This 23-page report examines several experiences with a home/neighbourhood-based nutrition programme for children called Positive Deviance (PD)/Hearth that is being implemented in different contexts around the world. As explained here, PD is based on the belief that in every community there are a few individuals and families whose particular practices enable them to have better health compared to their similarly impoverished neighbours. Hearth, suggesting a family around a fireplace or kitchen, is an implementation strategy that supports caregivers to learn and practice new health behaviours together in a safe environment (such as a home setting) and to rehabilitate their malnourished children. Trained volunteers assist participating parents or caretakers in preparing meals and snacks for their malnourished children with beneficial, locally available food using the results of a PD Inquiry (or, PDI). Participants practice positive child caring and active feeding techniques and feed malnourished children with extra energy-rich/calorie-dense supplemental meals. Even the standard 2-week participation period in a Hearth "often contributes to rapid improvement in nutritional status for children, as well as better long-term feeding practices by caretakers at home....PD/Hearth has enabled countless communities to reduce their levels of childhood malnutrition and to prevent malnutrition years after the program's completion."

As explained here, the programme methodology has been rapidly disseminated to the NGO community via manuals, studies, trainings, field visits, and consultant visits. In 2002, 14 CORE Group members, several multilateral and bilateral agencies and many local NGO partners were implementing PD/Hearth programmes in over 35 countries in Africa, Asia, Latin American & the Caribbean (LAC), and Eurasia. This number has rapidly expanded as new partners, consortium, and Title II food aid programmes have implemented the PD/H approach, often at large scale, around the world in various types of settings.

This Technical Advisory Groups (TAG) meeting was convened on February 6 2009 with these objectives:

1. Review PD/Hearth implementation and results from recent experiences.
2. Explore challenges and modifications in PD/Hearth implementation and make recommendations related to both essential elements and implications for scale.
3. Identify how PD/Hearth has been integrated into overall nutrition programmes and with other approaches.
4. Identify key messages and audiences related to communication for PD/Hearth.

First, the document provides background on past TAG meetings and PD/Hearth materials. Hosted by World Relief, the first TAG (1995) resulted in a BASICS publication on PD/Hearth experiences in 4 countries. TAG 2 (1999) developed the content for a BASICS-produced video and a pamphlet describing PD/Hearth. The CORE Group's Nutrition Working Group convened TAG 3 (April 2000) to follow up a request to create a "how-to" guide for field staff. TAG 4 (December 2002) reached a consensus on the essential elements of PD/Hearth for programme design and implementation, defining what it means to be called "PD/Hearth". Participants exchanged results related to impact measurement of PD/Hearth programmes. Information gaps to prioritise further research and generate knowledge were identified, and next steps defined. In February 2003 the CORE Group published the Positive Deviance/Hearth Manual. This TAG also resulted in the production of a paper entitled "PD/H in the Context of Other Nutrition and Child Survival Interventions" published in 2004. TAG 5 (December 2004) reviewed current information on PD/Hearth implementation around the world. Participants considered how to develop a standardised monitoring and evaluation system with minimum standards and suggested indicators for field and headquarters staff attention. The TAG made plans for a second publication, The Positive Deviance/Hearth Essential Elements: A Resource Guide for Sustainably Rehabilitating Malnourished Children (Addendum), which was published by the CORE Group in June 2005.

The February 2009 TAG meeting detailed in this document focused on whether PD/Hearth can be done differently than described in the Positive Deviance/Hearth Manual and Addendum, yet just as effectively infuse a behaviour change approach. Brief summaries of the case studies included in the report follow:

I. Malawi: A United States Agency for International Development (USAID) Title II consortium is jointly implementing the Improving Livelihoods through Increasing Food Security (I-LIFE) Project during FY2004-2009. The consortium, led by Christian Relief Service (CRS) includes: Africare, CARE, Emmanuel International, Save the Children, The Salvation Army, and World Vision. PD/Hearth is a key nutrition activity in the Title II project in Malawi. Based on challenges associated with facilitating PD/Hearth, such as lack of skilled personnel for growth monitoring (GM) and screening, poor-quality health education, and food shortages, the I-LIFE partners felt they could not scale PD/Hearth up sufficiently to reach the large population of mothers needing nutrition messages in the relatively small amount of time left in the project. It was decided to change the strategic direction of I-LIFE's nutrition activities to focus on behaviour change and prevention rather than rehabilitation, and increase coverage targeting more households. Needing a strategy to accomplish all this, I-LIFE turned to the Care Group Model being used by World Relief and Food for the Hungry in many African countries. Care Groups provide the means to provide health education with nutrition messages to sets of 10-15 households supported by a volunteer "lead mother" in turn supported by trained promoters. Mothers are reached through behaviour change strategies for nutrition counselling, and are encouraged to bring their children for immunisations, de-worming, vitamin A capsule (VAC) distribution, and community integrated management of childhood illness (C-IMCI). "Local chiefs' participation was emphasized, which proved important for mobilization and increased participation of mothers."

Modifications: Modifications to PD/Hearth itself were extensive, and are detailed in the document...

Results/Population Reached: The I-LIFE project has realised its goal for high coverage, with 69,290 parents in 662 Care Groups and 14,850 children enrolled in PD/Hearths. There has been a "huge increase" from 2,244 in 2006 and 1,109 in mid 2007 to close to 6,000 by the end of 2007 and 14,850 in 2008 in the number of children participating. Mothers are learning skills in food processing, preparation, and preservation techniques, and this is not limited to households with a malnourished child. Over 5,000 Lead Parents and 21 consortium staff had been trained in improved food processing and preparation methods. The Government of Malawi is interested in scale-up...

II. India: Since 2005, Christian Children's Fund (CCF) [now called "ChildFund International"] India has coordinated with 76 local NGOs in West Bengal, India, to implement a substantial scale-up of PD/Hearth.

Modifications: There is only one Hearth session per year per community. A volunteer is selected from her community where she initially works. She adds a new neighbouring community each month for a year, for a total of 12. "The volunteers gain significant expertise with the repetition, to where they can solve problems. They receive a stipend. Due to the target area's compactness, supervisors are able to use a bicycle for transportation. They do not necessarily have a health background. Volunteers conduct the one and only PDI in their own community as part of their training....Normal children and their mothers are invited to the Hearth together with malnourished children and their mothers/caregivers....The idea was to reduce stigma by inviting everyone....After the Hearth, the volunteers do home visits every 15 days for two months."

Results: In 2007, with just 22 of the 76 partners, there were 907 villages covered in which 20,927 malnourished children participated in Hearths. Of these, 7,600 moved to normal status. Only 4% of children did not improve their nutritional status. The cost per child for one month was six cents.

III. Tajikistan: Save the Children Tajikistan conducted PD/Hearth in 26 villages as part of a Child Survival and Health Grant Program (CSHGP) funded by USAID which ended in 2008. It operated only in villages with >30% malnutrition, most of which is mild.

Modifications: PDI is conducted in only 2 villages in each target district. For the session, mothers come for 2 days, then stay home on the third day, feeding their child the same menu at home as an extra meal. This cycle is then repeated. The actual days in attendance total 9 days out of a 12-day session.

Results: "Although there were many other complementary activities going on in the context of the CSHGP, it was PD/Hearth which brought the malnutrition rates down so significantly within one month of participation."

IV. Ethiopia: A PD/Hearth programme was begun by an unnamed NGO in 3 sites and later handed over to the government.

Modifications: Modifications were so significant the TAG discussed if the programme could truly be called "PD/Hearth". For instance, the project site is in a famine-hit area, so programme leaders chose not to ask mothers to bring any contributions to the Hearths; instead, NGOs provided most of the food.

Results: Of the 1,100 children monitored after 1 month: 64% gained <200g; 9% gained between 200 and 400g; 9% gained 400g; and 18% gained >400g.

V. Bolivia: An unnamed NGO working in Bolivia has made significant modifications to the PD/Hearth design; for instance, staff conduct PDIs in some villages, but not all.

Results: When there is a month with Hearth, the children grow well; however, they falter in the intervening 2 months. Children are more or less following an upward growth line, which is interpreted in Bolivia as indicative of success.

VI. Indonesia: In 2002, CARE, CRS, Mercy Corps, PATH, Save the Children, and World Vision (WV) formed the PD Network. Most of the children being addressed were living in urban areas; this challenged the PD/Hearth model, which is built on the concept of mothers preparing the food they feed their family. A qualitative study was carried out; each NGO provided its own monitoring data. Some specific anthropometric studies and final knowledge/practice/coverage (KPC) survey results were added.

Results: Of the total 9,997 children involved, only 4,847 had both initial and one-month follow-up weights recorded and were therefore included in the analysis. By looking at all 4 years of data from the 5 NGOs, the study found that the graduation rate across NGOs was 45% with a range of 35-54%, using as a graduation criteria gaining at least 400 g of weight. Average graduation rates per sites within NGOs ranged from 21-92%. Several of the NGOs were assisting District Health Offices (DHO); the average graduation rate for DHO sites implementing alone ranged from 23-70%. In spite of doing very well in terms of coverage and adherence to the methodology, graduation rates were surprisingly lower than expected. The evaluation examined key factors associated with success; amongst the communication-centred factors were: better understanding of the PD/Hearth by the kader (volunteer worker); belief in the process of PD/Hearth by puskesmas (health centre, or PKM) staff; level of understanding by community leaders on the causes and consequences of malnutrition; and community support of the Hearth (e.g., local leaders stopping by or providing food). In general, wherever NGOs put their behaviour change effort, they saw more change.

VII. Haiti: The Children's Nutrition Program of Haiti (CNP) is a small organisation working in Leogane, Haiti. In 2008, they had 299 children participating in PD/Hearths, down from 452 the year before. CNP has modified the standard PD/Hearth approach. They do not do complete PDIs in every community before every Hearth. Instead, 36 paid monitrices working in pairs with community health workers identify PD families, with whom they do 3 home observations before every Hearth. In the 6-7 years the project has been operating, there has been very little turnover. Monitrices have 6 weeks of intensive training up-front to learn menus, develop songs, explore the market, etc. The training is very hands-on. Two full-time supervisors supervise each Hearth 2-3 times during implementation. Monitrices receive 1-week refresher training each year so that they can practice doing role plays, incorporate new activities such as linkages to community-managed acute malnutrition (CMAM), and keep the project spirit up. Hearths themselves take place at a volunteer mother's house, where cooking pots and water are available. Various incentives keep these volunteers involved.

Results: CNP does follow-up after 6 months to check on progress. CNP results indicate that 62% of children had a rate of growth that increased 1 standard deviation (SD). In 2006 and 2007, there were 84% at or above the international standard median growth curve; in 2008, it was 86%. CNP's programme is integrated to other activities in several ways. For instance, CNP has a safe water programme and a point of use (POU) treatment component for which they subsidise Hearth families' participation. The cost of the CNP's PD/Hearth model is around US$10 per child, per month. People within Haiti are looking at PD/Hearth again.

Discussion of Essential Elements:

Introductory remarks touched on the requirement to have at least 30% moderately to severely malnourished children in a community in order to support PD/Hearth implementation. One participant stated that the power of PD/Hearth is in the caregivers "seeing" the change in their children and therefore being reinforced for changing their practices. Where there is very little malnutrition, or predominantly mild/moderate malnutrition, participants were concerned that it would be difficult for caregivers to see the change in their children. Additionally, where mothers do not perceive the malnutrition is a problem, motivating participation is difficult.

Essential Element 1: The PD/Hearth manual recommends that the community members and staff conduct a joint discovery to identify the unique practices and strategies used by caregivers with well-nourished children. Programmes modifying this step tend to either omit the PDI or conduct the PDI as formative research for development of the Hearth session by either volunteers or staff (there is then a tendency to see the process as too labour-intensive to conduct in every community where staff or volunteers are identifying the same solutions). Several participants expressed concerns that the discovery process by the community is critical to achieving ownership of the results, sustaining rehabilitation, and preventing future malnutrition. They discussed the importance of not thinking of the PDI as a formal questionnaire, but rather a more anthropological process.

Discussion touched on the importance of training and coaching related to conducting the PDI and analysing the results in order to help the community identify the real PD practices to be expanded. One recommendation made was to seek out staff with qualitative skills and consider splitting the work such that a social mobilisation team would engage the community in conducting the PDI and a separate monitoring team would support more quantitative work. It was noted that illiterate women can be among the best to conduct successful PDIs.

Discussion also included a number of suggestions related to conducting effective PDIs:

• A successful PDI will uncover a variety of behaviours by going beyond the immediate caretaker and looking at grandmothers and others in the community.
• The PDI should go beyond a search for a food item; it should extend and expand until all key practices are discovered or rediscovered.
• The PDI should be conducted more than once if there are seasonal coping issues.
• It is helpful to identify behaviours that are the easiest to change.
• Looking at barriers can be instructive.
• It is important to identify strategies and not just behaviours. An example was given of Bolivia where the practice was serving a meat soup, but the strategy was the way a mother scooped down to get to the nutrients at the bottom of the soup pot.
• The person behind a successful PDI will take a "non-expert" approach, and be willing to take the time - following the caregiver in the home or kitchen.
• Probing for enabling factors is important. An example provided was asking, "How do you manage to have fish in your house when your neighbor can't afford to?"
• The PDI should look at health-seeking behaviour and caring behaviours, and also consider practices related to food processing and preservation.

Essential Element 2: Utilise community women volunteers to conduct the Hearth sessions and the follow-up home visits. Some organisations are modifying this element to create a community-level paid cadre who moves from community to community. "Management and supervision are key to quality and retention and some level of skilled and paid staff are needed. CCF India partners successfully used non-health staff as supervisors covering up to thirteen community-based staff working in over 150 communities."

Essential Element 3: Prior to the Hearth sessions, de-worm all children, update immunisations, and provide needed micronutrients. TAG participants agreed on the need to integrate PD/Hearth into larger child-health programmes for a variety of relevant interventions.

Essential Element 4: Use growth monitoring to identify newly malnourished children and monitor nutritional status of participants who have graduated from the Hearth. The presentation from Indonesia raised a question about whether any children are being missed when programmes use only weight-for-age and if weight-for-height should also be considered as admission criteria. One participant proposed mid-upper arm circumference (MUAC) as a screening method or for entry/exit; other participants responded that MUAC identifies only those children who are malnourished and wouldn't identify those who are faltering.

Essential Element 5: Ensure that caregivers bring a daily contribution of food and/or materials to the Hearth sessions. TAG presentations touched on ways communities help supply the Hearth. For example, Care Groups in Malawi have, on their own initiative, developed communal gardens to harvest foods for nutrition programmes, and have developed innovative ways to preserve food. Bringing contributions provides parents/caregivers with practice obtaining foods or other supplies they do not commonly harvest or buy for their children, reinforcing the new behaviour(s). Negotiations with the community can determine appropriate ways to help reinforce the practice of the new behaviour(s).

Essential Element 6: Design Hearth session menus based on locally available and affordable foods. There was consensus that this element was important. The Ethiopian experience provided an example of where this element was modified: with externally created menus, mothers were unable to obtain listed ingredients.

Element 7: The TAG was in agreement that to effect rapid recuperation, it is essential the Hearth contain a nutrient-dense menu (meal and snack) with listed amounts of calories, protein, and micronutrients per child based on their age.

Essential Element 8: TAG participants agreed that it was essential to have both the caregiver and the child involved each day of the Hearth session.

Essential Element 9: There was general agreement that conducting the Hearth session for a concentrated, contiguous period of time was important. There were, however, several variations presented to the exact timing. "Both standard sessions and modified sessions have helped large numbers of children improve their nutritional status..."

Essential Element 10: Include follow-up visits at home for 2 weeks after the Hearth session (every 1-2 days) to ensure the average of 21 days of practice needed to change a new behaviour into a habit. The TAG reiterated the importance of follow-up visits for behaviour change, but explored modifications. For instance, in Indonesia, more frequent home visits after the PD/Hearth were positively associated with greater success (e.g. graduation).

Essential Element 11: The TAG reconfirmed the importance of community involvement. In the Indonesian evaluation, a local leader who stops by the Hearth or provides food was associated with increased graduation rates. CRS decided up front to work very closely with local communities as part of their integration strategy. Village chiefs' participation was an important element for increasing overall community involvement. The Care Groups themselves were community-owned structures. An additional thought that emerged out of discussion was the opportunity to use the PD messages more broadly in the community with other health and nutrition programming.

Essential Element 12: Monitor and evaluate progress. The Addendum describes a monitoring process that looks at attendance, entering, and one-month weights, and the percentage of children who graduate after one or two sessions. Discussion revolved around the need to target by age. A challenge raised was the children who register as malnourished on the growth chart, but who are on a good growth trajectory. The Addendum lists some ways to look at longer-term impact.

Essential Element 13: If a child doesn't gain weight after two 10-12 day sessions, refer the child to a health facility to check for any underlying causes of illness such as TB, HIV/AIDS, or other infection. TAG participants agreed that it would be better to address chronic or acute diseases at the beginning of the recuperation period if possible, rather than waiting for 2 unsuccessful sessions before referral.

Essential Element 14: Limit the number of participants in each Hearth session. The Addendum suggests limiting the number of participants to a maximum of 10 caregivers, with 6-8 being an ideal number. The network of NGOs implementing PD/Hearth in Indonesia, which graduated 45% of participating children, adhered to the standard and limited the number of caregivers, but several other organisations modified this element. CRS Malawi invited an average of 18 mothers to its sessions. Despite corollary benefits experienced (e.g., the "normal" mothers enriched a discovery process), large numbers can overwhelm the process.

Integration with Other Programmes: "PD/Hearth is not intended to be a stand-alone program....PD/Hearth goes beyond nutrition as a social change model. It acts as a springboard for women's groups, Care Groups, shared child-care groups, and the many groupings of women that are going on in a community. The Care Group model was presented as one vehicle for connection as a group of mothers who are already sharing experiences, thus providing a context for discovery to take place. PD/Hearth will be most successful when linked to other health and nutrition interventions for all families within target communities, taking care to keep messages simple and focused."

Scale Up: The TAG discussed several methods that organisations have planned for scale.

• NGO collaboration: "The Indonesia network, CCF India local NGO partners, and CRS Malawi demonstrate the scale that is possible when multiple NGOs apply the same approach in their different program areas."
• Government: "Experience has shown that there is a tendency for governments to short-cut program standards, which can lead to a drop in quality and results, and this is difficult to counteract if NGOs are no longer engaged."
• Incorporate PD practices in broader community programming: A third method of taking PD/Hearth to scale is to integrate PDI information with larger community efforts.

Recommendations:

1. Establish a learning agenda and analytic framework for PD/Hearth implementation and scale-up that would enable comparison of approaches and results. (For instance, maintain an inventory of PD/Hearth programmes, develop a set of key indicators for programme quality and related to scale-up methodology, and document and analyse learnings).
2. Develop a "lessons learned" paper that explores challenges and modifications made in current PD/Hearth programmes.
3. Document NGO and government experiences in integration of PD/Hearth into overall nutrition or other development programmes, and support organisations to publish their findings.
4. Provide guidance on measurement methods for PD/Hearth programmes in developmental, transitional, and emergency settings. Publicise evidence and documentation that may be forthcoming on how children grow, etc.
5. Develop a series of case studies that examine how PD/Hearth has been replicated and taken to scale by different organisations.
6. Rewrite the PD/H Essential Elements Guide using language that would enable adaptation and provide illustrative anecdotes to improve quality PD/H programming.
7. Develop and disseminate a series of short communiqués to various audiences to increase knowledge and awareness of PD/Hearth as an important strategy for sustainably rehabilitating malnourished children.

TAG concludes that "There are several new and creative adaptations of the methodology, ongoing experiments in scaling-it up by different actors, and a conveyed community need for effective nutrition interventions. As discussed, there is ample evidence available that can be analyzed to extend current knowledge, and there is need and opportunity for extensive communication activities."

Posting to the CORE Child Survival (CS) Community Listserv, June 8 2009.