Project 20: Midwives' insight into continuity of care models for women with social risk factors: what works, for whom, in what circumstances, and how.

INTRODUCTION: Continuity of care models are known to improve clinical outcomes for women and their babies, but it is not understood how. A realist synthesis of how women with social risk factors experience UK maternity care reported mechanisms thought to improve clinical outcomes and experiences. As part of a broader programme of work to test those theories and fill gaps in the literature base we conducted focus groups with midwives working within continuity of care models of care for women with social factors that put them at a higher chance of having poor birth outcomes. These risk factors can include poverty and social isolation, asylum or refugee status, domestic abuse, mental illness, learning difficulties, and substance abuse problems.
OBJECTIVE: To explore the insights of midwives working in continuity models of care for women with social risk factors in order to understand the resources they provide, and how the model of care can improve women's outcomes.
DESIGN: Realist methodology was used to gain a deeper understanding of how women react to specific resources that the models of care offer and how these resources are thought to lead to particular outcomes for women. Twelve midwives participated, six from a continuity of care model implemented in a community setting serving an area of deprivation in London, and six from a continuity of care model for women with social risk factors, based within a large teaching hospital in London.
FINDINGS: Three main themes were identified: 'Perceptions of the model of care, 'Tailoring the service to meet women's needs', 'Going above and beyond'. Each theme is broken down into three subthemes to reveal specific resources or mechanisms which midwives felt might have an impact on women's outcomes, and how women with different social risk factors respond to these mechanisms. CONCLUSIONS/IMPLICATIONS FOR PRACTICE: Overall the midwives in both models of care felt the service was beneficial to women and had a positive impact on their outcomes. It was thought the trusting relationships they had built with women enabled midwives to guide women through a fragmented, unfamiliar system and respond to their individual physical, emotional, and social needs, whilst ensuring follow-up of appointments and test results. Midwives felt that for these women the impact of a trusting relationship affected how much information women disclosed, allowing for enhanced, needs led, holistic care. Interesting mechanisms were identified when discussing women who had social care involvement with midwives revealing techniques they used to advocate for women and help them to regain trust in the system and demonstrate their parenting abilities. Differences in how each team provided care and its impact on women's outcomes were considered with the midwives in the community-based model reporting how their location enabled them to help women integrate into their local community and make use of specialist services. The study demonstrates the complexity of these models of care, with midwives using innovative and compassionate ways of working to meet the multifaceted needs of this population.

Introduction

Women with social risk factors such as those living in poverty and social isolation, seeking asylum or refugee status, experiencing domestic abuse, mental illness, learning difficulties, and substance abuse problems, have significantly higher rates of poor birth outcomes compared to their more advantaged counterparts (Draper et al., 2019; Biro, 2017; Lindquist et al., 2015; Blumenshine et al., 2010; Smith et al., 2009). In both the UK and the US women from black and minority ethnic backgrounds [BME] also experience unacceptably high rates of morbidity and mortality compared to their white counterparts, regardless of their socio-economic status (Knight et al., 2018). Recent reports and government policy in the UK have responded to these health inequalities by recommending models of maternity care which promote safety and personalised care (DOH, 2017, NHS England, 2016). The NHS ten-year plan (NHS England, 2019) set specific targets to ensure 75% of women from black and minority ethnic groups, and those living in social deprivation, receive continuity of care from a known midwife by 2024. This echoes international responses to health inequalities with the World Health Organisation (WHO, 2016) recommending midwife-led continuity of care for pregnant women in settings with a well-trained midwifery workforce. The recently updated Cochrane review of models of midwifery care (Sandall et al., 2016) found that women who received midwifery led continuity of care had reduced intervention, improved birth and neonatal outcomes, and increased satisfaction compared to those accessing standard maternity care. Non-randomised studies have also found benefits for women who have social risk factors, such as improved birth outcomes, neonatal outcomes, and more social and emotional support (Beake et al., 2013; Rayment-Jones et al., 2015; Homer et al., 2017). Improved access, engagement and screening, and birth outcomes have been identified for Aboriginal and Indigenous women accessing midwifery continuity models of care in Australia, (Kildea et al., 2019; McLachlan et al., 2017). The mechanisms for these improved outcomes are not fully understood, and less is known about the impact of continuity of care on women with social risk factors. Furthermore, there is huge variation in how continuity of care is operationalised within services and the associated issues of assessing whether it has been achieved. Symon et al. (2016) emphasized the need for research in models of maternity care to report not only the what and by whom, but also attempt to explain the why and how improvements in outcomes are seen to inform the implementation of effective care.

Despite the evidence base and clear policy direction, current maternity care in the UK is often fragmented with women reporting limited continuity of care and concerns about midwives’ awareness of their medical history (CQC, 2018). This is particularly concerning for women with social risk factors as they are known to struggle to access and engage with maternity services and often have complex medical histories (Ebert et al., 2011; Lindquist et al., 2015). A recent review of how women with social risk factors experience maternity care in the UK identified significant common barriers including difficulty accessing maternity care and interpreter services, inappropriate antenatal education, and a lack of continuity and practical support (Rayment-Jones et al., 2019). Many women experienced paternalistic care and discrimination from healthcare professionals and those who had a history of social care involvement often perceived health care services as a system of surveillance rather than support. A trusting relationship with a healthcare professional was thought to mitigate this perception and helped women regain a sense of control during their pregnancy and birth. This supports the growing evidence base that shows continuity of care enables a quality of mother-midwife relationship and level of trust that leads to improved clinical outcomes and increased satisfaction (Biro et al., 2003). However, recent hypotheses (Rayment-Jones et al., 2019) identify many more potential mechanisms which may lead to improved outcomes for women with social risk factors and BME women. These include consideration of: the potential impact of the location of maternity care; how midwives working in continuity models advocate for women and provide culturally responsive, individualised care; the value of external support services; community integration; and how to utilise the multi-disciplinary team without impacting on the mother-midwife relationship (Rayment-Jones et al., 2019). The concept of ‘candidacy’, that is, women's ability to engage with maternity services based on how they are structurally, culturally, organizationally and professionally constructed (Dixon-Woods, 2006) is an important consideration when exploring the disparities seen in service use and outcomes for this population.

This paper adds to the knowledge base by exploring how midwives provide continuity of care to women with complex needs, and what they believe works, for whom, in what circumstances. The findings will enable the refinement of the hypotheses - or programme theories - developed in the aforementioned review (Rayment-Jones et al., 2019), and provide practical guidance for those developing maternity services aimed at reducing health inequalities. The study forms part of a wider realist evaluation of two continuity of care models for women with social risk factors: Project20.uk

Methods

Aim

To explore the insights of midwives working in continuity models of care for women with social risk factors in order to understand the resources they provide, and how the model of care can improve women's outcomes.

Realist approach

This study was informed by the realist paradigm that assumes one external reality which can be explained through contexts, mechanisms, and outcomes, but that this reality is subject to change and volition which should be pursued by the evaluator (Pawson, 2013). The findings of the realist synthesis (Rayment-Jones et al., 2019), and potential gaps in knowledge, formed the focus group interview guide (see Appendix A) that aimed to highlight this change and volition in how the model of care works. Thematic analysis was deemed the most appropriate method of analysis of the focus group data to reveal potential mechanisms which may not have been apparent in the synthesis, contributing to theory development.

Sampling, recruitment, setting and participants

Purposive sampling was used to recruit midwives who were working in the continuity of carer models being evaluated as part of the wider Project20 evaluation. The two continuity models of care were chosen on the basis they had been implemented in areas with significant health inequalities (Public Health England, 2015) to provide care to women with social disadvantage. Many of the women accessing the two models of care have social care involvement. Social care in England is defined as ‘the provision of social work, personal care, protection or social support services to children or adults in need or at risk, or adults with needs arising from illness, disability, old age or poverty’ (Act, 1990). See Table 1 for descriptions of the two models of maternity care.

Table 1

Description of each model of care.

Community based model of care [CBM]

A team of 6 midwives provide continuity of care to women located in an area of social deprivation. Not all women under their care will have social risk factors. Each woman is assigned a named midwife who coordinates all care, multi-disciplinary communication, and referrals. The named midwife aims to provide the vast majority of clinical care, with others in the team providing care when she is not on duty. The midwives are based in a local community health centre and offer antenatal, intrapartum, and postnatal care in the home, community, or hospital setting.

Hospital based model of care [HBM]

A team of 6 midwives provide continuity of care to women with social risk factors only. Women living within the hospitals geographical boundary with one or more significant social risk factor are referred to the team. Each woman is assigned a named midwife who coordinates all care, multi-disciplinary communication, and referrals. The named midwife aims to provide the vast majority of clinical care, with others in the team providing care when she is not on duty. The midwives are based on the hospital site and offer antenatal, intrapartum, and postnatal care in the home or hospital setting.

The study inclusion criteria required the midwives to be working in the model at the time of the evaluation to enable all evaluation data to capture a similar time-point. Eleven out of a possible 12 midwives participated, five from a community-based continuity model of care [CBM] within an area of deprivation in London, and six from a specialist, hospital-based continuity model [HBM] for women with social risk factors in London. See Table 2 for data on the number of years each participant had been a registered midwife, and how long they had been working in the model.

Table 2

Participants’ time spent working within the model of care.

Participant	Number of years as a registered midwife	Time spent working in model of care
HBM1	8 years	<1 year
HBM2	6 years	2 years
HBM3	3 years	<1 year
HBM4	28 years	9 years
HBM5	5 years	<1 year
HBM6	25 years	4 years
CBM1	13 years	13 years
CBM2	<1 year	<1 year
CBM3	6 years	3 years
CBM4	4 years	<1 year
CBM5	6 years	<1 year

Data collection

Focus groups were considered the most appropriate method of data collection as not only do they seek opinions, values, and beliefs in a collective context, but they also provide insights into the mechanisms of complex behaviours and motivations (Jayasekara, 2012). Two focus groups were carried out, one per model of care. These were held in the clinical setting of each team and lasted up to two hours with six midwives in one [HMB], and five in the other [CMB]. They were conducted by lead researcher [HRJ] and facilitated by an academic colleague [ZK] who took notes on who was speaking, main topics or insights, and general time keeping. Using Manzano's (2016) guide to realist interviews, and the programme theories developed in the realist synthesis of women's experiences of UK maternity care (Rayment-Jones et al., 2019); a realist informed interview guide was prepared to elicit specific mechanisms of how each model of care was thought to work (see Appendix A). The term ‘programme’ has been changed to ‘service’ in the interview questions to reflect the language of the participants. Open questions were also used to clarify content or context, gain a deeper understanding of the midwives’ perspectives, and to stimulate the flow of discussion.

Analysis

Data from the two focus groups were analysed using thematic analysis (Braun and Clarke, 2006; 2013). This analytic approach to qualitative data involves inductive coding practices, which are both consultative and initially open (Braun et al., 2019). NVivo 12 was utilised for data management and analysis which followed Braun and Clarke's six-phase approach to thematic analysis (Braun and Clarke, 2006). In brief, these phases include familiarisation with the data, generation of initial codes, the searching for and review of themes, naming and offering explanations for each theme, and lastly producing a report. All data were coded by the lead author [HRJ], with a proportion coded by another author [SAS]. All codes and themes were subsequently ratified by all team members.

Themes were generated with a central organising concept to both explain and hold together each supporting quotation within each theme (Braun and Clarke, 2013). Regular discussions were held between all researchers to deliberate and, when required, revise aspects of the analysis, coding, or themes. This also helped ensure analytic rigour. When discrepancies occurred between researchers, these were debated until all were satisfied themes were fully explained and robust. We utilised existing models of sample size sufficiency (Morse, 2000), data adequacy (Vasileiou et al., 2018), and thematic concordance (Guest et al., 2006) to assess data quality and theme saturation – all of which were assessed to be excellent.

Results

Three main themes were identified: ‘Perceptions of the model of care’, ‘Tailoring the service to meet women's needs’, ‘Going above and beyond’. Each theme is broken down into three subthemes (Table 3) to reveal specific resources or mechanisms the midwives felt might have an impact on women's outcomes, and how women with different social risk factors respond to these mechanisms. Quotations from the midwives in each model of care have been given to add meaning and help identify differences and similarities between the two different models of care.

Table 3

Overview of main themes and subthemes .

Main Theme	Subthemes
1.0 Perceptions of the model of care	1.1 Variation in the perception of the aim of the model of care
	1.2 Belief the model of care is working
	1.3 Emotional investment
2.0 Tailoring the service to meet women's needs	2.1 Holistic care (multi-disciplinary working)
	2.2 Flexible working (early access and chasing)
	2.3 Community integration
3.0 Going above and beyond	3.1 Advocacy and disclosure
	3.2 Counteracting mistrust and fear of the system
	3.3 Trying to build relationships with those resistant to help

Perceptions of the model of care

Variation in the perception of the aim of the model of care

Midwives in both models of care gave varied answers when questioned about the aim of the model of care before discussing their uncertainty around a specific aim. Rather than give particular health outcomes they discussed social outcomes and the importance of being able to engage women in their maternity care and the impact on long term outcomes such as parenting. They acknowledged that this was something that they felt was important and not an official ‘aim’ or ‘key performance indicator’.

‘…better engagement with services. Trying to get you know, addicts off their, their, you know, life. Giving them the opportunities to see if they can parent, to be able to parent their children. Keep their children, if possible.’ (HBM6)

‘I don’t know, 18 years ago [when the service was set up] I don’t know what they would have been thinking. I think for us now I think a lot of it is engagement. (CBM2)

Some midwives indicated uncertainty around the specific mechanisms thought to improve women's outcomes.

‘So, my understanding is that its continuity of care for vulnerable women because vulnerable women have poor birth outcomes, we know continuity of care gives better outcomes so therefore stick those two together and hopefully we get better outcomes for vulnerable women. Less stillbirths.’ (HBM2)

Belief the model of care is working

Despite the variation discussed around the aim of the model of care, the midwives in both models were confident that their care has a positive impact on women.

‘I really do truly believe that we make a massive difference to people's social outcomes, I really, really do.’ (CBM5)

‘I have three women who lost babies [removed from parents to care of social services] in the past, I managed, you know, the care they received they were given an opportunity to keep their babies.’ (HBM6)

Midwives in both models of care revealed specific mechanisms thought to improve outcomes by highlighting the differences in how women experienced the continuity model compared to standard or traditional maternity care. These mechanisms included early recognition of abnormalities, and more disclosures of women's concerns

‘…getting them into the hospital sooner, and a plan made sooner, and, and a safety plan and maybe a delivery if that’s what’s needed. Whereas another lady [receiving standard care] like, who wouldn’t realise her symptoms, had no one she could contact, or felt she could contact, didn’t really go, missed an appointment, got sent a letter for two weeks later, by that point pre-eclampsia [worsens]’ (HBM3)

‘Because we have slightly longer appointments than traditional teams, we are able to talk to women for longer so might be able to find things that they need referrals for that other teams might not have the time to dig into.’ (CBM4)

Emotional investment

Midwives in the community-based model discussed the emotional investment they had in their women's wellbeing and how this motivates them to sustain their investment in the women they care for.

‘I think we also have that like emotional insight as well… I feel like we, as a team, we are quite invested in our women, and we do a lot for them and I think, when you have that investment in someone that you want to push for them and you want their outcome to be good.’ (CBM1)

‘…I think the fact that we see a lot of the women, you know repetitively throughout pregnancy we know them really well. And it just gives you that element of, like I want this to work for you.’ (CBM5)

Tailoring the service to meet women's needs

Holistic care (multi-disciplinary working)

Holism was referred to throughout each focus group. The midwives from both models of care were very clear about the importance of holistic, including culturally sensitive, care in comparison to the medical model of standard maternity care. The midwives described practical issues that women with social risk factors often face and how they spend time supporting and advising women on practical issues far wider than pregnancy or maternity care:

‘And it was even simple things of, because she’s been illiterate, you know she was given a bank card from the no recourse to public funds team from social services, but does she know how to use a bank card? Does she know how much things cost and things because she can’t read? And so there’s been quite a lot of other thinking outside the box that if someone were under a mainstream system of midwifery care … But also, being more just aware of kind of her general needs and what we’re thinking that she’s going to be needing after we’ve gone, as well. She was medicated. So that was a challenge, trying to make sure she knew which medicine to take because she couldn’t read the box.’(CBM1)

Both models of care reported having good relationships with their obstetric colleagues and named consultant. They felt that this relationship led to a level of respect that promoted multi-disciplinary working.

‘…And I think it’s really great that if we have just a general query about something, um, that comes up within an appointment…we can just email and, um, the named consultant will respond with whatever advice she would advise.’ (CBM4)

The midwives in the hospital-based model also spoke about their presence at women's obstetric appointments, and how this presence impacts on the obstetrician being able to provide more holistic care and encourage understanding of why women might make certain decisions:

‘I think that by knowing them [the team’s named obstetrician] then they help work with us…..to give the women the best care and the best, and maybe the, you know, the decisions they make are looking at the woman as a whole rather than just the obstetric concerns, they’re understanding the social impact of why she chooses…. I dunno, they can understand the whole picture, because we helped deliver that’ (HBM2)

Flexible working (early access and chasing)

Flexibility was discussed by the community-based midwives as an essential means of engaging women who struggle to attend appointments due to social factors such as caring responsibilities, financial and geographical barriers, unfamiliarity with the service, and mistrust.

‘And it works for the women. Like if you’ve got a woman that can only ever see you at 5 or 6pm then I can do that one day and then come in late the next day or whatever, like you have that flexibility’ (CBM1)

‘And I also think a lot of our women now, our particularly vulnerable women, really wouldn’t travel to the hospital for their appointments.’(CBM3)

‘We didn’t really stick to much of a pattern in terms of meeting her we could meet her when we could so there was a bit of a patch when we didn’t see her for a few weeks. Um, not necessarily like through want of not trying but like just door-knock her and she was moving between properties, so it was just a lot more difficult…but that could have ended very differently’ (CBM3) ‘….she could have entirely fallen off the radar.’(CBM1)

The hospital-based midwives discussed flexibility in terms of early access to pregnancy care and how this can impact on social care outcomes. They also felt that women with social care involvement are given a chance to demonstrate their ability to parent through referrals to parenting and rehabilitation programmes, whereas if they were going through the standard maternity care pathway, they may not have been referred to these programmes in time.

‘We see them quite early on [in pregnancy], we can recognise their needs and then send them to the relevant departments. So, when it gets to the time that we do go to core group meetings or strategy meetings, we’ve already referred them to relevant departments, we can already encourage our women to attend, or to be compliant with these programmes, erm, and once they’ve reached, the social services’ sort of decision about the care of their unborn, we can already demonstrate that these women have been involved in some sort of rehabilitation programmes for their care, where they probably wouldn’t have had that before'(HBM3)

Midwives in the community-based model of care discussed the time they spent chasing women and encouraging them to come to their appointments. They felt that this had an impact on the women's engagement, outcomes and overall safety. Neither model of care had administrative support for this aspect of care.

‘…we spend hours and hours and hours chasing people, and I think actually other services don’t perhaps know that we need to know things…it’s like other people’s awareness of what midwifery actually is and like safeguarding other children, because we seem to do a lot’ (CBM5)

‘So I think instead of them feeling like they might just be in a system of hundreds of women…they’re going to have to tell their stories again and again, um, whether it’s that aspect that they don’t, that they feel like they can engage with better. Or just kind of us having the capacity to almost … push people to come to their appointments and go to their scans’ (CBM2)

Community integration

When the midwives were asked about how engaged they felt to the local community there was a clear difference between the two models of care. Where the community-based model discussed a ‘learning curve’ they still felt they were well integrated into the community and knew about local services. They described a comprehensive but complex system of community support services that they have knowledge of through referrals and communication.

‘…she was a late booker, very like little support, or no support really for her. Um, living in very precarious situation when we met her. Um, and I think we were just able to, kind of build a bit of a team around her.(CBM2)

‘…although it’s been a massive learning curve with all these women coming through, and I know we’ve all learnt a lot about what’s available locally and what happens locally.’ (CBM3)

The hospital-based team midwives did not share this feeling- this did not seem to be solely based on their location and the size of their geographical area, but also cutbacks in services. They spoke about the enormity of the community, different cultures across the multi-ethnic geographical patch, and how this created difficulty in integrating women into local community support services.

‘There’s just too many communities. and it’s a very big catchment area, with very many different communities, multi-diverse, that actually sometimes it’s very hard to… get to know them all’ (HBM2)

‘…when I was a community midwife where I lived, I was known as the [name anonymised] clinic midwife, and when I’d go to the local high street they’d say hello to me and acknowledge me because they all, most of them had seen me in the clinic. But here, with the diversity and complexity of all the different ethnic communities that are going on, you just couldn’t integrate into them, it’s just impossible to do that because you can’t be everything to everyone, so you just have to be quite single in your care’(HBM1)

‘I think it’s a shame that, you know the erm, children’s centres, that’s shrunk, a lot. And I think that’s a real shame because when I very first started I felt we were more integrated into the children’s centres, and that’s gradually got less and less and less’ (HBM4)

‘They (health visitors) are very short (staffed) and its very difficult to get one very quickly’ (HBM6)

Midwives in the community-based model of care discussed how immersing themselves in the community setting enables them to integrate women into local services. This in turn helps women to feel supported and cared for by their local community.

‘I’m working with a young girl with learning difficulties at the moment and all of these incredible services have just come to light that I didn’t even know existed… Um, like we’re working with a support service for young people and people with learning difficulties, and they’ll like go round help them clean their flat, do a food shop, take them to their appointments, like it’s amazing what’s available, but I had no idea until this case came up.’ (CBM2)

‘…we use the Children’s centres a lot more now… and they’ll [outreach teams] see a lot of our families that just need a bit of help integrating into the community. So they’ll get them engaged in local services, get them coming along to the group sessions, meeting other parents’ (CBM3)

‘I think this, this location is what gets our women to engage and I hope that we set women and families up to actually believe that they deserve more. And that actually we’ve not been the only ones that care about them but actually the community cares about them, and I hope that we can make them feel that way about themselves. I think that’s important.’ (CBM3)

Going above and beyond

Advocacy and disclosure

The midwives in both models spoke about advocating for women by guiding them through a complex and often unfamiliar system. Advocacy was described in many examples of how the midwives supported women and their families and tried to give them a voice. This was discussed by midwives in both models in terms of the social care system, but only by the hospital-based team in terms of clinical care.

‘We've had quite a few interesting cases recently where social services have not deemed there to be a concern, whereas where we're having really regular contact with these women we are seriously concerned. And we push and push and re-refer and get a safeguarding lead involved from the hospital until we feel that, that family is safe. And I think having the time to do that, definitely as a traditional midwife you wouldn't have the time to do that. Um, so we are massively advocating for the safety of these families I think.’ (CMB5)

‘We attend the meetings. The social service meetings, the strat [strategic] meetings, the core group meetings, professional meetings. We’re there, and we are the ones that will go and represent our ladies, or the women in our care, so we know them personally rather than any midwife just turning up just with notes who doesn’t know them.’ (HBM1)

‘…we can navigate women through the process, through the system. It’s quite a scary system and I think by being here, by the relationships we’ve built around and between like the doctors and our medical colleagues and multi-professional teams, then we can kind of signpost and navigate a woman through easier, we will get her seen by a doctor early, so we know that she’ll be seen first and won’t have a 3 hour wait that other women might have and just to make it as kind of smooth as possible.’ (HBM2)

The midwives in the community-based model gave insight into how the trust they had built with women had impacted on women's disclosure of sensitive information.

'We’ve definitely had a, um, a few women that we’ve thought are not really a concern, like they might have come to us because of mild mental health, and that’s all we know about their history. And then actually it’s not until 25, 28 sometimes later weeks that they say, ‘Actually I’m in this really abusive relationship, or, ‘Actually I am technically homeless,’. I think it’s the, the building of trust…I think by then they feel maybe comfortable enough to disclose what they feel they need to. (CBM3)'

‘..it was all very routine and everything was normal, and I was thinking, oh like it's a really quick appointment compared to normal, so I said to her, ‘How's everything? Like how's your housing going, um, how's everything at home?’ and then she opened up about having a, quite a volatile relationship with her mum. And so that's then opened another, you know, can of worms that I wouldn't have discussed if, um, I'd had a 20 min appointment… because she hadn't disclosed it to me and we'd asked at booking and she'd said it was fine.’ (CBM4)

This last quote demonstrates not only the impact of flexibility with the length of appointments, but also how repeated contact with a known healthcare professional enables the development of a trusting mother-midwife relationship.

Counteracting mistrust and fear of the system

Midwives in both models of care felt that fear is the most common underlying reason behind women's resistance to help, particularly if they feel social care will become involved. They identified particular social situations where this fear contributed to the lack of trust and disengagement with services:

‘I think domestic violence can be a tricky one…there’s that level of fear and distrust I think of what will happen if the professionals get involved, if they do disclose, what will the outcome be?’ (CBM5)

‘because they are… scared. I think that underneath they are scared, they’re terrified’(HBM2)

They revealed that often this fear can be overcome through a trusting relationship and an ability to communicate how social care can provide practical support:

‘…I think for a lot of these women it’s the first time they’ve actually ever had someone take a proper interest in their lives, and be able to manage them for over a period of time and make sure they’ve got a plan going forward. Um, whereas that initially was a really difficult situation she then came to really understand and feel safer and more protected(CBM5)

When asked if the model of care works for all women, and if not, who does it not work for and why, the midwives in both models identified situations where they felt it was difficult to gain trust with women. Again, this lack of trust was often associated with social care involvement and women's perceptions of the aim of social care services. The midwives felt this had a direct impact on the woman's level of engagement and openness:

‘I’ve got at the moment who is terrified of social workers because she’s got two friends who’ve had a baby taken away… and now I’m trying to get a social worker involved and she’s having none of it. But I want it for support, I don’t want her baby [ to be removed], but she doesn’t understand that, she can’t’ (HBM6)

‘…they think that means their baby's going to be removed just like that, and actually it's more of an assessment and, yeah so I think that they have different views of what it is.’ (CBM4)

Midwives in both models tried to overcome this mistrust through various, innovative ways. The community-based midwives described having a ‘good cop, bad cop’ technique whereby the woman's ‘named midwife’ will provide midwifery care, and another midwife from the team will coordinate referrals to social care and attend child protection meetings. They felt that this preserved the trust between the woman and her named midwife.

‘We do have tactics that we use, so if someone has to break news to a woman about referring to social services or what the plan is, then we might make that maybe not, you know not the regular midwife they see.’ (CBM3) ‘Good cop bad cop. (CBM1). ‘Yeah, sometimes that works to keep them engaged.’(CBM3)

The hospital-based midwives described advocating social care to the women through explaining how they can provide practical support and give women an opportunity to demonstrate their parenting abilities. They felt that this has led to a reduction in the number of babies removed by social care.

‘So we also advocate social services to, to them, as well as for them to social services. Because as soon as someone says ‘social care’, ‘social services’ they immediately have this picture ‘they’re going to remove my baby’, but it, when we talk to them and say ‘we’ll be there, we’ll be there with you, we’ll make sure they’re, you know, they’re there to help and support you’ and they then actually start to engage a lot better..so, as in HBM6’s case women are managing to keep their babies, where before they didn’t engage, they fought against them [social services], and they lost their babies but by working with them they’ve kept their babies.’(HBM1)

Midwives in the hospital-based model also described a level of apprehension of the model of care for some women and reflected on one particular woman who felt like she was being stigmatised after being referred to the team. Again, they described ways of trying to overcome this through communicating the positive aspects of the model of care with women, but that for some women this doesn't work:

‘I think they can be quite apprehensive about it (the specialist model of care), but, I think if they realise they have to have a midwife anyway, having a midwife they know who will come to their house, who will be flexible with timings, who will work with their needs, and who will be there to support them, then I think it turns…it becomes a better experience. Because there’s a lot of women who don’t want full stop, any professionals involved, they kind of don’t even want to go into hospital, they’re going to do their own thing whatever’ (HBM2)

‘I did have one woman who declined our services because she felt that we were singling her out for special treatment and stigmatising her, so she didn’t want that’ (HBM1)

This concept was not discussed in the community-based model.

Trying to build relationships with those resistant to help

When exploring the issue of women who are more difficult to engage, the midwives from both models of care gave specific examples of social circumstances that led to a resistance to be helped:

‘Some of these cases though, you just aren’t ever going to win and that’s, well it feels like that. So some people are totally just going to disengage and no matter what we try, um, so they’re, I think it’s knowing that some we probably aren’t always going to help.’(CBM1)

‘Because like some women just see us as pests and that we’re interfering and … [Some agreement], I don’t know, they don’t want us so it, it would be impossible to … that’s the women rather than our service’(CBM4)

‘Some women have their own agenda, and no matter what you do or how you try, they will not … waiver from that. They have their own agenda, this is what they want and some of them will… will play you for what you want, for what they want, and to get what they want…’(HBM1)

One midwife described how some women access the model of care thinking that they ‘play the system’ to continue using drugs or alcohol:

‘and sometimes is actually the reason why they’ve come to us, so they may be dependant on, on drugs, or alcohol, and don’t want to get off of it, but will play the system, so they can remain using, or drinking, and still have their baby.’ (HBM3)

Discussion

Midwives working in both models of care were asked about how they provide care to women with social risk factors, and what aspects of their care they felt contributed to improved outcomes. There were many overlapping themes and similarities between the teams, but also some significant differences in how the teams worked and how midwives perceived the model to be working for different groups of women. It is important to bear in mind that although there was confusion around the aim of the models, all midwives believed the model of care they worked in was beneficial to most women and improved both clinical and social outcomes.

As expected, the quality of the midwife-mother relationship and importance of trust was often discussed theoretically and demonstrated through real life examples. As Hunter et al. (2008) highlight, the way in which maternity care is organised has a profound impact on midwives’ ability to form meaningful relationships with women. Continuity models of care have long been associated with increased trust between a woman and midwife, whereas fragmented, industrialised models of maternity care are far from conducive for the development of trust. Perhaps more interestingly though, this topic did not dominate the discussion and the midwives put forward a catalogue of other resources they employ to engage and support women with social risk factors. These resources often involved advocacy and guiding women through a fragmented and often unfamiliar system and using the flexible nature of the model of care to coordinate other professionals and agencies. This demonstrates that although the midwife-mother relationship is clearly integral to the model, a more complex system of mechanisms takes place ‘behind the scenes’, with midwives often planning care and orchestrating support for women when they are not physically with them. Insights such as this, raised throughout the discussions, have been formulated into programme theories to test in the wider evaluation of this model of care (Project20) - see Table 4.

Table 4

Additional programme theories for testing in realist evaluation of specialist models of care for women with social risk factors).

Programme Theories

If midwives are able to work flexibly, then they are able to meet women's individual needs and increase safety through spending time care planning and coordinating support that may not be available on demand (for example during an allocated appointment time in the standard maternity care model).

If midwives advocate social care to women through explaining their role and how they can provide practical support, then women's perception of surveillance may lessen leading to engagement, and child protection outcomes and maternal infant-bonding improve.

If the midwife-mother relationship is ‘two way’, that is the midwife also has trust in the woman then the many known benefits of the trusting relationship will be enhanced.

If models of care are based in the hospital setting or have large catchment areas, then midwives are less likely to have the knowledge and familiarity of niche support services that may benefit the women they care for.

If midwives are placed in the community setting, then they will be better able to place the individual needs of women before institutional norms because they feel a sense of obligation and responsibility towards the woman rather than the system.

If women do not have the time to form a trusting relationship with a midwife, then they are unlikely to disclose sensitive information and seek support for issues that may have long-term detrimental consequences for themselves and their families.

If women who remain resistant to help throughout their pregnancy despite continuity of care are known/handed over to primary care and early years services, then they will have a support network in place and will be more likely to be able to regain trust in the system over time.

Advocacy was discussed specifically and in more nuanced ways, but overall reflected the literature around its importance for this vulnerable population of women, particularly those with safeguarding concerns (Everitt et al., 2017; Woods, 2008). Midwives in both models spoke about advocating for social care services as well as for the women, in order to ease women's reluctance to engage with a service they may perceive as a form of unhelpful surveillance. This contributes to the hypotheses put forward by Rayment-Jones et al. (2019) that continuity of care mitigates this perception and helps women regain a sense of control. Whereas it was assumed that trust was the mechanism to improve women's engagement with social care, engagement may also be enhanced by how a trusted midwife conveys information and advocates the service to them. Lewis’ (2019) longitudinal qualitative work with pregnant women also identified the intricacies of the midwife-mother relationship, with trust being interwoven with women's agency and the importance of ‘two-way trust’ that includes the midwives trust in the woman. This reveals a level of trust and belief in the woman and a desire to extend this trust to other professionals. Trust as a generative mechanism may impact on far more than a woman's experience of maternity care. Dahlen and Aune (2013) described how women who perceived a trusting relationship with their midwife felt that this led to personal growth and development. Long term outcomes such as these are particularly significant for women who may lack trust in both the system and their own abilities as a mother. Although this ‘two-way trust’ was not explicit in this study it was alluded to when discussing how women with social care involvement can be encouraged to demonstrate their ability to parent by engaging with the system. This has the potential for improved maternal-infant bonding and a longer-term impact on social outcomes. This concept was also discussed by Ebert et al. (2014), who found that socially disadvantaged pregnant women did not feel safe to engage in discussions with midwives regarding choice or to seek control of their care. This resulted in midwives perceiving a lack of responsibility from the women and increased surveillance.

Midwives from the community-based model discussed multi-disciplinary working in terms of both hospital-based and community-based services. They described community services as comprehensive and complex, and constantly having to learn what was available, but felt that it was within their remit to communicate with services if they felt it would be beneficial for women. The hospital-based midwives on the other hand spoke about multi-disciplinary working in terms of their hospital-based, obstetric services. They reported a lack of community resources and short-staffed health visitor services. It was hypothesised that they may perceive a lack of community services due to the enormity of their catchment area. If the community-based midwives reported challenges in getting to know what is available locally, it would make sense that knowing and communicating with niche, local services is an impossible task for the hospital-based midwives with a much larger catchment area. In addition to this point, both the hospital-based, and the community-based midwives reported strong, effective working relationships with their named obstetric consultants, which involved frequent communication. Being based away from the hospital did not seem to impact on this. These are important points to consider when planning services to meet the needs of women with social risk factors who are often socially isolated. Midwives in the CBM felt that their community location impacted on how well looked after women felt, and demonstrates to women how their community cares for them. This ‘candidacy’ concept was discussed in Rayment-Jones et al. (2019) findings of how women experience maternity care. ‘Candidacy’ theory suggests that how a person interacts with health services is structurally, culturally, organizationally and professionally constructed (Dixon-Woods, 2006), and can give us insight into why women with social risk factors make less use of maternity services than their more affluent peers. This concept is described in Ebert et al. (2014) qualitative work with socially disadvantaged women in Australia, which found that without appropriate information and choice women believed they were outsiders to the maternity care culture. This resulted in women handing over their autonomy to those who they believe do belong in the culture: midwives.

Hyde and Roche-Reid (2004) reported conflicting communication ideologies between women and midwives, with midwives believing their role was empowering women, but in fact their communication reflected their employing institution's values. This study explored how this allegiance can shift in a continuity of care model, with midwives demonstrating how they aim to place the needs of the woman before the system's norms. This shifting of allegiance and different ideologies has been explored in the continuity of care literature over the past decade, with continuity of care being associated with a sense of obligation and responsibility towards the woman rather than the system (McCourt et al., 2006; McCourt et al., 2009; Hunter, 2004). In the current study, this seemed more apparent in the community-based model of care when midwives discussed holistic care, calling to question how the location of midwifery services might impact on midwives ideologies and communication methods. McCourt and Pearce's (2000) work with minority ethic women found that those receiving standard maternity care in the hospital setting had poorer experiences and felt that their care was not focused on them as a person. This begs the question that if midwives are immersed in the hospital environment are they more loyal to the needs and norms of the system than if they were on the ‘outside’ looking in alongside the woman?

The midwives in the community-based model gave insight into how the trust they had built with women had impacted on women's disclosure of sensitive information. Women they were caring for who may have been referred to the team for one particular social risk factor, often disclosed more complex and serious risks as they began to trust the midwives and understand their role. This in turn leads to referrals to support services and more individualised care plans. This insight begs the following questions:

How much are midwives working in standard maternity care models missing?

To what extent do women hold important information back through fear of disclosure to a system they do not trust?

What are the long-term consequences of this on the woman, the child and future children?

Perhaps the most insightful aspect of this study was the subtheme ‘Trying to build relationships with those resistant to help’ as it unpicked some of the complexity of looking after women who often live difficult lives with long-standing social, physical, psychological issues and mistrust in the system. The midwives in both models of care identified domestic violence, substance abuse, and social care involvement as particularly challenging factors in engaging women and building trust. Fear of the system was seen to be the main barrier and although midwives practised different techniques to try to remedy this, there was a general feeling that some women were too resistant to help for the model of care to have any effect. This demonstrates that continuity models of care are not a panacea for all poor health and social outcomes, and that the problems these women face are deep rooted and require more long-term multi-sector intervention. That said, continuity of care provides an opportunity to begin to focus on this resistance and work with primary care and early years services to ensure a support network is in place.

Strengths and limitations

When discussing the limitations of this study it should be taken into account that this method of theory building and refining, will be tested in the wider realist evaluation of the models of care using in depth qualitative and quantitative data from women with social risk factors. The ‘fragments of information’ gained during realist-informed qualitative methods (Emmel, 2013) will be re-tested to contribute to the interpretation and explanation of how the model might affect women's physical, emotional, and social outcomes.

The focus groups were undertaken by a realist-interview trained academic using Manzano's (2016) approach to generate data demonstrating the effectiveness of the model of care. This method helps to refine programme theory and improve rigour through the ‘teacher-learner’ relationship. In this case the interviewer presented theories extracted from a realist synthesis (Rayment-Jones et al., 2019) and asked the midwives to confirm, falsify, explain, and refine the theories. The midwives’ insights are not considered to be constructions, but ‘evidence for real phenomena and processes’ (Maxwell, 2013) that contribute to the overall evaluation of the programme's effectiveness. The realist-informed interview guide allowed for both the testing of pre-constructed theories, and new programme theories to be identified (Table 4).

Potential limitations of the study include the fact the participants knew this study is part of an evaluation of their service. These factors might have created a sense of being tested/assessed and therefore impacted on how the participants responded to demonstrate the success of the model of care. In the analysis however, less effective aspects of the models of care were apparent. Again, these insights will be tested in the wider evaluation of the model to increase rigour. A further limitation of this study is that it is urban based only, rural and remote models of care should be evaluated as the context is significantly different.

Conclusion/implications for practice

Overall the midwives in both models of care felt that the service was beneficial to women and had a positive impact on their outcomes. It was thought that the trusting relationships they had built with women enabled them to guide them through a fragmented, unfamiliar system and respond to their individual physical, emotional and social needs, and ensure follow up of appointments and test results. They felt that for women the development of a trusting relationship impacted on how much information they disclosed, allowing for enhanced, needs-led, holistic care. Interesting mechanisms were identified when discussing women who had social care involvement with midwives revealing techniques they used to advocate for women and help them to regain trust in the system and demonstrate their parenting abilities. This has the potential to reduce the number of babies removed from their mothers and greatly improve long term outcomes for children at social risk.

Differences in how each model provided care and its impact on women's outcomes were considered with the community-based midwives reporting how their location enabled them to help women integrate into their local community and make use of specialist services. The midwives in the hospital-based model described their extensive catchment area and location as a barrier to this. This has important implications for women with social risk factors who are often socially isolated and lack support.

Midwives in both models of care discussed how some women are more difficult to engage, with specific social risk factors intensifying their mistrust in the system. This should be taken into account when developing inclusion criteria for continuity models of care, and midwives’ workload.

The study demonstrates the complexity of these models of care, with midwives using innovative and compassionate ways of working to meet the multifaceted needs of this vulnerable population.

Declarations ethics approval and consent to participate

London Brent Research Ethics Committee (HRA) REC Reference 18-LO-0701.

Availability of data and material

Attached as additional file or contact the lead author HRJ.

Funding

This report is independent research supported by the (NIHR Doctoral Research Fellowship, Mrs Hannah Rayment-Jones, DRF-2017–10–033). Angela Harden was supported by the (NIHR) Collaboration for Leadership in Applied Health Research and Care North Thames at Bart's Health NHS Trust (NIHR CLAHRC North Thames), Jane Sandall is supported by the (NIHR) Applied Research Collaboration South London (NIHR ARC South London) at King's College Hospital NHS Foundation Trust and James Harris is supported by a Post-Doctoral Fellowship from funding agency ‘Wellbeing of Women’ (Award Ref PRF006). The views expressed in this article are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care.

Authors' contributions

Authors HRJ, JH, AH, JS contributed to the conceptualisation of the research question, and methodology. Author HRJ designed the interview guide and conducted the focus groups. Authors HRJ and SAS interpreted and analysed the focus group data. All authors read and approved the final manuscript.

Declaration of Competing Interest

None declared.

Question	Rationale
Can you tell me what your involvement in this specialist model of care is?	Realist evaluation assumes that people know different things according to their role. These answers will be used to tailor future questions according to the specific insight of the stakeholder.
What is the purpose of the service? /what do you think are the desired outcomes for women? Do you think the service makes a difference to these outcomes? Can you give examples?	Assuming that programmes have different outcomes for different groups, stakeholders, women and family members will be asked this question until the range of outcomes has been identified. Interviewer will prompt for evidence of the nature and extent of the outcome. If expected outcomes are not identified (improved access and engagement), Interviewer will prompt for those outcomes. If unexpected outcomes are identified, interviewer will prompt for greater description.These outcomes will be verified using the quantitative data analysis.
We are interested in how specialist models of care have an effect on women's outcomes. How do you think the service has caused, or helped to cause [outcomes identified earlier in interview]?	Initial question leading into exploration of mechanisms. When participants identify programme activities (for example flexible appointments, 24hr access to a known mw, safeguarding training) Interviewer will probe further – e.g. – So, what is it about being able to contact a known midwife 24/7? How did that help cause (the later outcome)?
Are the outcomes previously mentioning the same for all women? For example, women with different social risk factors? [using the specific sub-groups identified in the programme theories – specific disadvantaged groups/social risk factors and different cultures].In what ways have they been different?	This question is seeking more specific information about “for whom” the programme has and has not been effective (in what respects, to what extent). Interviewer will specifically probe in relation to sub- groups that are identified in the realist synthesis’ programme theories.
Do you think women with social risk factors want/are open to this model of care prior to accessing it? How might this differ for different groups of women (specific risk factors?)Do you think this specialist model of care changes the way women feel about maternity services? In what ways? Can you provide examples?	This theory-based question sets out to explore candidacy theory. Examples might be given of how women with particular social risk factors have reported their experience of maternity care (for example those who are unfamiliar with the UK system, or those who have social care involvement), to explore if and how the programme addresses these issues and what the outcomes of this might be.
There are lots of ideas about how specialist models of care actually work, and we think they probably work differently in different places or for different people. One of those ideas is (an example: that if women trust their midwife then they will engage with the services and be more open to disclosing concerns.)Does it work at all like that here? Can you give an example? Does this apply to all women? What about: (brief description of other mechanisms not previously identified)   - Engagement with the multi-disciplinary team   - Engagement with local community   - What other resources the service offers (practical support, interpretation services, access)	The subject of a realist interview is the programme theory. The aim is to get the respondent to refine the programme theory for the particular context about which they know. This question revisits the mechanisms (particularly those not identified before) but in a more specific way to test the programme theories and whether the programme works differently for different people. This (in conjunction with the women and family members responses) will help confirm or refute the initial programme theories.
We've seen that specialist models of care work differently in different places. What is it about this service that makes it work so well/less well? Do you think culture, the local community or other resources has an effect on women's outcomes? Can you give examples?	Realist evaluation assumes context does affect outcomes (by affecting which mechanisms fire). Interviewer will probe for aspects of culture, local resources/lack of them, local and family relationships/support, relationship between organisation and participants and so on.
If you could change something about this service to make it work more effectively here, what would you change and why?	This question aims to elicit understanding of why the programme has not worked as effectively as it might (i.e. mechanisms not firing, aspects of context) as well as strategies for improvement.
What else do you think we need to know, to really understand how the service works here?	This open probe that enables participants to comment on anything not covered by the interview. The structure of the question keeps the focus on ‘how the programme works’ and ‘in this context’.