The objective of the project is to develop and test an intervention to promote exclusive breastfeeding (EBF), to assess its impact on infant health in African contexts where a high prevalence of HIV is a barrier, and to strengthen the evidence base regarding the optimal duration for EBF.

Promotion of EBF is the most effective child health intervention currently feasible for implementation at the population level in low-income countries. It can lower infant mortality by 13%, and by an additional 2% were it not for the fact that breastfeeding transmits HIV. Only recently proven to be possible in hot and even dry climates, EBF without even offering water is still little appreciated by mothers or supported by health workers. EBF rates are especially low in Africa but the potential for rapid implementation may be high. A few studies elsewhere suggest that peer counselling can often achieve dramatic increases. Thus the investigators will run the first randomised trial to develop and test models for applying this approach in four African countries and to quantify health benefits, cost-effectiveness, and implications for the health care system.

While experts realize that the HIV threat ought not to present much of a biological constraint to promoting EBF, in heavily affected countries it does represent a cultural constraint. Overcoming this will require the development of a safe and effective means of promoting EBF that is HIV-sensitive by taking into account the need to minimise postnatal HIV transmission.

Another scientific constraint to the promotion of exclusive breastfeeding for six months, as recommended by the World Health Organization (WHO), is uncertainty about its impact on the micronutrient status of infants. In a substudy, the investigators will carefully follow markers of infant micronutrient status to see how they vary by feeding pattern, including EBF, for a longer period than has been examined previously.

Rationale

The single most effective way of saving the lives of millions of young children in developing countries would be the promotion of exclusive breastfeeding (EBF). Over a period of 10 years it could save the lives of an estimated 15 million children.

"Infants aged 0-5 months who are not breastfed have seven-fold and five-fold increased risks of death from diarrhoea and pneumonia respectively, compared with infants who are exclusively breastfed. At the same age, non-exclusive rather than exclusive breastfeeding results in more than two-fold increased risks of dying from diarrhoea and pneumonia." Lancet 2003; 361: 2226-34

This assertion, and related comments and findings, set out by the Bellagio Child Survival Group in a recent article in the Lancet summarising findings from international research of how best to reduce infant mortality in developing countries, is made despite our limited knowledge of how best to promote EBF. This proposal addresses the urgent need to tackle this limitation. How best to promote EBF needs attention because it will be necessary to secure high rates of support for EBF from young mothers, and we do not currently know the best way to bring this about, particularly in the cultural context of Africa. Unless we can find more effective strategies to combat poor feeding practices in both health facility and community settings, the child mortality rate will continue to be high.

Breastfeeding

The benefits of breastfeeding and the negative effects of artificial feeding in underprivileged environments were clearly brought to the public attention in the 1970s, backed by increasingly strong scientific support. As a result the International Code of Marketing of Breast-milk Substitutes and subsequent World Health Assembly Resolutions were created in the 1980s to avoid the marketing of infant foods, teats and bottles in ways that could interfere with breastfeeding. In the early 1990s, UNICEF started the "Baby-friendly Hospital Initiative" (BFHI), which decreased many of the practices in the health care system that had a negative impact on breastfeeding. Together with many of the other child health programmes during these decades these interventions contributed to a reduction in child morbidity and mortality.

Exclusive Breastfeeding - An Even Better Version of Breastfeeding

EBF means that the infant receives nothing else other than breast milk, not even water. That breastfed children do not need anything besides breast milk, not even water, and not even in hot climates, was discovered in the 1970s. A decade of research confirmed this, followed by an attempt to define breastfeeding into various patterns by WHO.

EBF is not a traditional concept; beliefs and practices regarding a wide range of early supplements are common around the world, and often seem to reflect beliefs expressed by early physicians. Nor is it widely understood by health care professionals, given the lack of information on it even in most paediatric basic training.

So if EBF is able to lower child mortality by 13%, why is it not promoted on a large scale already today? Two major reasons are: fear of HIV transmission through breast milk and uncertainty about the micronutrient status in children exclusively breastfed for 6 months.

Breastfeeding and Risk of HIV Transmission

When an HIV-positive mother breastfeeds for 1.5 years or more, this appears to lead to an additional 15% risk of postnatal HIV transmission on average. At a population level the epidemiological evidence is that exclusive breastfeeding should still be promoted even in countries with a high prevalence of HIV. For many HIV positive mothers the risk of transmitting HIV to their children is outweighed by the risk of the infant dying if not breastfed because of the poor socioeconomic and environmental conditions they live in. Limited evidence suggests that for these women and their children EBF is even more important, not just for minimising respiratory and gastrointestinal infections but also perhaps in reducing the risk of HIV transmission. However in reality the HIV epidemic has undermined breastfeeding promotion efforts especially in countries that would benefit the most. In addition, expensive approaches of dealing with HIV have removed the focus from breastfeeding as an inexpensive way of promoting optimal child health.

Intervention: Peer Counselling

The constraints discussed above have meant that efforts to promote EBF have been rudimentary so far, consisting mainly of the Baby-Friendly Hospital Initiative. But in Africa nearly all such efforts have come almost to a complete standstill in recent years. EBF promotion therefore needs to be reshaped in this era of HIV. The challenge is to urgently put back on track the promotion of exclusive breastfeeding for the benefit of the world's children without losing credibility by doing it indiscriminately. PROMISE will fine-tune and test an upgraded version of EBF-promotion that is HIV-sensitive and do it in four African countries with a high prevalence of HIV where it is most needed. In particular, it will be the first study to deal explicitly with each of these constraints, opening the way for a more widespread and successful promotion of EBF, especially in Africa, ultimately giving back to Africa one of the powerful and cheap interventions to prevent child mortality and morbidity.

Peer counselling is a proven cost-effective approach for changing behaviour, for example, preventing HIV infection. Several studies have also examined its impact on breastfeeding behaviour. Based largely on quasi-experimental studies peer counselling was shown to have an impact on the duration of any breastfeeding.

Potential Impact

This study will provide the first data from a randomised trial on the safety and efficacy of peer counselling as a way of increasing rates of exclusive breastfeeding in the African context. Outcomes include morbidity, growth, and micronutrient status.

PROMISE EBF will provide an example to the world of how efforts to promote exclusive breastfeeding and avoid postnatal HIV transmission can be done jointly, and what synergisms may result. Methods employed and costs will be documented and reported on to increase the ease of replication and sustainability.

In Africa, there is a growing awareness that strategies must be identified and tested that can lead to the highest possible infant survival rate in the face of the risk of postnatal HIV transmission. In 1997 UNAIDS, WHO and UNICEF changed their policy on HIV and infant feeding, and issued guidelines for implementing this new policy a year later. Whereas previously countries where most infant mortality was due to infection and malnutrition were advised to promote breastfeeding irrespective of the mother's HIV status, the new policy called for individual counselling on the advantages and disadvantages of breastfeeding vs. artificial feeding that would enable mothers to make an informed choice. UNICEF provided free formula at a number of pilot sites.

Methodology

Design

A prospective, multi-center, community-randomised, controlled, experimental design will be used to execute this trial.

Randomization

A number of over 30 randomisation units should be identified within a reasonable distance from the study office. The randomisation unit varies between sites but will typically be 1-2 villages with an average of 1000 inhabitants (about 35 infants born per year given a birth rate of 3%). Before engaging in the selection of communities, extensive information will be given to community leaders and the rest of the selection process will be done in close collaboration with community leaders. Within the district the suitable communities will be selected and groups considered similar, based on location, urban-rural, and socioeconomic status will be stratified. Care will be taken to allow for 'corridors' between selected communities to be randomised. In each stratum half of the communities will be randomised to intervention and the reminder to control communities. Local community leaders will be further informed about the study.

Each of the clusters should be visited and checked for size, for collaborative administrative leaders, and for not sharing important facilities (water supply, market/shop/trading centre, primary school, mosque, church) with any other potential cluster. If both urban and rural clusters are included no less than 30% of each should be included. The complete list of randomisable clusters will be checked and 24 clusters will be randomised in each site, if applicable in the rural/urban strata.

The study will be carried out in four African countries, Burkina Faso in French-speaking West Africa, Uganda in East Africa, Zambia in Central Africa and South Africa in Southern Africa.

In each country one or more districts have been selected as the intervention sites, Banfora in the South-west of Burkina-Faso, Mbale District in Eastern Uganda, Mazabuka in Southern Province in Zambia and three sites in South Africa: Umlazi in Durban, KwaZulu Natal, Rietvlei in the Eastern Cape and Paarl in the Western Cape. Pregnant women in the randomised communities (= clusters) will be informed about the study and asked if they are willing to participate. If they agree, an enrolment interview will be carried out and based on eligibility criteria the women will or will not be accepted into the study.

Sample Size and Power

Of the primary outcome measures the decrease in diarrhoea is the most difficult to catch. We have based our sample size calculation on the expected reduction in cases of diarrhoea (at the age of 3 months) over the last two weeks of one third from 12% to 8%. With a 95% confidence interval (alpha error 0.05) and power of 80%, an average number of infants of 35 per randomised community, and a coefficient of variation between the communities of 0.3, we will need to randomise 48 communities in each arm - a total of 96 communities. This will include 420 infants in each arm in each country, a total of 840 children per country and all in all 3360 children studied across the four sites. This sample size will very accurately give us the increase in EBF and be able to document the above decrease in diarrhoea morbidity.

Recruitment Procedure

The recruitment procedure will be adapted to the local circumstances. In the most common case, each cluster will have a recruiter ("pregnancy monitor") who will be a mature woman of good repute living within the cluster. The pregnancy monitors will inform the research centre about pregnant women that may be eligible to participate in the study. A data collector/research assistant will be sent to visit the woman to ascertain eligibility, obtain consent to participate in the study and conduct the recruitment interview.

In South Africa, the recruitment of women will be done at the health facility in the antenatal clinic, where a data collector/research assistant will interview the woman to ascertain eligibility, obtain consent to participate in the study and conduct the recruitment interview.

Inclusion Criteria:

• Lives in a selected cluster
• Is >= 6 months pregnant
• Has no plans to move outside of the cluster within 1 year

Exclusion Criteria:

• Reduced ability to collaborate for psychological/mental reasons
• Severely ill
• Has given birth more than 7 days ago
• Has given birth less than 7 days ago and it was a multiple birth; or the child has a severe malformation; or death of the mother or the baby.

Trial not entirely blinded.