Background
Breastfeeding is the backbone of childhood survival initiatives worldwide. However, infant feeding in the context of human immunodeficiency virus (HIV) poses major dilemmas for both carers and health care workers (HCWs), particularly in resource poor settings where safe and affordable infant feeding options are limited [
1‐
3]. Whilst breastfeeding carries some risk of vertical transmission [
4,
5], breast milk alone is adequate in meeting the nutritional needs of the majority of infants under 6 months of age [
6]. It also provides over half of the energy and nutrient intake of infants over 6 months of age, which can be crucial given the poor nutritional value of many complementary infant foods given in sub-Saharan Africa [
7‐
9]. Breastfeeding is also associated with reduced childhood infections and mortality, as well as better growth and neurodevelopment outcomes [
10]. In addition, maternal health is improved through child spacing [
10].
For infants born to women living with HIV in resource limited settings, exclusive breastfeeding where infants receive only breast milk in the first 6 months, is recommended by the World Health Organization (WHO), as it reduces the risk of breast milk transmission of HIV 2–4 fold when compared to mixed feeding in the absence of other interventions [
11‐
13]. Mixed feeding where infants under 6 months receive other fluids and/or semi-solids in addition to breast milk, is also associated with reduced HIV-free survival and is therefore not recommended [
13]. In relatively affluent urban African settings, trials of replacement feeding with formula demonstrated benefit in terms of HIV-free survival of infants born to HIV infected mothers [
14,
15]. However, data from more rural African populations suggest that the use of replacement feeds and/or rapid weaning off breast milk at 6 months of age are strategies that are nutritionally hazardous and detrimental to child survival, with limited benefit of HIV-free survival [
16‐
23]. The reasons for this include poverty, lack of access to clean water and adequate sanitation, and low literacy rates in many rural African settings [
7,
24]. With maternal combination antiretroviral treatment (ART) and infant prophylaxis with nevirapine or zidovudine, the risk of vertical transmission in exclusively breastfed infants can be reduced to less than 1% [
25‐
29], and HIV-free survival at 24 months increased to more than 87% [
20].
Considerable progress has been made in implementing these interventions for the prevention of vertical transmission in sub-Saharan Africa, with the universal adoption of option B+ by WHO [
30]. However, for many countries, reducing the vertical transmission of HIV has remained challenging due to both resource limitations and socio-cultural barriers to exclusive breastfeeding. Exclusive breastfeeding is often viewed as being nutritionally inadequate and harmful for the baby, and exclusive replacement feeding is culturally unacceptable [
31‐
34]. The uptake of combined ART during pregnancy by women living with HIV has also been suboptimal [
35].
A cross-sectional health survey conducted in Kilifi County in 2011 showed that although breastfeeding was common, exclusive breastfeeding was not the norm with only 22.4% of infants under 6 months exclusively breastfed [
36]. More than 70% of mothers introduced complementary foods by the fourth month of life [
37]. These foods included supplemental liquids such as water, fresh cow’s milk and semi-solid feeds such as thin
uji (maize meal porridge) [
38]. These patterns have important implications for vertical transmission. HIV is a major attributable cause of in hospital deaths for children between 6 to 60 months of age admitted to Kilifi County Hospital (KCH) with severe acute malnutrition (SAM) [
39]. When we reviewed two and one half years’ worth of routinely collected data from our HIV clinic of 486 HIV exposed and infected infants between 0 and 24 months of age, we found that on at least one visit 19% were severely wasted, 21% were moderately wasted and 18% severely stunted. Only 50% of these mothers reported practising exclusive breastfeeding under 6 months.
The aim of this study was therefore to describe the decision-making processes around infant feeding at a rural HIV clinic in Kenya in order to identify the constraints to the recommended infant feeding practices for infants of women living with HIV.
Discussion
Our study showed that women living with HIV in this community were constrained from adhering to the infant feeding guidance by three main factors: fear of going against cultural norms of infant feeding as exclusive breastfeeding is alien to this community, a lack of autonomy in the infant feeding decision-making and non-disclosure of HIV status to close family members.
In an era when the infant feeding recommendations on prevention of vertical transmission of HIV for sub-Saharan African mothers is backed by evidence of improved HIV free survival [
12,
13,
16,
20,
24], our study showed that the uptake was poor in this rural African community, as in other similar settings [
49]. In this paper, we have shown that the uptake of infant feeding recommendations has been hampered by a number of factors. Importantly, we found that the practice of exclusive breastfeeding was uncommon in Kilifi as in many communities in sub-Saharan Africa [
50‐
53]. Therefore mothers who adhered to exclusive breastfeeding were likely to be singled out as being HIV infected [
54], inadvertently disclosing their HIV status.
Mothers often lacked autonomy in infant feeding decision-making, despite being counselled on the best options for them. The communal nature of childcare in this rural community posed the risk that another family member would feed the baby on other fluids or solids in her absence. This elevated the risk of vertical transmission and undernutrition in HIV-exposed infants. We found that the family structure and dynamics greatly influenced infant feeding choices and that the decisions were made by dominant family members including mothers in law or husbands, who are rarely targeted in infant feeding or prevention of vertical transmission counseling sessions at health facilities. The role of grandmothers in infant feeding has been widely documented elsewhere in Africa, and their ideas often contradict recommended feeding practices [
55‐
60]. In a study conducted in South Africa in 2004 that predominantly recruited single young HIV infected mothers 25 (63%), for example, 20 (80%) of these mothers who had chosen to exclusively breastfeed had introduced other liquids in the first month of life due to pressure from family members [
61]. The majority of the women in the study are financially dependent on their mothers or mothers-in-law and therefore found it a challenge to protect their autonomy on the infant feeding decision-making [
61]. This highlights the importance of ensuring that interventions take into consideration the power dynamics within families and communities and therefore incorporate strategies that target those influential ‘others’. Such strategies can include home-based, and more family inclusive counseling sessions. There is also a need to empower mothers with skills to negotiate these power dynamics within the family [
56].
Our data also showed that the delays in turnaround time for the results of the infant HIV polymerase chain reaction (PCR) limited the autonomy of some mothers to opt out of breastfeeding, in their attempt to safeguard their HIV PCR negative infants against the continuing exposure to HIV via breast milk. Without these results, they often opted to continue exclusive breastfeeding but then later carried the guilt of transmitting HIV to their infants where positive results were returned. This is consistent with the findings in South Africa in 2004, where infant feeding decision making was reportedly influenced by the desire to protect their infants from risks of vertical transmission [
61]. In our context, the use of alternatives to breastfeeding would not be a viable option. Our findings therefore highlight the need for health care workers to utilize a more empathic approach to the issue of non-adherence in their prevention of vertical transmission counseling sessions, recognising the positive intentions and psychological stress that HIV infected mothers undergo in making infant feeding choices. Ultimately this approach may also help to promote adherence to recommended practices [
62].
Disclosure of their HIV status sometimes resulted in mothers receiving increased support from friends and significant family members, with positive implications for their adherence to safe infant feeding practices. This has also been reported elsewhere in Africa [
63]. Unfortunately, this was not the norm in our setting, as disclosure often had adverse effects on marital relationships resulting in the abandonment and rejection of mothers and their children leaving them destitute. This has been commonly reported in many African settings [
63‐
65]. Women therefore understandably chose not to disclose their HIV status. This led to significant challenges with adherence and families often opted to follow infant feeding ‘norms’ of mixed feeding from as early as 1 month of age.
Undernutrition and HIV often coexist in the context of broader structural factors such as endemic poverty and high food insecurity [
66]. In our study, some mothers associated breast milk volume with adequate maternal diet, and many mothers perceived they had milk insufficiency as food insecurity is common in this area [
67,
68]. This potentially contributed to many mothers practicing mixed feeding [
60]. Sustainable strategies that address food insecurity and poverty among these vulnerable mothers could enhance adherence to recommended infant feeding practices such as exclusive breastfeeding for the first 6 months of life [
37].
Beyond decision-making processes at the household/community level, health system related factors were also seen to impact on infant feeding choices by both carers and HCWs. HIV infected mothers expressed the value that they derived from a facility initiated peer group that helped to mitigate the challenges that they experienced with adhering to the infant feeding guidance. Peer support groups can be a very useful strategy for channeling prevention of vertical transmission recommendations to communities whilst providing psychosocial support for HIV infected mothers [
69,
70]. Mothers also reported that their interaction with healthcare workers was beneficial in helping them make informed choices and in enhanced adherence to the infant feeding recommendations. This is consistent with findings of a study conducted in South Africa in 1999 [
71], where healthcare workers were reported as an important influence on carers’ infant feeding choices. However, our data also showed that mothers of only less than 50% of infants at enrolment to the HIV clinic reported exclusive breastfeeding. This suggests that most mothers did not adhere to the infant feeding counseling in the antenatal and postnatal clinics due to the constraints that we have already stated and in reality, the proportion exclusively breastfeeding may be lower. Although, we were not able to elicit HCW factors associated with non-adherence from the mothers, studies from South Africa found that mothers did not trust healthcare workers and cited inconsistent and confusing advice provided by them [
72,
73].
There had been changes to the national PMTCT and infant feeding policies at the time of the study giving greater emphasis to exclusive breastfeeding in the first 6 months, in combination with prolonged use of antiretroviral drugs for mothers and infants [
13,
74‐
76]. The HIV infected mothers we spoke to did not report policy changes as being a barrier to the uptake of infant feeding recommendations because of more immediate pressing constraints. This suggests that irrespective of infant feeding policy changes, underlying factors affecting infant feeding decision-making processes need to be addressed to support mothers’ adoption of new practices, particularly where these deviate from community norms.
Although mothers did not raise direct concerns about infant feeding policy changes, HCWs expressed their discontent at the lack of consistent and timely updates on changes, attributing this to the lack of clarity nationally in technical aspects of MTCT, as well as other facility level challenges such as heavy workload and inadequate resources [
77]. This suggests that the implementation of new policies on infant feeding strategies in the context of PMTCT sometimes fails to adequately reach health care workers in rural and remote areas in Africa, which has an impact on the quality of infant feeding counseling [
78‐
80]. There is therefore an urgent need to evaluate innovative cost effective strategies before rolling out new infant feeding and PMTCT national policies. This could be done within a framework of continuing professional development where mentors and trainers interact with the frontline health care workers regularly either remotely or by visiting their work stations to provide ongoing mentorship. This would allow the health workers time to assimilate the new policies whilst reflecting on their infant feeding counseling practices and on how to improve on these skills [
81].
A key strength of this study was that it utilized multiple data collection techniques and sources, which allowed triangulation of the findings from both approaches [
82]. A limitation was that the data were collected in 2011. However, current literature suggests that the same challenges in infant feeding in the context of HIV persist and are still relevant [
37,
70,
72,
77,
78]. There was also the possibility of influence on the roles of the interviewers as health care workers in KCH, which may have contributed to the reported ‘positive’ effect of interacting with the health system by mothers on adherence to infant feeding guidance [
83].
Acknowledgements
The authors would like to sincerely thank all the health care staff and the patients at the HIV, antenatal, postnatal, maternal and child health clinics as well as on the pediatric wards at Kilifi County Hospital, for their participation in this study. We would also like to thank and acknowledge the support of Ms. Margaret Lozi, Ms. Jane Kahindi, Ms. Connie Kadenge, Mr. Shadrack Babu and Mr. Anderson Charo in conducting and transcribing the interviews. Our sincere gratitude also go to Dr. Kate Campbell for helping to develop the ideas for this study, Dr. Alison Talbert for her assistance with the logistics of the study and interpretation of findings and to Ms. Prinila Gwiyo for her advice on infant feeding in this context.