What influences feeding decisions for HIV-exposed infants in rural Kenya?

The study utilised qualitative methodology using a descriptive-exploratory approach to explore decision-making processes, constraints, knowledge, attitude and practice of infant feeding among carers and HCWs.

The study was conducted in Kilifi County Hospital (KCH), a government facility located on the Kenyan coast that serves a rural population of about 270,000. Over 80% of the population in Kilifi County live below the poverty line and rely on rain-fed subsistence farming. It is an area that suffers frequent food insecurity requiring emergency relief operations [40].

In 2014, the overall adult prevalence of HIV in the county was 4.4%, with higher prevalence among women (6.3%) than men (2.3%) [41]. In 2011, the antenatal clinic prevalence of HIV in KCH was approximately 6%, and about 24% of children admitted to the hospital with severe acute malnutrition were HIV infected.

Since 2003, KCH has been providing HIV services including the prevention of vertical transmission interventions and early infant diagnosis (EID). All nurses in the antenatal, postnatal and family planning clinics are trained in prevention of vertical transmission and almost all mothers attending KCH clinics are tested for HIV (> 90%). Women living with HIV are advised to give birth at a health facility and to register their infants for care soon after delivery at the KCH HIV clinic or the nearest health facility providing prevention of vertical transmission and EID services. In the KCH HIV clinic, 65% of the infants do not complete the 18 month follow-up period, with 43% of drop outs occurring within 2 months of enrolment [42].

At the time of this study, the national HIV infant feeding guidance was changing in line with the 2010 WHO guidance [13]. The prior recommendation had been exclusive breastfeeding in the first 6 months or replacement feeding only if it was acceptable, feasible, affordable, sustainable and safe (AFASS), and continuation of breastfeeding until the age of 12 months with the introduction of complementary feeds, if conditions of replacement of breast milk were still not met [43]. In addition, women living with HIV who were already on combined antiretroviral treatment (ART) were to continue with this, but for those who were not, it was recommended that they start zidovudine at 28 weeks’ gestation then, combine ART (zidovudine, lamivudine and nevirapine at the onset of labour with the continuation of zidovudine and lamivudine only up to 7 days [43]. For the infant nevirapine was to be administered within 72 h of birth plus lamivudine and zidovudine for 6 weeks [43]. This was subsequently revised in line with the WHO recommendations for the continuation of combined ART in the mother and nevirapine in the infant for the duration of breastfeeding [13].

In order to gather a range of views on experiences of infant feeding in the context of HIV, male and female respondents from both the hospital and community were purposively sampled. This included both carers and health care providers. Hospital respondents came from a range of departments within the hospital where nutritional counseling was offered to women living with HIV. These departments included the HIV clinic, maternal and child health clinic (MCH) antenatal (ANC) and postnatal clinics (PNC), and the paediatric wards. Respondents from the community in the hospital catchment area were selected to supplement the information from carers and health care workers; and to further explore community perceptions of infant feeding practices in the context of HIV. These comprised Kenya Medical Research Institute (KEMRI) Community Representatives (KCRs) who are carefully selected members who are typical of those communities and represent the geographical areas in which a diverse range of health studies are conducted [44].

A conceptual framework was developed based on the literature available on the factors that may influence infant feeding in low income countries (Fig. 1) [1, 2, 45]. We undertook structured observations of infant feeding counselling sessions at the KCH outpatient clinical areas including the HIV clinic. The information from these observations were then used to refine the conceptual framework which in turn helped guide the ideas to be explored in the focus group discussion (FGD) and in-depth individual interview (IDI. The information was also used in the analysis to contextualise some of the experiences that the respondents shared about infant feeding counselling sessions. These guides covered a range of topics including factors influencing infant feeding decision making and constraining adherence to infant feeding recommendations in the context of HIV. The observation tools and guides for the carers were translated from English to Kiswahili by an experienced translator, according to the principles of the WHO translation protocol [46]. An additional file shows the Kiswahili version of the observation tool (see Additional files 1 and 2). All interviews and group discussions were conducted by experienced facilitators and were undertaken in English, Kiswahili or in the local dialect, Kigiriama. Sessions were audio-recorded and additional notes of the interview dynamics and environment taken. In total 9 IDIs and 10 FGDs were undertaken. Interviews were transcribed verbatim and, when required, translated into English. Data were managed using NVivo (QSR International Pty Ltd. 2012) [47].

Data were analyzed using a thematic framework approach [48], which incorporated elements from the conceptual framework. This entailed extensive familiarization of the transcripts by HMN and JJ who independently immersed themselves in the data, made summary sheets of each transcript and took related notes of key emerging themes. These preliminary themes were then discussed amongst the lead authors HMN, JJ, KWM and MKM, who had independently reviewed two transcripts. Through a process of careful discussion, comparisons of emerging themes and emerging consensus, a coding scheme was developed. This coding scheme was used to code the data in NVivo. The coded data were drawn upon to develop charts of key data to support categorization and exploration of emerging issues and comparison across interviewees [48].

Mothers often lived with their extended families. In this context women living with HIV who had not disclosed their status to their family members found it difficult to undertake feeding practices that departed from the norm (such as sustaining exclusive breastfeeding for 6 months), even when they were motivated to do so. The practice of “communal” caring of young children in such homes increased the chances of another family member introducing other feeds to the baby in the mother’s absence. For instance, a young mother’s insistence on exclusive breastfeeding where this was contrary to the advice or instructions given by the senior woman risked damaging family relationships. The younger woman risked being perceived as stubborn, disrespectful, or too lazy to prepare “real” food for the baby.

. . . for example you have gone to fetch water from a place far away and you live with your mother in-law and leave the child with her, when you come back the child . . . has been given porridge . . . (Female HCP, FGD, HIV Clinic, P6).

Mothers often lacked autonomy in the decision-making processes around infant feeding in the context of HIV, as children in the study community were generally perceived as “belonging” to the father. Therefore, although fathers rarely attended infant feeding counseling sessions and had limited knowledge of the recommendations, they had to be consulted and their consent obtained for child health related matters, including decision making on new approaches to child care or feeding. This was particularly pertinent whenever financial resources were required, for example for mothers who had opted for exclusive replacement feeding.

In our set up a mother is supposed to breastfeed and if she opts for replacement feeding this can be an issue and [there is] the “mwenye culture” [child belonging to the father] if you are married you are not supposed to decide without your husband . . . (Female HCW, FGD, P2)

Conversely, when mothers opted to practice exclusive breastfeeding they did not feel empowered to go against their husbands’ instructions if he insisted on additional feeds. This contributed to mixed feeding through the introduction of complimentary foods such as porridge, even when the mother was aware of the increased risk of HIV transmission associated with this.

But some [fathers] even tell us to give porridge to boost the baby’s weight or to stop the baby from crying to allow both of you to rest . . . (Female HCP, FGD, ANC, P6).

In some instances, mothers suffered adverse consequences as a result of adhering to recommended infant feeding practices against the will of their spouses, such as being abandoned by the husband or receiving verbal abuse, leading to strained marital relationships.

To overcome challenges related to adhering to recommended feeding practices, mothers often disclosed their HIV status to particular close relations such as their birth mothers or partners, in order to get some support for their infant feeding choices.

. . . She knows [her older daughter knows about her HIV status]. . . She was sent to get water for my baby [by her sister] and she refused, I was told by a doctor not to give water to the child. . . its good if you would explain to somebody your condition and she assists you. . . (Female HCP, FGD, HIV clinic, P6).

However, not all mothers were able to disclose their status, particularly to their husbands. Some mothers therefore developed strategies to support their infant feeding choices for example, by framing those choices as general health advice that is given to all mothers at the health facility.

So we say it is advice that we (all mothers) were given at the clinic. . . (How infected mothers defend their choice to exclusively breastfeed their infants) (Female HCP, FGD, HIV clinic, P3).

Women living with HIV reported that the decision to breastfeed or give replacement feeding was primarily based on their desire to protect their child from HIV infection. At the time of the study, the turnaround time for confirmatory test results for infants’ HIV status varied from one to six months. Mothers reported that they considered breastfeeding because they were not sure of their infants’ HIV status soon after birth and that they would have considered not breastfeeding if they had been sure that their child was not infected from birth.

I plan to breastfeed my baby since it will be quite difficult to know the child’s HIV status right at birth . . . .I would not breastfeed him had I had the surety that, he is not HIV infected. It would be unfair if I don’t breastfeed him only to later realize that he is [after all] HIV infected . . . (Female HCP, IDI, HIV clinic, P2).

As per the counseling, I was informed that, it wasn't possible to know whether a child is infected on his very first day. I thus opt to breastfeed the baby since it isn’t possible to know the child's fate. I will exclusively breastfeed the baby for the first six months while awaiting the results. I will stop the breastfeeding when the results turn out negative. (Female HCP, IDI, HIV clinic, P2).

In contrast, but with the same intent, other mothers coped with this uncertainty by opting for exclusive replacement feeding over exclusive breastfeeding despite it being unaffordable and unsafe in this rural setting. This was because they worried that other family members who were unaware of the dangers of mixed feeding in the HIV context might give other feeds to the child increasing the risk of vertical transmission.

It’s because I am living with HIV and was given the option of either to exclusively breastfeed for six months or not to breastfeed at all. But I personally decided not to risk by breastfeeding so I didn’t . . . (Female HCP, FGD, MCH, P1).

Mothers living in the villages choose not to breastfeed at all because they fear someone giving the child things like water . . . (Female HCP, FGD, ANC, P1).

Contact with the health facilities especially during antenatal care visits was reported by carers to have been highly influential in their infant feeding decision making.

I started clinic when I was three months pregnant. I had no idea of the PMTCT program but upon enrolling for care, I met X and her colleagues who counseled me and I decided that I shall exclusively breastfeed my child for the first six months . . . (Female HCP, IDI, HIV Clinic, P1).

This was particularly true for multiparous women living with HIV who were keen to continue with infant feeding practices that had worked for them with previous infants, whether that was exclusive breastfeeding under 6 months or exclusive replacement feeding.

I will continue (exclusively) breastfeeding this one . . . I will follow the advice given as with the other child who eventually turned out to be negative and for this one I will also do that . . . (Female HCP, FGD, HIV Clinic, P6).

Although carers reported the importance of health system contact on their choices, several health workers felt that this did not impact on mothers’ infant feeding decisions as many mothers within the clinic still practiced mixed feeding. This observation suggests that socio-economic and cultural considerations can overwhelm stated intentions.

Support from peers also provided mothers with more information about the infant feeding recommendations, apparently influencing infant feeding choices. Women living with HIV enrolled in a mother-to-mother (M2 M) support group reported being better able to adhere to feeding recommendations.

Before the M2M program, we never had enough time to discuss about the infant feeding. It was just a short time of discussion with either a clinician or a counselor . . . before the M2M, most mothers were mix feeding their babies . . . (Female HCP, IDI, HIV Clinic, P2).

The majority of the women living with HIV chose to practice exclusive breastfeeding for infants under 6 months, stated that they did so because it was a cheaper feeding option.

If you breastfeed the child it helps in reducing the [household] expenses of buying powdered milk until the child is six months, then you can change to cow’s milk . . . (Female FGD, HIV Clinic, P14).

Financial constraints that were perceived to impact maternal nutritional status were also perceived to impede effective breastfeeding through reducing the quality or quantity of milk.

Others would like to breastfeed their babies but you find that because of dietary factors they are not able to feed well so they say because I’m not sure if I’m going to be able to feed well for the six months I will rather choose the supplements maybe because they are cheaper at their homes . . . (Female HCW, IDI, P1).

Job insecurity was linked to financial challenges, reportedly impacting infant feeding decision making because cost of the replacement feeds relies heavily on presence of a stable income over the 6 months period. Where this was not available, or when carers’ situations abruptly changed, this contributed to mixed feeding with consequences for vertical transmission.

There was a client who opted for replacement feeding and after some time she relocated and I think she got sacked from work or something and she could not afford buying the milk and she went back to breastfeeding . . . (Female FGD, HCW, P3)

Health care workers (HCWs) commonly reported that the ‘constantly changing policies and guidelines’ impacted on adherence to recommended feeding practices. Some HCWs reported being confused about the guidelines, specifically whether HIV infected mothers should breastfeed their infants or not. This implies the potential for mothers to receive contradictory health messages which could in turn impact on their infant feeding practices. The changing policy landscape was reported to reach HCWs in the more rural or peripheral facilities more slowly than those working in larger or more urban facilities.

This issue of infant feeding and HIV has brought mixing of issues you find we are well updated here at the [county] hospital but our colleagues, in more remote areas may not be as updated . . . (Female HCW, FGD, P9).

Health workers identified two challenges with regards to changes in recommended infant feeding practices with implications for maternal feeding choices: i) inadequate training and updates on the recommended infant feeding practices, both in general, and in the context of HIV and; ii) overwhelming workloads resulting in inadequate time for counseling and long waiting times for mothers. These challenges impact on the quality of nutrition advice, counseling and support that HCWs are able to give to the mothers.

. . . The challenge also comes with the workload so you don’t take time with a patient. You don’t have even time to demonstrate how to prepare the feeds . . . (Female HCW, FGD, P12).