Acceptability and feasibility of infant‐feeding options: experiences of HIV‐infected mothers in the World Health Organization Kesho Bora mother‐to‐child transmission prevention (PMTCT) trial in Burkina Faso

Abstract

In Burkina Faso, prolonged breastfeeding with introduction of ritual fluids from birth is a deep‐seated norm. We explored HIV‐infected mothers' views and experiences of the acceptability and feasibility of the World Health Organization's recommended infant‐feeding options within a mother‐to‐child‐transmission prevention trial. A qualitative study was conducted on 17 formula‐feeding and 19 breastfeeding mothers, from a larger cohort of 51 eligible HIV‐infected women, consenting to participate in separate focus group discussions in early post‐partum. Mothers opted for breastfeeding essentially out of fear of family rejection. Most of them were afraid of denigration for disrespecting tradition if they formula‐fed or being suspected of HIV infection. Achieving exclusive breastfeeding remained a difficult challenge as they engaged in a continuous struggle with close elders to avoid fluid feeding. Additional stress and fatigue were fed by their perception of a high transmission risk through breast milk. Exclusive formula‐feeding seemed easier to implement, especially as formula was provided free of charge. Formula‐feeding mothers more frequently had a supportive partner, a strong personality and lived in better socio‐economic conditions than breastfeeding mothers (76% had education and electricity supply vs. 42%, respectively). Exclusive breastfeeding for the first 6 months remains the most appropriate option for many HIV‐infected mothers in sub‐Saharan Africa. Its acceptability and feasibility urgently need to be improved by promoting it as the best feeding option for all infants. Other crucial interventions are the promotion of voluntary counselling and testing for couples, and greater partner involvement in infant‐feeding counselling.

Introduction

Post‐natal HIV‐1 transmission through breast milk is responsible for the 50% of infections among children in sub‐Saharan Africa in the absence of health interventions (De Cock et al. 2000). Mixed and prolonged breastfeeding are important risk factors. Thus, when replacement feeding with formula is acceptable, feasible, affordable, sustainable and safe (AFASS), the World Health Organization (WHO) recommends avoidance of all breastfeeding by HIV‐infected women. Otherwise, exclusive breastfeeding is recommended for the first 6 months of life (WHO 2007). Replacement feeding from birth is the only way to completely avoid post‐natal HIV transmission, but it is a high‐risk option for infant health and survival in developing countries (WHO 2000). Conversely, exclusive breastfeeding meets the infant's complete nutritional needs until 6 months and was associated with a reduction of the risk of transmission by more than 50% compared with non‐exclusive breastfeeding in large cohort studies from sub‐Saharan Africa (Iliff et al. 2005; Coovadia et al. 2007). Preliminary results from mother‐to‐child transmission prevention (PMTCT) and epidemiological studies highlighted the dramatic consequences of early cessation of breastfeeding on infants' health and survival and led the WHO to amend recommendations on early rapid weaning of HIV‐exposed infants (WHO 2007). Thus, at 6 months, if replacement feeding is still not AFASS, continuation of breastfeeding with additional complementary foods is recommended until a nutritionally adequate and safe diet can be provided without breast milk. However, in Burkina Faso, current PMTCT guidelines still recommend formula feeding from birth or exclusive breastfeeding for 4–6 months, followed by rapid breastfeeding cessation (Ministère de la Santé du Burkina Faso 2006). In most developing societies, both exclusive replacement feeding and short‐term exclusive breastfeeding are alien concepts that raise questions about their acceptability and feasibility (Becquet et al. 2005a; Abiona et al. 2006; Leshabari et al. 2007a). This public‐health dilemma is particularly acute in Burkina Faso, one of the world's poorest countries, where criteria usually required for formula‐feeding are seldom met. Basically, the traditional infant‐feeding pattern combines feeding with fluids from the first day of life with systematic and prolonged breastfeeding: 98 and 81% of infants are breastfed at birth and at 23 months of age, respectively; 83% of under 6‐month‐old infants receive semi‐solid food in addition to breast milk (INSD & ORC Macro 2004).

While qualitative data on infant‐feeding choice and experiences of HIV‐positive women in Eastern and Southern Africa have been published, limited data are available from West Africa. The present study aimed to explore the acceptability and feasibility of the two main infant‐feeding options in an urban context in Burkina Faso. For this purpose, we collected experiences from HIV‐infected mothers participating in a PMTCT research programme regarding infant‐feeding choice and adherence to exclusive mode, main difficulties encountered and coping strategies applied.

Key messages

• • The acceptability and feasibility of exclusive breastfeeding for 6 months urgently need to be improved for HIV‐free infant survival and mothers' welfare through strengthened promotion of exclusive breastfeeding for the general population in settings where this feeding mode is socially stigmatized.
• • Formula feeding was acceptable and feasible for some motivated women with a strong mindset and partner support in the favourable conditions of a PMTCT trial, with free provision of formula.
• • PMTCT programmes should promote couples' voluntary counselling and testing, and partner involvement in infant‐feeding counselling.
• • Better knowledge and understanding of HIV‐positive mothers' infant‐feeding experience, while taking into account the complexity of their socio‐cultural context, may inform interventions for improving and tailoring infant‐feeding counselling.

Methods

Focus group discussions (FGDs) were conducted with a subsample of participants from the WHO Kesho Bora study (KBS) PMTCT research programme in Bobo‐Dioulasso, Burkina Faso's second largest city (500 000 inhabitants).

Study setting

The overall goal of the KBS, conducted in five locations in Africa, is to optimize the use of antiretroviral drugs (ARV) drugs during the antepartum, intrapartum and post‐partum periods to prevent mother‐to‐child transmission of HIV and protect maternal health. HIV‐infected pregnant women were recruited to participate either in a randomized clinical trial or in a prospective cohort study with different regimens and duration of ARV prophylaxis or treatment according to their stage of HIV infection. Thus, all enrolled women received at least an ARV prophylaxis, whatever their infant‐feeding choice. All infants born to women enrolled in any of the study groups received one dose of nevirapine within 72 h of birth.

In Bobo‐Dioulasso, the KBS protocol included sustained individual infant‐feeding counselling coupled with psychosocial support throughout the study follow‐up carried out in the four study clinics by highly qualified female social workers, purposively trained by the WHO staff. Starting prior to inclusion (i.e. before week 32 of gestation), weekly sessions consisted of counselling women on the two main feeding modes. According to the WHO recommendations at the time of protocol development, antenatal counselling focused on feeding options: either exclusive breastfeeding and rapid weaning around 6 months or exclusive formula‐feeding from birth (WHO 2001). After an assessment of the woman's personal situation, assistance in the decision‐making process was provided and practical demonstrations on initiating the chosen feeding mode were conducted. A 6‐month supply of formula was offered free of charge to mothers opting for replacement feeding.

Post‐natal counselling included a home visit 3–5 days after delivery, then visits at the study sites every 2 weeks until 8 weeks post‐partum, then monthly until 1 year post‐partum, and finally every 3 months until 2 years post‐partum. The objective of these nine scheduled visits from birth to 6 months post‐partum was to support mothers in practising their chosen feeding option exclusively by reviewing appropriate feeding techniques and giving practical advice. From the fourth month, counsellors were also focusing counselling sessions on strategies for implementation of rapid breastfeeding cessation and subsequent introduction of adequate replacement food. Active detection and treatment of any disease or malnutrition episode in mothers and infants were provided free of cost by the KBS throughout the follow‐up period. Infant HIV‐testing was performed at 6 weeks and 12 months with a polymerase chain reaction (PCR) test. Mothers were invited to come to their study clinic in case of health problems (transportation refunds were ensured for both scheduled and unscheduled visits). Moreover, throughout the study follow‐up, mothers were encouraged and assisted to disclose their HIV status to partners and significant others and to seek support from community HIV networks.

Study participants

At the time of our study in April–May 2006, 51 women were fulfilling the inclusion criterion (i.e. 1–5 months post‐partum status) and were given individual information about the study's objectives and procedures. Special attention was given to the issue of confidentiality and the mothers' possible reluctance to face other HIV‐positive women from their community. Thirty‐six women gave consent and participated in the FGDs: 17 (out of 18) were formula‐feeding and 19 (out of 33) were breastfeeding.

Data collection and analysis

Six FGDs were conducted, three with formula‐feeding and three with breastfeeding mothers. Group discussions were led by a counsellor who was not usually in charge of these particular women (to minimize her impact on their accounts). All sessions were conducted in the predominant local language, using four thematic questions to guide the sessions' progression: infant‐feeding decision making, possible difficulties in practising the chosen mode, strategies for coping with difficulties and mothers' feelings about their infant‐feeding experience. All FGDs were audiotaped and transcribed verbatim into French by the leading counsellor. In order to maintain participants' confidentiality, names were not recorded. An interpretative approach to data analysis was used, which consisted of reading and rereading word‐processed texts, computerized coding, synthesizing and grouping sections of text from the FGDs that covered similar issues or experiences into exhaustive categories (Krueger 1998).

Participants' demographic characteristics at their inclusion in the KBS were extracted from the study database.

Ethical approval

The KBS protocol was approved by the Ethical Review Committees of the Burkinabe Ministry of Health and the WHO.

Results

Demographic characteristics of the FGD participants at the time of their inclusion in the KBS are shown in Table 1. The proportion of breastfeeding and formula‐feeding mothers was quite similar regarding their marital status (mainly monogamous union) and occupation. Most mothers were living with extended family. According to the database, seven formula‐feeding and six breastfeeding mothers were living with the nuclear family. However, ‘living in a nuclear family’ had been defined as no one other than the mother, partner and children residing in the household for at least the last 6 months. In West African societies, it is common that family members temporarily move to their relatives' homes for various reasons. As a result, only two women (one breastfeeding and one formula‐feeding mother) stated that they were actually living in a nuclear family at the time of the FGDs. The formula‐feeding mothers differed from their breastfeeding counterparts as a result of a higher education level and greater access to indoor tap water and electricity. At inclusion, which was concomitant with the infant‐feeding decision‐making process, about one‐half of the mothers in both groups had disclosed their HIV status to their partners. Infant mean age at the time of the FGDs was 4.4 (SD: 2.2) and 3.1 (SD: 1.8) months in formula‐feeding and breastfeeding groups, respectively. The following section presents the key themes that characterized the mothers' experiences concerning their infant‐feeding choice and practice.

Table 1.

Demographic characteristics of formula feeding (FF) and breastfeeding (BF). Burkinabe HIV‐infected mothers participating in focus group discussions at the time of their inclusion in the PMTCT study (32–34 weeks pregnancy)

Characteristics	Qualitative group
	FF (n = 17)	BF (n = 19)
Mean age (SD)	29.0 (±5.0)	25.6 (±6.1)
Education
No	4	11
Yes	13	8
Marital status
Married to partner	16	17
Not married to partner	1	2
Household
Living with nuclear family	7	6
Living with relatives in her home	5	7
Living at her parents' home	2	5
Living at her in‐laws' home	3	1
Main occupation
Unemployed	11	14
Formal employment	2	0
Informal employment	4	5
Drinking water source
Indoor tap	8	3
Community tap	9	16
Electricity supply
No	4	11
Yes	13	8
Owning a refrigerator
No	12	18
Yes	5	1
HIV‐status disclosure to partner
No	8	9
Yes	9	10

PMTCT, mother‐to‐child transmission prevention.

Breastfeeding mothers

Infant‐feeding decision making: an informed default choice

Breastfeeding mothers seemed particularly worried about the risk of transmission but were even more concerned about confronting social risks inherent in not conforming to the tradition of breastfeeding. Many of them chose to breastfeed out of fear of the anticipated negative reaction, mainly from their in‐laws and husband, or of the stigmatization associated with HIV if they formula‐fed. Others, despite support about their choice to formula‐feed from their husbands who were informed of their HIV status, encountered real opposition from the family when initiating replacement feeding and were forced to abandon it after a few days and started breastfeeding. A few stated that they made a feeding choice based on what they thought was optimal for their infant and themselves (Box 1).

Box 1. Infant‐feeding decision making: an informed default choice

‘They want to know what's going on inside you, why you don't breastfeed your infant! So they pass the time asking you and asking you over and over, to the point of ripping you open to see. And if you can't keep your mouth shut, you will sit down and confess everything. That's why I chose to breastfeed’.

‘I didn't want to give her the breast. But I was afraid because of the family. I asked for permission to go to my sister's. If I could have gone there and stayed awhile, I was going to give him artificial milk. But I didn't have permission’.

‘I had just learned to prepare the formula; I managed to buy a gas burner and a thermos. My husband also tested positive, so there was no problem with formula[. . .]but his mother said that those women who don't breastfeed are infected. We couldn't challenge her, which meant that I put away the equipment and ended up breastfeeding’.

Exclusive breastfeeding: an endurance race for mothers

The greatest difficulty reported by most of the breastfeeding mothers was the somewhat forceful pressure from elders to introduce other fluids in infant feeding (water, water from the infants' bath, ritual healing products), and, as the child grew older, traditional gruels. The husband was sometimes presented as supportive but unable to exercise enough influence. As social pressure in favour of non‐exclusive breastfeeding endured, the main challenge for breastfeeding mothers was to cope on their own while facing a long‐term continuous hardship using non‐conflictual strategies (i.e. keeping the infant with them, ‘listening to everyone but doing what they wanted’). Some felt confident in their ability to protect their child from fluids and other foods, but paid a price as constant efforts caused stress and fatigue that affected their behaviour and communication with their families. For others, social pressures, sometimes reinforced by their own culpability about not giving any fluids to the infant, could have potentially altered their capacity to breastfeed exclusively (Box 2).

Box 2. Exclusive breastfeeding: an endurance race for mothers

‘She[mother‐in‐law]told me to make decoctions for cleansing the baby of impurities that upset his bowels. She bought me nérétree bark. I peeled and boiled the bark in front of her and I took the pot and left. Now, it still sits where I left it, behind the wardrobe. This is my problem, they bother me too much’.

‘My mother‐in‐law wants to bathe my baby. If I let her bathe him, she is going to let him drink[water from the bath], so I don't want her bathing him. Because of this she's talking, saying that I think she is going to kill my baby, that I think she's a witch, that she has already bathed other children besides mine!’

‘My baby is always with me, my foot stuck to her foot[laughter]! If she isn't in my arms, she's in her father's arms. If people come in the house, they play with her, but they leave and leave her here. In the compound, they prepare dolo (fermented drink), if I leave her, even a moment, they will give her a drink’.

‘The wife of one of my brothers is a “doctor”. She brought her baby into the world, and for 15 months she never gave him traditional drugs. The baby has grown well. I said, “Wow, that's good! I'm going to do the same!” But because of what people say, I can't’.

‘My mother tells me that I walk in the sun and I don't give the child anything to drink. She says that one day I'll take my child off my back and I'll find her dead’.

Stress of infecting the child: mistrust towards breastfeeding

Mothers felt extremely anxious about being responsible for infecting the child at any time during the breastfeeding period. Perceived risk of transmission was quite high among participants, but it was unclear whether they associated this risk with their choice of breastfeeding per se or with the potential failure in adherence to its exclusive mode. As a consequence, they anticipated early breastfeeding cessation as the oncoming hardship with a mix of impatience and apprehension. Generally, they seemed highly motivated to stop breastfeeding in due time, although they feared negative or even violent reactions and worried about its effect on infant health and survival (Box 3).

Box 3. Stress of infecting the child: mistrust towards breastfeeding

‘They say that if you are infected, then there is no baby. That's what got me thinking. This disease will never leave me; I've already got it so I'm only concerned about my child. That's what scares me, you bring him into the world, but he doesn't have a long life’.

‘It was my first pregnancy and there I was, infected, with a problem child. I could die from worrying. I think about it every day and I go to the church and pray to God that she won't be infected’.

‘I breastfeed because of my mother‐in‐law, but deep down, I'm not calm, I am afraid’.

‘I'm afraid that my child will get the disease. I want the appointments[at the study site]to come quickly so that I will know how to wean’.

‘It isn't hard now; it's weaning that will be hard. My husband and I talked it over, but the others won't agree to it’.

‘What scares you about weaning the baby at this time is that we are in Africa. The others say that if you wean a child early, he will get “breast‐milk disease” and won't grow. That it can even kill the baby’.

Formula‐feeding mothers

Infant‐feeding decision making: determining the top priority

As formula‐feeding mothers did not want to take any risks of infecting the infant, defined as ‘the top priority’, the social risk was pushed to the background. Each had disclosed her HIV status to her partner at the time of the formula‐feeding initiation and benefited from his support. Secondary reasons were a desire to avoid the stress caused by early breastfeeding cessation on themselves and their infants and to remain autonomous in their movements (Box 4).

Box 4. Infant‐feeding decision making: determining the top priority

‘The baby is born and she is not infected and you say that you are going to breastfeed? Even though you are sick? If my child is healthy, the top priority is to give her the formula and nothing will happen to her’.

‘As for me, I give the baby formula because if he didn't get the disease during the pregnancy, he can only get it when he grows up[infection at adulthood]. It won't be me, his mother, who will be responsible for his infection’.

‘We had doubts about the breast. Why would you want to breastfeed your baby! And if it turns out that there are bleeding sores around the nipples, by nursing, the baby can catch this virus. But if it's formula, you stay calm’.

Formula‐feeding: the facilitator role of partner support

Whatever difficulties encountered when initiating formula‐feeding were generally described as not having serious consequences. ‘The beginning is difficult’ yet appears not to pose an ongoing problem. The husband was an actor, even leader, in not only developing strategies to resist the family circle but also sometimes in managing infant feeding. A number of pretexts (i.e. traditional medicine's failure to cure low breast‐milk secretion or an invented or genuine prior disease) allowed the couple to justify their choice based on criteria that others cannot control and presumably feel less able to question. As all of them had explained to their kin that they had had little or no milk at birth, they encountered no major pressure for occasional breastfeeding after a few weeks (Box 5).

Box 5. Formula feeding: the facilitator role of partner support

‘I had no problem during formula feeding because before I gave birth, my husband and I had intended to give the baby formula. And so, I asked him, if people ask us, what are we going to say? And he said that if people say something, then ask them “are you the ones buying formula for me?” ’

‘My father‐in‐law was constantly getting on my nerves. But my husband told me not to listen to him. He can do as he pleases, spending the whole day criticizing. Since we know what's going on[HIV infection], we don't respond to him. My husband has encouraged me’.

‘There's no problem at my house. Ha! Sometimes my husband gets up at night and prepares the milk and then he feeds the baby. As for me, I rest[laughter]!’

‘I'm so happy because my husband told them that I had some tests and that my breasts are not good; the pregnancy was only 5 months along. Then he lied, saying that I was even sad because when I deliver, I won't be able to breastfeed the baby’.

‘My father and mother came and they asked why I don't breastfeed the baby. We told them that I had a serious problem with my breasts. Since I had difficulties breastfeeding my first child, I breastfed but also gave him formula. They responded that if you can buy the milk then there is no problem’.

Self‐efficacy in formula‐feeding

Most FF mothers have come up against moments of doubt (i.e. guilt about not having provided maternal milk to the infant, fear that the infant will get sick if the milk is poorly prepared, etc.), but they felt satisfied with their infant‐feeding choice and proud of overcoming difficulties. Self‐confidence was reinforced by the relief from practising a feeding method that had no transmission risk and was viewed as adequate to satisfy the infant's nutritional needs. Their high level of self‐efficacy allowed them to see the consequences of their choice not as problems but rather as new stages that they must learn to surmount. Mental strength was thus a major factor contributing to their successful formula‐feeding practice (Box 6).

Box 6. Self‐efficacy in formula feeding

‘In the beginning, having to get up at night to prepare the milk when the baby cries makes it hard to sleep well. In fact when talking about difficulties, I especially want to talk about the fatigue, but now it's OK’.

‘Once I had decided to give my baby milk, it wasn't difficult for me. Each time, I hold on to the idea that the milk must be made at a certain time and think about how to get started with his feedings so that I could fill him with wonder’.

‘The babies also love the formula. He drank the formula from the bottle in the same way that he would have had to drink from his mother's breast. It's a big relief; our children are healthier than breastfed children. We go to the health meetings and we see the difference’.

‘It's the “know‐it‐alls” who bring about anxieties and difficulties, but if you think about it carefully, you have to push these stories out of your head. Since you alone know about the advantages of milk, you'll laugh at their words. You close your ears to the “know‐it‐alls”. If not, in Africa, this will never end’.

Formula‐feeding and breastfeeding mothers

Relationship to counsellor: between gratitude and dependence

All mothers were extremely grateful to the KBS, and all stressed the benefit of having received access to ARV drugs and free care for themselves and their infant. However, they particularly emphasized the intimate relationship they developed with their counsellor over time. Their accounts attested to a psychological dependence and a certain feeling of isolation. Breastfeeding mothers particularly valorized the knowledge they acquired about breastfeeding techniques and considered themselves ‘informed’ compared with other mothers in their situation. Among formula‐feeding mothers, psychosocial support provided by the counsellor was considered a determining factor to their current positive attitude (Box 7).

Box 7. Relationship to counsellor: between gratitude and dependence

‘As for me I thought that you simply breastfed the baby; I didn't know there was a way to breastfeed. You have to make sure the breast has no wounds that can transmit the disease to the baby’. (breastfeeding mother)

‘There are some who have not taken the test; today they don't know if they have this disease or not. They give other foods to the baby; they breastfeed but don't know’. (breastfeeding mother)

‘If you're at home alone with nobody to give you any advice, there you are, depressed. This can even make the disease worse. Discussing it together is a great remedy that is different from medicines for the disease. As for me, I am happy; I'm even proud!’ (formula‐feeding mother)

‘I beg God that the councellors (. . .) continue to take care of us because the disease is difficult (. . .). The new ones [participants who've just joined the project] are hard to handle; they're upset. Me, I've gone through that stage. But now I see (. . .) that many people have the disease. Many have died before me . . . but me, I have nothing, eh!’ (formula‐feeding mother)

Discussion

Our results suggest that even women living in very traditional and poor settings could clearly understand the stakes, risks and consequences of their infant‐feeding choice once they had been appropriately counselled. They were able to make their own ‘AFASS’ assessment and showed strong determination in following the counsellors' advice:

If someone shows you a path that can save your child, you must do your part to follow it so that luck is on your and your baby's side.

Over time, mothers built strong ties with their counsellors who were pushed far beyond their initial counselling responsibilities. Primarily, the counsellor was the one who informed the woman about her situation and gave technical advice. Then, she quickly became the key trustworthy person who could listen, understand, develop coping strategies based on the mother's context and sometimes act as a mediator with the partner or relatives to implement the appropriate strategy. Ethnographic research conducted in Southern Africa reported that even when HIV‐infected mothers gained good knowledge and understanding of infant‐feeding options through routine PMTCT services, they tended to disregard the counsellors' advice and give priority to their own feeding preferences and significant others' instructions (Buskens et al. 2007). In our sample, the trust and authority mothers attached to the counsellors' advice empowered them to engage in an awkard situation with their kin. To limit such a dependent relationship and alleviate mothers' isolation, counsellors were regularly encouraging them to make connections with HIV networks, but according to the mothers' accounts at the time of the FGDs, only two formula‐feeding mothers had participated in a local association. Such efficient and tailored support, which would be extremely difficult to implement in routine PMTCT services, in addition to fear of HIV‐related stigma could have diminished the mothers' motivation to seek support in HIV networks.

Exclusive breastfeeding: is it acceptable and feasible?

In our study, opting for breastfeeding seemed to be an informed ‘default’ choice for many mothers. This could explain the stress evident in the mothers' accounts about their responsibility in infecting the infant, particularly among those who were frustrated about their initial formula‐feeding plan. However, repeated failures in exclusive practice or overestimation of the transmission risk through breast milk could also be the cause. Actually, many mothers first knew about their HIV status and the risk of infant infection almost at the same time. In spite (or as a consequence) of the huge efforts expended to provide comprehensive and objective counselling, mothers in a state of psychological vulnerability may have particularly focused on the risks associated with breast milk and thus undermined this feeding option.

Achieving exclusive breastfeeding seemed to remain a difficult challenge for women, and it is likely that a disparity existed between their displayed determination and their actual practice. Conversely, formula‐feeding mothers stated they had no major difficulty with the exclusive mode mainly because they initially justified their choice by a health condition involving milk‐production problems. Evaluation studies on infant‐feeding counselling services in Southern Africa have shown that when high‐quality counselling is provided to the community (i.e. not targeted to HIV‐positive women only), it allows mothers to make the appropriate feeding choice and supports them in its exclusive practice (Piwoz et al. 2005; Bland et al. 2008). However, research conducted in West Africa in a PMTCT study context reported low adherence to the exclusive mode among breastfeeding mothers as compared with their formula‐feeding counterparts (Becquet et al. 2005b; Leroy et al. 2007). Our findings, consistent with these observations from Côte d'Ivoire, illustrate the continuum of efforts made by breastfeeding mothers to protect their infants from mixed feeding by close elders. Basically, the main challenge to exclusive breastfeeding was pressure to introduce fluids in the infant's daily diet, involving different bacteriological risks: from herbal or root preparations to water from the infant's bath, a traditional practice in Burkina Faso. In the first months of life, infants are bathed in water mixed with medicinal herbs, which is then partly given to the baby to drink (Alfieri 2000). In some cases of older infants, mothers were asked with insistence to feed their child the popular millet‐based gruel, bought ready‐to‐eat in the street and traditionally introduced early in the diet (Mouquet‐Rivier et al. 2008). In Burkina Faso, breastfeeding, gavages and enemas are considered ‘the bedrocks’ of baby care. These practices, perceived as highly healthy, remain a stronghold within traditional and mystical beliefs (Taverne 2000a). Infant feeding, particularly breastfeeding, is a family matter, especially entrusted to the close elders (mother, mother‐in‐law, aunts, neighbours) who are the guardians of tradition. Their role is to guide the mother, advising or imposing various practices to improve breastfeeding techniques or increase breast‐milk production (Taverne 2000b). In such societies where elder authority goes unquestioned, opting for exclusive breastfeeding of infants is viewed as a lack of respect for traditions and exposes the mother to denigration and isolation within her own family. Moreover, she may be judged as a ‘bad mother’ by depriving her infant of the traditional care and would be blamed in case of infant health problems or poor psychomotor development.

Ultimately, most of them expressed no particular pride or well‐being about their breastfeeding experience. Although they had the necessary skills and knowledge, they constantly run up against the pre‐existing concept that a child's health depends on following the social and religious system and rules. Low perceived self‐efficacy among some mothers could ultimately affect their willingness to persevere in exclusive breastfeeding when constraints prove too difficult. These experiences are consistent with observations from a qualitative study led in Tanzania (Leshabari et al. 2007b) reporting how BF HIV‐infected mothers could become obsessed by a constant family pressure to mix‐feed and the related risk of HIV transmission, and then finally resign themselves to either mix‐feed or cease breastfeeding earlier than intended. In the present study, decision making on early breastfeeding cessation was also viewed as a major dilemma: deciding to prolong an infant's exposure to HIV transmission risk could be painful for these ‘informed’ mothers. Conversely, they were also conscious that early cessation would expose them to social critics and could compromise their infant's growth and even survival. In sub‐Saharan Africa, scaling up PMTCT services, including infant‐feeding counselling, is likely to lead many breastfeeding mothers in this critical decision‐making process.

Our findings suggest that acceptability and feasibility of exclusive breastfeeding still need to be improved in communities where this feeding mode may be highly stigmatizing. This illustrates the important gap that can exist between two different conceptions, biomedical and popular, both aimed at protecting infant health. Moreover, it emphasizes the importance of exploring women's perceptions and understanding of transmission risk through breast milk and local baby‐care practices and of assisting them in having a positive breastfeeding experience. Finally, the study raises the dilemma of early breastfeeding cessation, which needs to be specifically addressed in PMTCT guidelines given recent knowledge on infant health and survival while also taking into account mothers' individual contexts.

Promoting partner involvement

Disclosure of HIV status to the partner is usually a major condition for replacement‐feeding choice (Kiarie et al. 2004; Brou et al. 2007). Actually, the disclosure rate at inclusion in the KBS was quite low, but at the time of the FGDs every formula‐feeding mother had disclosed her status to her partner and had his support. We assume that women who felt confident in obtaining their partners' support were more likely to choose formula‐feeding, but also that having opted for formula‐feeding was an incentive for them to disclose their status and thus gain the support of their partners. Partner support was not limited to confronting relatives and the community. It also helped women to accept their disease, stimulating a positive outlook and enabling them to plan for the future, subsequently contributing to increased self‐efficacy regarding infant feeding:

I have a husband who understands me. In any case, at the beginning it was very very hard. But in the end, he was the one who would console me. When I'm sad, he sees it in my face, and he comes and talks to me, he encourages me.

Research conducted in various settings has shown that HIV‐infected women must face a number of barriers when sharing their HIV‐test results with partners, including fear of stigma, violence and desertion (Nebie et al. 2001; Gaillard et al. 2002; Bwirire et al. 2008; Greeff et al. 2008). Disclosure‐related violence has even been reported in other studies among pregnant HIV‐infected women (Kilewo et al. 2001). In Bobo‐Dioulasso, research conducted among HIV‐infected pregnant women reported low rates of HIV disclosure to partners. However, it also showed that HIV‐status disclosure resulted mainly in the partner's indifference (70%) or support. No case of physical violence was reported from this study (Issiaka et al. 2001). Extension of voluntary counseling and testing (VCT) and HIV‐prevention services from the mid‐1990s and improved communication about HIV/AIDS are likely to have created better conditions for HIV disclosure to partners. Within the KBS, assistance to couples' VCT was provided subsequent to the pregnant woman's initial HIV testing. Many mothers reported that their husbands were also HIV‐infected. Such an approach could explain that disclosure outcomes were positive according to both breastfeeding and formula‐feeding mothers' accounts. Other studies, in settings offering individual or couples' VCT, reported that partners' reactions to the disclosure of their wives' HIV status were predominantly supportive (Maman et al. 2003; Sagay et al. 2006). Their subsequent involvement is expected to positively impact women's uptake of PMTCT interventions and infant‐feeding choice (Farquhar et al. 2004; Kiarie et al. 2004). It appeared that breastfeeding mothers who had shared their HIV status could not valorize their partners' support as much. A supportive husband seemed less effective to cope with the daily pressure associated with exclusive breastfeeding, particularly when it arose from his family. Providing assistance to couples' VCT as part of PMTCT programmes and promotion of subsequent partners' involvement in infant‐feeding counselling could represent a cost‐effective intervention for enhancing mothers' and fathers' empowerment in infant‐feeding decisions.

FF in very low‐income settings: feasibility imposes its own terms

FF is not a common practice in West Africa. However, the use of replacement milks (formula or whole cow's milk) may be considered in a few particular cases: maternal death, severe maternal illness and insufficient or no breast‐milk secretion in Burkina Faso (Taverne 2000b) and Senegal (Mane et al. 2006). Similar observations have been made in Nigeria where, in community perceptions, FF seemed acceptable provided that mothers had a health or breast problem, whereas exclusive breastfeeding was firmly considered to be an inacceptable feeding mode (Abiona et al. 2006). The collaboration between the mother, her partner and the counsellor in developing resistance strategies to confront relatives thus proved to be successful in our study. FF mothers took advantage of this limited tolerance under the pretext of health problems or insufficient breast‐milk secretion, thus making their feeding choice ‘acceptable’ to their close relatives.

FF mothers differed from their breastfeeding counterparts in some AFASS criteria (indoor tap water, electricity) recommended for adequate FF practice. However, none of them fulfilled all the usually required conditions (WHO 2005). Mothers were very conscious that their feeding choice was feasible in the particular conditions of the KBS only. In such a setting, providing free formula for at least 6 months can be considered a prerequisite to making an infant‐feeding ‘choice’. Most of them stated they would not have been able to afford formula, but they succeeded without major problems in justifying the use of expensive infant food with family and neighbours. It is likely that being autonomous in milk provision was an asset to face critics and increase self‐efficacy in practising FF. Mothers and infants also benefited from a sustained and excellent support and a care package within the KBS. Such positive FF experiences could probably not have been achieved in the national PMTCT context. However, given that conditions were the same for all participants in the KBS, when analysing perceived feasibility of FF a key factor was the mothers' personal skills and determination. These women were more educated and were probably less prone to popular beliefs and denigration, and thus decided to play their own part in transmission prevention:

If he didn't get the disease during the pregnancy [. . .]it won't be me, his mother, who'll be responsible for his infection.

Although their socio‐economic conditions were better suited to cope with social pressure and replacement‐feeding difficulties, ultimately, mothers who maintained their choice of FF were those who showed high levels of self‐efficacy in the difficult process of protecting their infant from HIV infection and facing adverse situations. At the end of inclusions in the KBS in Bobo‐Dioulasso (December 2007), only a minority (21.3%) among 300 participants opted for FF, despite the favourable context of the trial. This observation clearly establishes that breastfeeding remained the most appropriate feeding mode for many mothers.

Generalizability of findings

FGD analysis is not aimed at generalizing results to the population, but rather, to explore various experiences and perceptions around a specific topic in a particular context. Limitations of the study were that women were recruited within the framework of a PMTCT trial rather than a ‘routine’ context and that a larger proportion of breastfeeding women declined to participate compared with FF women. It is likely that women who did not wish to participate experienced even greater difficulties with their feeding choice than those who consented. Although findings may not be generalizable to all HIV‐positive women in West Africa, they highlight beliefs, constraints and practices surrounding infant feeding that are likely to be very similar in many West African populations.

Conclusion

In many sub‐Saharan societies, exclusive breastfeeding is considered by far the best feeding option for women of unknown HIV status and for most HIV‐positive mothers, but is challenged by low acceptability and feasibility. Mothers must deal with normative cultural beliefs about infant feeding with the underlying consequence of being stigmatized. Hence, strengthening and generalizing promotion of exclusive breastfeeding are a major public‐health concern both for the infant's HIV‐free survival and the mother's welfare while providing health benefits to all infants. In our particular research context, FF seemed to be acceptable and feasible for some women in particularly favourable circumstances, when formula was sustainably provided for free. Infant‐feeding counselling guidelines are mainly developed through the application of a global PMTCT strategy: a better understanding of HIV‐positive mothers' infant‐feeding experience, which takes into account the complexity of their socio‐cultural context, may inform interventions for improving and tailoring infant‐feeding counselling.

Source of funding

The study was funded by the Agence Nationale de Recherches sur le SIDA (ANRS), Paris, France, and is part of the ANRS 1271/Kesho Bora cohort study.

No conflicts of interest have been declared.

None declared.