The global pendulum swing towards community health workers in low- and middle-income countries: a scoping review of trends, geographical distribution and programmatic orientations, 2005 to 2014

The review methodology followed broadly the steps proposed by Levac et al. [12]. HS and UL developed the scoping review questions (and coding schemes) based on previous literature reviews. After an initial search, the volume of new literature in recent years immediately became apparent. Since we were primarily interested in trends and patterns, we decided to limit the scope of the search to the indexed literature and focus on a 10-year period. In October 2015, we searched the following electronic databases through EBSCOhost: Academic Search Premier, Africa-Wide Information, CINAHL, PsycINFO, SocINDEX and MEDLINE, for all English-language publications on CHWs, indexed from 2005 to 2014, and in countries defined by the World Bank as low- and middle income (http://​data.​worldbank.​org/​news/​new-country-classifications-2015). We also searched the Cochrane database for systematic reviews on CHWs.

Recognizing the wide diversity of forms and titles of CHWs across the globe, we developed an inclusive list of search terms (Table 1). However, we deliberately excluded certain terms, such as traditional birth attendants and facility-based lay counsellors, as these would have touched on significant other bodies of literature.

The search was conducted sequentially with all the EBSCOhost databases except MEDLINE searched together in the first step, followed by the search of MEDLINE in a second step. Each step yielded 5635 and 1445 hits, respectively.

From October 2015 to January 2016, two authors (HS and DO) independently screened the titles and abstracts obtained in these searches, based on the inclusion criteria (Table 1). In this initial process, we selected a total of 897 publications, which were entered in a database (Mendeley) and full texts downloaded. Entries were then independently coded by two authors (HS and DO) in an Excel spreadsheet following the scheme outlined in Table 2. In an iterative process that involved removing publications that did not meet the inclusion criteria, and adding relevant publications identified in the subset of review papers, a final total of 678 publications was selected for analysis.

Coding relied on the abstract in the first instance, with further verification based on the full-length article, if the abstract was not sufficient. We categorized each entry into year of publication, country and region, and type of publication—empirical, review or “analysis”. Empirical pieces reported research findings (qualitative or quantitative), and reviews were formal appraisals of the literature based on an identifiable search strategy. Some papers used “review” in the title in a more colloquial sense but were substantive reflections or commentaries, drawing on the literature, but not adopting a structured review strategy. We categorized these papers as “analyses”. Publications were also categorized by programmatic focus based on the conventionally accepted approaches (such as maternal-child health, malaria, reproductive health, comprehensive), drawing firstly on the title and abstract, and if this was not stated by scanning the full-length paper for a description of CHW roles. Additional file 1: Table S1 gives a detailed breakdown of the items included under each of the codes. We also noted the type of role such as treatment or prevention or both—performed by the CHWs. The coded items in the Excel spreadsheet were imported into Stata (Version 14) for descriptive quantitative analysis. A qualitative, thematic analysis of key roles within each programmatic area was done, drawing on the subset of review and multi-country articles in the first instance, followed by reading of individual papers if the reviews were judged not sufficient.

The findings reported are not a full inventory of all research and publications on CHWs but rather the trends and patterns of a delimited body of literature in the field, through one search process. Given the volume of publications, we did not conduct a grey literature search. However, we recognize there are significant and influential publications [9, 13‐15] and consensus statements [7, 16] in this sphere, whose insights have not necessarily made their way into the indexed literature.

Choices were made in the classification of the paper’s programmatic focus. For example, the prevention of mother to child transmission of HIV (PMTCT) interventions, because they overlap with general maternal health (breastfeeding, antenatal care), were classified under maternal-child health (MCH) rather than HIV/TB. On the other hand, studies evaluating intermittent preventive treatment of malaria in children (IPTc) were classified under malaria because they most often emerged from a malaria programmatic focus. However, integrated community case management interventions, combining pneumonia and malaria treatment of children, were classified under MCH. The specific choices are reflected in Additional file 1: Table S1.

To limit the scope of the review, we also excluded facility-based cadres as it would have meant assessing a growing body of work on task shifting within health facilities, especially in HIV-affected countries where lay counsellors have become an integral part of primary health care teams. Unfortunately, this also excluded significant developments in the field of mental health (see, for example, [17, 18]).

This review does not address a key preoccupation in the literature on the support and systems dimensions of CHW programmes, such as supervision, retention, motivation, monitoring and financing of CHWs.

Of the 678 papers, 604 (89 %) were empirical pieces, 55 (8 %) were reviews and 19 (3 %) were analyses. There was a nearly sevenfold growth in annual number of publications over the period, from 23 in 2005 to 156 in 2014 (Fig. 1).

The papers reported experiences in 46 countries (Additional file 2: Figure S1), with 17 countries contributing at least 10 publications each, amongst them the globally recognized national CHW initiatives (Table 3). Half of the publications came from the Africa Region, just under a third from the Asia/Pacific Region, and 6.5 % had a global perspective. Iran was the only country contributing experiences from the Middle East Region, and the Brazilian programme accounted for 80 % of publications from Latin America, possibly reflecting the English-language bias of the review (Additional file 3: Table S2). Three middle-income countries—South Africa, India and Brazil—each contributed 60 or more papers, together making up 30 % of the total publications.

The profile of programmatic foci in the publications, by region and country, is summarized in Table 3 and provided in full in Additional file 3: Table S2.

Programmatic emphases varied between regions (Fig. 3). In Africa, with high burdens of malaria and HIV, publications were more evenly distributed between HIV/TB, malaria and MCH. In Asia, MCH dominated as a programmatic focus, although with significant nodes of development in mental health and NCDs. The comprehensive orientation of the Latin American publications reflects the influence of the Brazilian Family Health Programme, which is delivered with the close support of health professionals and primary health care facilities.

Where this was identifiable (n = 615 papers), the roles of CHWs along the promotion-prevention-treatment continuum were coded. These were broadly clustered in four areas (Fig. 4): (1) diagnosis and treatment (notably in iCCM and malaria); (2) prevention and promotion, spanning the distribution of preventive technologies (such as contraceptives), to education (such as newborn health, breast feeding), to processes of social mobilization; (3) screening, referral and surveillance activities, such as early detection of cancers or chronic disease; and (4) counselling, care and adherence support for adults receiving treatment for chronic conditions (such as HIV/TB, mental illness). One quarter of the papers reported roles spanning two or more of these areas, with several papers suggesting the importance of combined roles in the community legitimacy of CHWs [43, 87].

Publications most often focused on evaluating or describing the work of one type of CHW. However, even in countries with recognized national programmes, the papers from this country, when brought together as a collection, portrayed a more diverse reality. In Ethiopia, for example, where the health extension workers (HEWs) are the recognized CHWs, at least two other community-based workers in communities were described: the community-based reproductive agents delivering contraceptive technologies [88] and community AIDS volunteers linked to NGOs and ART treatment programmes [53, 89]. Within the Health Extension Programme itself, HEWs relate to a cascade of community actors: they mobilize volunteer community health workers, also referred to as the Health Development Army, who, in turn, nominate female household members for training as “model households” [39, 90].

In Uganda, where community health workers are volunteers, and where roles and functions were less clearly defined nationally, the 34 papers in the collection described a plethora of disease- or programme-specific workers and interventions, including iCCM, maternal and new born health, reproductive health, malaria, onchocerciasis, antiretroviral therapy for HIV and palliative care.

Standing and Chowdhury [91] describe how community health workers in Bangladesh are positioned in dense and plural local health care environments, where they are but one player amongst the many informal, formal and traditional sources of care and healing which community members draw on. In such contexts, CHWs play a variety of different roles—as a generic provider linked to an agency (such as Building Resources Across Communities (BRAC)), specialized workers (e.g. reproductive health distributors), as agents that mediate relationships between households and the formal health system or as expert patients.