The expert panel consisted of 15 participants who completed the e-survey out of the twenty-five who consented (representing a 60% response rate). The expert panellists were from five SSA countries (Nigeria, South Africa, Malawi, Kenya and Namibia) and a variety of professional groups including obstetricians (1), paediatricians (3), nutritionists (5), and HIV program and treatment officers (3) and university academics (3). Twenty-five of the 57 invited to participate (43 purposive sample and 14 through snowballing), consented to participate, but only 15 completed the Delphi process, out of which four (2 nutritionists, 1 HIV program and treatment officer and 1 university academics) where recruited through snowballing.
Consensus agreement on the AFASS tool for feeding options
Consensus (>75% agreement = 12 or more) was reached for the questions enquiring whether the tool as a whole is appropriate, applicable and useful for SSA settings with 13, 12 and 12 experts agreeing respectively.
Table 1 shows the questions in the tool, number of experts agreeing and the percentage agreement obtained for each question. Three questions did not achieve consensus but one of these was still considered suitable for inclusion in the final tool based on synthesis of comments and review of current evidence. For all other questions, consensus was reached. Only questions where consensus was not reached or suggestions for improvement were made are discussed.
Only 11 of the experts agreed with the question in the section quantifying the increase in risk of HIV transmission associated. Twelve experts indicated that the risk of transmission associated with mixed feed is higher than the three-fold stated in the question.
In the acceptability section, seven experts disagreed with the appropriateness of the language used for the question assessing the ability of the mother to carry on with her choice of feeding. The majority (13) of the experts’ said the language was not appropriate as the question would be prone to misinterpretation, hence rephrasing before including the question in the tool was suggested. It should be noted that the tool was developed in English and it is clear that using the correct word is very important. This is something that will have to be taken into consideration when translating the tool into other languages as often a direct translation is not the best translation.
In the feasibility section, only 7 experts agreed with the question which enquired about the availability of the mother to be home for six months. The experts’ felt that a mother does not necessarily have to always be home with the baby before exclusive replacement feeding can be practised efficiently. It was argued that the question is more in support of exclusive breastfeeding, and even at that, six months is not feasible. It was further argued that a mother can express breast milk even if she is practising exclusive breastfeeding.
Other aspects of disagreement in the feasibility section related to the language or wording of the questions. The appropriateness of the language used to enquire whether the mother can prepare formula feed and feed her baby every 2–4 h, day and night for the first three months was supported by 11 experts. Most experts (13) suggested changing the three months to six months. They also pointed out the fact that the frequency of feeding will depend on the age of the baby. Rephrasing this question was suggested before including it into the final tool.
The appropriateness of the language used for the question assessing the availability of constant supply of electricity was supported by 11 experts. Summary of the experts’ opinion revealed misinterpretation of the question as 13 experts interpreted availability of constant electricity in relation to supply of cooking fuel for preparing formula feeds and suggested other local means of cooking fuel. This question had however been considered for inclusion in relation to preserving already prepared formula feed, indicating that rephrasing the question to make it clearer was needed.
The language used for the question assessing whether the mother would be able to buy formula milk any time it is required within her neighbourhood was supported by ten experts. Experts argued that mothers’ can obtain formula milk by other means and not necessarily within the neighbourhood. Hence, rephrasing the question to whether the mother can ensure a constant supply of formula feed at any time was suggested by 14 experts.
The appropriateness of the language used in enquiring about the availability of someone giving assistance with the care of the baby was supported by 12 experts. However, 11 experts suggested that the question should be expanded or made simpler by giving some examples of what is meant by ‘assisting in the care of the baby’, e.g. feeding, bathing, dressing and soothing and that emphasis should be placed on availability of assistance with the feeding of the baby.
Both questions in the affordability section, and the appropriateness of the language used in describing the questions achieved consensus, but the question enquiring whether the mother can afford the money to buy infant formula, utensils and cooking fuel required for at least six months was considered too complex by 11 experts. It was suggested that the question is broken down to a minimum of two questions, separating the availability of funds for purchasing the formula from the availability of funds for utensils and cooking fuel.
For the question enquiring whether the mother has a source of regular income, additional questions about whether the husband or father of the child has a source of regular income and the number of dependants was suggested by 12 experts.
Two out of the three safety questions assessing the safety of replacement feeding achieved consensus on both appropriateness of the question and of the language. However 14 experts suggested that the question enquiring whether the mother will be able to wash her hands after using the toilet and before preparing feeds, as well as sterilizing utensils should be split. It was suggested that ensuring washing of hands and keeping utensils clean are separate activities. Adding ‘clean water’ to the question was also suggested by seven experts, they argued that ensuring washing of hands and utensils must entail the use of clean water to achieve the required aim. Eleven of the experts agreed with the question enquiring about the availability of a water-borne latrine or flush toilet achieved. Four experts argued that measures to ensure clean hands and good hygiene after using the toilet was paramount rather than the type of sewage disposal system.
The best way to describe the father of a child in SSA was enquired from the participants and 12 experts agreed on using father, as opposed to husband or partner.
The most appropriate and easily understandable means of describing how often a baby is fed was also enquired and seven experts agreed on the use of hourly rate. They said the use of number of times/day may result in the mother squeezing in feeding times for her convenience and not necessarily at the right time when the baby is supposed to be fed. Those that agreed with using the number of times/day argued that using an hourly rate might be difficult for mothers who have poor idea of time. The remaining two suggested using either of the two as appropriate, depending on the educational level of the mother.
Whether to use clean water or tap water to enquire about the availability of safe water that could be used for preparing formula feed was also debated upon. Nine experts argued that clean water is more appropriate as not all households have tap water, but still have access to clean water. The remaining six argued that tap water is appropriate as this is the standard for a source of clean water.
Proximity of water access was also deliberated, whether to use a specific distance to describe proximity or leave it relative by using the word ‘close by’. Eight of the experts argued that the use of a relative term is better, as some mothers do not appreciate spatial orientation or metric distance. The rest of the participants supported using a specific distance as it is more objective.
Whether to specify cup feeding as the appropriate way of giving formula feed was also enquired, with 12 experts agreeing that this should be included in the tool. A minority (3) however said it should not be included as this might confuse mothers since cup feeding is not widely used and there is no evidence that proves it is a safer means of feeding.
All experts suggested avoidance of double-barrelled questions e.g. Would you be able to wash your hands after using the toilet and before preparing feeds, as well as wash and sterilize all utensils required for preparing? They argued that one part of the question might be a positive answer while the other may not, hence affecting the overall conclusion. Two-thirds of the experts also suggested including the demographic data of the respondents at the beginning the tool.
The cut-off points for feeding options
Only nine of experts agreed that the cut-off point for breastfeeding should be ≤8 (number of ‘yes’ responses). Also only 10 of experts agreed that the cut-off point for replacement feeding should be >12. Less than half (7) of the experts argued that a cut-off point of 12 for replacement feeding will make most women in SSA not eligible to practise exclusive replacement feeding.
Proposed AFASS assessment and screening tool
Twelve out of the 15 questions in the tool clearly achieved consensus from the panel of experts and hence have been included in the final AFASS tool. Of the remaining three questions not achieving consensus, one was corrected and the other two excluded from the final tool. Further language changes and splitting of double-barrelled questions were incorporated as suggested by the panel of experts.
The question in the preliminary section was corrected as suggested by the panel of experts and supported by the literature. Experts pointed out that the risk of transmitting HIV through breastfeeding when mothers mix feed is higher than the conservative estimate of three-fold employed in the tool. Studies have reported that mixed feeding can lead to a 2 to 10-fold increase in the risk of HIV transmission from an HIV infected mother to her child during the first six months of life, compared to exclusive breastfeeding [10–12].
The question in the feasibility section that enquires whether the mother will be home for the first six months and not be working full time has not been included in the final AFASS tool, as this did not reach consensus agreement. Experts argued that to achieve an effective way of replacement feeding does not necessarily require the mother home all the time.
The question under the safety section relating to the availability of water-borne latrine or toilet has also been excluded from the final AFASS tool. Experts argued that measures after using the toilet are more important than the type of sewage system in the home, although a counter argument can be made that the type of sewage system available in a community reflects the risk of communicable diseases.
The aspect that needed the most attention was the cut-off points for decision-making guides to which feeding option would be most advisable under the individual circumstances.
During the course of this study, a new WHO guideline for infant feeding in low income countries which recommends an extensive anti-retroviral (ARV) regimen to either infant or mother and mixed feeding was released . This is based on current evidence from recent studies that demonstrate that with ARV, MTCT can be reduced to around 2% and by extending a course of ARVs for either the mother or the infant, mixed feeding can be practised without the risk of transmission of HIV infection to the infant [31–33]. The WHO guideline still recommends that when ARV are not immediately available, the recommendations included in the 2006 HIV and infant feeding update (AFASS criteria) should provide useful guidance for mothers and health workers . This shows that the continuing use of the AFASS criteria remains relevant particularly in low income countries where regular supply of ARV is still a significant challenge. For example, over half of the 1.4 million pregnant women living with HIV are estimated to have received antiretroviral drugs to prevent transmission of HIV to their infants. An estimated 53% [40–79%] of pregnant women living with HIV received anti-retrovirals to reduce the risk of transmitting HIV to their infants, up from 45% [37–57%] in 2008 and 15% [12–18%] in 2005. However, a large proportion continued to receive the less efficacious single-dose Nevirapine regimen . It is also possible that the release of this guideline during the course of this study was a source of bias as this could have influenced some of the experts’ opinion.
Further possible bias might have arisen if respondents who consented to participate but did not complete the survey (making up two-fifths of initial respondents) were significantly different from those who completed the survey but an overview of location and professional group of respondents did not appear to suggest either of these two factors were significantly different.