Contradictions and conflict: A meta-ethnographic study of migrant women’s experiences of breastfeeding in a new country

The aim of this meta-ethnographic study was to explore the diversity of migrant and refugee women’s experiences and practices related to breastfeeding in a new country with a view to informing health care policy and practice and identifying ways in which migrant and refugee women can be better supported to achieve their breastfeeding goals.

The literature search was conducted between September and November 2011 and revised in May 2012 using the following databases: CINAHL, MEDLINE, PubMed, SCOPUS. The Cochrane Library search terms included: migrant, immigrant, Non English Speaking Background, ethnic, ethnicity, refugee, recent arrivals, breastfeeding, infant feeding, postnatal and postpartum. Inclusion criteria were: a primary focus on the breastfeeding experiences and/or practices of migrant or refugee women. The term migrant can be ambiguous [40, 41]. In this review, the term migrant was defined as women born in low or middle income countries, who had migrated permanently to a high income host country. Papers that included both migrant women and those born in the host country, including women from minority ethnic groups, were retained as long as the authors had identified the source of the data they reported. Studies were limited to those published in English between January 2000 and May 2012. While the prime focus of a meta-ethnographic study is on synthesising qualitative studies, we also included studies that reported both quantitative and qualitative data if a subset of the larger sample were interviewed or participated in focus groups yielding rich qualitative data. As not all databases enabled the use of search limiters, articles not satisfying these criteria were also removed during the first exclusion.

The search resulted in 3052 papers. Following removal of duplicates (2355 papers), 697 papers were reviewed by title and abstract to determine if they met the inclusion criteria. Thirty-three papers were identified and read in full. Twenty-two papers were then excluded as they did not meet inclusion criteria. Through ‘back-chaining’ [38], we identified and reviewed 10 additional papers, with five appropriate for full review (see Figure 1). Sixteen papers were included in the review.

Quality appraisal was undertaken using the Critical Appraisal Skills Programme (CASP) tool for qualitative research [42]. Sixteen papers were reviewed in full and five papers were excluded due to inadequate reporting of methods, for example: sample was not described; no detailed description of data collection and or analysis techniques; the paper reported limited data on breastfeeding; or did not report qualitative data adequately, that is there was no audit trail of the derivation of findings. All studies included in the review had received institutional ethics approval. This resulted in 11 papers being suitable for use in the meta-synthesis (8 qualitative and 3 studies reporting both qualitative and quantitative data) (see Figure 1). Of the 11 papers included, all came from English speaking, high income countries. Only one outlined a theoretical framework; the study by Groleau et al. applied critical theory to analysis and discussion [43]. Two studies indicated that they applied ethnography as a methodology and one described the methodological approach as ethno-nursing methodology [19, 29]. The remaining studies, other than those using a survey design, described their studies as either descriptive or interpretive qualitative studies. All the included studies provided an adequate audit trail of the derivation of study findings from the data. We found however, early in the process of analysis that the majority of themes identified in the 11 papers were descriptive in nature and had not been adequately abstracted or conceptualised. This influenced the approach we took to data extraction and synthesis described below.

Data extraction and synthesis was guided by the seven phases used in meta-ethnography articulated by Noblit and Hare [36], which include: identifying the area of interest, deciding what is relevant, reading and re-reading the studies, deciding how the chosen studies are related, translating the studies in relation to one another, synthesising translations and presenting the synthesis [44]. Noblit and Hare [36] suggested that translation involves examining the key concepts in relation to others in the original study and across studies. The way of translating key concepts or interpretive metaphors from one study to another involves an idiomatic rather than a word-for-word translation. Two authors (HO and VS) read each paper and then agreed on the index paper published by Rice and Naksook [19] from which the first set of themes were identified.

The majority of themes reported in the 11 studies were descriptive in nature for example, ‘Reasons for not breastfeeding’ [43]; ‘Health beliefs regarding breastfeeding’ [25] and ‘support for new mothers’ [29]. In four of the 11 papers [14, 19, 30, 45] the authors had included some themes or sub-themes that were more interpretive, most commonly using ‘in vivo’ codes or in a few instances a metaphorical statement. For example, the theme ‘Mother’s milk is best’ identified in this meta-ethnography was developed from themes in the work of Choudhry [14],‘“Maa ka Dood” – the mother’s milk‘; Rice [19], ‘Nom reak gerd: beliefs about colostrum’ and Chen [30] ‘purity of breast milk’.

Other authors have similarly reported that the level of abstraction offered in original qualitative research often lacks interpretation [46]. Atkins et al. [46] and Dickson-Swift et al. [47] suggest that to address this some meta-ethnographic studies make use of Schutz’s [48] notion of first, second and third order constructs. First order constructs represent participants views’ or perspectives of the phenomenon under investigation that is, the original quotes that the author’s use to illustrate or represent the theme. Second order constructs, typically used in meta-synthesis, are the thematic statements or abstractions reported as study findings or conclusions by the original authors that should provide more insight or explanation about the phenomenon. In meta-synthesis these themes or second order constructs are then synthesised to produce third order constructs.

Applying this approach the research team worked systematically through the papers, reading and re reading papers to create a list of themes and or metaphors; these were juxtaposed and examined to see how they related to each other [49]. We identified new themes within the first order constructs or original data and where possible we developed second order constructs juxtaposing the authors’ original themes. For example, the theme ‘Producing breast milk requires energy and health’ incorporated authors’ themes such as ‘Mothers' diet and breast feeding’ [19]; ‘Breast is not always best’; experiences of information and role conflict- role conflict’ [14]; and Mothers who bottle feed’ [43] as well as incorporating original data from other authors relevant to this theme. Interpretations by the authors of individual studies were also utilised to ensure the quotes were examined in context.

In almost all of the included studies there was consensus from participants that mother’s milk is best for the baby. In some studies women emphasised the health benefits, citing examples of information received from family [19, 30, 43] or related to common cultural beliefs in their country of origin. For example, one Thai woman living in Australia stated, ‘people in Thailand say that the mother’s milk has a lot of immunity, and that is why the baby does not become ill easily’ [19], p 14]. In many studies, breastfeeding was described as the ‘natural’ way to feed a baby [30]. Chinese women in Canada, for example, saw breastfeeding as part of ‘natural or dynamic laws by which humans live…the process of change that reflects and influences both a mothers’ health and that of her baby…’ [30], p. 1023].

In the majority of studies, migrant women took the decision to breastfeed for granted. For example, Latino women living in the United States [45], p. 27] stated, ‘I never doubted [breastfeeding]’, ‘you just know you will breastfeed’, and ‘it's the way it's supposed to be’. For Chinese women living in Canada, breastfeeding was part of the ‘natural work’ of a mother. As one woman stated, ‘the crucial thing I give is devotion, my love for my baby. I perform a mother’s duty…I do not think about nutrition and I know little about it’ [30], p. 1023]. These women were surprised that so many women in their host country did not breastfeed, ‘I don’t understand…after delivery, they directly formula fed their baby. They told me they do not have breast milk’ [30], p. 1024].

Breastfeeding was also important to mothers for the mother-child relationship [19, 27, 30]. In many of the studies, women described breastfeeding as ‘a gift’ to the child and as ‘…a beautiful feeling to see him look at me from my breast…It is the most peaceful rewarding moment in my life‘ [27], p.346], and one woman felt that with breastfeeding, ‘(the baby has) more ‘tendresse’ to you. It is the best relationship, the baby likes you so much, only wants the mother’ [27], p. 346]. Giving breast milk was likened by Thai women in Australia [19] and Pakistani grandmothers in the UK [29] to sharing one’s blood with the infant:

…breast milk comes from our blood. The blood changes into milk. My mother told me about this. She used to say to me that ‘Do you know that every drop of milk that you take is from my blood’ and she would ask me if I love her. I am now teaching my kids this belief too. I am very proud of this close relationship between my kids and me [19], p. 15].

So important was the breastfeeding relationship with the child that one Latino mother decided to formula feed after birth as she was to return to work. She spoke of not wanting to breastfeed so that her child would not feel deprived, ‘it is better to give formula quickly so that the baby doesn’t suffer when you have to leave them to be cared for by another person’ [25], p. 323].

For many mothers in the included studies, breastfeeding was also associated with the positive development of the baby’s personality [14, 19, 25, 29]. Women stated that breastfeeding ’deliver(s) goodness and mak(es) the baby grow up to be a good person’ [14], p.79], and ‘oh he/she’s like that because he had the mother’s milk. . . such as the way he develops his personality or intelligence’ [14], p.79]. As one Thai woman in Rice’s Australian study described, ‘in my hometown, people say a bottle-fed infant will be stubborn because he or she is fed with cow milk, not human milk’ [19], p. 14].

Only in two studies was formula milk described by some participants as equal to, or better than, breast milk [13, 14]. In the UK, Choudhry and Wallace reported ambivalence among women from countries such as India, Pakistan and Bangladesh about the value of breast milk and breastfeeding [14]. In this study, women spoke about the infant feeding practices they had observed, for example, ‘in Pakistan most people would breastfeed but here (referring to the UK) everyone bottle feeds’ [14], p. 81] and another added:

I thought if they’re (peers from the Anglo Culture) doing it and there’s no harm then there’s no need for me to go through the pain of trying to breastfeed when I could give them formula [14], p. 81].

Some women in Choudhry and Wallace’s study perceived their breast milk to be inadequate or less nutritious, a message promoted by some family members. In four studies women reported that formula feeding was more convenient, offering more freedom to ‘get on with our lives’ [50], p. 392], an easier option and a way to share the responsibility of infant care [14], a way to manage the daily hassles and stresses of life with a baby in a new country with limited support and at times, to address breastfeeding difficulties [13, 14, 25]. It was rare however, for women in these studies to report experiencing pain with breastfeeding. Pain was only reported in three studies and was not directly related to participant’s personal experiences [14, 19, 25]. For example, one woman living in the US stated, ‘I think women give formula because of the pain, your breasts can suddenly hurt a lot when breastfeeding’ [25], p.324]. Some women also reported formula feeding if they had to return to work, ‘I was unable to breastfeed for a longer period because I had to work after childbirth to assist with family finances’ [13], p.274].

Two studies highlighted the implications of visibility and affordability of infant formula in the host countries in this case, Australia and the US [13, 25]. Latino women in Vaughan’s study observed that:

In Mexico it would be more likely that you would give your child breast milk, compared to here; simply because there the milk (formula) is much more expensive and it is not like here where the government gives out milk (formula) [25], p. 324].

In many of the included studies, migrant women talked about receiving mixed messages about different breastfeeding practices, such as giving colostrum, length of a breastfeed, age at which breastfeeding should cease and breastfeeding in public. At times contradictions or conflict were easily resolved and women felt comfortable to maintain their traditional breastfeeding practices whilst at other times this conflict resulted in heightened confusion or distress.

In eight of the 11 studies women emphasised the importance of adhering to traditional postpartum practices related to diet, rest and body care to ensure that their body had the energy to produce enough breast milk. Women in Chen’s [30] and Groleau et al.’s [43] studies who had migrated to Canada prioritised care for the maternal body over specific practices related to caring for the infant including feeding. Stated simply by Latino women in the US, ‘if you are healthy, the baby is healthy’ [45], p. 28]. The importance of diet and rest was expressed by many of the mothers. As one woman in the study by Condon et al. in the UK explained, ‘you need well prepared food, time to feed and rest yourself, if you don’t have these you won’t be able to produce enough milk for the baby’ [27], p.346]. Chen reported the importance of diet for Chinese women stating:

In traditional Chinese culture, the mother’s body is weak after delivery. She must have tonic soup (Chinese herb soup) in order to recover her vitality and energy… Then the breast milk naturally comes out [30], p. 1024].

Study participants also emphasised the importance of consuming hot foods, resting and exposing their body to ritualised steam baths, massages and having a constant source of heat under the bed (for example, a pot of burning charcoal or an electric blanket) so that the body of the postpartum mother will recover, closing pores and reconstituting the lost blood and heat and thus regain the strength and equilibrium necessary for health [19, 30, 43]. This was most evident in the studies that included women who were born in Vietnam, Thailand and China [13, 19, 30, 43]. As Chen described,

During this period, after eating the tonic soups [prepared by family members], I had a lot of breast milk. When my first baby sucked my breast on this side, milk naturally flowed out from the other breast [30], p. 1025].

The importance of women resting for at least one month in the postnatal period was emphasised. Respondents in Groleau et al.’s study explained that to maintain heat ‘mothers are given a traditional steam bath twice a day usually performed by an elderly woman in her family, usually her own mother with ingredients such as lemon, guava or grapefruit’ [43], p.520]. This ritualised ‘time out’ lasts longer for primiparous mothers (100 days) as opposed to one month for multiparous mothers. According to Vietnamese tradition, breastfeeding is contraindicated for women experiencing postpartum fatigue, because the production of milk requires a lot of ‘vital energy’ [43], p.520].

Women in many of the included studies reported becoming anxious if they could not adhere to traditional practices [13, 19, 27, 29, 45]. For example, in the study by Groleau et al. [43], none of the Vietnamese women were exclusively breastfeeding and only those who were multiparous initiated breastfeeding. Groleau et al. [43] reports that initially women explained they were not breastfeeding because they had to return to work or that they did not have enough milk yet; however, when probing deeper, the women spoke about their concern about losing heat from their bodies because they were not participating in traditional postpartum practices. In contrast, the Latino women in Schlickau and Wilson’s study were somewhat ambivalent about these practices. While acknowledging the traditional ways of keeping healthy, they simultaneously faced the problem of returning to work. As one woman stated, ‘I was on the farm with the cows and the children and could not stay in bed for 40 days…You have to take care of yourself, though, during that time’ [45], p.28].

As a consequence, some women prepared for insufficient milk, ‘you prepare for not having enough milk, so you buy bottles and powdered milk’ [27], p.346] and others reported that they were more likely to bottle feed:

A mother’s body is the foundation of breastfeeding. I breastfed my previous child but this time I fed formula to this baby because my body was very weak after pregnancy. I also did not receive good quality Chinese traditional postpartum care [30], p. 1025].

I only breast fed my baby for two months because I was concerned about my body and due to this I did not produce enough milk for her..I do not eat those things that people say will help to make a lot of milk, so my milk was diminishing’ [19], p.18].

Messages from others including health professionals can also result in the mother assuming she won’t have enough milk. This was evident in Straub, Melvin and Labbok’s (2001) study in the US when a Cambodian woman stated:

I think the nurse brought it, the formula. And they asked me [if] I want to breastfeed and I said yes, she asked if I wanted formula and I said yes…Well, I was thinking that maybe I wasn’t producing enough milk…I just didn’t know if I had enough [51].

In most of the studies the importance of being cared for by family was emphasised. In three studies [14, 19, 25] women described the role that their partners played in supporting breastfeeding. Other women indicated that their partners were not involved and indeed were often a hindrance [14, 25]. Most support came from the woman’s own mother and or mother-in-law. As one participant in Chen’s study described:

Babies are bigger and stronger in China. They look healthier. Probably in China a baby eats well [and has good quality of breast milk from mother] and [also] has good quality of care from the grandmother, grandmother-in-law. But in Canada the babies do not look healthier and stronger; they look slim. Westerners do not spend a lot of time taking care of their children and they do not have many family members to help them [30], p.1026].

This was also well illustrated in Groleau et al.’s [43] study where women who had previously given birth in refugee camps spoke of the support they received from other female relatives or community members who were living with them in Vietnam. However, this was not their experience in Canada. Groleau et al. [43] found that the constraints of living in a new culture brought about by the absence of close family and the support it normally provides meant that many Vietnamese women decided to bottle-feed rather than breastfeed. This was seen as the best option for preserving their own and their baby’s health. As one woman in Groleau et al.’s study describes:

In Vietnam, she received a lot of visitors [after the birth]. In the Philippines [in the refugee camp], she received family members. In Canada, she doesn’t receive anyone; she is all alone with her husband [43], p.521].

The decision to formula feed appeared to be influenced by views of family members living in the new country, such as her mother-in-law. This is also well illustrated in Choudhry and Wallace’s study, where women living in the UK made statements such as, ‘my mother-in-law encouraged bottle feeding for my own benefit, I was able to share the feeding with her which was good for me’ [14], p.79] and:

I only breastfed for a while, because when I started I realised that my mother in law was right, I wasn’t able to cope with everything myself. Formula-feeding meant that whilst I was doing other work, my mother in law was able to help out and feed my baby [14], p.80].

These persuasive views of a mother-in-law may also contradict views of the woman’s own mother who may be still living in the country of origin. One mother in Choudhry and Wallace’s study in the UK explained, ‘When I lived in Pakistan my mum always encouraged me to breastfeed my baby but after I got married and came here I was told something different by my mother in law’ [14], p.80]. Another woman stated, ‘my parents advised that I should breastfeed but I found it really hard to fit into my life and the things I have to do around the house, like housework and looking after my mother in law’ [14], p.80].

The findings of this meta-ethnographic study are limited by the nature and depth of data collection and analysis in each study. As discussed, most studies were descriptive in nature, and did not identify metaphorical statements in their original work. This made it difficult to conduct reciprocal translation. To facilitate the synthesis we adopted the approach described by Atkins [46] working with first order constructs presented in the papers.

It is also important to emphasise that in this meta-ethnography we have synthesised studies reporting the experiences of women who come from diverse cultural backgrounds and it is inappropriate to attempt to create one picture of migrant and refugee women's breastfeeding experiences and practices. We also note that there are inconsistencies in the approach and terminology used in the original papers when identifying and categorising the ethnicity or cultural background of participants. In some studies, the ethnic background of participants was only identified by country of birth [13, 45]. Alternatively, some researchers asked participants to identify their ethnic or cultural background when completing a demographic questionnaire but it is not clear if these were open ended or pre categorised responses [25]. Two studies [14, 27] explicitly stated that women were asked to self identify their cultural background or ethnicity. Others made no mention of how these details were obtained [30, 43, 50]. Commentators have emphasised the complexity of assigning ethnicity to study participants [40, 41] indicating a preference for the value of self assigned ethnicity, although recognising that this adds a level of complexity to data collection and interpretation. This complexity often results in researchers using inappropriate labels to describe cultural background. For example, the terms Asian and South Asian were used in a number of papers to describe people from the Indian sub-continent, many of whom, if asked, would vary in how they described their ethnic identity.