SECTION C
WHAT HAVE WE LEARNED?

We analyse and interpret here the main findings of the in-depth case studies and desk reviews in order to draw out the main lessons learned from the experiences of the programmes. The programmes’ strengths, weaknesses and constraints, as well as opportunities they offer, are also presented. An effort is made to identify those features that are common to all or many of the programmes, and to highlight some interesting findings of individual programmes.

1. Lessons learned

i) Macrocontextual factors

• A strong, supportive policy environment is crucial to the success and sustainability of a community-based nutrition programme. However, the mere existence of a national nutrition policy, or the fact that the government is signatory to international declarations and codes are not in themselves factors that will guarantee success. What is needed rather is public awareness of the importance of good nutrition, and of food and nutrition security as a human right for all, as well as a national commitment to improving nutrition as an integral part of the development process. Commitment itself must be backed up by a serious financial commitment from the national budget. In the absence of such financial commitment, policies and declarations will continue to gather dust on the shelves of government ministries. Examples of programmes that have benefited from strong political support are the Philippines’ LAKASS programme, Sri Lanka’s Samurdhi programme, Brazil’s Child Pastorate programme, Mexico’s PROGRESA and Thailand’s national poverty alleviation programme. All of these programmes are funded, exclusively or almost exclusively, by their governments. The policy environment need not specifically relate to a food and nutrition policy. Strong policies, backed up by good programmes, addressing rural development or poverty alleviation can easily accommodate nutrition priorities and have a positive impact on nutrition. Sri Lanka’s Samurdhi programme, Honduras’ rural development programme and PROGRESA in Mexico are examples of such programmes.
• International agencies and donors can help to create awareness of the need to address nutritional issues. One method is by supporting high-visibility and high-impact programmes, such as the one we have examined in Madagascar (SEECALINE). However, when such large-scale, costly programmes are planned, the funders themselves have a responsibility to ensure that a clear plan of financial handover is in place for the programme’s recurrent costs. While such a handover may be gradual and span many years, it needs to be agreed upon at the start of the programme, and adhered to rigorously for the duration of the programme.
• Funders must not impose their own priorities or time scales on countries. In the long run, such imposition is likely to lead to wasted funding. In recent years we have witnessed a series of efforts, largely donor-driven, to resolve the problem of micronutrient deficiencies using short-term approaches and targeting single nutrient deficiencies. Many have indeed achieved good results, but these results (and programmes) are often not sustained. Many national programmes have started with good intentions, with a combination of supplementation, fortification and food-based approaches, but few have paid much more than lip service to the food-based strategies⁸. The combination of such short and long-term strategies is indeed essential, but we also need to recognize that attacking a single nutrient deficiency is not the answer. Few nutrient deficiencies occur in isolation; almost all reflect an inadequate diet, inadequate both in quality and quantity. Factors other than poor diet are also important contributors to deficiency, most importantly health-related factors, and these need to be addressed within the context of an integrated approach to improving nutrition. Thailand is an example of a country that has successfully implemented such an approach.
• Intersectoral collaboration is an essential feature of a successful integrated nutrition programme. However, in the absence of strong political commitment, it may be difficult to achieve at the start of a programme. Multisectoral planning and coordination was espoused and promoted strongly during the 1970s. At that time, many countries established national multisectoral committees to address food and nutrition issues. Few were successful, and still fewer, if any, are active today. The literature on the topic is substantial. More recently, the 1992 International Conference on Nutrition encouraged a resurgence of interest in the approach, and many countries have once again established multisectoral committees with little regard to reasons for past failures.

Nutrition is indeed a cross-cutting issue, and a multisectoral approach therefore seems to be the sensible choice. In practice, however, collaboration and coordination among sectors is difficult to achieve at the central level, for a wide variety of reasons. It is often much more successful at the local level⁹. Community-level committees in particular are more likely to take a holistic view of development. They are not hampered by sectoral policies, priorities or budgets. At least some of their development priorities and felt needs will directly address nutrition (increased food production, supplementary feeding for the vulnerable, for example), others will indirectly impact on nutrition (access to markets, adequate water supply, sanitation and health care, employment opportunities, for example). Broadly, therefore, there are two options. On the one hand, it might be wisest to abandon efforts to achieve multisectoral collaboration at the top, initially at least, and to root the programme firmly within one ministry (for example, Ministry of Health or Ministry of Agriculture) at the central level, then use existing development structures at the district or community level (development committees, councils, community groups, for example) to achieve the desired collaboration. Later in the life of the programme it may be appropriate to extend collaboration upwards. This, to a large extent, planned or unplanned, is what Zimbabwe achieved through its Community Food and Nutrition Programme. Such an approach, however, by no means eliminates the need for strong support from the top. Thailand, on the other hand, has achieved a multisectoral and interdisciplinary approach. However, this did not come easily or quickly: “It may take more than a decade to get things off the ground, another decade for continual assessments and adjustments to be firmly established within the system, and another decade before the impacts can be clearly seen and evaluated”¹⁰. Public opinion and awareness are important contributors to government action, and these need to be harnessed to ensure the collaboration of a range of sectors: “Thailand’s experience has indicated that policy decisions which bring about deliberate actions are often in response to political concerns, public opinion and awareness”¹¹. Good nutrition must be viewed as an outcome to strive for, an indicator of development and progress. Historically, nutrition in Thailand was viewed as a medical issue, a disorder or disease such as malaria. Time and effort were needed to convince other sectors of the multifactorial nature of malnutrition: Their contribution was not an option, to be accommodated if it fitted in with their own priorities but a crucial prerequisite for national development.

• Strong technical institutions and expertise are needed to achieve a quality programme. In the absence of such national institutions and expertise, it is the responsibility of any large-scale programme to seek to build capacity. If it fails to do so, quality is jeopardized during the life of the programme, and it will become unsustainable once external support (financial and technical) ends. Capacity building needs to take place at all levels. Most programmes provide training for local-level workers (community health workers, agricultural extension staff, for example), but few consider the need for support for serious tertiary-level training (undergraduate and postgraduate programmes), other than short courses. Such tertiary education is generally considered to be beyond the mandate of the programme and adding inordinately to the cost of the programme. Yet the reality is that if the country has no well-trained, senior nutritionists to assume responsibility for the programme, then that country will continue to rely on external expertise and consistently fail to achieve self-reliance. Brazil, Mexico, Thailand and the Philippines all have strong research institutes and tertiary-level nutrition training programmes to train national and regional nutritionists. In the latter three countries these institutes have collaborated closely with the programmes examined for this report. The other issue to consider with regard to the availability of technical expertise is the administrative level at which it is available in relation to where it is most needed. Most countries have at least some nutritionists at the central level, and perhaps some at the middle level, but very few have adequately trained nutritionists at the local level. This, for example, was a problem faced by Zimbabwe’s programme: the absence of good-quality technical advice at the district or community levels, although Zimbabwe does at least have nutritionists at the middle (provincial) level. With the general trend towards decentralization, countries need to consider the availability of well-trained nutritionists at subnational levels where technical decisions will increasingly be taken.
• Events at the national or subnational levels, such as economic recession, political instability, civil war or natural disasters will affect the achievements of the programme. Most of the programmes studied, through desk reviews or in-depth case studies, had weak monitoring and evaluation systems¹². Claiming credit for a positive impact in the absence of a good research design or representative sample is questionable. Events beyond the control of a programme can affect both programme performance and impact and achievements (or lack of achievements) must be interpreted in the light of these events. There is little doubt, for example, that Thailand’s programme has benefited from years of political stability and a healthy national economy. Structural adjustment policies, of the 1980s in particular, compelled many developing countries to cut back on social sector spending. Thailand did not. This is not to belittle Thailand’s achievements in nutrition, but merely to place them in perspective: Thailand’s nutrition programme is impeccable and its achievements remarkable, but they benefited from an enabling macroeconomic and social environment. It is notable also in the Honduran case study that Southern Lempira, the department where the programme was implemented, was able to withstand the effects of a natural disaster (hurricane Mitch); and in Kenya, there was no decline in nutritional status despite drought and food shortages. LAKASS, the Philippines programme, claims significant improvement in the nutritional status of children despite a decline in the national economy. For a programme with national coverage, a research design that includes control groups or areas is not possible (unless there is a pilot phase of the programme). For these programmes, we must rely on qualitative and anecdotal data and other data sources (for example, economic data, data on the prevalence of AIDS) for information on events that may have affected programme performance and impact.
• The community-based programme should seek to establish collaborative links with other programmes (nutrition or otherwise) that could enhance its own programme impact. In a sense, this is an extension of multisectoral collaboration, and if such collaboration is achieved at the central level, the likelihood is also high of achieving interprogramme collaboration. Such collaboration could lead to a sharing of resources, both human and physical. It may, for example, not be necessary to employ community mobilizers if a rural development programme employs agricultural extension workers, or a health programme employs community health workers. With training, and assuming no conflict in priorities, staff can be shared. There are dangers in this approach: improving nutrition may be seen as a secondary activity or conflicts in supervision may arise. However, these can be minimized if an integrated approach to development prevails, or if communities are sufficiently empowered to demand that their needs for technical advice and other support be met. This leads us again to the importance of achieving good political and public awareness of nutrition as an outcome indicator of development, as in the case of Thailand.
• Partnerships with institutions outside the government sector can make valuable contributions to programme achievements and sustainability. Initially at least, NGOs (generally international NGOs) had tended to run their own development programmes, often with little involvement of national expertise, employing local staff but mostly not in high positions. This situation has changed in recent years. We have seen the growth of local or national NGOs who have become increasingly involved with their country’s development efforts, leaving international NGOs to focus more on emergency situations. Some of these national NGOs are quite large, while others work in fairly limited geographical areas. Partnerships with national NGOs are gaining in popularity, and in many cases their involvement constitutes a contribution to sustainability much as the involvement of a government ministry would. In the Bangladesh programme, for example, large national NGOs are contracted to manage the programme, with the collaboration of the Government. The programme began with two modalities operating in different regions of the country: government-led and NGO-led. Over time, the programme has moved to one modality, namely NGO-led. The Kenya project also is operated by an NGO, and NGOs are mentioned as partners in most of the programmes studied. Partnerships of community-based nutrition programmes with national training and research institutes have existed for many years. Such partnerships can bring valuable expertise into the programme, expertise for training and capacity building as well as expertise to run small and large research projects that can examine in detail aspects of the programme. Thus, for example, the Mexico programme has established partnerships with national (and international) research institutes, the outcome of which has been the availability of much useful data from studies on programme impact, programme management, and a cost-benefit analysis. National research and training institutes are partners also of the programmes in Kenya, the Philippines, Thailand, and Viet Nam. Partnerships with the private sector are also becoming more common, with the private sector being asked to undertake responsibility for specific components of the programme. The studies found such partnerships in Brazil, the Philippines and Zambia. Brazil also presents an example of a unique partnership, namely the Catholic Church. Indeed it was the Church that initiated the programme and continues to operate it. Partnerships can be valuable: they can provide services and products that are beyond the human and physical resources of government ministries. They can offer a flexibility free from the constraints of bureaucracy, and a diversity to match the diversity found in different regions of the country. The problems lie in the area of coordination and quality control, especially if too many partners are engaged in the programme. Some local NGOs have much good will but limited technical expertise. These problems were mentioned in the case study from Madagascar. These problems can be overcome with the selection of the right partners, strong programme management and a good monitoring and evaluation system.

ii) Community-level factors

• A community-based programme is not necessarily one that employs a participatory approach. By their very nature most nutrition programmes are community-based. At the very least, they require community dwellers, urban or rural, to ‘receive’ the services offered by the programme. Indeed, community members are often referred to as beneficiaries rather than participants. However, few community-based nutrition programmes are truly participatory in nature, engaging communities in decision-making and the selection of activities to answer their felt needs.
• “Community participation is an orientation, which should not be taken for granted; it should be an overriding and conscious concern, so as to avoid degenerating into mere rhetoric”¹³. The literature on participation and participatory strategies is large, and growing daily. Most people-oriented programmes today will naturally mention community involvement in some form. Indeed, all except one (Kenya) of the programmes examined for this report have community participation, involvement or development specifically mentioned as an objective of the programme. Yet none can be said to have achieved full participation (Level 6), as judged by the typology of participation provided in Annex 1: “Self-mobilization: People take initiatives independent of project staff. They develop contacts with external institutions to access technical expertise and funding, but retain control over decision-making”¹⁴. It is unrealistic perhaps to expect a nutrition programme to achieve such a level of participation, especially since the primary objective of the programme must be to achieve better nutrition, and not community development. Three programmes at least (Madagascar, Bangladesh and Sri Lanka) fall far short of achieving acceptable levels of community involvement, except as recipients of services, welfare or incentives. Interestingly though, two of these programmes (Bangladesh and Sri Lanka), almost despite themselves, are achieving a measure of empowerment in some communities at least, simply through the need to organize communities to participate in their activities. In Sri Lanka, for example, deviations from pre-set activities are now permitted, as the process of community participation develops. In Bangladesh, as in the communities where the programme first began, villagers are becoming more vocal in their views and demands, which are increasingly accommodated by local leaders; these communities are thus graduating to “functional participation” (Level 4 of the typology given in Annex 1).
• A fully participatory approach to programme design and implementation is difficult to reconcile with donor procedures for seeking programme funds and also with programme durations that are normally acceptable to donors. A truly participatory programme would need a blank cheque from donors and an unspecified time frame. It needs a lengthy preparatory phase¹⁵ during which community mobilizers work with community groups (or establish groups if none exist) to help them identify their problems, needs, priorities and to plan activities. The nature of these activities cannot be predetermined, as this contradicts the basic principle of a participatory approach, nor will they necessarily be the same for all communities in a region, let alone in a country, especially in a country that is ecologically and culturally diverse. The pace of the process of achieving community empowerment varies also from country to country and within a country. Many factors may influence the rate of achieving empowerment:
• the stage of development of the community at the start of the exercise¹⁶;
• the level of literacy;
• the position of women;
• the economic conditions of the community: a very poor community for example, struggles to survive, and has no time to devote to community efforts;
• the pre-existence of a strong representative community group¹⁷;
• a culture of working together for the common good rather than as individuals;
• the degree of homogeneity of the community: a community where most members are experiencing the same problems because of similar socio-economic conditions or culture is more likely to function as a single unit;
• the degree of geographic or social isolation of the community¹⁸.

These are some of the factors that may promote or delay a participatory approach. They have emerged from experience and are recorded in the literature or in the case studies of this exercise. The reality is, therefore, that full community participation demands a considerable investment of time and resources with little to show in terms of impact on nutritional status at the early stages of the programme. This is generally not acceptable to donors (although some donors do now recognize and accommodate the need for community engagement), nor indeed to governments. Faced with poverty, food insecurity and malnutrition, communities too may opt for shorter-term, high impact solutions¹⁹. The one approach, however, does not necessarily negate the other. Community organization and empowerment can proceed while at the same time addressing the more urgent, life-threatening concerns of individual community members through traditional top-down approaches. The problem arises when the latter becomes the only mode of operation and the objective of achieving community participation is forgotten or receives no more than a token gesture. Finally, what is needed from donors is a more flexible approach to programme design and funding: an acceptance that it may not be possible, indeed it is not desirable, to define precisely the nature of the activities to be undertaken by communities (and hence also the cost of these activities), but rather to allocate funds to broadly defined activities that will be specified more clearly during the process of community planning. This flexibility was achieved in the case of the Zambia project and the project in Central America (Ismail, 1999).

• Community participation for nutrition improvement cannot function in the absence of a supportive macroenvironment, at the national and subnational level, which ensures good quality nutrition support services and which responds to community demands in a timely fashion. It is tempting for government ministries to abdicate responsibility in a participatory programme, where communities are expected to make their own decisions and seek their own solutions. Community participation should not be viewed as a way out of unsuccessful nutrition programmes. Communities will continue to need access to services provided by the government, NGOs, private sector or other agencies and institutions. Indeed, if community participation is successful, demand for such services will rise and the insistence upon quality will also rise: better access to good health care and nutrition services, education, access to markets, safe water supply and good sanitation. As the example of Thailand has shown us, we must link the top with the bottom. Community participation will fail if its demands and needs consistently remain unmet. The programmes of Brazil and Mexico mention the threat of increasingly unmet demand for services.
• Capacity building and easy access to sound technical advice are essential components of effective community participation. Communities have a wealth of local knowledge which should not be ignored (see below). Equally there are large areas of inexperience and lack of knowledge that need to be addressed by the programme. Of primary importance in achieving community participation and development is an understanding of community organization, good planning and management skills, and the ability to undertake a situation analysis and prioritize action. Capacity-building is thus a crucial component of a successful programme. The programmes in Bangladesh, Honduras, Kenya, the Philippines, Thailand and Zambia have recognized this. Linked to capacity-building is the ability to recognize when to seek technical expertise and knowing how to obtain access to it. It thus becomes crucial that communities be able to have access to sound advice easily. Many small local NGOs are unable to provide the quality of technical advice that communities need, leading to mistakes and wasted resources. This factor is closely linked to factors mentioned previously: availability of nutritional and other expertise at the local level where it is needed, and quality control. The absence of such expertise is highlighted in the studies of programmes from Kenya, Zambia and Zimbabwe.
• The utilization of existing community groups, provided these are active, may be preferable to the superimposition of new groups. Communities are complex entities, with pre-established hierarchies and linkages, through friendships, historical events and experiences, and kinships. Even a seemingly homogenous community will have depths and networks that cannot be appreciated in the time available for programme preparation. Communities also have a leader (African villages for example almost always have a chief) and varying degrees of self-organization, often as a result of their cultures and centuries of tradition. We ignore these complex community dynamics at our peril. The temptation is to superimpose new structures and new organizations for programme implementation, with the excellent intention of creating community groups that are truly representative of all members of the community, that are focused specifically on the programme and are answerable to it. The alternative is to work with existing community groups and leadership, imperfect though these may be, and strive to transform and improve them slowly through capacity-building and other programme activities. Such an approach may improve chances of institutionalization of the activities, community ownership of the programme, and ultimately, sustainability. The programmes in Kenya and Zambia, for example, have successfully used existing community groups. The dangers of this approach include poor leadership, poor representativeness of the group, domination by the few wealthier members of the community and the marginalization of women. However, there is no guarantee that new groups established by the programmes will be any different. Either approach is likely to demand a time-consuming process of education and sensitization, with due recognition of local cultural realities.

Most programmes have allowed communities to select their own ‘volunteers’ (or mobilizers or promoters; the terminology varies, and not all are volunteers; in many instances they are paid employees of the programme). Thailand for example, employed a complex ‘sociogram process’ that was agreed upon by all members of the community. Programmes generally provide a minimal level of technical training but few provide a course on leadership and group dynamics as did the Zambia project. Choosing and training the right volunteer is crucial to the success of the programme, so time devoted to this is not time wasted. It is also important to remember that the volunteer too, like any other community member, will have his or her own ‘niche’ within that community, and friendships and networks. He or she will have aspirations, and indeed the decision to volunteer may be motivated by hopes and expectations that are not congruent with the programme. Such mistakes have occurred and will continue to occur in all programmes that use a participatory approach. The important thing is to have built-in ways of identifying mistakes and acting to correct them. This highlights the need for good supervision of the volunteers, a need recognized by many of the programmes examined. It also emphasizes the need to have a built-in and funded system of social recognition of the achievements and contributions of the volunteers²⁰. Such a system is very important for motivation, to strengthen the link with the central level, and to demonstrate the government’s support for the programme.

• Local cultural practices and knowledge can provide useful entry points for community action. Communities possess a wealth of knowledge and experiences that can be harnessed to become a part of the programme. In Honduras, the project was in fact built around a soil conservation technique developed in the local village of Quesungual. Communities that have lived with food insecurity have developed a number of coping strategies to improve their chances of survival. Many of these can be retained and supported. Local cultural practices, such as Zunde raMambo in Zimbabwe, offer entry points and strategies that are more acceptable to communities than unfamiliar, externally-imposed strategies. In Kenya the tradition of women’s groups and regular community meetings was used by the project to encourage participation.
• Sharing of experiences and information among communities can lead to greater empowerment and the ability to have better representation at the next level (district or municipality). In the Central American project (Ismail, 1999), pilot communities passed on to neighbouring communities skills in food security planning and project development learned through the project, such that the project expanded from the original eight communities to forty. These communities were then able to make stronger representation to municipal committees. In Zambia, the Luapula Valley project illustrated that stronger communities can help others. Some community groups shared their newly acquired groups skills with other groups that were not as developed. Such events cannot necessarily be initiated by the programme but should be supported when they happen.
• The programme must seek to ensure that funding and technical support is available to support the microprojects that communities select and design to meet their felt needs. In Zambia, for example, despite arrangements to fund microprojects identified by community groups as part of their planning process, funding frequently did not materialize, leading to disillusionment and a sense of abandonment by the project. In many cases too the technical advice provided was questionable. Again in the Central American project, communities developed plans, established priorities, and designed projects to address the problems they felt needed most urgent attention. The project, in an effort to encourage self-reliance and sustainability, had decided that communities should seek their own funding from local NGOs, international agencies and municipal budgets. Despite assistance provided by community facilitators this proved difficult in many cases as communities came up against donor priorities that were not in line with their own. An incomplete decentralization process meant also that municipal budgets were limited and inflexible. Communities were thus compelled to re-align their priorities to accommodate those of donors.

iii) Programme design features

Many programme design features are implicit in or follow from the lessons learned on macrocontextual and community-level factors listed and discussed previously. We describe here those key features that complement these factors, that are essential to a good programme design, and that are (or should be) within the control of a programme planner. This subsection is not intended to be a comprehensive guide to programme formulation; rather it highlights the findings and lessons learned from the nine in-depth case studies and three desk reviews that form the subject of this report.

• Developing a conceptual framework:
  A conceptual framework of the causes of malnutrition in a country sets the scene for action. It identifies clearly those factors, nutritional and otherwise, that need to be addressed if the nutrition situation is to improve. It leads logically to an integrated approach and identifies the sectors that need to contribute to the programme. The absence of such a framework led, in the cases of Bangladesh and Madagascar, to short-term interventions and a curative rather than preventive approach which did not address the root causes of malnutrition. Thailand’s conceptual framework, on the other hand, was a valuable guide to decision-making and action.
• Setting objectives:
  It is against clear measurable objectives that a programme’s success or failure will be evaluated. A community-based nutrition programme must at minimum include objectives related to improved nutrition and to encouraging community participation. Ultimately, such a programme must aim to improve nutritional status (anthropometric status and/or micronutrient status, as appropriate) but there are a number of food-related and health-related objectives that can be seen as intermediary objectives, and contributing to the ultimate aim of improved nutritional status. Thus, improving access to food or food security, or increasing diet diversity or improving weaning diets are all food-related objectives that can contribute to improved nutrition but will not necessarily achieve improved nutrition unless other contributors to malnutrition are also addressed, such as health, poor sanitation and water supply. This highlights again the need for a conceptual framework and an integrated approach to tackle malnutrition.

If it is accepted that community participation is a key contributor to sustainability (see Section C, subsection iv) on sustainability), then a sustainable, as opposed to transitory, positive impact on nutrition will only be achieved if the community is thoroughly engaged in the programme. Hence promoting community participation, and all that this entails, must also be an essential objective of the programme.

The findings from virtually all the studies have illustrated the importance of a supportive macroenvironment for the programme. In the absence of such an environment, it becomes the responsibility of a national nutrition programme to seek to create it. Thailand’s experience has shown that creating political and public awareness of nutrition as a crucial developmental issue can contribute substantially to engaging political support and commitment. Madagascar, Zimbabwe and the Philippines have illustrated how community-based nutrition programmes can act as advocates for nutrition and help to raise awareness. Thus, an objective relating specifically to developing political and public awareness is essential in a national programme for a country where a supportive macroenvironment is absent. Other objectives of the programme will relate to the specific focus of the programme: poverty alleviation, improving food production, raising incomes of the ultra poor, improving the quality of life, improving access to health care and education, are some examples found in the programmes examined.

• Links to other programmes:
  Intersectoral collaboration has been discussed above. However, an aspect of such collaboration that is often forgotten, is the linkage with other programmes and projects, some of which may themselves have nutrition or nutrition-related objectives. Such linkages can provide useful technical support and a sharing of resources. The link should be viewed as two-way: nutrition inputs can strengthen an agricultural programme (essentially the Honduras project, for example, was an agricultural programme) or a poverty alleviation programme (such as the Samurdhi programme) just as much as these non-nutrition programmes can contribute to a nutrition programme. In Viet Nam, collaboration of its Household Food Security Programme with a health programme was beneficial to both programmes. Such linkages generally need to be established at central level, which emphasizes again the importance of achieving good intersectoral collaboration.
• Targeting:
  Most programmes have explicit or implicit targeting of one form or another. Targeting can take various forms: geographic targeting, such as that found in the Kenya project (arid and semi-arid lands) or the Honduras project (focus on an area with a specific environmental problem); vulnerable area targeting (targeting to areas where the density of poor communities is high); or socio-economic targeting (selection of households below a poverty line, such as found in the Samurdhi programme). In the Mexico programme a more elaborate dual form of targeting was employed: vulnerable area targeting to select programme implementation localities, then socio-economic targeting to select participating households. Targeting can also occur by default: wealthy households are unlikely to avail themselves of public health services if better private health care is on hand.

Targeting, or the need for targeting, should be considered at the start of a programme. Good targeting can save resources, on the other hand elaborate screening procedures entail unnecessary bureaucracy and high administrative costs (for example, the case of Mexico). They are also open to political manipulation (such as was found in the programmes of Bangladesh and Sri Lanka) and corruption.

• Provision of basic services:
  Much of the success of the Thailand experience relates to its insistence on the access of all households to a package of basic minimum needs. This is indeed a desirable condition and one that most countries strive towards. However, the reality of many of the world’s poorest countries is that both access to such services and also the quality of services available to poor communities are grossly inadequate. When this is the case, firstly it becomes the responsibility of a national nutrition programme to ensure that, at the very least, the nutrition services that are available to poor communities are indeed accessible to all and are of good quality. Secondly, the programme needs to advocate for improved services in other areas, especially health and agriculture so as to reduce morbidity and improve household food security. Many nutrition programmes address some health issues that impact directly on nutritional status, such as diarrhoea, and also engage in community or household food production activities. Ultimately, however, their achievements will be limited unless other relevant sectors are able to improve the services they provide.
• Programme monitoring and evaluation:
  A major constraint experienced when analysing the case studies was determining if the programme had had any significant impact on nutrition. Where data were available, it was generally of questionable quality with results that were clearly wrong or not credible. In Madagascar for example, the data were obtained from the programme’s growth monitoring activity. Since participation in the activity was less than 50 percent, and linked to the availability of the food supplement, these data cannot be accepted as an indication of positive impact or otherwise. The reductions in infant and maternal mortality rates reported by the Brazil programme are also barely credible. The Mexico programme, however, seemed to have an acceptable monitoring and evaluation system and was able to offer good data on nutritional impact.

All programmes should have a built in monitoring and evaluation system, and there should be a budget allocated to this component of the programme. The programme monitoring should be designed to provide information on ‘process’ so as to improve programme management. Information from this process monitoring system should be available on a timely basis so that action can be taken quickly to correct any operational or management errors. Indicators for impact evaluation must be identified at the start of the programme, as well as the frequency and methodology for data collection. The indicators selected must relate to the objectives of the programme, so that the programme can clearly say whether or not it is meeting its objectives. Neither component of the monitoring and evaluation system need be elaborate; a few simple indicators, collected routinely can provide the desired information. If the country has an adequate food and nutritional surveillance system which few have then this information can be used for at least some of the information needed. On the whole, it is not recommended that growth monitoring data be used to assess the impact of the programme on nutritional status, unless coverage is close to 100 percent, the data quality is good and there is no bias in the age of children attending growth monitoring sessions²¹. Growth monitoring data are also generally limited to weight-for-age as the only nutritional status indicator. This indicator does not distinguish between wasting and stunting, a distinction that nutritionists are increasingly recommending since the causes of the two conditions may be different²². If the country has a good research institute, it might be best to follow the example of Mexico and contract out the monitoring and evaluation activity of the programme.

• Participatory monitoring:
  Participatory monitoring refers to the self-monitoring undertaken by communities to assess their own progress towards achieving their developmental goals. It should be developed and designed by them using indicators that relate directly to their activities and aims and in a form that is easily understood by community members, taking into account literacy levels in the community. In many cases, indicators are represented pictorially, using simple graphs and maps (to show for example which houses have acquired latrines). In some cases communities may choose to include information from community growth monitoring as part of their system but this should be their choice. Programme managers should not impose any uniformity on such monitoring systems; this is information gathered by the community and for the community, and should be an integral part of the process of achieving participation. It provides the essential feedback needed to motivate communities. None of the in-depth case studies made reference to a participatory monitoring system. The Zambia and Central American projects attempted to persuade communities to establish participatory monitoring but were unsuccessful. In the Central American project the concept was not well understood and therefore not promoted by the facilitators (Ismail, 1999).
• Management issues:
  Management problems were mentioned specifically only in the case study report for the Bangladesh programme. There are, however, a number of other issues that were mentioned that can come under the heading of management: operational delays in release of funding and resources (Bangladesh, Honduras, Philippines), poor supervision and quality control (Madagascar); absence of feedback to communities (most programmes), political interference in programme implementation and selection of programme staff (Bangladesh and Sri Lanka), excessive control over community leadership (Brazil), absence of feasibility studies for income-generating activities (Sri Lanka) or failure to secure funding for the activities (Kenya), and difficulties with credit schemes (Viet Nam).

Many programmes had elaborate management structures, with committees at all levels and frequent meetings. The challenge with good management is to establish a structure that promotes transparency, that defines roles and responsibilities clearly, that permits quick response and limits bureaucratic procedures, but that at the same time is able to check misuse of programme resources and is also not inordinately time consuming. Perhaps one of the most important features of good management is the ability to maintain a committed and motivated staff. For this, frequent feedback is needed and a recognition of achievement and good performance.

iv) Sustainability

Sustainability is a complex issue. Are we concerned about the sustainability of the programme or of its achievements? At one level sustainability can defined as the ability to maintain the positive impact of a programme, once that programme has achieved its objectives. This definition has many implications. If a programme has achieved its objective of reducing malnutrition by 30 percent, then this achievement must be sustained for generations to come. However, is this really enough? Should we not be saying that the objective for the next phase should be to reduce malnutrition by a further 30 percent, and so on? Future generations will need good basic services just as much as the communities of today, so the thinking that goes into a programme’s design must take that into account. A community-based nutrition programme should not be viewed as a ‘one-off’ activity, but rather as a continuing commitment.

Our case studies are interesting in that they present a range of objectives, not all relating directly to nutrition. Mexico’s PROGRESA programme aims to break the cycle of poverty, the argument presumably being that this would allow development to take place. If this is so, and the programme achieves its aim, arguably then the sustainability of the programme itself is not an issue. It will have served its purpose of ‘jumpstarting’ the development process, and support to this process would then need a different approach. Along similar lines, Sri Lanka’s Samurdhi programme sees itself as a transition from a welfare approach to a development approach. If the transition is successful, the programme will have done its job. Honduras’ PROLESUR programme, on the other hand, focuses on a transfer of technology, although its objectives are much broader than this, relating to the improvement of the quality of life of the Southern Lempira Department’s rural population. It has clearly succeeded in transferring the technologies, and these appear now to be institutionalized and hence sustainable (although still in need of some continuing technical support), but it has been less successful in other areas (health services for example).

With the exception of Sri Lanka’s programme, all the programmes have nutrition or food-related objectives, and if their achievements in these areas are to be maintained, indeed heightened, then the programmes themselves will need to be sustained in some way. This then raises the issue of continued funding, even if it is assumed that some of the programmes’ activities can and should become incorporated and institutionalized in sectoral plans and activities. Ministries of Health and Agriculture can indeed assume responsibility for much that relates to nutrition, but there will continue to be a need for a specific focus on nutrition activities, and for this a country will continue to need a nutrition programme. The governments of the Philippines and Zimbabwe, and to a large extent Brazil, have shown a clear commitment in this regard, and national funding has been made available and sustained for many years. In these countries, the supportive macropolicy environment is translated into a tangible investment in nutrition.

This has not happened in Bangladesh, which continues to rely heavily on external donors. In the poorest countries of the world perhaps sustainability lies in the willingness of donors to continue to provide funding. There is a danger in such reliance: the case of Zimbabwe (and many other countries) has shown how political events can lead to the withdrawal of donor support. There is also the danger of donor-fatigue: simply put, the donor’s decision that it is time to move on to something else or somewhere else. In Tanzania and Zimbabwe, for example, nutrition, and related institution building, enjoyed the support of Swedish development aid for many years, since 1973 for Tanzania and since the early 1980s for Zimbabwe. In 1999, SIDA decided to stop the assistance for no obvious technical or developmental reason. This raises the issue of the time frame needed for sustainability. Few donors will continue to support a programme for as long as SIDA did in Tanzania and Zimbabwe. Madagascar’s programme, for example, will end in 2003, and there is no evidence of any effort on the part of the government to assume responsibility for it.

Sustainability lies also in the continued availability of technical expertise. Increasingly, many developing countries are able to provide this, but often only to a limited extent. Bangladesh, for example, has strong technical institutes, and many well-trained nutritionists, but apparently not enough to provide the human resources needed for the Bangladesh Integrated Nutrition Programme (BINP). Hence the programme had to employ staff with inadequate technical expertise. Capacity-building, at all levels (including the community level), thus becomes a responsibility of the programme (as discussed above under Lessons Learned, Section C, 1). There is increasing reliance on local NGOs to provide the needed expertise and the continued existence of NGOs itself provides a measure of sustainability. Almost all the programmes collaborated with NGOs (see Table 1), contracting out to them specific components of the programme or, as in the case of Bangladesh, almost the whole programme. This can lead to difficulties with adequate supervision and poor quality control.

Community participation is a precondition for sustainability but does not in itself guarantee sustainability. Community participation encompasses community engagement, a sense of ownership and a sense of the ability to control and influence one’s environment. It will not be found in programmes which view community members as mere recipients, as beneficiaries (indeed calls them beneficiaries) asking them to passively donate their time and resources, with little or no powers of decision-making. On the other hand, a programme that succeeds in empowering communities but then fails them by not providing the basic services they are entitled to expect, or the technical and financial support they require for their chosen development activities, can only lead to alienation and disillusionment. Thus successful linking of ‘top to bottom’, of the macroenvironment to the microenvironment, is as important as achieving community empowerment, as Thailand has demonstrated.

Ultimately then, sustainability lies in the recognition that good nutrition and food security are crucial elements of development, and that therefore nutrition activities must become institutionalized within the government structure at all levels. Political and public awareness of the need for an integrated approach to nutrition, and commitment to achieving the goal of a healthy, well nourished population then become the most important elements of sustainability.

2. Results of the SWOC (strengths, weaknesses, opportunities and constraints/threats) analyses performed for the in-depth case studies

Tables 3 to 6 summarize and bring together the nine SWOC/T analyses carried out by the regional consultants for their allocated programmes. There may be a methodological problem in comparing the analyses of the different programmes: the analyses inevitably summarize the wealth of information gathered for the case study, and this leads to decisions on what to include and what to exclude. When, for example, we find that only the African programmes have weak monitoring and evaluation systems, this may reflect a decision to mention this issue as it relates to these programmes, rather than a real regional difference. These tables should therefore be viewed as a minimum list of strengths, weaknesses, opportunities and constraints or threats: if a feature is not checked for a particular country, it does not necessarily mean that it is absent.

Table 3: Strengths of nine in-depth case studies (^*)

Table 4: Weaknesses of nine in-depth case studies (*)

Table 5: Opportunities presented by, or available to, nine in-depth case studies (^*)

Table 6: Threats and constraints to nine in-depth case studies (^*)

⁸ It is recognized that supplementation is essential in some situations; faced with a child showing signs of vitamin A deficiency, for example, you must provide the appropriate supplement.
⁹ Countries vary in their administrative structures and the nomenclature of their geographical or administrative subdivisions. For the sake of simplicity, this report will use the following terms, although it is recognized that each level (central, middle and local) is often further subdivided: a) central or national level; b) middle or provincial level (includes regions, departments or municipalities); and c) local level (includes districts, communities and villages).
¹⁰ Quote from: Integrating Food and Nutrition into Development: Thailand’s Experiences and Future Visions. Edited by P. Winichagoon et al. (p.229). UNICEF.
¹¹ Ibid.
¹² Monitoring and evaluation is discussed further under Section C, 1) iii) Programme design features.
¹³ Quoted from the case study report of the Philippines LAKASS programme.
¹⁴ Quoted from: A typology of community participation (see Annex 1).
¹⁵ The Zambia study stresses the importance of an adequate preparatory phase during which the characterization of nutritional vulnerability and food insecurity in the project area is undertaken. The Viet Nam study points out the importance of gaining a clear understanding of community conditions, dynamics and behaviour. To achieve such understanding of local conditions for a large, national programme is not feasible in the preparatory phase of programme development, given the ecological and cultural diversity found in most countries. It thus needs to be included as part of the programme’s initial activities. This issue is discussed further in Section D.
¹⁶ In the Central American project (Ismail, 1999), communities in Nicaragua were more organized and prepared for the community approach, largely because of their history of civil war, the position of women which itself was influenced by the war, and the level of literacy. In Guatemala, on the other hand, abject poverty and the struggle for survival hampered community development efforts.
¹⁷ In the Kenya case study, it is reported that the tribal Kamba culture requires that every woman, of any age, religion or social class, be a member of a group (10-15 members per group) known as “Mwethya”. Also in the same region of Kenya, community decisions are commonly taken through open discussion at community meetings called “Baraza”.
¹⁸ The Philippines case study mentions the geographic isolation of some communities as a constraint to full community involvement in the programme. The Zambia desk study highlights the difficulties not only of reaching the poorest households, but also of reaching the most vulnerable individuals within the household.
¹⁹ The mother of a malnourished child cannot wait; she is concerned about the survival of her child, not the organization of her community or the sustainability of the programme.
²⁰ Thailand made good use of this approach and found it essential to the smooth running of the programme. Social recognition can take the form of badges or t-shirts, public recognition through various ceremonies, or training opportunities even outside the community. Arguably, funding spent on a social recognition programme is more effective than funding spent on salaries for the volunteers.
²¹ Coverage of most growth monitoring programmes does have an age bias: coverage is good in the first year of life when mothers attend clinics to obtain immunization for their children, but falls dramatically thereafter.
²² Victora (1992) suggests that wasting is linked to a deficiency of energy, while stunting occurs when diversity is poor, namely that the diet is lacking in specific micronutrients.