Evaluating the role that Care Groups play in providing breastfeeding and infant feeding support at community level: a qualitative study in Dedza district in Malawi

Aims and objectives

Our research project, focusing on Malawi’s health and nutrition policy implementation regarding exclusive breastfeeding and infant feeding support at community level, was guided by the following aims and objectives:

Ethical statement

Ethical approval for this research was obtained from Dublin City University Research Ethics Committee on the 11th January 2021 (approval reference: DCUREC/2020/268) and from the College of Medicine Research Ethics Committee (COMREC) on the 4th of February 2021 (approval reference: P.02/21/3266). During the application for ethical approval, the researchers were directed by COMREC to pay each interviewee the equivalent of US$10 in Malawi Kwacha in cash (Nyangulu et al., 2019), with which we complied. Written informed consent was obtained from all interviewees.

Study design

For the field research, we used a qualitative approach, conducting one-to-one interviews in Dedza District. The research focused firstly on mothers with children under two, who were likely to have recently received breastfeeding and infant feeding support. To examine the role that healthcare providers, community leaders and district health officials’ play in supporting and integrating Care Groups into the communities, we interviewed village headmen, HSAs and a number of district- level health officials, as well as an NGO representative. Using focus group discussions was ruled out to reduce the risk of COVID-19 and associated restrictions in place at the time of data collection.

The Care Group interventions that were the focus of this study were implemented by an international NGO that had been actively implementing the SUN programme in Dedza since 2012. Their willingness to facilitate the research and the opportunity this offered to study two Care Group interventions led us to focus on these particular interventions. Both interventions used the Care Group model. One intervention was implemented across the entirety of Dedza District, while the second intervention, providing additional support to Care Groups volunteers, was limited to half of the district’s traditional authorities (TAs). In advance of the field research, the implementing NGO explained that since they supported the establishment of Care Groups in Dedza in 2016, several communities had witnessed a collapse of care group activities, due to the fact that despite repeated efforts to train and incentivise, volunteers were unwilling to carry out the tasks expected of them. This provided us with the opportunity to explore the support for EBF in areas where no Care Groups were operational.

We made efforts to identify Care Group interventions that were established without the support of an international NGO or donor-funded intervention but were unable to find any that we could include in our research. A report commissioned by the Department of Nutrition, HIV and AIDS at the time of our research was to demonstrate that very few Care Groups have been started without external support in Malawi (GoM, DNHA, 2021), despite the national guidelines, which set out how this can be done.

Research setting

Our field research was conducted in Dedza District in the Central Region of Malawi. Poverty remains widespread in Malawi, where 61.7% of the population was reported to be multi-dimensionally poor, according to a 2021 research report (NSO et al., 2021). 20% of the country’s population live in extreme poverty, defined as an inability to satisfy basic food needs, while 70.3% live below the international poverty line of $1.90 a day (World Bank, 2020). Food insecurity is high, and a significant portion of Malawi’s population require food assistance annually, with malnutrition remaining a challenge nationally as a major contributor to preventable child deaths in Malawi (UNICEF, 2019). Dedza District has higher than the average mortality rates for the Central Region with a neo-natal mortality rate of 33 deaths per 1,000 live births and an infant mortality rate of 53 deaths per 1,000 live births, compared to national rates of 19 and 29. In Dedza, 18% of children are born with a low birth weight (below 2.5kg), compared with national average of 12.3% (NSO & ICF, 2017).

Sampling and data collection

In order to answer our first and second research objective, ‘to explore how and from whom women received information about (exclusive) breastfeeding - before delivery, immediately after delivery and in the days and weeks thereafter’ and ‘to evaluate, in places where care groups operate, if and how CGVs played a role in visiting new mothers to support infant feeding’, we interviewed mothers with children under the age of two (including a total of nine women who had a child who was born prematurely or with low birthweight) and asked her primarily about the support for EBF that she may have received after the delivery of her most recent baby/ies. We also interviewed other community members, including CGVs, HSAs, and District Health Authority personnel, in order to explore (research objective three) ‘if CGVs appeared to be working in an integrated manner with HSAs or other community-level structures, as Malawi’s national policy dictates’. The interviewees at district-level were identified by the District Medical Officer, to whom the research team first reported to introduce themselves at the start of the data gathering visit in Dedza District. Standard operational procedures were developed for the researchers to guide purposive sampling. The research team were to arrive at a study site, find and interview the Village Headman, ask to be introduced to a HSA, interview the HSA, and then ask the HSA to bring them to houses of women who had a child under two years of age. The number of women who fitted the research participant criteria and were willing to be interviewed at each location, determined the total number of interviews conducted.

In total 36 women with at least one child under two were interviewed.

An overview of research participants:

Each of the women was interviewed using semi-structured interview guides, which were prepared in English and then translated into Chichewa by the researchers who conducted the interviews. The guides were developed from our aims and objectives and drew on a realist synthesis of Care Group literature, they (Pieterse et al., 2021)²⁷ were piloted with five women before the field research took place. The original interview guides were shortened, and some language was simplified after two members of the research team (MC and JM) used them to interview five women with children under two, in two villages close to Blantyre.

All data were collected by three members of our research team (EC, MC and CM), who were joined by two additional research assistants (JM and JC). All researchers were Malawian and used Chichewa to conduct the interviews, using the Chichewa version of the interview guides. Interviews were audio recorded and later transcribed and translated into English. The audio recordings were deleted off all recorders after transcription. The translations were conducted by all Malawian research team members, who double checked each other’s translations against the Chichewa transcriptions.

Data analysis

Two researchers (PP and JM) coded all data using qualitative data analysis software Nvivo version 12 (https://lumivero.com/products/nvivo/). (A freely accessible software package, Taguette (https://www.taguette.org/), could also be used to run the same analysis), ensuring that 15% of the interviews were coded by a second person (any member of the research team) during the first round of inductive, open coding. We used Braun and Clarke’s (2006, 2021) approach to thematic analysis to guide our multiple rounds of coding: The manuscripts were read several times to familiarise ourselves with the content. The first round of coding was done inductively, where all interesting and relevant sections of text were coded openly, without any guidance, codebook or preconceived structure. The generated codes were them loosely grouped. The second round of coding focused on the development of themes, based on the open codes generated in the first round of coding. Team discussions took place to agree on the formation of themes. Further coding took place to explore differences in the characteristics or themes across the different interventions or across variations in context (Bazeley, 2009). The codes were then agreed upon by the research team and named and ordered into several clusters of related findings. This allowed for the writing of a feedback report for the NGO that facilitated the research (available on the project website), and academic publications. The codebooks are available in the open access data repository established for this research. Table 1 shows the trustworthiness table for the qualitative data.

Sources of breastfeeding support (objective one)

Overall, in the areas where Care Groups were purportedly active, only eight out of 26 interviewees (women with children under two) reported receiving breastfeeding support from Care Groups. Those who received Care Group support reported anything from having been encouraged to breastfeed exclusively during a group meeting, to home visits whereby CGVs engaged with a mother and her newborn baby on a one-to-one basis:

All 36 women who were interviewed reported delivering their baby at a health facility. On the question regarding where, or from whom, women received breastfeeding support (excluding support from family and friends), all women reported receiving some support at the facility where the baby was born, (Figure 1). ‘Support’ was often no more than an instruction from somebody at the health facility that a mother should breastfeed her newborn infant; this was interpreted as support: ”At that time, they the nurses said the baby was hungry and needed to feed; she also said that the baby was cold so I needed to warm her.” [04MCA].

Figure 1. Source of breastfeeding support (outside of family and friends).

24 out of 36 women also reported receiving EBF support during their ante-natal visits:

Seven women mentioned receiving EBF support during the ‘under-five clinics’ (monthly baby check-ups usually conducted by a health facility out-patient team that provides vaccinations, growth monitoring, etc.) and six out of 36 women reported receiving support from a HSA. From the responses it appears that the HSAs tended to be the individuals who conducted the under-five check-ups at the local clinics, so it is hard to know whether ‘receiving support from the HSA’ means during the under-five check up at the local clinic, or during a home visit. In some locations both seem to happen, in other places, interviewees only reported seeing HSAs when they brought their babies for vaccinations and check-ups:

Non-professional support for breastfeeding was provided most often by mothers or mothers-in-law (n=19), followed by fathers (n=17), who were often mentioned as being supportive, taking on chores, and reminding women to make sure the baby gets breastfed often. Other family members (n=14) were also mentioned as providers of breastfeeding support, while neighbours and friends did less so (n=8). All women mentioned at least one source of ‘non-professional’ EBF support, some mentioned two or three.

Care Group activities (objective two)

The Care Groups in Dedza work with NGO promoters, who train CGVs. CGVs pass on the promotors’ health and nutrition lessons on to so-called ‘cluster leaders’, who are tasked with home visits. Findings show that 18 out of the 26 interviewees living in areas with Care Groups mentioned interacting with them. Home visits were mentioned, as were group activities. The group activities, especially cooking demonstrations, appear to have been well received by the women who were interviewed. Several women credited Care Group cooking demonstrations for teaching them how to add nourishing ingredients to the porridge they feed their small children. Some linked this to their infants’ improved growth [18MCB] and mentioned learning “from the group”. Group meetings that involved cooking and sharing food appeared to live up to some mothers’ expectations of what NGO-supported interventions do, while receiving counselling, or information, only, was not always appreciated in the same manner. A CGV recounted being told by a family “we don’t see a reason why you come to visit us because there is nothing that you give us” [52CGVA], while several mothers explained during our interviews that they had received ‘no support’ from Care Groups, which later turned out to mean no material support, while the same women further expressed their satisfaction about the advice they have received from Care Group cluster leaders or volunteers [e.g. 02MCA].

Care Group interaction with HSAs and District officials (objective three)

The CGVs, cluster leaders, and promoters interviewed for this research did not make any references to HSAs. Clearly, the entire Care Group structure was operated by the NGO, with little or no integration into the local healthcare structures. Even though Malawi’s integrated Care Group policy¹⁷ suggests that the volunteers should receive monthly behaviour change lessons from local frontline service providers, this was clearly not happening in Dedza. The NGO coordinator explained why:

At a higher level, it was clear that District Health Officials knew the Care Groups are supposed to be integrated in the district health and nutrition activities. Supervising the Care Group activities should be an integral, and motivational, part of the District Authorities’ responsibilities. However, the resources to support these activities are often lacking, which means that district officials have to rely on the NGOs to facilitate the monitoring of the NGO-supported Care Groups:

The interview data contained just one reference from a mother, suggesting that Care Groups worked with HSAs in her area [17MCB], which demonstrates that on an ad hoc basis, care group volunteers and HSAs do work together at community level. Most of the interviewees, however, did not indicate seeing HSAs and Care Group members together or appeared to associate HSAs and Care Groups. Even among the women with preterm or low-birthweight babies, there was no evidence of a coordinated follow up by Care Groups, which could have been directed by a concerned HSA. Only two out of the nine women with preterm or low-birthweight babies had received EBF support from Care Groups, and in one case this was because the local Care Group volunteer was also an aunt of the new mother who was interviewed. None of the Care Group team (paid or unpaid staff) made any reference to coordination or collaboration with HSAs or other Ministry of Health staff at District level. One of the HSAs, who described himself as a HSA supervisor, gave an elaborate description of newborn mother support provided by a mixture of HSAs and Care Group members:

Yet during the interviews with the mothers, none of them appear to have received the integrated care or the referrals from HSAs to the CGVs as is described by this HSA.

Our research showed that the HSAs were not always fully informed of what was happening in the community. Whilst the HSA cadre was originally designed to be men and women who live in the rural communities they serve; evidence suggests that this is not always the case. HSAs may in fact live in a nearby town and focus primarily on tasks based at the nearby health facility or hospital whilst spending only limited amounts of time at community level. It is therefore not surprising, according to HSAs, that events such as twin or triplet births, or preterm babies brought home to a village after delivery, do not automatically come to a HSA’s attention:

From the perspective of the local hospitals where mothers deliver their babies and where vulnerable newborns such as premature babies or twins may spend several weeks before they are discharged, the referral system appears to be equally unclear, according to a District level interviewee:

Overall, the level of support for EBF provided by any member of a professional or volunteer health workers cadre, was found to be limited, especially considering that a one in four of our interviewees were mothers of vulnerable babies, who included those in need of support with expressing breastmilk and managing this hygienically, whilst their babies needed the protection and nourishment from EBF all the more.

Care Group efficacy

Our research focused on whether or not women had received breastfeeding support from members of the Care Groups working in their community, given that the Care Group members were mandated to engage with all pregnant women and women with children under two, and focus on spreading key nutrition messages, including the promotion of exclusive breastfeeding. Finding that only eight out of 26 of the women participants had received EBF support from Care Group members meant that their performance was well below expectations. Care Group interventions have often been successful in bringing women together and promoting important behaviour change messages such as those regarding breastfeeding and handwashing, which are no- or low-cost interventions that can be adopted by the large majority of women who live within the reach of Care Group activities (Davis et al., 2013; George et al., 2015). The same level of success does not seem to have been achieved and sustained by the interventions focused on during this research.

Where Dedza’s Care Groups were active, they were reported to have reinforced all-important EBF messages, but overall, it appears that breastfeeding support was not a priority for the Care Group interventions we studied. 18 out of the 26 women in areas with Care Groups reported receiving support of some kind (other than EBF) from care group members. In most cases this related to cooking demonstrations. Such cooking demonstrations were done in groups, as opposed to ‘passing on/receiving Care Group messages’, which the NGOs’ second tier CGVs (called ‘cluster leaders’) were expected to do on a house-to-house basis. Several references to what the community’s expectations Care Group support entails, might explain why some volunteers were reluctant to engage in visiting households empty-handed with ‘just advice’.

While volunteer attrition is an issue that is highlighted by most researchers and practitioners of Care Groups, it was noted that the NGO that implemented the Dedza-based interventions reported high numbers of drop-out rates, in the earlier (pre-2019) phases of the Care Group interventions we studied; between 32% and 16% of the trained CGVs and/or cluster leaders reportedly becoming inactive before the end of the programme (United Purpose, 2022). Given that this reported high attrition rate predated the COVID-19 pandemic (during which a lot of Care Group activities, including motivational visits and training sessions, had to be cancelled), we suspect that the lack of integration into the community-based HSA system may have undermined the Care Groups’ perceived legitimacy and standing within the community, thereby demotivating the volunteers. A lack of budget for supportive supervision by the district health authorities also deprived Care Group volunteers of the validation such exposure to higher level officials would bring. Research into HSA motivation shows that attrition is also a challenge among HSAs. Finding high drop-out rates is therefore not surprising (Guenther et al., 2019), yet, the budget shortage at district level that leads to a lack of opportunity to supervise and motivate both cadres is regrettable, as HSAs and CGVs could play a significant role in visiting newborns and supporting EBF, potentially saving lives.

Another factor that may have hampered Care Group efficacy is the focus on household visits over group activities. The NGO whose project we visited explained that they interpreted Malawi’s national Care Group approach as one that requires cluster leaders to visit households on a one-to-one basis, creating significantly more work per volunteer than the traditional Care Group approach. The Malawi Care Group guidance emphasises group learning when a frontline health worker or NGO promoter passes on health, hygiene, or nutrition lessons to CGVs, but it explicitly mentions that cluster leaders should be passing on hygiene or nutrition lessons by visiting households individually. The traditional Care Group model emphasises sharing all lessons with the mothers at community level as group activities, as group meetings provide the additional benefit of creating social cohesion, the opportunity to learn from other group members and, in many cases, a sense of camaraderie Guenther et al., 2019; Kok & Muula, 2013; Perera, 2015). Our realist synthesis of the Care Group literature (Pieterse et al., 2021)²⁷ uncovered that much of the CGVs’ motivation was derived from meeting with neighbourhood group members (local mothers), whose social bonds with their CGVs often provided them with affirmation and a sense of self-worth. Care Group leader visibility at community level accords status to volunteers and earns respect from peers (Greco et al., 2017). This study showed that group activities, especially the cooking demonstrations, were referred to much more often than the one-to-one visits conducted by cluster leaders, and should perhaps be prioritised, not only for cooking demonstrations but also for general knowledge sharing, and especially for passing on messages about EBF.

A lack of collaboration between HSAs and Care Groups

The evident absence of communication and collaboration between HSAs and Care Group members can only be perceived as a missed opportunity. Given that one HSA explained in detail how a system of integrated and delegated care for women with newborn babies should have worked, there appears to be an intention on part of the HSA structures to collaborate more closely with Care Groups, but that these plans were simply not followed through on. The NGO’s programme documents also suggest that greater engagement with government structures was wished for, at District level specifically, but that government officials’ heavy workloads made scheduling meetings with the relevant individuals challenging. Similar challenges have been noted in other Care Group interventions in Malawi (World Relief, 2004). It appears to be difficult to receive district-level buy-in for a well-integrated Care Group intervention, which translates into a lack of higher-level facilitation of the HSA-care group relationship that could have created the types of linkages that Malawi’s national Care Group strategy called for.

HSAs have been lauded with significantly expanding Malawi’s vaccine coverage, women’s access to antenatal and post-natal care and growth monitoring (Kok & Muula, 2013; Mziray et al., 2017). However, the expanding list of demands placed on HSAs have exacerbated the pressure on their time, which appears to have increased the risk of HSAs not being able to carry out all the tasks that are expected of them (Guenther et al., 2019; Kok & Muula, 2013; Ngwira et al., 2021). This context, whereby HSAs are overburdened and a substantial volunteer cadre (which the Care Group volunteers effectively are) has been in place since 2012, could have lent itself to task sharing and, for example, the delegation of ‘mother and newborn baby home visits’ to care group volunteers in Malawi. There has been no evidence of this in the literature to date. Instead, the NGO staff implementing the Care Group intervention clearly struggled to convince HSAs (and those in charge of managing the HSA programme) that proper integration of the CGVs would be to the benefit of the HSAs, in terms of workload reduction, and to the benefit of women and their infants, who do currently not receive sufficient EBF and infant feeding support. Such a sharing of community-level tasks is taking place in some other LMICs, for instance in Ethiopia, where volunteer ‘women/ health development army’ members carry out a significant amount of the community-based household visits, increasing the efficiency of Ethiopia’s Health Extension Workers “in reaching households with actionable health messages” (Yitbarek et al., 2019, p 2). In a series of papers examining the latest trends in the increasing deployment of community health workers (CHWs) in LMICs, Perry et al. (2021) highlight an “emerging dual cadre of CHWs in which professionalized CHWs are supervising lower tiers of CHWs, who are often volunteers responsible for a small number of households”.

The lack of integration of the Care Group members into the community-based healthcare system in Malawi means that overburdened HSAs have not been given the opportunity to systematically delegate some of their home visiting responsibilities to the Care Groups, which results in a continued lack of check-ups and support for newborn babies and their mothers at community-level. In Burundi, where a randomised-controlled trial showed that Care Group-inspired interventions can be successfully adapted to be managed by Ministry of Health staff, the key to its success appears to have been the active involvement of the HSAs (or their equivalents) at community-level from the outset. This allowed the HAS to clearly see that there were potential advantages for them, in terms of being able to delegate certain time-consuming responsibilities (Weiss et al., 2015).

Our targeted search for, and reading of, all of the Government of Malawi’s health and nutrition-focused policy and strategy documents and key aid donor reports, led us to the identification of just three documents that reference Malawi’s national Care Group strategy (GoM, DNHA, 2012; GoM, DNHA, 2017; World Bank, 2019). These include the strategy document that first included the Care Group approach, the national guidelines (in English and Chichewa, counted as one document) and a World Bank report. At the tenth anniversary of its adoption into Malawi’s national policy, it is timely to reflect on this nutrition policy component and suggest ways that can improve the efficacy of Malawi’s adapted Care Group approach. Updating and improving Malawi’s national Care Group strategy and guidance could increase the chance of community-level integration that could see Care Group members develop into a more established volunteer cadre. This would be an opportunity to conceive of CGVs as lower tier Community Health Worker (CHW) cadre, as described by Perry et al. (2021), who are responsible for a small number of households (similar to current CGVs) supervised by professionalised CHWs (very much like HSAs). According to the World Health Organisation, worldwide health workforce shortages are leading many low-income countries to consider ways in which CHWs can be adopted into national human resources for health strategies, inspired by examples from countries where this has been successful, for example the two-tier CHW model from Ethiopia (Rieger et al., 2019; WHO, 2020).

Recommendations for further research

Despite there being a significant demand for care provision that is local and easily accessible, community-level healthcare provision remains a challenge in LMICs, not least because healthcare budgets are often tight. Whist there is a wide range of literature on CHW programmes, many of those are based on pilot projects, short-term projects or interventions that are donor-driven and have not been tested in terms of long-term sustainability. The Malawi government’s efforts to integrate the Care Group approach into national policy is one of the few examples where a clear intention exists to integrate lay or volunteer health worker systems into the existing community-level healthcare structures, yet until now, this approach had not been reviewed or published about. More research is needed into ways in which national-level or sub-national level efforts are made to integrate volunteer, lay and ‘stipended’ community-level healthcare providers in LMICs (such as Zulu et al., 2015), as are more explorations of what shape such integrations could take in future (such as Naimoli et al., 2015; Schneider & Lehmann, 2016).

Limitations

Our study of two Care Group interventions was focused on one district in Malawi only and was based on one implementing NGO that used the same type of Care Group approach for its interventions. While this limits transferability to other Care Group interventions in Malawi, the authors of this study did further consider at least five published programme reviews and two peer reviewed articles on Care Groups in Malawi programme reviews (Best et al., 2017; Crespo et al., 2009; Robins & Best, 2009; van Haeften et al., 2013; World Relief, 2004), plus peer reviewed sources (George et al., 2015; Wilner et al., 2017) and the findings from our realist synthesis of Care Group literature that encompasses 42 texts from a range of LMICs (Pieterse et al., 2021) in our interpretation of these findings. The focus of our study was breastfeeding, which meant that we examined the efficacy of the interventions through that specific lens, although we did consider all other mentions of our interviewees’ interactions with Care Groups. Finally, we acknowledge that we have examined interventions that are based on group interactions and people learning from face-to-face meetings to pass on behaviour change messages, in the middle of a pandemic that necessitated the cancellation of many care group activities. It can therefore be assumed that overall efficacy of the examined interventions would have been better than what we recorded, had this not been the case. However, we managed to draw most of our conclusions from a mix of field research, as well as existing programme design, policy guidelines, and pre-COVID-19 Malawi-focused research reports, on which the pandemic had no bearing.