Mainstreaming nutrition into maternal and child health programmes: scaling up of exclusive breastfeeding

Abstract

Table of Contents
Summary	5
Background	6
Methods and search strategy	6
The scale‐up process	7
Issues that need to be addressed while scaling up a programme for exclusive breastfeeding	7
Conclusions	13
Acknowledgements	14
References	14
Appendix 1	18
Appendix 2	22

Summary

Interventions to promote exclusive breastfeeding have been estimated to have the potential to prevent 13% of all under‐5 deaths in developing countries and are the single most important preventive intervention against child mortality. According to World Health Organization and United Nations Children Funds (UNICEF), only 39% infants are exclusively breastfed for less than 4 months. This review examines programme efforts to scale up exclusive breastfeeding in different countries and draws lesson for successful scale‐up. Opportunities and challenges in scaling up of exclusive breastfeeding into Maternal and Child Health programmes are identified. The key processes required for exclusive breastfeeding scale‐up are: (1) an evidence‐based policy and science‐driven technical guidelines; and (2) an implementation strategy and plan for achieving high exclusive breastfeeding rates in all strata of society, on a sustainable basis. Factors related to success include political will, strong advocacy, enabling policies, well‐defined short‐ and long‐term programme strategy, sustained financial support, clear definition of roles of multiple stakeholders and emphasis on delivery at the community level. Effective use of antenatal, birth and post‐natal contacts at homes and through community mobilization efforts is emphasized. Formative research to ensure appropriate intervention design and delivery is critical particularly in areas with high HIV prevalence. Strong communication strategy and support, quality trainers and training contributed significantly to programme success. Monitoring and evaluation with feedback systems that allow for periodic programme corrections and continued innovation are central to very high coverage. Legal framework must make it possible for mothers to exclusively breastfeed for at least 4 months. Sustained programme efforts are critical to achieve high coverage and this requires strong national‐ and state‐level leadership.

Background

The Lancet series on child survival identified breastfeeding interventions to have the potential to prevent 13% of all under‐5 deaths in developing areas of the world, ranking it as the most important preventive approach for saving child lives (Jones et al. 2003). Exclusive breastfeeding, that is, administering only breastmilk and no other liquids or foods (WHO 2004) for the first 6 months of life confers important benefits to the infant and the mother – it protects infants against many common childhood diseases, including repeated gastrointestinal infections and pneumonia, and thereby against some of the major causes of childhood mortality (Feachem & Koblinsky 1984; Jason et al. 1984; Habicht et al. 1986; de Zoysa et al. 1991; WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality 2000; Arifeen et al. 2001). A large volume of literature exists regarding efficacy trials that have clearly demonstrated the beneficial impact of exclusive breastfeeding (Green 1999).

Unfortunately, exclusive breastfeeding is not practised universally. Global monitoring indicates that only 39% of infants are exclusively breastfed for less than 4 months following birth (WHO/UNICEF 2003). Some reasons for failure to breastfeed exclusively for the recommended first 6 months of life include community beliefs that result in delay in initiation of breastfeeding after birth, giving other fluids and/or foods to breastfeeding infants, the lack of social support for women in resolving breastfeeding difficulties and lack of health system support (Green 1999).

The current international optimal feeding recommendations include exclusive breastfeeding for 6 months; these are based on a review of scientific literature to define the optimal duration of breastfeeding (Butte et al. 2002; Kramer & Kakuma 2002; WHO 2002). These are also included in United Nations Children Funds’ (UNICEF's) Facts for Life ‘Key Message: What every family and community has a right to know about breastfeeding’ (UNICEF 2002).

In the early 1990s, the World Health Organization (WHO) and UNICEF launched the Baby‐Friendly Hospital Initiative (BFHI) for strengthening maternity services to support breastfeeding practices (WHO/UNICEF 1992; table 3).

Although the BFHI includes a community component, its implementation in developing countries is weak and support is only available in institutions before and for a short period after delivery. In several regions a large proportion of deliveries occur at home. Interventions at the community level are, therefore, very important for supporting exclusive breastfeeding.

Identifying effective and sustainable means for scaling up of proven interventions into child health programmes remains a major problem. The purpose of this review is to define a process for scaling up exclusive breastfeeding based on experiences of programmes or pilot projects for promotion of exclusive breastfeeding implemented and evaluated in different countries. The term ‘scale up’ is commonly understood to mean reaching a larger number of beneficiaries with a given intervention. However, apart from coverage, maintaining quality as larger numbers are reached and effectively reaching out to those difficult to reach are also important considerations (The Core Group 2005). The issues addressed herein pertain to policy, programme design, implementation strategy, resources and costs management, monitoring and evaluation, and the possibility of continuing innovation and testing of novel approaches.

Methods and search strategy

We conducted a systematic literature search with no time limits defined, and used data from published literature, programme reports and monographs describing effectiveness or cluster randomized trials or implementation programmes. The databases searched included PubMed, the National Library of Medicine's search service and Cochrane Library (Cochrane Controlled Trials Register) using the following keywords: breastfeeding or exclusive breastfeeding or breastfeeding promotion or breastfeeding intervention or breastfeeding programme or breastfeeding education or breastfeeding effect or impact or evaluation. We also reviewed reference lists of identified articles and hand searched reviews, bibliographies of books, and abstracts and proceedings of national and international conferences or meetings. Additionally, we contacted key researchers and organizations/agencies for unpublished material or narrowly disseminated reports.

Consistent with the programmatic focus, the examples quoted in this review are demonstration projects that reached large populations or cluster randomized intervention trials. The emphasis is on developing countries and as some of these countries are burdened by HIV/AIDS, this review includes this context as well.

The scale‐up process

The scale‐up planning process must follow a logical order. Several key issues that need attention are summarized in Table 1. It is apparent that many of the features of the process are generic to neonatal and child health programmes, but there are issues in the area of exclusive breastfeeding that requires additional attention, particularly in the HIV context.

Table 1.

Steps in scaling up

Assess situation, create a policy environment conducive to exclusive breastfeeding

Define roles, relationships and responsibilities of all partners, establish agreements

Review technical support

Define programme strategy

Mobilize resources

Provide training and technical assistance

Develop and use monitoring and evaluation systems

Monitor coverage and quality

Measure impact and cost

Provide for testing novel approaches and continuing innovation

Adapted from Knippenberg et al. (2005) and Gonzales et al. (1998).

Issues that need to be addressed while scaling up a programme for exclusive breastfeeding

Policy framework

The Global Strategy on Infant and Young Child Feeding was adopted by a World Health Assembly resolution (55.25; WHA 2002). The global strategy builds on the BFHI (WHO/UNICEF 1992), the International Code of Marketing of Breastmilk Substitutes (WHO 1981) and the Innocenti Declaration on the Protection (WHO 1989), Promotion and Support of Breastfeeding in the overall context of national policies and programmes on nutrition and child health. The aim of the strategy is to improve the feeding of infants and young children and increase the commitment of governments, society groups and international organizations to promote the health and nutrition of children. The strategy emphasizes the need for comprehensive national policies on infant and young child feeding, on an urgent basis including guidelines on ensuring optimal feeding of infants and young children in exceptionally difficult circumstances and the need to ensure that all health services protect, promote and support infant and young child feeding (Gupta 2002; WHO/UNICEF 2003).

Once the decision to scale up a programme for exclusive breastfeeding is made in a region or state or country, the logical starting point is to review the relevant, existing national policies and to assess whether these address varying circumstances under which women deliver and live such as urban or rural residence, working women and place of delivery. Specific features to note are the status of the implementation of the BFHI in facilities where women deliver, access and support to mothers with young infants by trained personnel during the post‐natal period and beyond, support for exclusive breastfeeding in community setting, legislation for working mothers that ensures leave of appropriate duration, an enabling environment at the work place within government and private settings, and the policies in place for discouraging promotion of breastmilk substitutes.

Existing policies are optimal in some countries (The Norwegian Government 1993; ILO 1998) but lacunae exists in others; for example, exclusive breastfeeding for 6 months is a policy in India but the duration of maternity leave for government employees is 135 days (Government of India 1997). In Norway, however, a mother is entitled to over 6 months of maternity leave (The Norwegian Government 1993). Wherever weaknesses exist, a decision needs to be made on how these will be addressed during scale‐up.

Prior to scale‐up, a policy framework for implementation needs to be developed. This framework should include a vision, consensus on technical issues, on programme strategy, roles, relationships and responsibilities of implementing partners, measurable goals within defined time lines and predictable adequate funding to achieve these (The Core Group 2005). It should address the capacity of all implementing partners and the approach to fulfilling capacity gaps at all levels particularly at the community level. The policy framework should emphasize breastfeeding promotion through community‐based programmes. Political commitment and ownership by relevant government departments at national, state and district levels have also been identified to be critical for effective scale‐up (The Core Group 2005).

Technical and management support and other partnerships

Creating a policy framework, designing a potentially effective programme and an implementation strategy require multiple capacities which may not always be available within government departments or other concerned agencies.

Technical and managerial support may need to be mobilized, more so during the planning and early implementation phase. If this support is not available, involvement of external talent – national or even international – may be required. Once the programme is implemented the external support could be phased out gradually as local capacity is built.

Forging partnerships between government ministries, donors, local non‐governmental organizations (NGOs) and universities was a feature of many successful programmes (Appendix 1; Van Roekel et al. 2002; Bhandari et al. 2005; Quinn et al. 2005). In the LINKAGES projects (Appendix 1), partnership with Programa de Coordinacion en Salud Integral (Collaborative Program for Integrated Health; PROCOSI) and NGOs in Bolivia, UNICEF, NGOs and the national government in Ghana, and Intersectoral Action Group for Nutrition and Jereo Salama Isika in Madagascar added value and quality to programmes, resulted in effective pooling of resources and expertise and improved programme design and management (LINKAGES 2002; Van Roekel et al. 2002; Quinn et al. 2005).

Programme design and strategy

This requires a systematic description of steps and processes that help achieve high coverage, quality, equity and sustainability; capacity building and programme management are particularly important features (The Core Group 2005). Experience so far suggests that for majority of developing countries, scale‐up for exclusive breastfeeding will not be possible unless intervention delivery at the community level receives emphasis (Appendix 1). As an increasing number of mothers deliver in institutions, mixed models will be required even in developing countries.

An initial step is to carefully examine the existing opportunities as vehicles for promoting exclusive breastfeeding. If use of existing opportunities is insufficient to achieve high coverage with good quality, then additional approaches would need to be evolved in a phased manner. In Bangladesh, the strategy of peer counsellors was used as the majority of deliveries occur at home and a community health worker is not routinely available in that setting (Haider et al. 2000). In India, on the other hand, village health workers already exist so programmes are often designed around them (2003, 2004, 2005).

Hospital/institution‐based strategies

Efforts to scale up the BFHI in countries have shown mixed results (Table 2).

Table 2.

Proportion of facilities with Baby‐Friendly Hospital Initiative (WABA 2007)

Region	Median (range)
West and Central Africa	0.13 (0.00–0.59)
Eastern/Southern Africa	0.23 (0.00–1.00)
Middle East and North Africa	0.14 (0.00–1.00)
South Asia	0.07 (0.01–1.00)
East Asia and Pacific	0.005 (0.00–0.87)
Americas and the Caribbean	0.10 (0.00–1.00)
Central and Eastern Europe/Commonwealth of Independent States	0.25 (0.04–0.93)
All Industrialized Countries	0.07 (0.00–0.97)

In regions where hospital deliveries are the norm, scale‐up efforts through the BFHI (WHO/UNICEF 1992; Table 3; Perez‐Escamilla 2007) have been successful and the strategy helped increase rates of exclusive breastfeeding (Appendix 1). However, in many countries, after initial years of enthusiasm, the scale‐up effort seems to have plateaued, perhaps reflecting changing emphasis of donors and government. The initial high rates of exclusive breastfeeding may not be sustained after mothers are discharged from facilities and return to their homes (Coutinho et al. 2005). The challenge, therefore, is to continue at home the post‐delivery support available in institutions through public and private providers to sustain exclusive breastfeeding or develop other mechanisms that support recently delivered mothers to exclusively breastfeed once they are discharged from hospitals (de Oliveira et al. 2006). Countries have addressed this in different ways. In Australia, a cadre of breastfeeding counsellors exists (Australian Breastfeeding Association 2005). In pilot programmes in India, exclusive breastfeeding was promoted through health worker home visits, at immunization clinics, growth monitoring sessions and at sick‐child contacts (Bhandari et al. 2005) in addition to the BFHI. Home visits are particularly important in regions where coverage through other channels is low or not feasible and a cadre of community workers exists.

Table 3.

The 10 steps to successful breastfeeding for hospitals and birth centres

• Maintain a written breastfeeding policy that is routinely communicated to all health‐care staff

• Train all health‐care staff in skills necessary to implement this policy

• Inform all pregnant women about the benefits and management of breastfeeding

• Help mothers initiate breastfeeding within 1 h of birth

• Show mothers how to breastfeed and how to maintain lactation, even if they are separated from their infants

• Give infants no food or drink other than breastmilk, unless medically indicated

• Practise ‘rooming in’– allow mothers and infants to remain together 24 hours a day

• Encourage unrestricted breastfeeding

• Give no pacifiers or artificial nipples to breastfeeding infants

• Foster the establishment of breastfeeding support groups and refer mothers to them on discharge from the hospital or clinic

Community‐based strategies

Pilot efforts in countries or regions where home deliveries predominate have shown success through vertical as well as integrated approaches; however, programmes integrated into ongoing activities are more likely to be more sustainable (The Core Group 2005). In these programmes (Appendix 1), the messages were delivered through either a single existing or new channel, or multiple delivery channels already in place within the primary health‐care system. The decision to use one or multiple channels was made based on the ability to cover the target population. In some countries, high coverage was feasible only with multiple, facility and community channels (Appendix 1). Multiple channels also help achieve consistency in advice. The credibility and acceptability of a given channel by the community was another factor in their being used. Routine activities of channels were modified in some instances to incorporate exclusive breastfeeding counselling. In India, for instance, it was believed that immunization sessions could not be used as a counselling opportunity but better organization and the presence of another available local worker during the session enabled the vaccinator to counsel mothers (Bhandari et al. 2005). Another illustration is the adaptation of the Integrated Management of Childhood Illness (IMCI) programme of the WHO/UNICEF (2002) to the Integrated Management of Neonatal and Childhood Illness (IMNCI; Government of India 2006) by India. The IMNCI includes three home visits for all births in the first 7 days of life by an existing community worker, the Anganwadi worker (ICDS 1982) to promote essential newborn care which includes exclusive breastfeeding. In other countries, such as Bangladesh, facility‐based IMCI was sought to be improved in its impact by greater emphasis on community component of IMCI, that is, C‐IMCI (Gwatkin 2004).

Programme delivery

Recognizing that facility‐based interventions are insufficient, countries have experimented with a variety of local workers such as peer counsellors in Bangladesh (Haider et al. 2000), lady health workers in Pakistan (Government of Pakistan 2003) and the ‘Accredited Social Health Activists’ (ASHA) worker in India (Government of India 2005) to promote exclusive breastfeeding by itself or as a part of essential newborn care. Programmes that used peer counsellors as a single channel have been shown to be effective in several efficacy (Green 1999; Lewin et al. 2006) and effectiveness trials (Morrow et al. 1999; Haider et al. 2000; Lewin et al. 2006). However, while peer or lay counsellors have been effective in the short run, long‐term sustainability of this approach on a large scale is as yet not established.

The possible reasons stated for the short‐term success of the lay health workers approach include payment to the counsellors for performing a specific piece of work, the counsellors’ own perception of benefits of their work and their recognition by the community as being skilled. These workers also felt this opportunity to have the potential for opening up of further education and training (Haider et al. 2002). An important issue would also be linking new workers with the existing primary health system as pilot experience showed that these workers may not be able to resolve complex breastfeeding problems (Haider et al. 2002).

Recruitment of ASHA in India provides a unique opportunity to promote exclusive breastfeeding (Government of India 2005). Currently being recruited all over the country, these workers will focus on essential newborn care including exclusive breastfeeding through home visitation in the first month of life and community‐based activities.

While these efforts in different countries at improving coverage of the target population through local workers are of value, the voluntary or semi‐voluntary nature of the workers and ill‐defined models of supervision raise issues about sustainability and possibly high attrition rates.

Programmes that used multiple channels reported broader coverage (Bhandari et al. 2005; Quinn et al. 2005). In India, formative research showed that immunization sessions and private providers were the channels to which infants were most commonly (∼80%) exposed. Following constitution of a breastfeeding‐promotion programme, caregivers most often received exclusive breastfeeding counselling at immunization sessions (∼40%), weighing sessions (∼25%) and home visits (25–30%), Government physicians or private providers rarely (∼2%) counselled despite being trained (Bhandari et al. 2005). When all the opportunities were considered together, 55–60% of caregivers had been counselled at least once, by at least one of the channels in the last 3 months, indicating that multiple channels reached more families (Bhandari et al. 2005). The frequency of contact with caregivers or the intensity of counselling influences the cumulative benefit of the counselling; for example, six contacts with mothers resulted in higher adoption rates for exclusive breastfeeding as compared with three (Morrow et al. 1999). In India, the number of channels at which caregivers were counselled was positively associated with the rates of exclusive breastfeeding at 3 months of age (Bhandari et al. 2003). However, the costs of training and the effort will increase more than the number of channels used and, therefore, this issue requires a balanced view.

Another issue of interest is the value of counselling for the antenatal period itself. Antenatal counselling has been used in several programmes but its effect is difficult to delineate by itself as it was usually followed by post‐natal counselling (Green 1999). In the HIV context, however, the importance of antenatal counselling is obvious.

Communications approach

An effective communications approach was a feature of most programmes, and involvement of journalists, radio announcements and other mass media from the programme outset extended the reach of nutrition messages even further (Appendix 1; Quinn et al. 2005). Mass media were not used in some programmes where a comparison group was present for risk of contamination (Bhandari et al. 2005). Utilization of existing or newly formed community groups produced the same effect (Bhandari et al. 2005; Quinn et al. 2005), and community mobilization programmes, such as street plays, rallies, radio programmes and songs, helped disseminate messages and increase community involvement (Bhandari et al. 2005; Quinn et al. 2005).

Effective counselling requires assessment of feeding practices, recognition of problems and negotiation with the mother in choosing improved practices that are feasible (WHO/UNICEF 2002); this requires time and patience. Programmes may choose one‐on‐one counselling during opportunities in which more time is available such as home visits, and group counselling for opportunities with lesser available time (Bhandari et al. 2005). One of the reasons that peer counsellors may have been effective was that their sole responsibility was counselling mothers through home visits (Morrow et al. 1999; Haider et al. 2000). The more diverse the functions of the local worker, the lesser will be the emphasis on a single component such as exclusive breastfeeding.

Message development

The development of messages for interpersonal communication and for mass media needs careful attention to design and content. Successful programmes included an element of formative research to gain an in‐depth understanding of the communities where these programmes were implemented with regard to practices pertaining to breastfeeding, the reasons for the current practices, the common breastfeeding problems and the barriers to adoption of ideal behaviours in that cultural setting. Formative research in north India, for instance, showed general unwillingness to offer water to infants in summer months. This barrier was overcome by building specific explanatory messages using local terms within the communication strategy (Bhandari et al. 2003).

Once the design features and broad strategy are in place, attention must then shift to the operationalization of the strategy.

Training and supervision

Prior to scale‐up, the training needs will need to be defined. Fortunately for promotion of exclusive breastfeeding, extensive experience is available on training of different cadres of health workers and even for lay counsellors. Most experiences used the adapted versions of the UNICEF/WHO Breastfeeding Counseling Course (WHO/UNICEF 1993) or the IMCI module (WHO/UNICEF 2002).

Training strategy and its quality merit attention. In successful programmes, the duration of training extended from 3 full days to 10 part days for different workers. The training strategies, as described, were ‘concise, action oriented’ and included modules that aimed to equip trainers with practical skills and confidence to counsel effectively (Appendix 1).

Training modules for areas with high HIV prevalence also exist (WHO 2006).

While the training modules used in studies worked well, a gap remains for referral. Trained workers at the community level require that referral options available are easily accessible, prompt and preferably available free of cost to mothers with complex breastfeeding problems.

Another important issue to address is handling training of new staff; in most developing countries, staff turnover and absenteeism rates are high.

Monitoring and evaluation

The programme strategy will need to include systems for monitoring and evaluation of the programme. Possible markers could include the rates of exclusive breastfeeding, worker performance, referral and action. As for all other national programmes, there will be a need for periodic reviews and achievable targets by time.

Costs and resources

In all pilot programmes and projects, services were provided to mothers free of cost. In many developing countries if the government cadre of health workers is used to promote exclusive breastfeeding as a part of their other responsibilities (e.g. the IMNCI) the basic services will remain free. However, a system that requires setting up a cadre of trained lactation counsellors (e.g. in an urban setting or as referral sources for breastfeeding problems in rural areas) will require additional resources. A cost‐effective analysis of the BFHI was conducted in three countries: Brazil, Mexico and Honduras (Horton et al. 1996). Savings on health care resulting from reduction in diarrhoeal and respiratory morbidity were estimated as a result of the effect of BFHI on breastfeeding. The study showed that breastfeeding promotion through a BFHI‐like model is very cost‐effective (1992 US$ 2–19 per disability adjusted life years; Horton et al. 1996; Perez‐Escamilla 2007). Examples of cost‐effectiveness analysis of community‐based programmes using lay or professional support in different settings could not be identified.

Testing of novel approaches and continuing innovation

Scaled‐up programmes must have inbuilt provision to test novel approaches and the ability for continuing innovation.

Scaling up of exclusive breastfeeding in the context of HIV

The policy and programme design and implementation issues relevant to exclusive breastfeeding scale‐up discussed above are relevant to HIV‐prevalent areas, but there are other unique features that merit consideration.

Creating policies for exclusive breastfeeding promotion in countries with high HIV prevalence is particularly challenging. The revised recommendations encourage exclusive breastfeeding for the first 6 months and continued breastfeeding for 2 years and beyond for infants whose mothers are HIV negative or whose HIV status is unknown. For women who are known to be HIV positive, the most appropriate infant feeding option for HIV‐exposed infants depends on individual circumstances including consideration of health services, counselling and support. Exclusive breastfeeding is the better option for such infants unless replacement feeding is acceptable, feasible, affordable, sustainable and safe for both the mother and baby. This recommendation is based on the finding that exclusive breastfeeding for up to 6 months in three large cohort studies is associated with a three‐ to fourfold decreased risk of transmission of HIV compared with non‐exclusive breastfeeding (WHO 2006). The recommendations may need further revision if new findings demonstrate that exclusive breastfeeding should be the feeding of choice for HIV‐infected women (Iliff et al. 2005).

Some pilot programmes for exclusive breastfeeding promotion in high‐HIV‐prevalence countries are summarized in Appendix 2.

In general, exclusive breastfeeding promotion formed a part of the prevention of mother‐to‐child transmission programmes. LINKAGES/Zambia integrated improved counselling on infant feeding, maternal nutrition and antiretroviral therapy during prophylaxis in both health‐care and community services to enable women to make and act on informed choice to feed their infants optimally in the context of high HIV prevalence.

Formative research was critical to develop locally appropriate and feasible infant feeding recommendations for individual infants and their families based on HIV status. In Zambia, the research included focused group discussions, key information interviews, observations of household food preparation and feeding, market survey of replacement foods and breastmilk substitutes and household trials of improved feeding and caring practices (LINKAGES 2005).

High quality of health workers is another critical factor for sensitive and accurate individual counselling. Programmes incorporated behaviour change communication methodologies in training (LINKAGES 2005). It is important to develop skills of community members as well, for improved services, referral, outreach and follow‐up care and support of HIV‐positive mothers and their infants.

A comprehensive communications strategy to support scale‐up was a notable feature of the LINKAGES programme in Zambia. In 2004, an intensive effort was launched using brochures, posters, radio and television spots, billboards and reference guides for health providers (LINKAGES 2005). Messages were targeted to relevant stakeholders – pregnant and lactating mothers, health providers, youth and community leaders, and all the materials were nationally disseminated.

Monitoring and evaluation, another important feature of effective scale‐up, was a feature of pilot programmes (Appendix 2, Ntabaye & Lusiola 2004; Rutenberg et al. 2004) in Zambia and Tanzania. Monitoring and evaluation not only provides a measure of success, but identifies factors that may have contributed to it and to lack of success to enable mid‐course corrections. The impact of some pilot programmes on exclusive breastfeeding is summarized in Appendix 2. In most programmes, exclusive breastfeeding rates improved. In Zambia, exclusive breastfeeding rates increased from 50% in 2002 to 64% in 2004 in Livingstone, from 57% in 2000 to 74% in 2004 in Ndola. In South Africa, an antenatal clinic‐based intervention led to 83% of infants born to HIV‐infected mother initiating exclusive breastfeeding. In Zimbabwe, an intervention based on antenatal counselling resulted in a relatively lower exclusive breastfeeding rate of 24.6% at 3 months (Piwoz et al. 2005).

Several barriers to success were identified. These include inadequate skills of workers and non‐integration of HIV counselling and tailoring into antenatal care services (Ntabaye & Lusiola 2004) which is the backbone in which infant feeding counselling is built. Women reported that mothers not breastfeeding their children were discriminated against by the community and they faced the threat of domestic violence. Programme manager's understanding of barriers to effective programme implementation and impact were emphasized as being important. Government support, partner's coordination, increased human resource and improving the supply system were identified to be essential features to scale up programmes in the HIV context (Rutenberg et al. 2004). Low overall exclusive breastfeeding rates prompted Piwoz et al. (2005) to conclude that women and their partners should have been reached more frequently during the antenatal and post‐natal period.

Conclusions

In a large number of programmes for improving exclusive breastfeeding in developing countries, including in areas with high prevalence of HIV, a positive impact on exclusive breastfeeding rates has been achieved.

Some of the factors that contributed to successful scale‐up are summarized in Table 4. Several limitations of the programmes reviewed merit attention. Most programmes involved special and generous funding, concentrated efforts and multiple contacts. These programmes were evaluated over relatively short periods of time, usually 2–5 years, and the long‐term sustainability has not been really tested. In the absence of specific measures, there is likely to be dilution of quality, as the programmes are scaled up further from their initial size. The experience thus far leaves important questions to be addressed, particularly on how programmes can be sustained over time as they are scaled up at affordable costs.

Table 4.

Lessons learnt about scale‐up; key factors in success

• Political will, advocacy, enabling policies

• Championship by health ministries at national, state and district level

• Technical consensus on feeding guidelines

• Programme strategy and measurable short‐ and long‐term goals

• Long‐term financial commitment

• Partners with defined roles, assessment of their capacity to deliver and plans to cover gaps in capacity

• Emphasis on community‐level interventions, and not only facility‐based programmes

• Formative research as the backbone of effective scale‐up programmes, particularly in the context of HIV

• Careful selection of channels and the number required to achieve timely coverage, consistency of messages

• Communication strategy aided by formative research, pre‐tested, validated messages and tools

• Quality trainers and training centres proportionate to required scale

• Monitoring and evaluation

• Programme redesign when relevant through research, analysis and innovation

A key factor relevant to sustainability is to plan on a long‐term basis, a period of 10–20 years, rather than 5–10 years, from the outset. Experience with other child health programmes has shown that performance in immunization and oral rehydration therapy use rates plateaued over time in many countries.

Programme scale‐up requires a programme management capacity proportionate to the scale of operations, periodic assessment of barriers (Table 5) and an ability to revise programme strategy to overcome these. Another key feature to sustainability of scale‐up is the extent to which capacity has been built at the district, block and community levels. National‐ or state‐level driven interventions begin with enormous enthusiasm which is difficult to sustain. This factor is particularly important as most developing countries are going through a process of decentralization with increasing responsibility for planning, implementation and evaluation being delegated to lower levels of the health‐care system. Building capacity at that level requires time, resources and patience and a longer‐term perspective.

Table 5.

Barriers to scaling up of health interventions

Community and household levels

Lack of information, women's education, physical, financial, women's decision‐making power

Health services delivery level

Shortage of qualified staff, weak technical guidance, programme management and supervision, inadequate supplies, equipment and infrastructure, poor accessibility

Health sector policy and strategic management level

Weak and centralized system for planning and management, weak drug policies and supply system, inadequate regulations, lack of intersectoral action and partnership, weak incentives to use inputs efficiently and respond to user needs and preferences, reliance on donor funding

Public policies cutting across sectors

Government bureaucracy, poor communication and transport infrastructure

Environmental characteristics

Governance and overall policy framework

Corruption, weak government, weak laws, political instability and insecurity, low priority to social sectors, weak structure for public accountability and opportunities for public opinions, lack of free press

Physical environment

Climatic and geographic predisposition, physical environment unfavourable to service delivery

Adapted from Ranson et al. (2003).

Building capacity emerges to be a key factor. Trainers and training centres vary in their quality when training health workers and their supervisors or programme managers. As the burden of training increases, availability of good‐quality trainers becomes a limiting factor (WHO/UNICEF 2002; Government of India 2006). An effective strategy must plan for these incremental requirements including ensuring secure and sustained funding.

Identifying barriers to success is a key actor. The barriers to scaling up of health interventions have been elegantly described by Ranson et  al (Table 5). Most of these are important for exclusive breastfeeding scale‐up as well.

Overall, within the limitations discussed above, a number of programmes in countries with varying HIV prevalence show that exclusive breastfeeding on a large scale is feasible and pilot programmes have given valuable guidance related to policies, planning and implementation of large‐scale programmes. The challenge now is to build on these experiences when planning national programmes.

Appendix 1

Case studies of effectiveness studies or pilot programmes

Study description	Population	Type	Intervention	Partnerships	Communication strategy	Training	Impact on exclusive breastfeeding	Comments
Lutter et al. (1997)	Brazil. Low‐income urban women delivering in hospitals	Breastfeeding promotion in hospitals compared with no breastfeeding promotion	Promotion of breastfeeding in selected hospitals				Mean days of exclusive breastfeeding 75 days vs. 22 days (P < 0.01)
Rea & Berquo (1990)	Women with a child aged <2 years in Sao Paulo	National breastfeeding campaign, 1981–1986, BFHI	Mass media campaign				11% infants 5–6 months exclusively breastfed in 1988 and 37% in 1987	Confounding factors over the 6‐year period not measured
2000, 2001) PROBIT trial	Belarus	Cluster randomized, 34 hospitals and their polyclinics	Training in BFHI (WHO/UNICEF 1992) vs. routine practices and policies				Exclusive breastfeeding at 3 (43.3% vs. 6.4%) and 6 (7.9% vs. 0.6%) months significantly higher	Implemented through BFHI in a setting where hospital deliveries are common
Perez‐Escamilla (2004) BASICS II	Mexico, Honduras, Brazil		MADLAC low‐cost management information system to use an evidence‐based approach to improve service performance Interviewing of women at hospital discharge post delivery using a questionnaire that takes 5–7 min to administer and contains 19 breastfeeding promotion/counselling indicators Data entered and results discussed 3‐monthly by hospital committee				Hospital implements changes in breastfeeding support based on analysis of questionnaires	It is a cost‐effective intervention that can help improvements in the BFHI and is a valuable tool for process evaluation
LINKAGES project in Africa and Latin America (Quinn et al. 2005; LINKAGES 2002;WHO 2003a)	Bolivia (1 million), Ghana (3.5 million), Madagascar (6 million). Resource‐poor setting	Pilot programme. In Bolivia and Ghana programme focused on breastfeeding and complementary feeding. In Madagascar, programme promoted breastfeeding within 7 essential nutrition actions and subsequently through IMCI	Built on formative research and sustained through programme activities that focused on bringing about the desired behaviour change. Mix of interventions reached women individually, or in groups in health facilities, homes or community settings Consistent messages and material across all programme channels	Bolivi: PROCOSI NGOs Ghana: UNICEF national nutrition office, NGOs Madagascar: Intersectoral Action Group for Nutrition Jereo Salama Isika Program designed with partners	Radio in all three sites	Materials developed to suit needs of local health workers and community members. Great emphasis on development of counselling and negotiation skills practical sessions	Significant increases in exclusive breastfeeding rates of 0–6‐month‐old infants; from 54% to 65% in Bolivia, 68% to 79% in Ghana and 46% to 68% in Madagascar. In Ghana and Madagascar significant results observed within a year of initiation of community interventions	Approach was effective and achieved significant increases among large populations as early as 9–12 mo after initiation of community activities
Van Roekel et al. (2002); WHO (2003a)	Honduras	Growth monitoring and promotion known as Atención Integral a la Niñez (AIN) or Integrated Child Care Initially, primarily a growth monitoring programme ultimately through IMCI	Programme on growth promotion – assessment of monthly weight gain and focus on illness, feeding practices and general child care in infants under 2 years. Implemented at government facilities and community. Started in 1992, strengthened in 1995–1997 with greater focus on child feeding. In 1997, programme introduced with IMCI (WHO/UNICEF 2002)	Major partners: UNICEF, PAHO, American and Honduran Red Cross, CARE	Radio broadcasts, religious leaders	IMCI	Mid‐term evaluation showed 39% of mothers in intervention communities exclusively breastfed compared with 13% in the control sites
Bhandari et al. (2003);WHO (2003a)	Rural communities in Haryana, India	Cluster randomized trial, 8 communities (total population ∼40 000)	Promotion of exclusive breastfeeding and appropriate complementary feeding Developed through formative research – Problem identification – Participatory design of intervention implementation and evaluation Implementation within an existing primary health‐care programme Opportunities used to deliver messages were immunization clinics, weighing sessions, sick‐child contacts and monthly home visits	Researchers with local government, ICDS system and NGO	Theatre, songs. Mass media not used for risk of contamination. Materials developed for health workers; posters, flip books, cards	Breastfeeding counselling section of the IMCI (WHO/UNICEF 2002)	Exclusive breastfeeding in the previous 24 h at 3 months of age 79% vs. 48%. The positive effect remained up to 6 months of age (42% vs. 4%)	Total participation of the district health system from inception of the programme, in developing implementation strategy, tool development and monitoring in addition to involvement of an NGO were important features that contributed to the success of the intervention
Morrow et al. (1999)	Mexico City	Cluster randomized trial. Total population 30 000. Two communities in the intervention, one comparison. 130 women	Home‐based peer counselling Peer counsellors made 3 or 6 home visits to the mother Key family members also included			By La Leche League of Mexico	Exclusive breastfeeding at 3 months post‐partum 67% (6 visits), 50% (3 visits), 12% (comparison)	First cluster randomized community‐based trial that used peer counsellors to promote exclusive breastfeeding
2000, 2002	Dhaka. Low middle and low‐socio‐economic population	Cluster randomized trial. 20 zones in intervention and 20 in control	Community‐based peer counselling in Dhaka 15 home‐based counselling visits by peer counsellors			40 h (4 h × 10 days). WHO/UNICEF breastfeeding counselling course (WHO/UNICEF 1993)	Exclusive breastfeeding at 5 months 70% in intervention and 6% in the control	Community‐based breastfeeding promotion strategy was effective in Bangladesh where most deliveries occur at home
Bryce et al. (2004); WHO (2003b)	Brazil, Peru, Tanzania, Uganda	Multicountry evaluation of IMCI that included effectiveness, cost and impact (Bryce et al. 2004)	Breastfeeding promotion as a part of IMCI IMCI is a strategy for improving child health and development through the combined delivery of essential child health interventions. IMCI began with a set of case management guidelines for the management of sick children at first‐level health facilities and the guidelines were designed for adaptation at country level and below. Over time the strategy expanded to include a range of guidelines for delivering child survival intervention at household, community and referral levels			IMCI module (WHO/UNICEF 2002)	Results on exclusive breastfeeding available from Tanzania where no significant impact was seen	Despite substantial improvement in the quality of care delivered by IMCI at health facilities, key family practices such as exclusive breastfeeding did not change as community IMCI had not been implemented by that time (WHO 2003b)
S. Arifeen, personal communication	Bangladesh, Matlab	Community‐based cluster design	Evaluation of facility‐ and community‐based IMCI. Intervention package includes feeding assessment and counselling on mothers of young children seeking care at facilities for illness and home‐based counselling on improved feeding by community‐based health and nutrition workers				At baseline in the year 2000, a little more than half the 0–5‐month‐old infants in the study area were being exclusively breastfed. By 2005, this increased by 30% in the IMCI area. Data not available from the comparison area as yet	Lack of comparable data from the comparison area prevents a definitive conclusion as to whether this truly is an impact of IMCI

BASICS II, Basic Support for Institutionalizing Child Survival Project; BFHI, Baby‐Friendly Hospital Initiative; ICDS, Integrated Child Development Services; IMCI, Integrated Management of Childhood Illness; MADLAC, Monitoreo de Apoyo Directo con la Lacatancia Materna; NGO, non‐governmental organization; PAHO, Pan American Health Organization; PROBIT, Promotion of Breastfeeding Intervention Trial; PROCOSI, Programa de Coordinacion en Salud Integral (Collaborative Program for Integrated Health); UNICEF, United Nations Children Funds; WHO, World Health Organization.

Appendix 2

Case studies in scale‐up efforts in HIV/AIDS

Study description	Population	Intervention	Communication strategy	Impact on exclusive breastfeeding	Comments
LINKAGES (2005)	Zambia	Programme through health‐care and community services Reached 60 sites in 6 districts by 2005 Integration of infant feeding counselling, maternal nutrition and antiretroviral drugs prophylaxis Training of health workers and community service providers to counsel women on safety options for their situation	Behaviour change. Communication interventions in the first few years Material developed ‘Act Now’ Media campaigns included brochures, posters, radio and television	Assessed through baseline and follow‐up household and clinic surveys in selected districts Exclusive breastfeeding rate increased from 50% in 2002 to 64% in 2004 in Livingstone and from 57% in 2000 to 74% in 2004 in Ndola	Integrated demonstration project. Had several strengths in terms of the use of formative research to develop messages, training and building up skills of programme managers, health‐care providers and community workers, advocacy for national policy, strong behaviour change communication component, strengthening community capacity for counselling and referrals and monitoring and evaluation with local partners
Piwoz et al. (2005) ZVITAMBO trial	Harare, Zimbabwe	Education and counselling programme delivered antenatally 14 110 mother–baby pairs enrolled from 14 maternity clinics within 96 h of delivery and followed up at 6 weeks, 3 months, and 3‐monthly intervals Infant‐feeding patterns determined Exclusive breastfeeding recommended for HIV negative, unknown or HIV positive who chose to breastfeed	Group education, individual counselling, videos, brochures	Mothers who enrolled when programme was fully implemented were 8.4 times more likely to exclusively breastfeed than mothers who enrolled before the programme began Overall exclusive breastfeeding rates till 3 months was low at 24.6%	Formative research used to design a culturally sensitive education and counselling programme Low overall exclusive breastfeeding rates prompted authors to conclude that women and their partners should have been reached more frequently during antenatal and post‐natal period
Thior et al. (2006); Rollins (2007); Cohen (2007)	HIV‐infected pregnant women attending antenatal clinics in south Botswana Investigations by CDC of a diarrhoea outbreak in January‐February 2006 in the same population	Maternal and infant antiretroviral therapy to ascertain impact on HIV transmission	Antenatal counselling	Case–control study to identify risk factors for diarrhoea; not breastfeeding was the biggest risk factor, adjusted odds ratio 50 (95% confidence interval 4.5–100)	Botswana has the best primary health‐care structure, water supply and sanitation systems in South Africa. Formula provided free by the government to infected mothers Lower mortality in the breastfeeding group at 7 months demonstrates risk of formula feeding and need for alternate strategies (Thior et al. 2006) Authors conclude that health systems in which these programmes are implemented should be able to cope with diarrhoea and malnutrition that results because of formula feeding (Rollins 2007)
Coovadia et al. (2007)	KwaZuluNatal, South Africa	2722 HIV‐infected and uninfected pregnant women enrolled into a non‐randomized intervention cohort study from antenatal clinics Infant feeding data obtained weekly, blood samples from infants taken monthly		83% of 1372 infants born to HIV‐infected mothers initiated exclusive breastfeeding from birth Breastfed infants who received solids were significantly more likely to acquire infection compared with exclusively breastfed (HR 10.87, 1.51–78, P = 0.018) as were infants who at 12 weeks received breast and formula milk (1.82, 0.98–3.36, P = 0.57) Cumulative 3‐month mortality in exclusively breastfed was 6.1% (4.74–7.92) vs. 15.1% (7.63–28.73) in infants given replacement feeds (HR 2.06, 1.0–4.27, P = 0.05)	The study shows that exclusive breastfeeding can be successfully supported in HIV‐infected women The authors call for further revision of UNICEF/WHO and USAID infant feeding guidelines

CDC, Center of Disease Control; HR, hazards ratio; UNICEF, United Nations Children Funds; USAID, US Agency for International Development; WHO, World Health Organization; ZVITAMBO, Zimbabwe Vitamin A for Mothers and Babies.