Participants and recruitment

Study participants were selected using purposive, non-probability sampling. Respondents were selected if they had detailed knowledge of community MNCH services (eg, They were involved in community MNCH policy, implementation, or supervision, or were recipients of CHW services) before, during, or after Ebola, and/or with the community-level Ebola response. Respondents included representatives from the MoH, United Nations agencies, international non-governmental organizations (INGOs), national NGOs, and traditional healers (THs) unions. At the district level, respondents included key implementing partners, UNICEF staff, district health personnel, local government authorities, and TH associations. At the community-level, informants included community leaders, caregivers of children under five, CHWs, traditional birth attendants (TBAs), officers in charge of health facilities, maternal and child health aides in health facilities, members of community health committees (CHCs), and EVD survivors.

Data collection

Data collection was conducted between February and August 2016 (February-March in Liberia; May-June in Sierra Leone; and July-August in Guinea). Data were collected through IDIs and FGDs. Field researchers used semi-structured IDI and FGD guides that were broadly based on four study objectives given above. Within each of these broad themes, questions and probes were informed by the iCCM Benchmarks Framework [14], which ensured that key service delivery components (coordination and policy setting, costing and financing, human resources, supply chain management, service delivery and referral, communication and social mobilization, supervision and performance quality assurance, monitoring and evaluation and health information systems) would be captured in the data. Separate guides were used for each type of respondent, but key themes were addressed in each interview and FGD, which allowed analysis of themes across participant groups and fieldsites. IDI and FGD participants were identified by UNICEF, MoH, and district health officials or by community leaders based on their experiences with the Ebola outbreak and community- and facility-based health services. The research was designed to facilitate input from multiple stakeholders using a phased approach, so that issues raised by one group of interlocutors could be discussed with other groups of stakeholders as appropriate. This ensured the collation of in-depth material and the rigour of its validation and triangulation. Research teams were expected to complete approximately 44 interviews and 16 FGDs (with 6-8 participants each) per country.

Data collection was carried out by one lead researcher from Anthrologica (a research organisation specializing in applied anthropology in global health), and one local research assistant in each country. The lead researcher led the interviews and FGDs, with the assistant providing translation when necessary and taking notes. Research assistants were experienced qualitative researchers who had previously worked with Anthrologica, and were trained prior to data collection on the study objectives and procedures.

Data analysis

At the end of data collection, audio recordings of all interviews and FGDs were transcribed and translated into English as necessary. The transcripts were reviewed for accuracy and were cross-referenced with the research team’s field notes. Any areas of inconsistency were resolved after additional review of the original audio file.

Dominant themes were identified through the systematic review of interview and FGD transcripts. Thematic analysis was primarily deductive, based on the research questions and service delivery components (eg, policy, coordination, training, supply chain, supervision, social mobilization). Additional themes were identified inductively from the data. All qualitative data were coded and analyzed by hand. Two researchers coded an initial subset of the data. The two sets of codes were then compared and any discrepancies were discussed and consensus was reached on the codes that were used on the rest of the data.

Service delivery

Prior to Ebola, the governments of Guinea, Liberia, and Sierra Leone recognized the importance of community-based MNCH services to improve access to essential care. National policies were launched in 2011 in Liberia and in 2012 in Guinea and Sierra Leone that defined the roles, responsibilities, selection criteria, training, supervision, and reporting for CHWs. The policies were intended as a framework for harmonizing programming under MoH authority and oversight. However, because of a reliance on donor funding and INGO implementing partners, in reality there was a complex landscape of stand-alone projects and limited MoH oversight.

CHWs in the three countries worked as volunteers, but received small travel allowances and some non-financial incentives, such as boots, rain gear, and flashlights. Although CHWs worked in their home communities, they usually covered several additional communities, often long distances from their home.

The services provided by CHWs varied by country, district, and implementing partner. The most common services delivered included iCCM; community sensitization and health promotion; screening for childhood undernutrition; childhood immunization; distribution and promotion of insecticide-treated bednets; and referral for danger signs in children and pregnant or postpartum women. However, most CHWs were not implementing the full package of services. In Guinea for example, CHWs were providing community case management (CCM) of malaria, but were not treating pneumonia or diarrhea.

A number of challenges were highlighted that hindered effective implementation of community health services. These included:

Supply chain management and supportive supervision were frequently highlighted as key challenges in all three countries. Common supply chain bottlenecks included shortages of drugs at the national level, lengthy bureaucratic processes to procure drugs from national warehouses, and health facilities choosing to keep medication at the facility rather than supply CHWs. As a result of persistent stock-outs, iCCM had effectively ceased in three of the four study districts in Liberia prior to the Ebola outbreak. In Guinea and Sierra Leone the situation was less severe, with generally functioning iCCM/CCM programs that faced occasional drug stock-outs. Supervision proved challenging because of shortages of health staff and the fact that many CHWs were required to travel long distances to attend supervision meetings at the health facilities and they often incurred travel costs that exceeded their monthly incentive.

CHWs frequently reported that they were not adequately incentivized. Besides the lack of financial incentive, many CHWs had not been supplied with basic equipment such as rain gear and ID cards. A CHW in Guinea explained the challenges that CHWs faced in carrying out their work:

CHWs also faced challenges in balancing their non-CHW paid work with voluntary CHW activities. Some CHWs admitted that they were not always available for community health duties because of other commitments. Both health workers and NGO staff suggested that they were not able to hold CHWs to account for their community health responsibilities because they were not being paid. National- and prefecture-level stakeholders remained concerned that CHW interventions were unsustainable without some form of increased incentive.

Care seeking and referral compliance

In many communities, caregivers appeared well informed about CHW services and reported that they would first seek care from a CHW if their child was ill. Conversely, in communities where CHWs did not have a routine supply of medication, CHWs were considered to primarily provide first aid before referring the child to a health facility for treatment. In communities that did not have a resident CHW, caregivers had less knowledge about and experience of their services.

Many barriers to referral compliance were reported. These included distance from home to the health facility, poor road conditions, limited transport, and the cost of transport. A perception of poor quality care provided at the health facility also caused community members to be reluctant to seek care outside the community. Frequently caregivers did not follow referral advice but sought care from alternative community-based providers, such as THs. Stakeholders described several community-based strategies that had been adopted to facilitate referral, including the CHW accompanying the caregiver and child to the facility, and the community pooling funds to pay for transport, but the impact and scale of such initiatives remained unclear.

Other community-based actors

Although TBAs traditionally carried out deliveries in their communities, each country had enacted policies to promote facility-based delivery. TBAs were prohibited from assisting deliveries in the community, and were instead enlisted as “maternal health promoters” to encourage women to seek family planning, antenatal care, and plan for facility-based deliveries. However, despite these policies, community members and TBAs acknowledged that routine deliveries had continued to be performed in the community, especially in communities that were further from a health facility. Although TBAs were recognized as an official health cadre in all three countries, they received limited support, not only in terms of incentive, but also with regards to training, supervision, and material resources. TBAs and CHWs frequently reported that they worked “hand-in-hand” and perceived their roles as being distinct but complementary, and it was broadly understood that CHWs had a supervisory role over TBAs.

Communities, particularly in rural areas, strongly believed in “traditional” practices and THs were often the first person caregivers would seek treatment from for ill children. THs were largely excluded from government policy, were not engaged by the health system, and did not formally interact with CHWs. Many expressed frustration at what they considered to be neglect by the government.

Each country had some version of CHCs that were recognized in MoH policies, but whose roles were not clearly defined and whose level of activity in communities varied. In communities with CHCs, they were made up of community leaders, religious leaders, women’s leaders, and youth leaders. Their activities included health promotion, community mobilization, selection and supervising of CHWs and TBAs, liaising with health facilities, promoting facility-based deliveries, promoting the use of CHWs, accompanying ill people to facilities, and pooling community funds to finance emergency transport and treatment. Where CHCs were active, they reportedly worked closely with the CHWs, supporting their activities and amplifying their health promotion messages. CHCs tended to be more active if they were supported by an NGO for a specific project and stakeholders expressed concern that CHCs were only active if financially motivated.

Community leaders played an important function brokering relations between the community and CHWs, and CHWs strongly emphasized the need for community leaders to advocate for them. As one CHW in Sierra Leone affirmed, “The chief is going to be your ambassador and has to introduce you to the community.” Community leaders also enforced the community bylaws, including the prohibition of home deliveries.

Service delivery

The extent to which community-based MNCH services continued during Ebola varied by country, district, and phase of the outbreak. The majority of stakeholders engaged in this study concluded that CHWs were willing to continue their regular activities during the Ebola outbreak, motivated by a commitment to their communities and a desire to contribute to “a cause larger than themselves.” However, the ability of CHWs to continue providing services was largely dependent on the level of support they received. In all three countries, case management services largely stopped or drastically declined in the early months of the outbreak. Reasons for withdrawal or reduction of iCCM/CCM services included CHWs’ fear of becoming infected, reorientation towards Ebola-related activities, lack of support, and/or because of a clear directive from an implementing partner or health facility to cease providing treatment.

In recognition of the need to restart community-based services and the fact that CHWs were at high-risk of exposure to Ebola, the WHO and UNICEF, in collaboration with the MoH in each country, developed a “no touch policy” that recommended that CHWs avoid all physical contact with patients and base their assessments solely on patient history and observations [15]. The “no touch policy” was adopted in all three countries and training of CHWs on the protocol began in late 2014. Many CHWs experienced a long period between when the outbreak was declared in their district to when they received training. In Macenta, Guinea, for example, the December “no touch” training of CHWs came eight months after the first Ebola case was reported in the district.

There was also a great deal of confusion over the intent and practical implementation of the policy. Many interpreted the policy as an instruction to cease treatment entirely and cited this as a reason for stopping services. A number of district- and national-level stakeholders engaged in this study also conceded that the policy had been introduced too late in the response and only after a number of CHWs had died having been infected by Ebola during the course of their work. CHWs also displayed significant variation in their understanding of the implications of the “no touch policy” on their clinical decisions within iCCM. For example, because of the overlapping symptoms of Ebola and iCCM illnesses, there was a lack of clarity for CHWs on when to treat children with fever or diarrhea in the community vs when to refer them to a facility. They felt that the “no touch policy,” and in particular the prohibition on the use of rapid tests for malaria, limited their ability to differentiate children with malaria or diarrhea from suspected Ebola cases. Some CHWs chose to refer all patients to a facility, while others treated all fevers presumptively for malaria and only referred if the patient did not recover after 24-48 hours. Alternatively, some CHWs only referred fever cases if there were other signs or symptoms of Ebola (eg, diarrhea or vomiting). As one CHW in Sierra Leone reported, “The caregiver was telling us the condition of their child and we could give them drugs based on their explanation, but it was difficult to diagnose the actual sickness the child was suffering from.”

As discussed above, iCCM services in three of the study districts in Liberia had largely ceased prior to Ebola. Quantitative data from Bong (Figure 2), the one district that had ongoing service provision at the start of the Ebola outbreak, showed that iCCM and immunization services largely ceased during the height of the outbreak in the district, from July to December 2014. By January 2015, there were no more Ebola cases in the district, and the number of treatments recovered to the pre-Ebola level.

Figure 2.  Number of integrated community case managemen (iCCM) treatments and children immunized at community level in Bong, Liberia, January 2013 – December 2015.

In Sierra Leone CHWs explained that continuation of routine services was largely dependent on the extent to which their community had been affected by Ebola: in communities with a high Ebola caseload, CHWs tended to focus mostly on EVD-related duties. Quantitative data from the four study districts (Figure 3) show that iCCM treatments and malnutrition screening declined, but did not cease entirely, from the early period of the outbreak in June 2014. iCCM services started to recover around December 2014, which was still during the height of the outbreak in the districts, while malnutrition screening continued at lower levels throughout the period observed. In contrast, the number of home visits and referrals to health facilities did not seem to be affected by the outbreak.

Figure 3.  Number of integrated community case managemen (iCCM) treatments, children screened for malnutrition, antenatal/postnatal visits, and referrals at community level in Kenema, Tonkolili, Kailahun, and Bombali Districts (aggregated), Sierra Leone, January 2013 – December 2015.

In Guinea, CHWs in Macenta, a district that was affected by Ebola earlier in the outbreak, and Kérouané, which bordered Macenta, were instructed to stop all community-based treatment of malaria and refer patients to local health facilities. Dubréka and Forécariah did not record active Ebola transmission until August and September 2014, respectively. By then, implementing partners in both districts had benefited from time to prepare and adjust their activities. Consequently, CHWs were advised to continue CCM using the “no touch” approach and CCM programming largely continued. MNCH promotion activities during Ebola followed a similar pattern; in the districts that were affected earlier in the outbreak, communication and social mobilization activities largely shifted from MNCH to Ebola infection prevention control (IPC). In the districts that were affected later, CHWs reported that they continued MNCH sensitization activities. As one CHW concluded “Our regular activities on the sensitization of maternal health and child health were not disturbed at all, because any time we go to the field, we start first the sensitization on our malaria, pneumonia, and diarrhea activities, then we add on the Ebola sensitization.” By early 2015 the intensity of Ebola transmission had eased in Guinea. While the Ebola response was ongoing, partners began re-committing resources to strategies aimed at increasing accessibility and utilization of essential MNCH services. The subsequent expansion of community health services included training CHWs to provide iCCM and screen for malnutrition. From March 2015, CHWs were trained on iCCM with “no touch” modifications across three of the four study districts.

In all three countries, the supply chain and supervision issues that existed prior to Ebola continued or were exacerbated during the outbreak. In areas where CHWs were instructed to continue MNCH services, program implementers faced a number of service delivery challenges in addition to those that existed prior to Ebola. These included NGO policies restricting staff movement for safety reasons; the inability of NGO workers to enter many communities due to hostility and community resistance; the closure of health facilities due to quarantine, illness, or the death of health workers; and an overwhelming preoccupation with Ebola-related activities. CHWs who had not been consistently supervised during the outbreak expressed a keen sense of abandonment. As one CHW in Sierra Leone asserted, “We went nine months without seeing the partner staff. They used to supervise us and were responsible for paying our incentives. If those things are not provided then how can we work?”

Care seeking and referral compliance

Because of their ties to health facilities, communities displayed elevated levels of fear and mistrust toward CHWs. Many caregivers of ill children avoided CHWs and used traditional remedies or sought care from drug vendors. As one caregiver in Sierra Leone recalled, “We were afraid of using all medicines from the health facility and the CHWs because we heard that there was an Ebola tablet and if you drink it you will have Ebola. So we were using the traditional herbs instead.” Caregivers were also afraid of what they perceived to be indiscriminate referral to health facilties or Ebola treatment units (ETUs). As a CHW in Sierra Leone explained,

It should be emphasized, however, that according to many stakeholders, the long-established close and trusting relationship between communities and their CHWs was significantly more resilient than the relationship communities had with facility-based health workers. Although CHW service utilization declined, caregivers in the three countries confirmed that they were still more likely to seek care from CHWs than from health facilities. Community leaders also emphasized that they were primarily fearful of outsiders, and the provision of effective treatment by CHWs helped to dispel rumors and rebuild community trust in CHWs. As one CHW in Guinea concluded,

Stakeholders reported that as communities gradually adopted Ebola IPC measures, the rate of utilization for MNCH services also increased. Stakeholders attributed this to a combination of strategies, including intensive and persistent sensitization by community and youth leaders, religious leaders, health facilities, and CHWs. A CHW in Sierra Leone explained, “We kept talking to the community that they should please accept the medicine and give it to their children so that they would not die. And they all eventually accepted the message.”

Other community-based actors

As pregnant women started to avoid health facilities, TBAs reported that they conducted increased numbers of home deliveries. However, most TBAs reported that they were given neither training nor IPC materials. The following testimony provided by a TBA in Liberia was broadly illustrative of the role TBAs played in assisting community-based deliveries during the Ebola outbreak.

Authorities perceived that THs could be “super-spreaders” of Ebola, and in response the governments called for a cessation of all traditional healing practices. In Sierra Leone and Guinea, this was bolstered by the implementation of bylaws which prohibited traditional practices. Some THs were trained on Ebola warning signs, encouraged to avoid physical contact with patients, and instructed to refer all ill people directly to health facilities. Nevertheless, many THs highlighted their frustration that they were not incorporated and utilized in the response more quickly, despite often being the first point of contact for patients. According to THs, it was not until “Ebola was almost finished” that the governments and Ebola response partners started to train and actively include them in the Ebola response. As one TH in Guinea concluded,

When Ebola started we really saw difficulties because we were not aware of which type of sickness it was.

Other community-based actors

At the community level, the mobilization of “Ebola task forces” or “community watch committees” was seen as a useful mechanism to quickly coordinate response efforts and enforce protocols at the community level. Members of the task forces included community, religious, womens, and youth leaders; THs; CHC representatives; and CHWs. Members were specifically chosen by the community to ensure that those most resistant to Ebola response activities were included. As one district health representative in Guinea explained, “We attacked the problem at the source by training and including...youths, since youths were throwing stones at Ebola workers. And we included women leaders since women were often against safe burials conducted by the Red Cross.” Amongst other actions, the task forces oversaw community-based surveillance (for example monitoring for strangers or symptomatic people entering the village), active case finding and referral, promoting safe and dignified burials, establishing hand washing stations, promoting compliance with the isolation and referral of ill people, and taking care of orphans. All of these activities included social mobilization on the prevention of Ebola, and dispelling or debunking “Ebola rumors.”

CHCs were a natural platform for the establishment of Ebola task forces and subsequent coordination activities, and stakeholders suggested that communities with a CHC in place prior to the Ebola outbreak appeared to be more resilient and were more likely to self-mobilize and organize community-led solutions. CHC members described their roles during the response as similar to those of CHWs. The main difference was that CHC members undertook these activities largely without formal support. As one CHC member in Liberia emphasized,

Despite the trust that communities felt for TBAs, they were rarely included in the formal EVD response because they were considered not sufficiently skilled to take on Ebola-related activities. Similarly, THs were sidelined throughout much of the formal response. In some hotspot areas, THs were eventually trained and supported to work as social mobilizers, but this collaboration came late in the outbreak.

Service delivery

As the countries transitioned into the post-Ebola recovery period, Ebola-related activities were scaled back and implementing partners focused again on MNCH programming. Partners worked quickly to strengthen community-based MNCH services where they had continued throughout the outbreak and to restart services where they had ceased. While some organizations had returned to “business as usual,” with implementing partners continuing to run short-term programs, often in parallel to the public health system, other partners were scaling up more comprehensive community health packages in line with new national policies.

Implementing partners discussed a number of ongoing challenges in implementing community-based MNCH services after the outbreak. Drug stock-outs remained a key challenge to service provision, as before Ebola. Stakeholders in all three countries confirmed that when CHWs were re-activated and when new iCCM services were established post-Ebola, CHWs were given initial stocks of drugs for the first month or so. However, once the initial stocks were depleted, CHWs experienced frequent and persistent stock-outs. CHWs in most of the study areas reported that widespread drug stock-outs prevented them from providing community-based treatment at the time of the study.

Stakeholders also confirmed that challenges in supervision and monitoring of CHWs were similar in the post-Ebola period to those experienced before the outbreak, particularly because some of the management structures that were put in place during the response had since been removed. Program implementers highlighted that the lack of dedicated personnel to supervise CHWs was detrimental to the successful implementation of iCCM.

Care seeking and referral compliance

With the restart of routine services, the MoH and partners implemented social mobilization and community engagement strategies to rebuild trust, increase utilization of services, and improve referral compliance. Respondents suggested that the mistrust communities had harbored for CHWs during the Ebola outbreak had largely resolved. In a number of villages, participants reported a greater depth of respect and trust for their CHWs because of the efforts made by CHWs to protect them during the Ebola outbreak. Many CHWs asserted that care seeking for childhood illness had almost returned to pre-Ebola levels in areas where CHWs had iCCM drugs available. A CHW in Sierra Leone concluded,

Officers-in-charge at health facilities also reported that patient numbers had returned to normal and most caregivers confirmed that they would attempt to comply with referrals made by their CHW.

However, the drug supply issues described above severely impacted both CHW motivation levels and community perceptions of CHW services. Many caregivers cited drug stock-outs as a determining factor in their decision to seek care from THs or other informal providers. Community-level stakeholders emphasized that the lack of medicines held by both CHWs and health facilities persistently caused delays in treating ill children. As one community leader in Guinea concluded,

In Guinea, the continuation of a cost-recovery policy meant that community members were required to purchase antibiotics, zinc, and oral rehydration salts from CHWs. While CHWs appreciated the financial benefit, they were concerned that this strategy was affecting caregiver perceptions of CHWs and utilization of their services. In some areas, caregivers reportedly turned to purchasing drugs at lower cost in the market. Some CHWs also reported that caregivers would “boycott” CHWs in protest of having to pay for drugs. Highlighting the distrust this policy caused, one CHW in Guinea explained,

CHW remuneration and motivation

The issue of remuneration for CHWs was one of the most hotly debated topics among stakeholders at all levels in this study. In recognition of concerns about the sustainability of community health services, all three countries had developed new community health policies and strategies that call for national cadres of well-trained, paid CHWs providing a broad package of services. The policies in Sierra Leone and Liberia call for monthly payments of around US$ 25 and US$ 70, respectively, for salary and transportation costs. In Guinea, the policy did not specify an amount, but a key MoH representative confirmed that the government intended to implement a policy requiring all CHWs be paid a monthly payment of US$ 60. Stakeholders at all levels were in strong support of the proposal to pay CHWs, but they also expressed concerns over the feasibility of implementing this policy. The current model of external donor funding was not seen as a viable or sustainable option for supporting a large-scale, institutionalized community health program. A number of national stakeholders suggested that CHW salaries would have to be funded primarily through the government payroll, which would also strengthen accountability and oversight. However, some participants also questioned the viability of paying CHWs through the government payrolls, which did not yet have the capacity to remunerate facility workers regularly, and were plagued by corruption and a lack of transparency.

For their part, CHWs clearly articulated their dissatisfaction with having to return to volunteerism, especially after the experience of receiving higher payments for Ebola-related activities. Many CHWs suggested that they would consider leaving the position if they were not better incentivized in the future. As one CHW in Guinea concluded, “A time will come when we have to first take care of our families and make sure they are fed and survive, so we will one day see someone dying and we will not attend to the person because we will be attending to our families.”

They also resented that the government had afforded them little recognition for their contribution to the Ebola response. As one CHW in Sierra Leone explained,

Less controversial were discussions regarding non-financial incentives. These included preferential selection for paid work to implement health campaigns, opportunities for scholarships to facilitate career development, financing education for CHWs’ children, and micro-financing for small businesses. CHWs regarded all these initiatives to be positive and desirable.

Other community-based actors

Across the three countries, many stakeholders, particularly community leaders and CHWs, emphasized the need to formally recognize TBAs and CHC members and their important role in the provision of community-level health services in the post-Ebola period. As part of the recovery process, there was a renewed focus on establishing or re-activating CHCs to ensure that communities have a role in the planning, implementation, financing, and monitoring of community health initiatives.

Many TBAs who had performed an increased number of home deliveries during the Ebola outbreak had ceased conducting deliveries. However, home deliveries were still taking place, particularly in the hardest-to-reach communities. TBAs emphasized that after the outbreak was declared over, they had not received any formal acknowledgement of their role during the response. As one TBA in Liberia concluded, “We were not even called to say thank you.” They were keen to continue working, particularly with CHWs, but stressed the need for recognition and remuneration.

THs also resented their marginalization during the Ebola response, the ongoing decrees prohibiting use of THs, and persistent efforts of CHWs to discourage the utilization of THs. They expressed a firm desire to be integrated into the health system and maintained that they complemented the efforts of health workers and CHWs by offering treatment for conditions not suited to biomedical interventions, including psychological illness. District-level stakeholders also recognized that community members were likely to continue seeking care from THs for a range of illnesses, and acknowledged the need for them to be included in community health plans. They suggested, for example, that THs had a role to play in community-based surveillance or community sensitization.