Care Groups I: An Innovative Community-Based Strategy for Improving Maternal, Neonatal, and Child Health in Resource-Constrained Settings

Care Groups use volunteers to motivate mothers to adopt key MCH behaviors. The volunteers meet as a group every 2–4 weeks with a paid facilitator to learn new health promotion messages. Key ingredients of the approach include: peer-to-peer health promotion, selection of volunteers by the mothers, a manageable workload for the volunteers (no more than 15 households per volunteer), frequent (at least monthly) contact between volunteers and mothers, and regular supervision of the volunteers.


INTRODUCTION
T here is a recognized need to accelerate progress in reducing maternal and child mortality in the 75 countries of the world where 95% of the world's maternal and child deaths take place. [1][2][3] The Millennium Development Goals (MDGs), established in the year 2000, called for achieving by the year 2015 reductions of threefourths and two-thirds, respectively, in maternal and child mortality based on 1990 levels. 4 These goals will not be achieved by the great majority of these countries, particularly in sub-Saharan Africa, where only 5 of 44 countries are on track to achieve the maternal health MDG and only 14 are on track to achieve the child health MDG. 5 One of the important reasons for lack of progress has been the low population coverage of key interventions proven to be effective for reducing maternal and child deaths. Although the median population coverage of immunizations and vitamin A supplementation is in the range of 80%, coverage of other key interventions is 60% or less, and for a number of interventions, the median range of coverage is 30% or less. 5 In some countries, levels of coverage are less than 10%. 5 There is a lack of evidence that facility-based services by themselves in resource-constrained settings with high mortality can achieve high levels of population coverage of key maternal and child health interventions and mortality impact, and some evidence that they cannot. [6][7][8] Expanding coverage of key interventions and achieving documented reductions in maternal, neonatal, and child mortality will require approaches that are not only low-cost and effective on a shortterm, pilot basis in small populations but also low-cost, effective, and feasible at scale over the longer term. This requires, among other things, approaches that engage the community as partners, empower women and communities, and reach a high proportion of households with health education that encourages healthy behaviors and appropriate use of health facilities. 9 Interest in and experience with community health workers (CHWs) is growing rapidly, and CHW programs are expanding in many countries. 10,11 CHW programs in some countries (such as Bangladesh, Brazil, Ethiopia, and Nepal) have been widely credited with achieving high levels of population coverage of key maternal and child health interventions and marked improvements in child survival, while in other countries such as India and Pakistan progress has lagged behind in spite of large-scale CHW programs. 5 There is a wide diversity of CHW programs and cadres, with some using ''professionalized'' workers with 1 year or more of training, a broad scope of preventive and curative skills, and a full-time government salary while others engage volunteers working only a few hours a week on a highly specialized activity such as immunizations, HIV/AIDS control, or distribution of bed nets. In all cases, CHWs are formally trained and engaged with the health system but work at the community level outside of facilities and receive no formal professional or paraprofessional certificate and no tertiary education degree.
This article describes the Care Group approach-a delivery strategy for expanding coverage of maternal and child health interventions using volunteer CHWs. We describe what Care Groups are, the key ingredients (the ''secret sauce'') that seem to be important for their successful implementation, their history, the field experience with use of this delivery strategy, modifications that are emerging in Care Group implementation, and how Care Groups might be integrated into government health programs. Finally, we highlight Care Groups as one example of the growing importance of participatory women's groups in improving maternal and child health. A companion article in Global Health: Science and Practice summarizes the evidence on the effectiveness, cost, and cost-effectiveness of Care Groups to improving child survival. 12

WHAT ARE CARE GROUPS?
The formal definition of a Care Group is the following 13 : A Care Group is a group of 10-15 volunteer, community-based health educators who regularly meet together with project staff for training and supervision. They are different from typical mother's groups in that each volunteer is responsible for regularly visiting 10-15 of her neighbors, sharing what she has learned and facilitating behavior change at the household level. Care Groups create a multiplying effect to equitably reach every beneficiary household with interpersonal behavior change communication.
A representation of a Care Group intervention delivery system is shown in the Figure. The system is established initially by identifying 1 volunteer (called a Care Group Volunteer) who is responsible for about 12 mothers (pregnant women and mothers of young children, usually 0-59 months of age or 0-23 months of age). The Care Group Volunteer is often selected by the mothers themselves; sometimes community leaders participate in the selection process. Supervisory field staff are recruited and trained to set up Care Groups in collaboration with community leaders so that: (1) Care Group Volunteers are in place and are responsible for about 12 mothers who are their neighbors, and (2) all pregnant women and mothers of young children are linked to a Care Group Volunteer.
Depending on the size of the population covered by the project or program, several layers of paid program staff are required so that a Care Group Facilitator (referred to in the Figure as a Promoter and who is a low-level paid project staff person) can meet with each Care Group every 2-4 weeks for 2 hours or so. At that time, the Care Group Facilitator teaches one or a small number of health promotion messages for the Care Group Volunteers to share with the women for whom they are responsible. The Facilitator uses participatory learning, including role play and composition of songs and skits, to convey messages. During the following 2-4 weeks (depending on the schedule established by the program), each Care Group Volunteer meets with each of the women for whom she is responsible (and other family members who may be present, such as grandmothers, husbands, and older children)-either by visiting the woman at her home or by meeting with her and a few neighbors as a small group. At the subsequent Care Group meeting, the Care Group Volunteers discuss their experience in sharing the previous messages and learn a new set of messages. In most Care Group programs, the Care Groups Volunteers also report births and deaths to the Care Group Facilitators/ Promoters, who report this information upward through the health information system. (A manual outlining how to monitor mortality using the Care Group approach is available at: http://www.core group.org/resources/271-the-mortality-assessmentfor-health-programs-system.)  The Facilitators/Promoters along with higherlevel supervisory staff meet together every few months to learn the health promotion messages that they will later convey to the Care Group Volunteers. The Facilitators/Promoters are also taught participatory methods for behavior change promotion, including demonstrations, role plays, stories, and songs (often composed by the Care Group Volunteers themselves) to convey these messages.
The educational content focuses on key knowledge about maternal and child health, important household practices for promoting maternal and child health, and indications for use of health facilities, including danger signs for which medical care should be sought. Messages are often based on results of formative research such as positive deviance studies and barrier analysis 14 studies that identify behavioral determinants of key behaviors.
Further details regarding what are considered to be essential criteria for the Care Group approach are shown in Box 1. Additional suggested criteria are shown in Box 2. The rationale for these lists is described in greater detail elsewhere. 13 The degree to which each criterion in Box 1 and Box 2 is required for the Care Group approach to be effective and whether or not the approach would function if one or more criteria were not met is unknown at present, but fidelity BOX 1. Essential Criteria for Application of the Care Group Approach 1. Peer-to-peer (mother-to-mother for maternal and child health and nutrition behaviors) health promotion is essential.
2. Care Group Volunteers are selected by mothers within the group of households they will serve or by the leadership in the village.
3. Because they are volunteers, the Care Group Volunteers must limit their workload. There should be no more than 15 households per Care Group Volunteer. The terrain and distance between households may affect the number of households for which Care Group Volunteers can take responsibility. In settings where households are dispersed or the terrain is difficult to traverse, Care Group Volunteers may need to take responsibility for fewer households.
4. There should be between 6-16 members in a Care Group.
5. Contact between the Care Group Volunteer and her assigned beneficiary mothers is monitored; contact should occur at a minimum once a month, but preferably twice monthly.
6. The plan is to reach 100% of the households in the targeted group on at least a monthly basis and to attain at least 80% monthly coverage of households within the target group. Coverage is monitored.
7. Care Group Volunteers collect vital events (data on pregnancies, births, and deaths). This is very important as they can discuss and solve problems as a group related to what kind of follow-up is needed and how similar deaths might be prevented in the future.
8. The majority of what is promoted through the Care Group is directed toward reduction of maternal and child mortality and malnutrition (e.g., Essential Nutrition Actions and Essential Hygiene Actions). It could be a useful strategy to include other topics but this is what Care Groups were originally intended to do.
9. Care Group Volunteers should use some sort of visual teaching tool (e.g., job aids, flip charts) for health promotion at the household level.
10. Participatory methods of behavior change communication are important. This is not specific only to Care Groups, but Care Groups should use best practices for behavior change.
11. No more than 1-2 hours should be spent in a meeting of Care Group Volunteers. Such "drip training" involves small amounts of information, and then the information is applied. The Care Groups can fit the time needed for the meeting into their schedule. Care Group Volunteers then bring the information and messages to the women in their catchment area as they have time.
12. Supervision of Promoters and of at least one of the Care Group Volunteers should occur at least monthly. This is an important part of the Care Group approach and part of the cascade.
13. All of a Care Group Volunteer's beneficiaries should live less than a 1-hour's walk from the Volunteer's home.
14. The implementing agency needs to successfully create a project/program culture that conveys respect for women, for the Care Group Volunteers, and for the beneficiaries.
to most, if not all, criteria seems to be essential at this stage in the development of the approach. Over time, with further experience and evaluation, variations in the approach could be tested. Box 3 provides a case study of how one Care Group project in Mozambique functioned.

HOW DID CARE GROUPS EMERGE?
The 5. The ''care'' in Care Groups may not be quantifiable, but it is supremely important for an entire system based on volunteerism. Care and respect need to be modeled from the senior leadership all the way through to the volunteers.
6. Ensuring that each Care Group Volunteer does not have more than 3-4 hours of work per week is essential in order not to overload and overburden her.
7. Peer selection of Care Group Volunteers helps to identify the most effective recruits.

Field Supervisors and Care
Group Facilitators/Promoters need support from their supervisors on a regular and intensive basis, and they need to be provided with the tools to make their work effective, including transport and educational materials/visual tools such as flip charts.
9. Participatory methods of behavior change communication should be used, including songs, role play, games, and stories. Oftentimes Care Groups make up their own songs to use. 10. Flexible management and supervision that enable the Care Group principles to be upheld while adjusting to local contextual realities and implementation hurdles is recommended. The number of Field Supervisors and Care Group Promoters as well as the number of Care Group members needs to be adjusted based on local geographical conditions and ease of transport. mortality, as estimated by expansions of coverage of key maternal and child health interventions using the Lives Saved Tool (LiST), was one of the highest achieved up to that time by USAIDsupported child survival projects. 15 Later, the CORE Group provided a grant to carry out an independent assessment of the mortality impact of the project and to prepare a manual describing the Care Group delivery strategy. 16 As evidence of the effectiveness and feasibility of Care Groups began to accumulate among NGO child survival projects, other NGOs began to try the approach. Again, under the leadership of Tom Davis, Curamericas Global in Huehuetenango, Guatemala implemented the Care Group approach. World Relief, Food for the Hungry, and Curamericas Global also began to apply the Care Group approach at their other project sites in different countries. The success of these projects, the technical support offered to other organization by Melanie Morrow (then working with World Relief) and Tom Davis (then working with Every 2 weeks, each Promoter met with each of his/her 5 Care Groups. Each Care Group meeting lasted about 1.5 to 2 hours. At each meeting, 15-20 minutes were spent reviewing the previous activities since the last Care Group meeting 2 weeks previously. The Leader Mothers also informed the Promoter of any births or deaths that had occurred during the previous 2 weeks. Then they learned 3-4 new key messages to share with their households and practiced among themselves how to best present them to the beneficiary mothers.
Over the next 2 weeks, the Leader Mothers met with each of their 12 or so beneficiary mothers, either individually or in small groups. Over time, these beneficiary mothers ''graduated'' as their child reached 2 years of age, and newly pregnant women or those with a newborn became a beneficiary mother. Leader Mothers used a flip chart to illustrate the lesson being given.
When it became apparent that neonatal deaths were frequent, the Leader Mothers began to visit newborns as soon as possible after birth and then on a daily basis during the first week, 3 times a week during the second week, twice during the third week, and once during the fourth week. Leader Mothers received training in danger signs of neonates, used a checklist, and counseled mothers during these visits using flip charts and a checklist for newborns.
One Leader Mother in each Care Group was selected as a C-IMCI Leader Mother and received 5 days of training. C-IMCI refers to Community-based Integrated Management of Childhood Illness. She served as a resource and a referral point for the other Leader Mothers in the Care Group.
The project had only one vehicle, which was used to transport the project leadership team from the project headquarters in the city of Beira to and from the project area. Each of the 5 District Supervisors had a motorbike and each of the Promoters had a bicycle.
Food for the Hungry), and the availability of a very useful implementation guide 16  The fact that all the early Care Group projects were funded by the USAID Child Survival and Health Grants Program meant they all had baseline measurements of population coverage of key interventions obtained from household surveys as well as end-of-project measures using similar survey instruments since this was required of all grantee projects. The projects were therefore able to assess changes in practice and coverage over the course of the projects. This success was shared with the NGO community And in all cases, Care Group projects promote use of MOH services and programs, including facility-based services. At present, based on available information, we estimate that 1.3 million households have been reached using the Care Group implementation system, and at least 106,000 Care Group Volunteers have been trained.
Almost all the Care Group projects implemented so far have been in rural areas of low-income countries. To our knowledge, there is only one example of Care Group implementation in an urban or peri-urban setting, 18 although several rural projects have had ''pockets'' of peri-urban populations. Projects often vary in terms of the specific interventions implemented (such as nutrition, diarrhea control, newborn health, immunizations, and so forth) depending on the local epidemiological context. The details of supervision and training also vary from project to project and from NGO to NGO.
International NGOs have initiated implementation of all known Care Group projects to date. The major donors for these projects have been the USAID Child Survival and Health Grants Program, in-country USAID missions, the USAID Food for Peace (Title II)  Community Humanitarian Aid Office and still referred to as ECHO), the United Nations Children's Fund (UNICEF), and private NGO funds.
There is now early experience in applying the Care Group delivery strategy within MOH rural health care delivery systems. Concern Worldwide has carried out an operations research project in Burundi comparing the effectiveness of the traditional NGO Care Group project structure (in which the Care Group Facilitators/Promoters are paid by the NGO) with an alternative approach in which Care Group Facilitators/Promoters are MOH CHWs (who are unpaid volunteers). The findings indicate that-with NGO technical and financial support during the time of the study-the effectiveness of the Care Group strategy using MOH CHWs as Care Group Facilitators/Promoters is similar to the traditional NGO implementation of the approach. 19  duties, but they each supervised only 2 Care Groups rather than the usual 5-9 Care Groups in the typical NGO implementation. (A short user's guide on how to integrate Care Groups into MOH health systems is available at http://www.fsnnetwork.org/ sites/default/files/resource_uploads/integrating_care_ groups_into_moh_systems_a_users_guide.pdf.)

OTHER EMERGING MODIFICATIONS OF THE CARE GROUP APPROACH
One important example of a modified Care Group approach is in Rwanda. 20 There, 3 international NGOs (Concern Worldwide, The International Rescue Committee, and World Relief) worked with the MOH to modify the Care Group delivery system to fit within the current role and functions of government CHWs (where the CHWs are volunteers who receive performance-based incentives). In this case, there were 4 CHWs in each village: Two (one male and one female) were assigned to carry out integrated community case management of childhood illness Another was responsible for maternal health (a female CHW) The fourth was responsible for social affairs (either a male for female) The 4 CHWs were responsible together for the entire village of 60-80 households, but each focused on his/her specialty area. The Care Group methodology was modified so that the CHWs from 2-5 neighboring villages were organized into CHW Peer Support Groups with each group having up to 20 CHW members, about half of whom were men.
The CHWs divided up households so each CHW was responsible for 15-20 households, and they visited each household monthly to provide behavior change communication. The project helped to organize and oversee training of the CHW Peer Support Groups in behavior change interventions. CHW Peer Support Groups elected unpaid CHW Cell Coordinators to help with the training and supervision of the CHW Peer Support Groups.
There were about 100 CHWs working in a health center's catchment area. Each health center had an MOH employee in charge of supervising the CHWs working in the health center's catchment area. This Community Health In-Charge at the health center also functioned as a CHW Peer Support Group Facilitator. This approach was implemented in a catchment area of 1.7 million people, reaching 18% of Rwanda's population. Appropriate treatment for malaria, point-of-use water purification, and the proportion of caretakers who increased fluid or food during diarrheal episodes doubled or almost doubled. 21 Care seeking for respiratory symptoms more than tripled, and skilled attendance during childbirth increased from 39% to 91%. 21 Care seeking for fever, diarrhea, and symptoms of pneumonia in children were significantly greater in the catchment area with CHW Peer Support Groups than in the non-project districts. 20 Another example of a modified Care Group approach is in Mozambique, where World Relief has just completed a 5-year Care Group project focused on tuberculosis (TB) control in the same locale where Care Group projects had been previously implemented. For the TB project, World Relief used many of the same Care Group Volunteers and similar supervisory structures as those that had been established over the previous 2 decades in 6 districts of Gaza Province. The project was able to achieve marked improvements in awareness about TB, its treatable nature, and the availability of free treatment at government health centers. 22

INTEGRATION OF CARE GROUPS INTO MINISTRY OF HEALTH DELIVERY SYSTEMS
As previously mentioned, the field experience with Care Groups has been primarily with projects implemented by NGOs. Opportunities for incorporating the Care Group approach into existing MOH programs would help give the approach a long-term, sustainable ''home'' that NGOs are not usually able to provide. There is room for new creative partnerships between NGOs and MOHs to implement the Care Group approach, from A Care Group Volunteer in Burundi shares a health message during a home visit. Care Group Volunteers make home visits to each woman for which they are responsible on a monthly or twice-monthly basis.
Incorporating the Care Group approach into Ministries of Health would help make it sustainable.
contracting out service delivery to engaging NGOs for training, monitoring, or quality assurance.
The essential link needed in MOH programs is the creation of formal postings for what we refer to in the Figure as Promoters (who meet with, teach, and support Care Group Volunteers) and Supervisors (who meet with, teach, and support the Facilitators/Promoters). Simply adding the duty of the Care Group Facilitator/Promoter to the existing duties of currently functioning CHWs would seem to be fraught with high risk of failure, as would adding the duty of the Supervisor to the MOH supervisor of MOH CHWs since these persons are already likely overburdened with too many responsibilities. However, the Concern Worldwide experience in Burundi revealed that in a situation in which CHWs were already overloaded with responsibilities that included community mobilization, integrated community case management (for pneumonia, diarrhea, and malaria), and home visits, giving them responsibility for supervising Care Groups actually lightened their workload-better prevention led to fewer cases and earlier home-based treatment and, at the same time, earlier referral led to fewer seriously ill children for CHWs to manage. 19 One option-aside from creating new posts specifically for these functions (which we acknowledge is quite difficult within the government system)-might include recruiting more MOH CHWs and supervisors of CHWs so that the additional Care Group workload would be manageable. Whatever strategy might be adopted, there would need to be additional resources devoted to high-quality, community-based delivery. Even though the costs of Care Group programs are quite modest (as is discussed in the companion article 12 ), the success of Care Group implementation rests in large part on having well-trained, highly motivated, and wellsupported field workers.

CARE GROUPS AS AN EXAMPLE OF PARTICIPATORY WOMEN'S GROUPS
Care Groups are an example of how programs are gradually learning to harness the power of women working together to improve their own health and the health of their children. Women's groups have been in use now for decades, but well-delineated methods for engaging them and mobilizing them to deliver key evidence-based interventions that result in scientifically demonstrated improvements in either population coverage of these interventions or improved populationlevel health outcomes have been lacking until recently.
A similar but nonetheless distinct approach to engaging the power of groups of women is women's participatory learning and action (PLA) groups. In this approach, a facilitator meets with women in a village, and together they discuss health recommendations for pregnancy, birth, and neonatal care and how they could apply them in their own particular situation. 23 Although pregnant women and those with newborns are targeted, anyone in the village can attend the meetings. This approach has benefited from rigorous implementation research in a variety of settings, all being led initially by the same research group based at the University of London. Robust evidence finds the approach can reduce maternal and neonatal mortality if there are an adequate number of facilitators to ensure high levels of service coverage. 24 A key difference between Care Groups and PLA Groups is that with Care Groups, there is a systematic approach to reaching every household where there is a pregnant woman or young children with specific, carefully crafted messages. With PLA Groups, the focus is on the PLA Group discussing key health messages, formulating how they might incorporate these messages in their context, and then, through spontaneous dissemination, engaging other women who do not attend the facilitated sessions. 25 Most certainly, other approaches are emerging now or will emerge in the future to harness the potential of participatory women's groups.
The enthusiasm for and the demonstrated results of Care Groups and PLA Groups indicate this is a fruitful area for further field experimentation with rigorous evaluation and broader implementation.

CONCLUSIONS
Although not widely known about outside of NGO child survival and food security networks, Care Groups are a rapidly growing innovative approach to implementing maternal, neonatal, and child health and nutrition interventions. Care Groups are able to motivate women volunteers to assist their neighbors in adopting positive health behaviors and seeking health care from the formal health system when needed. The NGOs that have implemented the Care Group approach in a variety of field settings throughout the world have been uniformly enthusiastic about the effectiveness of the approach in changing behaviors, improving appropriate health care utilization, achieving demonstrable benefits in the health of mothers, neonates, and children, and empowering women and their communities. The Sustainable Development Goals, scheduled for adoption at the 2015 meeting of the United Nations General Assembly, call for the achievement of universal access to quality essential health services and the end of preventable deaths of newborns and under-5 children by the year 2030. 26 Care Groups and related approaches hold great promise in helping to achieve these goals.