches to evidence synthesis across a wide array of research questions, making the application of a standardized quality criteria inappropriate.

Throughout this report, we use the term CHW although many review articles and individual studies used different terms such as close-to-community provider or trained traditional birth attendant.

Results

From 4 139 unique references identified in our search, 122 reviews met our inclusion criteria (Fig. 1). Additional file 3 provides an overview of the included reviews, which can be searched and filtered for regional focus, review type (non-systematic, systematic, meta-analysis), population focus, health issue, nature of the intervention, findings on CHW capacities and/or intervention outcomes, and AMSTAR rating. Additional file 4 presents a summary of the main findings of all included articles. Additional file 5 presents complete references of included and excluded articles.

Fig. 1
figure 1

Diagram of review selection process

Full size image

Of the 122 included reviews, 75 were systematic (including 34 meta-analyses) and were assessed using the AMSTAR quality criteria (Additional files 2 and 3). Seven of the 11 AMSTAR indicators of quality were met by the vast majority of the systematic reviews included in our study, while the remaining four AMSTAR quality indicators (duplicate data screening and data extraction, gray literature searched, publication bias discussed; and included and excluded studies listed) were less commonly met.

Most of the reviews focused on LMICs (n = 83) and a range of primary health care (n = 14), child health (n = 13), and maternal and child health (n = 14) interventions. High-income country reviews (n = 29) tended to focus on non-communicable diseases (n = 12) and reaching specific underserved groups (n = 7) (Table 1).

Table 1 Health topics discussed in the included reviews
Full size table

We now present findings from the reviews on considerations for CHW programmatic design and operation in health systems. We first present evidence on CHW functions and their contributions to improving health outcomes. We then report on health system enablers that can support CHW functionality, including optimal approaches to training, support, supervision, remuneration, and health system integration.

CHW roles and capacities

CHWs perform a variety of health system functions, which can be clustered into six general categories (Table 2).

Table 2 Health system functions of CHWs
Full size table

The number, complexity, and range of functions CHWs perform vary substantially among programs according to context-specific needs and opportunities; functions also evolve over time [13]. While there is no optimal set of tasks or workload level that maximizes CHW productivity, one review [14] cited studies that found that too many responsibilities reduce CHW productivity and service quality, and CHWs in these situations are forced to choose which tasks to perform based on factors such as feasibility, remuneration, or preference. The authors of this review conclude that CHWs are more likely to succeed when they have a clear role and a limited number of tasks. In LMICs, CHWs commonly provide curative services, and there is some evidence that being tasked with curative tasks as opposed to solely providing health education or psychosocial support may increase CHW motivation in LMIC settings [15].

Among the reviews that assessed CHW contributions to addressing specific health issues, most found that CHWs can improve health outcomes (Table 3) but many noted concerns about the low quality of included studies and emphasized the importance of health systems enablers such as training and support, discussed in later sections of this article. The reviews were heterogeneous, examining diverse CHW programs and analyzing effectiveness across a range of health outcome measures. As a result, a meta-synthesis across the reviews was unfeasible. Thus, while Table 3 summarizes evidence on CHW contributions to health outcomes, we encourage readers to refer to each individual review in Additional file 3 and Additional file 4 for details.

Table 3 CHW capacities for delivering specific health interventions
Full size table

As shown in Table 3, CHWs can make important contributions to improving health, particularly in extending care to underserved groups, and can successfully handle complex health counseling and biomedical tasks. However, CHWs can only meet their potential in performing these roles and improving health outcomes when supported by a range of health system enablers, discussed next.

Training

The proper amount and type of training required by CHWs must be understood in relation to the health system context, the CHWs’ pre-existing capacities, and the roles that CHWs are expected to play. Table 4 presents findings from the review literature on core considerations in CHW training domains.

Table 4 Summary of findings on CHW training
Full size table

Training should seek to impart both technical competency and socially oriented capacities such as skills in communication and counseling as well as awareness of the importance of confidentiality [15,16,17]. Awareness of the social and political determinants of health [18] and problem-solving skills were also identified as being important [19]. One review noted that theoretical, classroom-based competency-oriented CHW training to promote immunization in India is an inappropriate approach [20]. Other reviews suggest that some competencies such as record keeping or correctly interpreting malaria test results can be introduced in the classroom but require supportive supervision and hands-on practice to be implemented properly in the field [20, 21].

Training increases CHW knowledge and skills [22] and can positively influence CHW motivation, job satisfaction, and performance [23, 24]. However, there was no direct evidence linking training to health outcomes in one review that looked for it [25], nor is there evidence that different aspects of training or different training approaches affect CHW performance [24]. One pathway through which training can contribute to CHW motivation is by increasing community confidence in their CHWs and ultimately increasing CHWs’ confidence in their capacity to perform their duties [20, 24]. Relatedly, short and insufficient training erodes CHW confidence and reduces community trust and uptake of their CHW’s services [26].

Supervision

Supervision was often mentioned as critical for the effectiveness of CHWs, and there is some evidence regarding the benefits of supervision on CHW performance [14, 15, 23, 27, 28]. However, few details of the supervisory structure (type of supervisor, frequency of supervision, and type of training and support provided to supervisors) contributing to success were mentioned [15], and few studies have tested which approaches work best or how they are best implemented [15, 29,30,31]. Poor-quality supervision and low recognition from the health system can undermine community embeddedness and reduce CHW motivation [32,33,34]. Negative interactions of CHWs with higher-level health system actors (such as punitive supervision styles) can discourage and demotivate CHWs [33]. Supervision is often one of the “weakest links” in a CHW program, and CHW programs commonly give inadequate attention to ensuring high-quality supervision [14, 35], with negative implications for CHW empowerment [36]. Table 5 summarizes findings from the review literature on support and supervision.

Table 5 Summary findings on supervision for CHWs
Full size table

Level of education prior to becoming a CHW

There is some evidence that CHWs with higher levels of formal education prior to becoming CHWs are more effective (for example, in record-keeping, diagnosing childhood illness, and appropriately counseling clients), but more highly educated CHWs may also be more likely to drop out after deployment [24]. One review concluded that completion of primary school should be a minimum educational requirement for entering CHW training to meet the needs of underserved communities far from health centers [35].

Performance measurement

The reviews included in our study provided very little evidence linking routine supervisory performance appraisal to CHW performance as measured by researchers [15]. However, formal supervisory checklists may increase the efficiency of identifying CHWs who are most in need of further training or supervision [20].

Logistical support and supplies

Regular provision of logistical support and supplies (such as drugs and educational materials) is essential to maintain CHW program effectiveness, productivity, and respect of CHWs by the community [26, 37]. Lack of supplies is demotivating for CHWs [14, 15, 35, 38]. Table 6 summarizes findings from the review literature on logistics and supplies.

Table 6 Summary of findings on logistical support and supplies
Full size table

Remuneration and incentives

Monetary remuneration (such as salaries, financial incentives, or income from selling commodities) and non-monetary incentives (such as respect, trust, recognition, and opportunities for personal growth, learning, and career advancement) are important motivators for CHWs [15, 19, 23, 33, 39]. In Kok et al.’s [15] review on intervention design factors that influence CHW performance, 25 of the 81 studies with information on incentives reported that CHWs were dissatisfied with their incentives. Satisfaction (or dissatisfaction) with incentives was closely linked to CHW motivation and performance (or lack thereof). Improved financial remuneration can reduce attrition among CHWs in LMICs [23, 40]. CHW rights and the need of CHWs for reliable financial remuneration were discussed in only one review, which highlighted Indian CHWs’ consistent (and unmet) demand for salaried positions [41]. Table 7 summarizes findings from the review literature on remuneration and incentives.

Table 7 Summary findings on remuneration and incentives
Full size table

Deployment

There is no simple formula for determining the optimal size of a CHW’s catchment population. Instead, decisions about catchment area population should be based on a variety of context-specific considerations: frequency of contact required; nature of the services provided; expected weekly time commitment from the CHW; and local geography (including proximity of households), weather, and transport availability [14, 15, 24]. One review [42] found that for interventions consisting of home visits only, there was no consistent effect of the size of the catchment population and neonatal mortality impact. However, when the interventions involved community mobilization as well, the reduction in neonatal mortality was greater when the catchment population for the CHW was smaller. Another related finding was that a high workload can lead to CHW demotivation [23].

Community embeddedness

Fourteen reviews highlighted aspects of community embeddedness as important enablers of CHW program success [14, 15, 19, 23, 34, 35, 37, 40, 43,44,45,46,47,48]. CHWs are embedded in communities when community members trust and respect them and feel a sense of ownership over the program, such as can be achieved by giving communities a role in CHW selection and definition of CHW activities [19]. The community’s acceptance of CHWs and their sense that the CHW program is locally appropriate and “owned” is associated with CHW retention, motivation, performance, accountability, and support, and ultimately with the acceptability and uptake of CHWs’ health-related work [14, 15, 19, 23, 38, 40, 47, 49]. Locally trusted CHWs can serve as an effective link between health facilities, health workers, and communities [50], and CHWs who are embedded in their communities can provide services to difficult-to-reach populations [20, 40]. However, CHW embeddedness can lead to CHWs being caught in tensions between the community and the health system as well as between social and biomedical issues [51]. Table 8 summarizes findings from the review literature on community embeddedness.

Table 8 Summary findings on community embeddedness
Full size table

Cost-effectiveness

Research from LMICs has found that shifting aspects of HIV care from higher-level health workers to CHWs is cost-effective [50, 52, 53]. There is some evidence of cost-effectiveness for community case management of malaria by CHWs compared to standard malaria treatment at a health facility [21, 33], for the provision of mental health care by CHWs in LMICs [54], and for the delivery of multiple primary health care interventions [55]. However, one review noted that costing methods varied across studies, making it difficult to generate clear conclusions. The same review also noted that the opportunity costs borne by CHWs for volunteering their time were inadequately accounted for [33]. Table 9 summarizes findings from the review literature on cost-effectiveness.

Table 9 Summary findings on cost-effectiveness
Full size table

In HICs, interventions delivered by CHWs to reduce triggers for childhood asthma brought cost savings [56, 57]. Another HIC study reported cost savings associated with peer support for breastfeeding [58]. Three reviews found inconclusive or no evidence on cost-effectiveness: vaccination promotion in LMICs [59], control of vascular diseases in HICs [60], and outreach to underserved groups in the USA [25].

Integration into health systems

The integration of CHW programs into the health system is reported in many reviews to be a key enabler [14, 15, 17, 19, 23, 24, 26, 32, 34, 35, 38, 61]. Pallas et al. [23] highlight that the integration of CHW programs into the agendas of the ministry of health, NGOs, and international donors can strengthen CHW programs and can also help bolster programs in times of political upheaval, loss of external donor funding, and reduced prioritization by the ministry of health. Integration that fosters respectful collaboration and communication between CHWs and higher-level staff can enable the health system to benefit from the unique, practical knowledge that CHWs have and can support CHW retention; this integration can enhance the acceptability and credibility of CHW programs [14, 15, 19, 24, 38, 44]. Table 10 summarizes findings from the review literature on health system integration.

Table 10 Summary findings on health system integration
Full size table

Discussion

CHWs perform many roles in high-, middle-, and low-income country health systems and contribute to improving a range of health outcomes. However, their capacity is directly contingent on the support they receive from the health system. This review of reviews identifies a number of broad health system supports that optimize CHW programs and can be considered in light of context-specific factors to support health policy decision-making. It finds that CHW tasks should be clearly defined and should require a time commitment appropriate to the incentives/remuneration and support provided. Training should seek to impart both technical competency and socially oriented skills such as communication and counseling, including on confidentiality. Training appears to be more effective in imparting competencies by integrating hands-on practical components rather than just providing classroom learning and should be closely linked to ongoing supportive (rather than punitive or bureaucratic) supervision. Regular provision of supplies, such as medicines, communication tools and teaching aids, and transportation support, is essential for maintaining CHW program effectiveness. The review finds strong support for ensuring community embeddedness, as this is associated with CHW retention, motivation, performance, accountability, and support -- and ultimately affects the acceptability and uptake of CHWs’ health-related work. Linking CHWs to a supportive and functioning referral facility is often vital to CHW program effectiveness. Furthermore, programs must develop appropriate financial and non-financial incentives that take into account a range of factors, including the health system’s resource availability, CHW needs, rights, and expectations, and the tasks and time commitments required. The size of a CHW’s catchment population should be determined in response to the local reality, including population density, travel required, and workload.

As many countries are in the process of implementing new national CHW programs or strengthening current ones, the evidence synthesized in this review can help optimize these efforts. Ultimately, CHW programs are highly context specific. There are no standard blueprints that can be used to design and implement a CHW program. When developing programs, decisions must be made based on national, sub-national, district, and local realities.

This review also enabled the identification of several gaps in the review evidence. Relatively more (and higher quality) evidence is available on the effectiveness of CHWs in delivering specific health interventions than on effective approaches and cross-cutting strategies to integrate and support CBPs in health systems and optimize their performance [62]. There is little discussion in the review literature on the rights and needs of CHWs (with notable exceptions [36, 41]), on effective approaches to training and supervision, on CHWs as community change agents, as multisectorial actors, and on the influence of health system decentralization, social accountability, and governance.

Effectively addressing population needs for Universal Health Coverage with realistically available resources requires harnessing opportunities from the education and deployment of CHWs as members of inter-professional primary health care teams [1]. Countries should develop policies and mechanisms to integrate CHWs with the health system so as to enable these cadres to benefit from health system support and to enable the health system to achieve optimal benefit from CHWs [63]. Health system integration should foster respectful communication and collaboration between CHWs and other health system actors.

Integration of CHWs with health systems requires their inclusion into public policies, including those related to national human resources for health planning, governance, legal frameworks, and financing for health services. The requisite inputs of human and financial resources should be factored in at planning and budgeting stages and should be reflected in national health workforce and health sector strategies.

Policy dialogue about creating a strong role for CHWs in health systems must also address human and labor rights issues surrounding the CHW workforce [64], the favorable consequences of employment of large numbers of CHWs for economic growth and social development [64,65,66], as well as for achieving the Sustainable Development Goals [67].

This review faced some limitations. In including a range of study types—meta-analyses, systematic reviews, and non-systematic reviews (e.g., scoping, narrative, realist reviews)—and synthesizing findings across a broad range of issues and contexts, it was not possible to assess the overall risk of bias in the findings or to systematically account for the variable quality of the included reviews. Furthermore, presenting findings synthesized from a range of reviews necessitated a high level of abstraction and limited our capacity to present specific and important details and findings from individual studies. We encourage readers to examine AMSTAR quality scores and strength of evidence assessments in Additional file 3 for specific articles and to return to the source materials referenced for more information on topics of interest. Our definition of CHWs may not match definitions used by other teams, leading to inclusions and exclusions that may not fit the needs of all readers. In including research from high-, middle-, and low-income countries, some findings from drastically different settings may be difficult to transfer and apply. In addition, we focused only on academic, peer-reviewed literature, likely missing out on important findings from the gray literature.

Conclusion

The findings from this review can be adapted to national contexts, where the available resources to support CHW programs are highly variable. Developing and strengthening CHW programs will involve taking into account existing evidence of CHW program effectiveness, weighing options in light of a country’s existing primary health care system and needs, making informed decisions involving all stakeholders, designing and implementing the best program possible, and then adjusting course on the basis of experience, monitoring and evaluation, and findings from rigorous implementation research. Future progress in improving CHW programs will depend not only on synthesizing existing evidence but also on supporting and funding research to continually advance the contextualized evidence on how to design and implement CHW programs to maximize effectiveness [68]. CHWs can play a key role in strengthening health systems to provide universal, comprehensive, and people-centered care that is equitable, culturally appropriate, and economically feasible [1].