The midwife–woman relationship in a South Wales community: Experiences of midwives and migrant Pakistani women in early pregnancy

Abstract Background In 2015, 27.5% of births in England and Wales were to mothers born outside of the UK. Compared to their White British peers, minority ethnic and migrant women are at a significantly higher risk of maternal and perinatal mortality, along with lower maternity care satisfaction. Existing literature highlights the importance of midwife–woman relationships in care satisfaction and pregnancy outcomes; however, little research has explored midwife–woman relationships for migrant and minority ethnic women in the UK. Methods A focused ethnography was conducted in South Wales, UK, including semi‐structured interviews with 9 migrant Pakistani participants and 11 practising midwives, fieldwork in the local migrant Pakistani community and local maternity services, observations of antenatal appointments, and reviews of relevant media. Thematic data analysis was undertaken concurrently with data collection. Findings The midwife–woman relationship was important for participants' experiences of care. Numerous social and ecological factors influenced this relationship, including family relationships, culture and religion, differing health‐care systems, authoritative knowledge and communication of information. Marked differences were seen between midwives and women in the perceived importance of these factors. Conclusions Findings provide new theoretical insights into the complex factors contributing to the health‐care expectations of pregnant migrant Pakistani women in the UK. These findings may be used to create meaningful dialogue between women and midwives, encourage women's involvement in decisions about their health care and facilitate future midwifery education and research. Conclusions are relevant to a broad international audience, as achieving better outcomes for migrant and ethnic minority communities is of global concern.


| BACKGROUND
Non-UK-born communities continue to grow within the United Kingdom. 1 Indeed, a recent report from the migration observatory suggests that one in seven (13.1%) of the UK population in 2014 were born abroad, 2 and in 2015, over a quarter of births (27.5%) in England and Wales were to mothers born outside of the UK. 3 Furthermore, during 2015, the number of births to non-UK-born women in England and Wales increased by 2.5% from the previous year, whilst births to UK-born women decreased by 0.4%. 3 In the UK, minority ethnic and migrant women consistently report lower maternity care satisfaction [4][5][6][7] and less choice in their maternity care 4,8 than their White British counterparts. In addition to poor experiences of maternity care in the UK, a wealth of research details poor pregnancy outcomes for these women, including an increased risk of complications during pregnancy, 9 unplanned caesarean section 10 and having their baby cared for in a neonatal unit. 11 Substantially higher maternal mortality rates are also observed for minority ethnic and migrant women. 12 For example, between 2011 and 2013, the estimated mortality rate for White women in England was 7.8 deaths per 100 000 maternities; 12 for Black women, this rate was more than tripled at 28.3 12 and was also significantly higher for both Pakistani and Bangladeshi women, 15.9 and 14.7, respectively. 12 Furthermore, the risk of maternal mortality seems to be increasing for some migrant women in the UK: between 2011 and 2014, the relative risk of maternal mortality for Pakistan-born women living in the UK increased from 1.53 to 2.24. 12,13 Data consistently suggest that minority ethnic and migrant women are also at an increased risk of perinatal mortality.
In 2013, mothers of Black ethnic origin were twice as likely to have a stillbirth than mothers of White ethnic origin, 14 and women of Asian or Asian British ethnic origin had up to 64% higher stillbirth rates than their White counterparts. 15 Previous research has attempted to ascertain why these differences exist, and suggests that risk factors for poor outcomes such as access to health care, racism, cultural beliefs and poor underlying health are likely to impact migrant women to a higher degree than UK-born minority ethnic women. 16 For example, Hayes et al 16 suggest that women who have recently arrived in the host country and who do not know how to, or cannot, legitimately access care are those most at risk of negative maternity outcomes. This view is echoed by women 17 and health-care professionals 18 alike; UK-born minority ethnic women felt that being born in the UK allowed them a better understanding of how to access care and information, 18 and maternity care professionals suggested that women's language competency and familiarity with the system were "key advantages" in care provision. 17 Poor social networks, commonly seen in new migrant populations, 16 have also been proposed as a risk for substandard maternity outcomes. 16,19 Associations have also been suggested between poor pregnancy outcomes and factors such as genetic risk, 20,21 differences in socioeconomic status, 22 language barriers 23 and stereotyping/ racism. 24 However, ethnic inequality in these outcomes remains even once these contributing factors have been accounted for. 25,26 Consequently, it would seem that additional, largely unexplored, explanations may exist for the observed differences in pregnancy outcomes. 26 Existing literature suggests an association between midwifewoman relationships and pregnancy outcomes, 27 impacting not only on uptake of antenatal care, 27,28 but also influencing the quality of care received once services have been accessed. 17,[29][30][31] There is a strong suggestion in the literature that a poor relationship may result in poor outcomes for women. 29 Despite this, research exploring midwife-woman relationships is limited. This is surprising considering the importance placed on the quality of the relationship by both women and midwives in previous research. 26 The limited research suggests that midwives may have more difficult relationships with migrant and minority ethnic women, compared to their White British counterparts. 32,33 It is possible, therefore, that poorer quality of midwife-woman relationships for minority ethnic and migrant women may present an alternative contributing factor towards ethnic inequalities in outcomes, a perspective that has largely been under-explored in previous studies. 26 As such, this study was designed to address the paucity of literature examining midwife-woman relationships for migrant women by exploring relationships between first-generation migrant women and midwives in the South Wales region of the UK, focusing on identifying the factors contributing to these relationships, and the ways in which these relationships might affect women's experiences of care.
The focus was on Pakistani women specifically, as mortality reports published at the time of study design suggested that Pakistani women were at significantly increased risk of both infant mortality 34 and maternal mortality 29 when compared to all other ethnic groups in the UK.

| METHODS
Given the exploratory nature of the enquiry and the limited existing evidence base, a focused ethnographic approach, using qualitative research methods, was taken to data collection (a full data collection schedule is presented in Figure 1). The ethnographic approach is of particular relevance to the current research question, as it was developed as a way to understand the social life of humans within specific cultures 35 and analyse cultural norms, 36 allowing for cross-cultural comparison and providing a better understanding of behavioural differences and intergroup conflicts. 35,37 Focused ethnography differs to the traditional ethnographic approach, in that it is conducted within a discrete community or context, whereby participants have specific knowledge about an identified phenomenon. 38 In this way, focused ethnography aims to explore participants' beliefs and practices by viewing them within the context in which they actually occur. 39 As such, proponents of focussed ethnography argue that it is especially well suited to studying the practice of health care as a cultural phenomenon and to understanding the meaning that members of a subculture or group assign to their experiences. 40 The focused ethnographic approach has, therefore, been increasingly adopted in health services research, 41 as it can assist health-care practitioners to identify and meet the needs of individuals from a certain culture by giving insight to behavioural differences. 42 In line with the ethnographic approach and to enhance thickness of data, 39,41,43 several data collection methods were utilized ( Figure 1). These included the following: (i) preliminary fieldwork in the communities under study; (ii) reviews of relevant media (ie newspaper articles, policies); (iii) semi-structured interviews; (iv) non-participant observations of antenatal booking appointments; (v) reflexive fieldnotes (written throughout study design, recruitment data collection and analysis).
Preliminary fieldwork consisted of approximately 80 hours of participation in activities and events in migrant Pakistani communities and local maternity services over a 3-month period, which were designed to facilitate embeddedness in the cultural worlds of both participant groups (migrant Pakistani women and UK midwives), in order to "interpret the world in the way they do" [Ref. 35 , p.8]. As recommended by other health service researchers, 32,41 more formal observation periods would be undertaken at a later stage of the study, alongside indepth interviews, to facilitate strategic data collection as the research questions became progressively more focused. Interviews and observations were repeated to capture any changes in midwife-woman relationships over time, 44 and reflexive accounts were written and shared with a project support group (described below) to ensure that all potential personal and interpersonal influences were explored and considered appropriately.

| Participants and recruitment procedures
The study site was a maternity unit in South Wales, which provides care to around 6000 women annually. Services include an alongside midwifery-led unit, and tertiary foetal medicine and neonatal services.
The health board employs around 14 000 staff in total, including around 300 midwives. 45 In the city served by this health board, 80% of the population report their ethnicity as "White British." The other main ethnic groups are "Other White" (4%), "Indian" (2%), "Pakistani" (2%) and "African" (2%). 46 For interviews, participants included 7 first-generation migrant Pakistani women receiving maternity care in South Wales and 11 midwives with experience of providing maternity care to migrant Pakistani women. Inclusion/exclusion criteria are outlined in Table 1. Naturally occurring interviews resulted in data also being included from the mother of one of the participants (n = 1) and a migrant Pakistani interpreter (n = 1). The number of interviews was not predefined but was limited to an extent by the planned duration for data collection; however, no new concepts were emerging from the data when recruitment stopped. 47  Migrant Pakistani women were interviewed at two time-points, once after their first antenatal appointment, and then again after their second/third antenatal appointment. Interviews lasted between 20 and 90 minutes, and a flexible topic guide was used to guide the conversation. The aim of the first set of interviews was to explore women's initial expectations of maternity care in South Wales, and whether these expectations were met during their first contact with their midwives.
The second set of interviews explored how women's initial expectations of maternity care in South Wales were managed throughout their pregnancy, and how women perceived their midwife-woman relationship to have affected the pregnancy so far. Language interpreters were offered to all women at all points of engagement with the study. In cases where an interpreter was used (n = 2), the researcher would ask a question, which the interpreter would translate into the required language (in both cases this was Urdu). The woman would then reply in Urdu, which the interpreter would translate back to English for the researcher. Anonymized audio extracts of interviews were sent to an independent interpreter for validation.
To recruit midwives, invitations to participate were initially sent out by the Head of Midwifery, with a request to contact the researcher to express interest. A snowballing approach was then used to recruit other midwives eligible for participation. Interviews with midwives lasted between 20 and 60 minutes and explored their experiences of working with migrant clients and with Pakistani women specifically.
Midwives were asked about their relationships with these women and were asked to discuss the barriers and facilitators to establishing these relationships.
A total of 15 observations of antenatal booking appointments (20-60 minutes each) took place in antenatal clinics across the local health board over a period of 3-6 months. In line with the focused ethnographic approach, the enquiry became progressively focused on specific research questions as data collection progressed. 41 In practice, this meant that flexible "observation guides" (prompts about what to focus observations on) were gradually developed to collect data more purposively. 38 These guides were informed by initial observation data, interview data and previous observation research in maternity care. 49

| Analysis
Thematic analysis 50 of the data was undertaken, resulting in a datadriven inductive approach. The first step of the analysis was to personally transcribe and listen to interviews in their entirety, at least 3-4 times. Next, early themes and topics of importance to participants were highlighted, and comment boxes were used to record corresponding notes. In the third stage of analysis, transcripts were imported into the data-management software NVivo 10, where selections of text were coded to represent instances of a concept. 51 Codes were reviewed in terms of their relationship to other codes and combined to create more developed themes. 51 From this analysis, distinctions between the different levels of themes appeared (eg main overarching themes and subthemes within them).
Data extracts were regularly shared with members of the project support group to discuss interpretation of the data and to confirm the emerging themes -improving consistency and reliability of findings. 52 In order to maintain a progressive, iterative, process of analysis, 53 data were collected and analysed concurrently, allowing emergent findings to guide, where relevant, the next piece of data collection.

| FINDINGS
Migrant Pakistani participants had a mean age of 27 years old, and mean length of residency in the UK was 7 years. Social characteristics varied, with women from a range of social and economic backgrounds; for example, some had partners in well-paid professional jobs, whilst others had partners who were manual labourers. Only one of the women was employed. Further demographics of participants are presented in Table 2. All midwife participants were UK-born and worked in the community. It was decided that further demographic information for this population (such as time in practice or seniority) would act as possible identifiers in such a small sample, so this information is not reported.
Both participant groups placed high importance on midwifewoman relationships. Not only was this relationship viewed as significant in ensuring the best outcome for women, but it was also suggested that difficult relationships could potentially result in higher

| Family relationships
Mothers-in-law and domestic partners were seen to play significant roles in women's maternity care by all participants. Whilst women found these family members to be a source of support, midwives, however, perceived this involvement as having a negative impact on midwife-woman relationships. This perception is typified in the following midwife interview extract where mothers-in-law are described as "dominating" antenatal clinic appointments, preventing midwives from "really" getting to know women. The pregnancy and childcare advice given by mothers-in-law was also perceived as a potential barrier to midwife-woman relationships, as this advice often conflicted with information provided by midwives.

| Culture and religion
Although midwives tended to express either positivity or neutrality towards Islam, many expressed concern regarding traditional pregnancy and post-natal practices rooted in the Muslim culture.
Some practices were viewed as unsafe and/or unhygienic; for example dressing newborns with glass/string bracelets, fasting whilst pregnant, shaving the newborn's head or placing honey on the newborn's tongue immediately after birth. Midwives perceived such practices as impacting negatively on the midwife-woman relationship.

Pregnant women shouldn't fast. And I always find that if they are fasting then I'm kind of lecturing them "noyou shouldn't be fasting" and that kind of thing…and they do get a bit funny about it… you can see that they're not
happy that I'm saying "no you shouldn't".

Mary (M)
When speaking about practices which they knew to be negatively viewed by health professionals, women tended to discuss the behaviours and attitudes of others. For example, whilst claiming they would personally adhere to the midwife's advice during pregnancy, women suggested that "other" Pakistani women were more likely to agree with the midwife superficially, whilst continuing practices out of the health professional's sight.

Whatever [midwife] say, [Pakistani women] won't follow
you. They will say "ok yes we will do" in front of you…but when they go back home they won't follow you! They will follow whatever the elders say -they will follow that!

Hana (W)
Continuation of practices against advice was also cited by midwives as a barrier to good midwife-woman relationships. Midwives expressed anxieties about balancing professional accountability with providing choice, personalized services and safe care to women and babies.
Even without personal experience of women's traditional practices or views leading to adverse outcomes, midwives continued to negatively stereotype migrant Pakistani women. For example, midwives described how they thought Muslim women were more likely to accept a stillbirth as "God's will" and were therefore less likely to seek medical help for pregnancy concerns. Women, on the other hand, suggested that this belief provided them with a way of coping with negative outcomes, but that it did not negate the need for antenatal care.
Interestingly, data suggested a bidirectional effect between midwives' views on traditional pregnancy practices and the establishment of the midwife-woman relationship: when a positive midwife-woman relationship was established, midwives tended to view traditional practices such as head shaving more positively. However, the extract below demonstrates that if a positive relationship was not yet established, midwives could harshly judge some of the women's decision making.

If you haven't built up a relationship with somebody during
pregnancy…you tend to actually be very hard on some of the decisions they make, and I think we need to be honest about it.

| Understanding different health-care systems
Whilst midwives spoke about UK maternity care as the "gold standard," some women stated a preference for the Pakistani maternity care system, particularly citing the lack of woman-initiated contact at the beginning of pregnancy as a potential problem for the on-going relationship with midwives.
We can't contact our midwife, or visit them frequently…. [

| Connecting the themes
Despite commonalities in the maternity care issues identified by women and midwives (family involvement, culture and religion, differing health-care systems), differences were observed in the way midwives and women prioritized these topics in terms of their influence on the midwife-woman relationship. For example, midwives most commonly attributed poor relationships with migrant women firstly to misunderstanding of, or lack of adaptation to the health-care system generally and secondly to misunderstanding the nature of the midwife-woman relationship specifically (eg that family members were not expected to contribute when midwives met with women). In contrast, women tended to centre their accounts around the importance of family involvement in their maternity care and their expectations that services adapt to their needs.
Two interweaving themes were also identified from the data, which linked the main themes. The first of these was labelled "authoritative knowledge" and was used to refer to instances where competing sources of knowledge seemed to disrupt mid-  about decisions regarding practices they are unfamiliar with, compared to more familiar decisions made by UK-born women. This may explain research findings which suggest that midwives describe their job and relationships with women as more demanding, difficult and stressful when working with migrant women. 33,64 Consequently, these findings suggest that more needs to be done to address midwives' concerns around the safety of unfamiliar practices, for example having standardized evidence-based information on such practices available to give to women. Such actions may allow midwife-woman tensions to be reduced and positive relationships to be maintained.

| DISCUSSION
The lack of lone contact with migrant women was also highlighted as a concern by midwives, who suggested that the presence of male partners not only negated the possibility of conducting routine enquiry but also prevented the establishment of a good midwifewoman relationship. Minimum standards for midwifery in Wales posit that midwives should ensure lone contact with all women at least once in their pregnancy and that women should be alone when asked about domestic abuse (a routine enquiry which should occur at least once during antenatal care). 65 Such standards are therefore likely to frame the way in which midwives in Wales view partner involvement and could account for the anxieties expressed by midwives in this research.
Despite the priority given to this issue by midwives, women seemed unaware that their partner's involvement in their antenatal care might affect their relationships with midwives and instead framed the partner's involvement as positive and caring. It is our view, therefore, that addressing these expectations of lone midwife-woman contact with both parties at an early stage of maternity care engagement could go some way to reducing the potential for tension or misunderstanding regarding partner involvement in care and therefore improve relationships between midwives and migrant women.
Expectations of the UK maternity care system also differed between midwives and women, and this was especially apparent when discussing women's navigation of maternity care. Indeed, late or non-attendance at antenatal appointments was seen by midwives to be one of the biggest influences on their relationships with women.  64 and can therefore become dissatisfied with care which differs from these expectations. 64 Overall, midwife-woman relationships appeared to be influenced by divergent expectations and priorities placed on aspects of maternity care, such as family involvement, culture and religion, and navigating health-care systems. However, it is important to note the existence of individual differences in terms of expectations and priorities and to acknowledge that some pairings of midwives and women will converge more closely on these factors than others. Therefore, it is our recommendation that expectations of UK maternity care are addressed and managed not only at a global level (ie all midwives and women) but also on an individual level (ie exploring and managing expectations for each individual pairing of midwife and woman).

| Strengths and limitations
The methodological approach taken by this research provides one of the most in-depth ethnographic studies of this topic area since Bowler's work investigating South Asian women's maternity experiences in 1993. 32 As such, the current study delivers an updated insight into the lived experiences of, and relationships between, midwives and pregnant migrant women in the UK. Furthermore, this study expands knowledge in the underresearched area of maternity care experiences in the Welsh context, where maternity policy and health inequality policies differ to those in England. [72][73][74] As with all studies, results should be interpreted in the context of limitations. This research was conducted in a single health region, with recruitment of participants linked to the services provided in this region. It is therefore possible that the findings of this research may not be generalizable to other geographical areas, especially those outside South Wales. However, in keeping with an ethnographic approach, the richness of data was prioritized over generalizability, providing a nuanced, empirically rich, holistic account of two specific "cultures," that takes into account contextual factors such as local characteristics. 75-77

| CONCLUSIONS
Ethnic and migrant inequality in pregnancy outcomes is an increasingly important area of study as achieving better outcomes for migrant and ethnic minority communities is of global concern. In the UK specifically, growing numbers of migrant women are accessing maternity care each year, 3 and their relative risk of maternal death appears to be increasing. 13 Better understanding of the relationship between midwives and migrant woman at this key time in a woman's life may contribute to addressing some of these challenges.
Findings from this study provide new theoretical insights into the complex factors contributing to the health-care expectations of pregnant migrant Pakistani women in the UK and the ways in which these expectations influence midwife-woman relationships for this population. The differences seen between midwives and women in the perceived importance of these factors suggest that, in order to understand how midwifewoman relationships are created and maintained, more needs to be done to recognize and manage differing expectations of maternity care.