Multi-site implementation of a promising innovation in low income communities: support for childbearing women

Existing evidence

A rapid review of studies carried out prior to commencement of this research explored ‘doula support’ including systematic searches of the following databases: MEDLINE, EMBASE, Cochrane and the Cumulative Index to Nursing and Allied Health Literature. The search was not limited by country, date, methodology or language. Support during labour from trained doulas is associated with reduced length of labour,14 less pharmacological pain relief and oxytocin augmentation, and fewer instrumental or operative births. 15 In particular, instrumental and operative births are associated with increases in the risk of morbidity for women and their babies. This morbidity includes postpartum haemorrhage,16 genital tract trauma for the mother17 and increased risk of intracranial haemorrhage for babies. 18

In addition to positive impacts on labour outcomes, there is also evidence of positive impacts on breastfeeding,10 including increases in the proportion of women initiating breastfeeding and continuing with exclusive breastfeeding. 11 It is particularly noteworthy that these positive impacts have been achieved in groups where rates are frequently lower than national figures, including low-income, first-time mothers. These findings reflect the wider evidence base of breastfeeding support by peers19 and resonate with contemporary policies that encourage the implementation of peer support for breastfeeding. 20

Positive benefits for women’s psychosocial well-being include more positive feelings about labour and less anxiety,11 increased feelings of control and confidence as a mother,21 and fewer women experiencing postpartum depression and anxiety. 22 Evidence suggests that doula support during labour may also have potential positive effects on parenting behaviours and the relationship between a woman and her child,23 including increased acceptance of a baby immediately after birth and an increase in behaviours such as stroking, smiling and talking to their babies,24 and more positive parenting when babies are 2 months old. 25

All of these findings resonate with important aspects of the policy context and many also offer potential benefits to the NHS from reduced resource use, including shorter inpatient stay following normal birth than assisted birth and fewer referrals to specialist services, including mental health care. Evidence of benefit from doula care is particularly striking for women in situations of social or economic disadvantage, those with lower educational attainment and where supportive contact starts during pregnancy. There are also suggestions that the provision of doula support is associated with increased use of required health-care services. 26

The UK NHS spent £1.6B on maternity services during 2008. Part of this cost is attributable to the high rate of caesarean sections (C-sections), which increased from 12% in 1990 to 24% of all births in 2008, each costing between £1197 and £3194. 27 It was further estimated that the cost to the NHS for maternal care due to smoking in pregnancy is between £8M and £64M per year (depending on the costing approach);28 a further £12M to £23.5M per year is spent treating infant conditions attributable to smoking during pregnancy. Another study estimated that the cost of neonatal care for low birthweight babies was between £12,344 and £18,495 per child in English hospitals. 29 These items reflect those in the Quality Innovation, Productivity and Prevention (QIPP) Metrics included in the NHS Outcome Indicators for Maternity.

The impacts of doula care described above are derived from quantitative data generated by randomised controlled trials and included in systematic reviews. There is a relative dearth of qualitative evidence to enable understanding of the experience of receiving doula support. The evidence that is available from women who received doula support indicates a greater sense of participation during labour. 30 A study of the experience of receiving doula support in Sweden identified continuity, the ‘natural’ nature of the support provided and a human dimension to the birth experience as the key characteristics of doula support. Private doulas are available in the UK;31 these are usually accessed by women from higher-income groups who can afford to pay for their services, with the resulting potential to perpetuate inequalities in health and social support.

Although existing evidence from a range of countries identifies important benefits to the provision of lay support in labour, key questions remain. There is a paucity of UK evidence and doula support is rare in the UK, especially for disadvantaged women. Existing studies have as their major focus lay support in labour, yet there may be advantage in providing such support throughout the childbearing episode.

The original volunteer doula project

The first volunteer doula project was established as a social enterprise initiative and has provided support to women in situations of social disadvantage since 2005. The project developed in an area with high levels of social and economic deprivation, poor education, housing difficulties and health states lower than the general population. Women are referred to the service by health professionals, interpreters, social services workers and the Teenage Pregnancy Support Services. Support can be offered at any stage but commonly starts around the sixth month of pregnancy and continues until 6 weeks after birth. Following an initial facilitated meeting, subsequent contact occurs approximately fortnightly during pregnancy until the last month, when contact occurs weekly. This project therefore differs from many of the studies of doula support identified, several of which were limited to care in labour and the immediate postpartum period.

The original project also differs from others that have been reported elsewhere in what the doulas are trained to do. Women who volunteer to provide the doula service, who are themselves usually women from the local area with children, receive training for the role, accredited by the Open College Network (OCN). Topics included in the training are preparation for birth and the birthing process, breastfeeding, child protection, domestic abuse awareness training, cultural diversity and communication skills. The doulas are expected to work closely with existing services and to optimise women’s use of both health and social services, for example attending smoking cessation clinics, accessing Healthy Start and attending clinic appointments. Signposting women to other services is a key part of the doula’s role. Women referred to the service are matched with doulas according to personality, background, locations and availability. Volunteer doulas receive reimbursement of expenses, for example travel and childcare during training sessions. There are systems in place to provide ongoing support for the doulas through, for example, local project workers.

Descriptive data from the early years of the service indicated a range of benefit when compared with the whole population of the city; under normal circumstances, women with the deprivation profile of those cared for would expect substantially worse outcomes. There were also suggestions that experience as a volunteer doula had enabled subsequent access to employment and higher education, indicating a community development aspect to this work. 32 Descriptive data such as these informed the Department of Health’s (DH’s) decision, in March 2009, to provide 3 years’ funding (£267,000) to support roll-out and replication in up to eight additional sites. This funding supported the provision of a portfolio that informs establishing and running a volunteer doula service, including consultancy expertise for 1 year; support with issues related to human resources, volunteer recruitment and induction; ‘training the trainer’; promotional material and support; training for the first cohort in each roll-out site; and access to accredited training materials. Sites have to provide and fund their own staff. Identification of replication sites was slower than expected. By February 2011, four sites, which have substantially different service and demographic contexts from the original site, had confirmed service funding for replication.

Summary

The original doula service appeared a promising innovation, developed from an international and high-quality evidence base with demonstrable health benefits that aimed to maximise the use and efficiency of existing health and social care services. The technology, that of doula support, had been transferred from other countries, systems and settings and had been established in England for 4 years, where it had apparently been well received by women and well integrated with existing health and statutory services. It therefore met the criteria specified in the National Institute for Health Research Service Delivery and Organisation programme call for promising innovations in health care. DH funding had previously been awarded and four replication sites indicated a willingness to participate in this research. This research therefore offered the potential for an increased return on original DH investment while relating to several important policy areas.

An outline proposal was submitted to the National Institute for Health Research Service Delivery and Organisation programme on 30 March 2010 and a full proposal in October 2010. Funding was confirmed in January 2011. The research commenced in October 2011 and was carried out over a 21-month period in five settings in England.

Objective 1: implications for the NHS

1. To determine clinical and public health impacts for women and their babies, including type of delivery, low birthweight and admission to neonatal unit; to determine method of infant feeding planned during pregnancy, infant feeding initiated at birth and baby’s feeding method at 6 weeks of age; to determine impact on mothers’ smoking behaviour; and to compare these for women who have received the volunteer doula service with data for the general Hull Primary Care Trust (PCT) population, designated statistical neighbours and England averages.

2. To identify the impacts on and experiences of NHS maternity care services and providers (midwives and heads of midwifery).

3. To identify impacts on other NHS services including referral to and uptake of smoking cessation services.

4. To determine the actual and potential impacts on NHS maternity resource use of roll-out of doula support at scale.

5. To determine potential savings to the NHS through clinical events averted by the service.

Objective 2: health and psychosocial impacts on women

1. To identify underlying beliefs and theories about how the service works and the contexts in which it has more or less impact.

2. Based on this, to identify key outcomes which will allow the theories to be tested.

3. To identify the views, experiences and psychosocial impacts on women who have been recipients of the service.

4. To examine the characteristics and reasons of women who disengage from the service.

Objective 3: impact on volunteer doulas

1. To identify the views and experiences of the volunteer doulas and the impacts on their life course.

Objective 4: implementing and sustaining the service

1. To provide an independent assessment of the costs of providing a volunteer doula service, including training.

2. To identify the challenges, facilitators and barriers experienced by the manager and staff of the original initiative in establishing and maintaining the service.

3. To identify the process of agreeing funding for service costs and the main factors responsible for the positive decision.

4. To examine facilitators and barriers to implementation in the roll-out sites and the extent to which these differ between sites and from the original service.

5. To investigate the experiences of the replication package at the roll-out sites.

Conceptual framework

This study takes a realistic evaluation perspective,33 in recognition of the complex intervention being investigated in a real-life setting. 34 The focus is therefore not so much on addressing the question ‘does it work?’ but rather on the subtler question of ‘what works for whom in what circumstances?’. The identification of Contexts, Mechanisms and Outcomes (CMOs) at an early stage of the project is key, as they were required to generate hypotheses about the circumstances under which the intervention works. These CMOs are generated both from the literature and from the beliefs of key informants, in this case doula managers and project workers in all sites and doulas and service user representatives in the original site. This approach meant that later data collection could not be fully specified until after the identification of CMOs.

Accordingly, following a description of the literature review, our methods for data collection will be described in two stages. We will first describe the methods used in the initial stage of the research and the outcomes of this stage, including the generation of CMO configurations (CMOcs). We will then present the methods for the second stage of the study, which arose from this first stage and involved some modification to our original plans.

Doula service staff

Information about the research was provided directly to current and previous managers and subsequently to other doula service staff at all sites. Following provision of informed consent, digitally recorded individual interviews were carried out with the current and former managers of the original doula project and a group interview with project workers at the original site whose role includes matching women and doulas. Interviews mapped how the service works in practice and explored underlying beliefs about how it works, the contexts in which it has more or less impact and the enablers and barriers to establishing a doula service.

Individual or group interviews were then carried out with the managers and project workers in each of the roll-out services, exploring the development of the service to date and aspirations for the coming year.

Across the five sites, 25 individuals were identified who contributed to the running of the service, either in the office or in a managerial role, hereafter referred to as ‘service staff’.

Volunteer doulas

A focus group was considered to be the best way of eliciting the views and experiences of volunteer doulas so that their different perceptions could be discussed between group members. Participants were identified with the assistance of project staff in the original doula service with the aim of representing a range of views and experiences including volunteers from different training cohorts and with different sociodemographic backgrounds. Participant information leaflets were forwarded to potential participants by staff of the doula service to maintain confidentiality, with the opportunity to raise questions either via the service staff or directly with the research team. Following permission to contact/expression of interest, the focus group was arranged by the research team and held independently of project staff. Written consent was obtained from participants and the digitally recorded discussion addressed a number of areas including what might constitute a ‘good match’, which women might particularly benefit from doula support and impacts on doulas themselves from volunteering. Eight volunteer doulas (current and past) from the original doula service attended and participated in the focus group. These doulas subsequently agreed to be a doula user panel that acted as a source of advice in subsequent stages of the research, joined in a second focus group by a further three.

Service users

As with volunteer doulas, a focus group was considered to be the best way of eliciting views and experiences. Identical processes were used as for the volunteer doulas, described above. Seven women attended and participated in a focus group. These women agreed to be a women’s user panel that acted as a source of advice in subsequent stages of the research. Three women who had taken part in the first women’s focus group participated in a second group discussion.

All recordings were fully transcribed and thematically analysed to support the development of CMOcs.

Data sources

Reference groups

Process

As no data dictionary was available, we met the administrative team in the Goodwin Volunteer Doula project in Hull to understand the database. This allowed confirmation of fields to be used and the interpretation of them. After a preliminary analysis and presentation of results, data quality issues were identified that required resolution prior to further analysis. Data quality checks were carried out to highlight missing data, obvious errors and conflicting observations such as the method of birth or feeding. Because of the amount of work required and resources available, data items were prioritised for checking and correction. A number of additional fields were added to the doula service’s standard report, including a flag when the mother had disengaged from the doula service, a common reason for missing data; person identification (ID); and previous number of children.

The data set on which the final analysis was based was extracted on 27 December 2012.

The data are reported from the doula database as three data sheets: ‘referral’, ‘mother’ and ‘outcome’. Data linkage across tables was possible through the creation of a unique identifier consisting of the person ID and another number based on the number of times the mother had used the service. On extraction, the fields in the doula database are represented as a ‘0’ (no), ‘1’ (yes), ‘–1‘ (not known) or blank (not known). Items were collated on the outcomes sheet, including the mother’s age, postcode district, ethnicity and previous number of children.

Additional data items were created. These included:

• summary fields indicating feeding method at birth, feeding method at 6 weeks and mode of birth.

• a field indicating financial year based on the baby’s date of birth; where the date of birth was missing, an assumed date of birth was created based on the date of matching. Previous analysis had indicated that, on average, the mother was matched with a doula approximately 3 months before giving birth

• a grouping based on ethnicity where those defined as ‘white British’ were called ‘white British’ and all others were grouped into ‘non-white British’

• mother’s age category, summarised as ‘under 20 years’, ‘20 to 24 years’, ‘25 to 29 years’, 30 to 34 years’, ‘35 to 39 years’ and ‘40 years and over’.

As the database included records for mothers who had not yet given birth, a cut-off point was established of 31 August 2012. This was the last month with almost complete data and was included in the data cleanse. Births after that date were ignored to ensure the data quality was as high as possible. Analysis of the number of live births showed they were consistent over the period reported and it was felt unnecessary to make any adjustments for this.

Process

Data collected directly from women

Use of data collected by services

The information collected by doula services on their databases and paper-based service documents (including initial referral forms and service evaluation documents) provided an additional source of data that we had not been aware of at the time of writing our proposal.

In practice we did not find that these sources added materially to our data and they will not be reported in the body of this report, but further information about our use of these data and an overview of the findings are available in Appendix 5.

Process

To maintain confidentiality within doula services, we could not approach volunteer doulas directly. The process reported for contacting women (see To determine the impacts of volunteer doula support on key clinical and public health outcomes for women and their babies, Process) was also used to develop the sampling frame and for sending questionnaires to doulas.

The research team provided doula service staff with prepared packs for forwarding to doulas following the addition of an address label by service staff. The packs included a covering letter from the research team, questionnaire and reply-paid envelope for direct return of the questionnaire to the research team, thus maintaining confidentiality of response. The covering letter contained all aspects pertaining to participation in the study, including the voluntary nature of participation, and contact details for the research team to enable the potential participant to ask any questions. Centre codes (denoting the five sites) were printed on questionnaires.

We had originally planned for an initial and second wave of data collection to include doulas who completed training during the life of the research but for a number of reasons, including minimising demands on doula services and the protracted nature of governance processes, it became clear that a one-stage approach to this element of data collection, later than originally scheduled, was optimal.

The questionnaire asked volunteer doulas if they would be willing to take part in a telephone interview to explore any topics in more detail. Those who were willing completed a reply slip when returning the questionnaire providing their contact details and indicating their informed consent to contributing further. Respondents were not asked to provide contact details unless they wished to be contacted for participation in an interview. The sampling strategy for interviews was, therefore, a purposive one that enabled more detailed exploration of new issues identified in a questionnaire response.

Reminder postcards were sent to all volunteer doulas by doula service staff, 3 weeks after the mailing out of the questionnaires. This thanked those who had already completed questionnaires and invited completion from those who had not yet done so. If selected for participation in a telephone interview, doulas were contacted directly by a member of the research team using the details they had provided. The telephone interviews were digitally recorded (with one exception) and fully transcribed.

Doula service managers and staff

All staff employed by the doula services were identified by virtue of their professional role and invited to participate by a member of the research team. Doula service managers and staff were interviewed early in the evaluation, as already described, and a subset were reinterviewed a year later to identify changes and factors related to sustaining those services. In the proposal we had intended to limit second interviews to the roll-out sites, but it was recognised that issues of sustainability were equally important in the original site, so they were included in this aspect of data collection. A total of 11 staff contributed data.

Clinical and public health outcomes for women and their babies

Microsoft Excel (version 2010, Microsoft Corporation, Redmond, WA, USA) was used to summarise the data in pivot tables. Generally filters were used to remove mothers who were not supported, those who had disengaged from the service and those who gave birth (or had been expected to give birth) after 31 August 2012. Analysis of some data also required the use of ‘postnatal support only’ as a filter.

Confidence intervals were calculated using approved methodologies for public health observatories. 42 The doula service data were plotted onto bar charts alongside available comparisons, including those of designated statistical neighbours. Ninety-five per cent confidence intervals were included to indicate whether or not there were any significant differences between the doula data and comparators. Chi-squared tests were used to compare percentages of births by caesarean sections, breastfeeding at delivery and smoking in the doula population with data available from the maternity system at Hull and East Yorkshire Hospitals NHS Trust. Comparisons were made between outcomes derived from the doula database and those obtained from the maternity system at Hull and East Yorkshire Hospitals NHS Trust. Unfortunately it proved impossible to obtain all of the requisite data from the Trust data set, so we were unable to control for certain variables, such as parity.

The findings of these analyses are presented in Chapter 3, Clinical and public health impacts for women and their babies and Outcomes for women and babies, and further utilised in Chapter 7, where they have informed the economic analyses. Full details of all data sources and analytic methods for calculation of doula service and NHS costs are included in Chapter 7.

Quantitative analyses: questionnaires completed by women and doulas

On receipt of a questionnaire, each participant was assigned a unique study identity code number, for use on study documents and the Statistical Product and Service Solutions (SPSS, version 20, IBM Corporation, Armonk, NY, USA) database. Separate databases were created for women and for doulas. The women’s database included both supported and unsupported women, whether they had completed the questionnaire themselves or with assistance, with or without an interpreter. The doula database similarly included all doulas whether or not they had yet supported a woman one-to-one.

Data were analysed using SPSS. Descriptive statistics (means, frequencies, etc.) are presented where appropriate. Within-sample comparisons (e.g. of sites) were carried out as sample size allowed, using chi-squared tests for proportions and t-tests or analysis of variance to tests differences in means. When expected cell sizes were small, exact tests were used instead of chi-squared.

Qualitative data

Data collected using interviews, focus groups and free text data in questionnaires were analysed using content analysis43 to identify themes within the data, underpinned by the framework of realistic evaluation.

Each data source was treated independently. The transcripts from the interviews and focus groups were read and reread to gain a detailed familiarity with the overall accounts, and then systematically coded line by line to identify emerging themes. These themes were grouped and collapsed into higher-order conceptual themes with subthemes. For data related to women and doulas, emerging themes were considered in the light of CMO configurations. The open text questionnaire responses were tabulated using an Excel spreadsheet to show horizontally all of an individual’s responses to the questions and vertically all of the responses received to any question. This facilitated coding of themes identified in the responses on a question-by-question basis, identification of disconfirming responses and the exploration of linked patterns between questions.

For data collected from women and doulas, the final stage of the analysis was to integrate the findings of the qualitative and quantitative analyses to provide a comprehensive and integrated narrative of their experiences.

The findings related to the experiences of women and doulas are presented in Chapters 4 and 5 respectively, and those relating to implementing and sustaining the doula services in Chapter 6.

Quantitative data: costs of doula services and NHS costs

Full details of data collected and analytic method for doula service and NHS costs are included in Chapter 7.

Women supported by the original doula service

The aim of the following analysis was to identify the characteristics of the mothers who had been referred to the original volunteer doula service. Data have been pooled to produce an overall profile. At the point of extracting the data, a total of 603 mothers had been referred to the original service, of whom 599 were matched to a volunteer doula and 507 subsequently supported.

Method of referral

There were several routes of referral into the doula service. Referral sources were grouped together, as shown in Figure 1. Just over 43% of the 603 referrals were from a NHS professional and 23% by community midwives. The next largest category was ‘self-referral’.

FIGURE 1.

Method of referral summary, Hull doula mothers, all years, percentage. Source: developed from Goodwin volunteer doula database.

Most of the referrals by ‘other support services’ were from ‘asylum seekers/refugees support’ (3.3%) and ‘teenage pregnancy support’ (1.8%).

Mother’s postcode

Figure 2 shows the postcode districts of the 507 mothers supported by the original volunteer doula service in Hull. Most mothers lived within postcode districts HU1 to HU9, which match the boundaries of Kingston upon Hull local authority. Comparison data from HEY and the Office for National Statistics have been restricted to these postcodes.

FIGURE 2.

Mother’s postcode district, Hull doula users compared with female population, deliveries in Hull PCT and births in Hull and East Yorkshire Hospitals NHS Trust, all years, percentage. Sources: Goodwin volunteer doula database (all years); Hull and East Yorkshire Hospitals NHS Trust; Hospital Episode Statistics (HES) Copyright © 2013, reused with the permission of The Health and Social Care Information Centre. All rights reserved; Office for National Statistics (Census 2011). ONS, Office for National Statistics.

The Index of Multiple Deprivation (IMD) 2010 is a commonly accepted measure of deprivation. Kingston upon Hull is one of the 20% most deprived local authorities in England. Within Kingston upon Hull, as with most local authorities, there are differences in levels of deprivation across the area. Table 4 shows the percentage of women aged 16–44 years by postcode district in each national deprivation quintile.

TABLE 4 - Percentage of mothers aged 16–44 years living in each deprivation quintile

Sources: IMD (2010),44 2010 mid-year population. 45

Postcode district	Deprivation quintile
	1 (most deprived)	2	3	4	5 (least deprived)
HU1	99.8	0.0	0.2	0.0	0.0
HU2	100.0	0.0	0.0	0.0	0.0
HU3	85.5	14.0	0.5	0.0	0.0
HU4	43.6	14.5	24.1	6.7	11.2
HU5	16.6	44.9	26.5	8.9	3.1
HU6	53.4	5.9	24.8	15.9	0.0
HU7	49.7	13.3	18.7	18.2	0.1
HU8	37.7	22.3	19.1	13.3	7.6
HU9	71.6	13.5	14.8	0.0	0.0

The majority of the population of HU1, HU2 and HU3 are in the 20% most deprived areas. It would suggest that the Goodwin volunteer doula service is used by mothers from areas of deprivation. The exception to this is the HU9 postcode district. Only 14.4% of Hull doula mothers come from this area compared with approximately 18.1% of both the deliveries in Hull PCT and babies being born in HEY to mothers from Kingston upon Hull. The children’s centre from which the Goodwin volunteer doula service operates is based in HU3 and close to HU1 and HU2.

Mother’s age

The ages of women referred to the doula service ranged from 15 to 43 years (n = 599, mean 26.7 years, mode 26 years). Figure 3 compares the age of mothers using the Hull doula service with mothers who had a child in Hull PCT and with those who lived in Kingston upon Hull and had a baby in HEY. The age profile of women offered doula support reflects the overall profile of that for Hull PCT and HEY. Mothers under 20 years of age accounted for 14.8% of all doula service users compared with 10.8% of deliveries in Hull PCT and 11.3% of deliveries to Hull mothers in HEY. Mothers aged 35 years and over were slightly over-represented among doula referrals, while those aged 20–35 years were somewhat under-represented.

FIGURE 3.

Mother’s age, Hull doula mothers compared with Hull PCT and HEY, percentage. Sources: Goodwin volunteer doula database (all years); Hull and East Yorkshire Hospitals NHS Trust, Hospital Episode Statistics (HES) Copyright © 2013, reused with the permission of The Health and Social Care Information Centre. All rights reserved.

Ethnicity

Figure 4 shows the ethnicity of mothers referred to the service. Although greater detail is available, for comparison the mothers were grouped into ‘white British’ and ‘non-white British’. ‘Non-white British’ mothers are all those mothers who are not ‘white British’. Thirty-seven per cent of women referred to the doula service were from a non-white British background. This is over twice as many as the local population rates.

FIGURE 4.

Ethnicity, Hull doula mothers compared with Hull PCT, HEY and 2011 Census, all years. Sources: Goodwin volunteer doula database (all years); Hull and East Yorkshire Hospitals NHS Trust, Hospital Episode Statistics (HES) Copyright © 2013, reused with the permission of The Health and Social Care Information Centre. All rights reserved; Office for National Statistics (Census 2011).

Further analysis of the doula service database showed the most frequent non-white British groups were ‘white – other’ and ‘black – African’, which each accounted for approximately 10% of the doula mothers. While definitions may vary slightly, as a comparison the 2011 Census indicated 4.0% of females (all ages) in Kingston upon Hull considered themselves as ‘white other’ and only 1.2% of females (all ages) considered themselves as either ‘black/African/Caribbean/black British: African’ or ‘mixed/multiple ethnic group: white and black African’. 46

Parity

Figure 5 shows the number of previous children born to the 387 mothers who were supported and were engaging with the doula service in Hull. The number of previous babies ranged from none to eight. Almost 45% (173) of supported and engaging users were first-time mothers. A further 28.1% of mothers had one other child.

FIGURE 5.

Previous number of children, Hull doula mothers, all years. Source: developed from Goodwin volunteer doula database.

Among non-white British mothers, 46% were expecting their first baby, compared with 44% of white British mothers.

Disengagement

Almost 20% of mothers who were referred to the doula service were recorded as having disengaged from the service. As a result, there were fewer records available for analysis of outcomes than expected. The majority of these analyses are based on a sample size of around 330 mothers.

Chi-squared tests were used to examine whether or not disengagement varied by ethnic group or parity (Tables 5 and 6). Mothers who disengaged were more likely to be ‘white British’ than ‘non-white British’ [χ² = 20.3, 1 degree of freedom (df); p < 0.05] and to be first-time mothers than women with other children (χ² = 10.6, 1 df; p < 0.05).

Similar comparisons found no relationship between disengagement and postcode district.

A chi-squared test was used to compare the proportion of mothers who disengaged with those who continued with the service by age group. Because the numbers were small the categories ‘35 to 39 years’ and ‘40 years and over’ were combined into ‘35 years and over’. A summary of the data can be found in Table 7. Percentages are in brackets and may not add up to 100% because of rounding. Unknown ages were removed.

The age distributions of the disengaged and engaged mothers were significantly different. Disengaged mothers tended to be younger, particularly under the age of 20 years (χ² = 15.7, 4 df; p < 0.005). Mothers under 20 years old accounted for 27.4% of disengaged mothers compared with 10.6% of those who engaged with the service.

Epidural

Figure 6 shows the use of epidurals in all births (including C-section) for mothers using the Hull doula service. The use of epidurals ranged from 22.6% in 2008/9 to 35.5% in 2011/12.

FIGURE 6.

Epidurals given for all deliveries with 95% confidence intervals, Hull doula mothers, all years. Source: developed from Goodwin volunteer doula database.

Figure 7 compares the epidural use of mothers using the Hull doula service and mothers resident in Hull PCT. The data include normal and instrumental births only and suggest that epidural use by doula mothers is lower than that by mothers from Hull PCT. However, these differences reach significance only in 2010/11.

FIGURE 7.

Epidurals for normal and instrumental deliveries with 95% confidence intervals, Hull doula mothers compared with Hull PCT, all years. Source: Goodwin volunteer doula database; Hospital Episode Statistics (HES) Copyright © 2013, reused with the permission of The Health and Social Care Information Centre. All rights reserved.

Method of delivery

Figure 8 shows method of delivery by Hull doula mothers, pooled for all years (n = 327). It shows that normal birth (operationalised as all vaginal births excluding those assisted by forceps or ventouse and also excluding all caesarean sections) accounts for 71.3% of births. A further 21.1% of births took place by C-section and 4.6% were instrumental deliveries.

FIGURE 8.

Delivery method with 95% confidence intervals, Hull doula mothers, all years. Source: Goodwin volunteer doula database.

Caesarean section

Figure 9 illustrates a comparison of C-section rates for mothers using the doula service with those for mothers resident in Hull PCT and England. The percentage giving birth by C-section was lower for doula service mothers than for the other two groups, except in 2009/10. However, differences are not statistically significantly in any year, as indicated by the wide confidence intervals around the rates for the doula service mothers.

FIGURE 9.

Comparison of C-sections with 95% confidence intervals, 2008/9 to 2011/12, percentage of all deliveries. Sources: Goodwin volunteer doula database; Hospital Episode Statistics (HES) Copyright © 2013, reused with the permission of The Health and Social Care Information Centre. All rights reserved.

Full interpretation of C-section outcomes requires information about parity to be available. We had planned to conduct further comparisons of C-section rates by parity. We applied to the trust for these data but unfortunately a data set containing this variable could not be provided.

Babies admitted to the neonatal unit

Figure 10 compares the percentage of women using the doula service whose baby was admitted to a neonatal unit (including special care baby unit or intensive care unit) and the percentage of mothers from Kingston upon Hull whose baby was born in Hull and East Yorkshire Hospitals NHS Trust and admitted to the neonatal unit.

FIGURE 10.

Babies referred to neonatal unit with 95% confidence intervals, Hull doula mothers compared with HEY, 2009/10 to 2012/13 (quarters 1 + 2; pooled). Sources: developed from Goodwin volunteer doula database; Hull and East Yorkshire Hospitals NHS Trust.

The percentage of babies of doula-supported mothers admitted to neonatal units was higher than those of mothers resident in Hull who gave birth in Hull and East Yorkshire Hospitals NHS Trust. The small numbers result in wide confidence intervals and the difference is not statistically significant.

Smoking at delivery

Figure 11 shows the percentage of doula mothers smoking between 2008/9 and 2012/13 (partial year); generally around 15% to 20% of mothers smoked at time of delivery, with 2011/12 as an exception.

FIGURE 11.

Smoking at time of delivery with 95% confidence intervals, Hull doula mothers, all years. Source: developed from Goodwin volunteer doula database.

Rates of smoking at delivery for doula service mothers were compared with those for Hull PCT and the statistical neighbours, as represented in Figure 12.

FIGURE 12.

Smoking at delivery with 95% confidence intervals, Hull doula mothers compared with Hull PCT and statistical neighbours, percentage of mothers. Sources: Goodwin volunteer doula database; Health and Social Care Information Centre Return on smoking status at time of delivery.

The percentage of mothers using the Hull doula service and smoking at time of delivery is lower than in the statistical neighbours of Hull PCT, except for Salford PCT and Stoke-on-Trent PCT in 2009/10. The confidence intervals show that the differences are not statistically significant except for the comparison with Middlesborough PCT in 2010/11 and for all services combined in 2011/12. Smoking at time of delivery has traditionally been very high in Hull PCT compared with the rest of the country. Since 2005/6, Hull has always been in the top 10% of PCTs for percentage of mothers smoking at time of delivery. During the years for which doula service data were available, it was often in the top 5%. 39

Figure 13 presents the comparison of smoking among mothers using the Hull doula service with Hull PCT and mothers from Kingston upon Hull having babies in HEY between 2008/9 and 2012/13. The overlapping confidence intervals indicate that the differences are not statistically significant except for 2011/12.

FIGURE 13.

Smoking at delivery with 95% confidence intervals, Hull doula mothers compared with Hull PCT and HEY, 2009/10 to 2012/13 (quarters 1 + 2). Sources: developed from Goodwin volunteer doula database; HEY; Health and Social Care Information Centre Return on smoking status at time of delivery.

As identified earlier in Table 4, deprivation varies by postcode district.

Figure 14 shows smoking at delivery by postcode district. Hull doula mothers are compared with mothers resident in Kingston upon Hull giving birth within HEY.

FIGURE 14.

Smoking at delivery by postcode district with 95% confidence intervals, Hull doula compared with HEY, all years. Sources: developed from Goodwin volunteer doula database; HEY.

The percentage of mothers smoking at delivery is lower for doula-supported mothers across all postcode districts in Kingston upon Hull, except for HU5 and HU7. The confidence intervals indicate that these differences are statistically significant in postcode districts HU2, HU3, HU8 and HU9. As stated previously, postcode districts HU2, HU3 and HU9 have very high proportions of the populations living in the 20% most deprived areas.

Low birthweight

Figure 15 compares babies born with a low birthweight (less than 2500 g) to mothers who received the Hull doula service and to mothers resident in Hull PCT. Although only half as many babies with a low birthweight were born to doula mothers, the small numbers lead to wide confidence intervals and the difference is not statistically significant.

FIGURE 15.

Low birthweight, Hull doula mothers compared with Hull PCT, percentage of babies with 95% confidence intervals, 2009/10 to 2011/12 (pooled). Sources: Goodwin volunteer doula database; Hospital Episode Statistics (HES) Copyright © 2013, reused with the permission of The Health and Social Care Information Centre. All rights reserved.

Feeding method at birth

Data reflecting feeding method recorded at birth for Hull doula mothers were pooled. The most common method of feeding was breastfeeding (71.8% of mothers), with 17.9% of mothers bottle feeding and 7.0% mixed feeding. Feeding data were not recorded for 3.3% of babies. Figure 16 shows the rates of breastfeeding initiation between 2007/8 and 2012/13 (partial year); rates exceed 70% in four of the six periods for which data are available. The wide confidence intervals reflect the small sample size.

FIGURE 16.

Breastfeeding initiation with 95% confidence intervals, Hull doula mothers, 2007/8 to 2012/13 (quarters 1 + 2). Source: developed from Goodwin volunteer doula database.

Figure 17 illustrates the comparison for breastfeeding initiation between Hull doula mothers and Hull PCT and its statistical neighbours. It shows breastfeeding initiation is higher among Hull doula mothers than in the comparison groups for each year considered, but the confidence intervals indicate that it is statistically significantly different for all groups only in 2011/12. For 2008/9, doula-supported women were significantly more likely to initiate breastfeeding than all neighbours, with the exceptions of Salford and Stoke-on-Trent; for 2009/10, a similar pattern occurred, with Salford as the exception; for 2010/11, doula-supported women were significantly more likely to initiate breastfeeding than mothers in Hartlepool and Middlesbrough PCTs; and, for the first two quarters of 2012/13, doula-supported women were significantly more likely to initiate breastfeeding than mothers in Hartlepool, Middlesbrough and Salford PCTs.

FIGURE 17.

Breastfeeding initiation with 95% confidence intervals, Hull doula mothers compared with Hull PCT and statistical neighbours, 2008/9 to 2012/13 (quarters 1 + 2). Sources: developed from Goodwin volunteer doula database; Department of Health, Integrated Performance Measure Return.

When compared with data from Hull PCT and the HEY (Figure 18), it appears breastfeeding initiation is significantly higher among women supported by the doula service, with the exception of 2010/11.

FIGURE 18.

Breastfeeding initiation with 95% confidence intervals, doula mothers compared with Hull PCT and HEY, 2008/9 to 2012/13 (quarters 1 + 2). Sources: developed from Goodwin volunteer doula database; HEY; Department of Health, Integrated Performance Measure Return.

Mothers from the Hull doula population were more likely to breastfeed at delivery (74.3%) than mothers from HEY (55.2%) (χ² = 46.40, 1 df; p < 0.0005).

Further information about breastfeeding and postcode district is available in Appendix 8.

Breastfeeding at 6 weeks

Breastfeeding at 6 weeks was compared with published data related to total and partial breastfeeding at 6–8 weeks. Figure 19 provides information for total and partial breastfeeding for doula mothers compared with Hull PCT and its statistical neighbours. Figures are percentages of maternities where the feeding method is known. At the point of extracting the data from the site A doula database there were 20 mothers whose feeding method was not known for 2012/13. The percentage of doula service mothers breastfeeding at 6 weeks is significantly higher than all the statistical neighbours in each year.

FIGURE 19.

Total and partial breastfeeding at 6 weeks, doula mothers compared with Hull and statistical neighbours, 2008/9 to 2012/13 (quarters 1 + 2). Sources: developed from Goodwin volunteer doula database; Department of Health, Integrated Performance Measure Return.

Summary of findings: clinical and public health outcomes

Table 8 provides a summary of findings of comparisons for clinical and public health outcomes between mothers who received the doula service and comparison data sets.

Midwives

Focus groups were held with midwives in four of the five trusts providing maternity care for women offered volunteer doula support. A total of 31 midwives participated. In the fifth trust it did not prove possible to convene a group within the required time scale. Midwives who participated held a range of roles and were of varying seniority, from newly qualified midwives to managers and consultant midwives. Awareness of, and personal contact with, doula services varied within groups. Our aim had been to achieve a sample with experience of providing care to women who had received doula support. However, while interested in the topic, at several sites some participants were without such experience. Negative perceptions of paid-for doulas were voiced spontaneously at all sites. Midwives’ quotes are coded for site and participant (MW, midwife).

Heads of midwifery

Telephone interviews were carried out with heads of midwifery (HoMs) in the five trusts providing maternity care for women offered volunteer doula support. Interviews were digitally recorded, transcribed verbatim and subjected to content analysis to identify themes. Interview duration varied between 35 and 50 minutes. To maintain anonymity of participants in unique roles, all quotes are attributed only to a head of midwifery and the site identifier omitted. All five heads of midwifery who participated in the interviews contributed quotes in the following section.

The extent to which the questionnaire sample is representative of women introduced to the doula service

Site A questionnaire data were broadly comparable to the Goodwin database data for ethnicity but the questionnaires under-represented primiparous women (33.7% vs. 49.1%). Consistent with this, the mean age at introduction to the doula service was higher in the site A questionnaire sample (mean 28.1 years, SD 6.0 years; n = 95) than in the database (mean 26.7 years, SD 6.4 years; n = 595), even allowing for the time elapsed.

Year of introduction

Year of introduction to the doula service ranged from 2006 to 2013 (Figure 20) with the majority between 2010 and 2012 (used, 107 of 131; did not use, 26 of 29). Introductions prior to 2010 were only at the original site (used, 23; did not use, 2), reflecting the histories of the services.

FIGURE 20.

What year were you introduced to the service?

Finding out about the service and getting in touch with the service

Women had found out about the service through a diverse range of sources and commonly through more than one route; thus, where someone had contacted the service on the woman’s behalf it was not always possible to determine the source of referral.

The most common way of finding out about the doula service was being told about it by a midwife (total sample: 76 of 159; 47.8%). This was particularly pronounced at site W (15 of 21; 71.4%), where the doula service is based within the NHS maternity service. At the only site where midwifery was not the most common route (site Y), finding out about the service predominantly occurred through doula service outreach or contact with the host organisation’s affiliated services (e.g. antenatal classes: 7 of 18; 38.9%). Being told by a family member or friend was the second most common across sites (26 of 159) and comprised approximately half of women who used the service at sites X and Z (in both of these sites the Mechanism ‘word of mouth’ was identified as important in the early data collection with key informants). Only one woman reported having heard about the service via local media (newspaper/radio) and she had also seen a leaflet. Key informant data identified this route for promotion but particularly in the early stages and for recruiting doulas, yet this does not appear to have been an effective Mechanism for obtaining referrals.

Defining support

Early in the questionnaire, women were asked to identify whether or not they had received support for the following: during pregnancy (antenatal), labour/birth (intrapartum) and after the birth (postnatal). There were some discrepancies between these responses and subsequent endorsement of various support behaviours that were listed for each stage. Of 22 reporting no antenatal support, nine selected antenatal support behaviours, including four women who showed a similar discrepancy concerning postnatal support. Of 53 reporting no postnatal support, 32 selected postnatal support behaviours. There was some suggestion of having misunderstood the initial questions; English was a first language for one of nine supported during pregnancy and 14 of 32 supported after the birth. There was also a suggestion that discrepancies may reflect the woman not having felt supported during that period, with one woman writing ‘only had 2 visits’; this will be discussed further in (Dis)satisfaction with the service, Did the service help you in the way you hoped? of this chapter.

Stages supported/combinations of support

The distribution of combinations of support is shown in Appendix 11. Receiving doula support in all three stages (antenatal, intrapartum and postnatal support) was the most common combination, reported by 47.8% of respondents. All of the women who received postnatal support only (n = 9) were introduced to the service after the birth; the majority of these (eight of nine) were supported at site A. Of the 127 women who were introduced to the service prior to the birth, 75 (59.0%) had their birth attended by a doula; however, this varied significantly by site (A, 58.5%; W, 78.6%; X, 88.9%; Y, 42.9%; Z, 25.0%).

Late matches, operationalised as having intrapartum support without antenatal support, identified 11 women based on the initial support questions and six women based on support behaviours, three of whom were common to both. The 11 women came from site A (n = 5), site W (n = 5) and site X (n = 1), whereas the six came from site A (n = 4) and site X (n = 2). Women matched later were predominantly born outside the UK and, had they not had doula support, would have been birthing alone. Information on the combination of stages supported by site is presented in Appendix 12.

The most important stage

Both women and doulas were asked their views on the most important stage to have doula support (before the birth, during labour/birth, after the birth). Women were asked ‘If you could only have had support at one stage, which would it be?’ Responses are presented in Table 13.

Of the 128 women who completed the question, 118 selected one stage, as requested, the most common of which was labour/birth (50.8%), followed by antenatal (30.5%) and postnatal (18.6%). The remaining women selected antenatal and labour/birth (n = 4), antenatal and postnatal (n = 1), labour/birth and postnatal (n = 2) and all three stages (n = 3). Women were not asked to provide reasons; however, one woman selecting all three stages wrote:

I can’t pick it was all essential – removing any one part would make service pointless.

Woman Q457A

Of the 36 women who selected antenatal, 35 had received antenatal support; the remainder had postnatal only. Similarly, 21 of the 22 who selected postnatal had received postnatal support; the remainder had antenatal only. In contrast, 10% of the women favouring intrapartum support (6 of 60) had not received intrapartum support; this comprised one antenatal only, two postnatal only and three antenatal and postnatal only. Intrapartum support was the most commonly endorsed stage across all sites except site Z, where antenatal was most common. Of note, at site X all women selected labour/birth, including one who additionally selected antenatal.

Intensity of support

Women who had used the service and self-completed the questionnaire (n = 119) were asked the average weekly number of hours of doula support received.

Women on average reported having 2 hours’ support each week (median = 2.0 hours, interquartile range = 1.5 hours; n = 98); however, this ranged greatly between 0 and 10 hours. This was one of the items with the most missing data. There were some issues of recall (where women wrote ‘can’t remember’), which may also offer insights into the perceptions of support.

Some comments suggested that the service was individualised and needs-led rather than a prescriptive amount of time per week. However, this item was commonly the place where any discontent with the service was expressed:

My doula kept cancelling the day before so I only saw her twice. [Amount not reported]

Woman Q307A

I only saw her every 2–3 weeks for about 3 hours. [Amount not reported]

Woman Q385X

Was too late saw doula twice. [0 hours/week]

Woman Q449A

She only spoke to me on the phone. She never visited me. [0 hours/week]

Woman Q313A

Intensity of support and recipient characteristics

Key informant data (site A) had identified three contexts which might require more intensive doula support: (1) women in particularly complex social situations, (2) women who self-referred – seen as keen but less in need – and (3) women who lacked social support. However, none of these factors was found to be statistically associated with the intensity of support.

The distribution of average weekly hours by site is shown in Appendix 11. There was no significant difference between sites.

Support during pregnancy

It was striking that none of the support behaviours was endorsed by all women who were supported in that period. During pregnancy, the most common type of support was home visits (87.6%). Telephone support was reported by 65.3%. Information giving was also common (71.9%) and birth preparation was reported by 70.2% of women including 60.0% (30 of 50) of those who received antenatal support without intrapartum support. Giving information was more commonly reported by first-time mothers, but, interestingly, birth preparation did not differ by parity. This may relate to suggestions that a high proportion of the multiparous women supported had had previous bad birth experiences.

Intrapartum support

The listed practical and emotional support measures were all endorsed by the majority of women receiving intrapartum support (range: 63.8–85.9%); the most common was encouragement. In addition to the listed practical intrapartum support behaviours (see Figure 26), women listed other support measures that often spanned practical and emotional dimensions (see Figure 27):

Got a cool flannel and wiped it over my forehead because I felt too hot.

Woman Q333A

She helped to have a shower after birth and she take a photo for me and baby and daddy together.

Woman Q314A

Held my hand, kept me calm with speech.

Woman Q456A

My doula is pain killer for me.

Woman Q356Y

Talking to me, reassuring and explaining midwife speak and getting my points across.

Woman Q315A

Held my hand, was another female presence in the room.

Woman Q385X

Postnatal support

Figure 28 demonstrates that, during the postnatal period, the most common form of support was again home visits (92.9%). Attending appointments and going ‘out and about’ were only half as likely as in the antenatal period. Breastfeeding support was reported by half of the women. Other support that was described included bathing the baby (n = 4) and visiting the woman in hospital (n = 2).

The main differences in support measures across sites concerned accompanying the woman on walks and trips to cafes and introducing women to other services. Accompanying on walks and trips to cafes was most common at site A, particularly in the postnatal period, where 18 of 76 reported such support compared with none of one (W), one of six (X), 1 of 12 (Y) and neither of two (Z), although this may have reflected cultural patterns at sites W, X and Z.

Helping women to use other services

Women were asked directly about doula support in using other services. Of 83 (64.8%) who were told about a service, 22 said that their doula contacted a service on her behalf and 23 said that she accompanied them to a service, including 12 who did both. Of 45 who did not indicate that their doula had told them about a service, nine said that the doula had contacted a service on her behalf and eight said that she accompanied them to a service; none said both.

There were notable differences between sites concerning signposting. Although ‘Told me about a service’ was relatively balanced across sites (A, 55 of 86; W, 7 of 13; X, 7 of 9; Y, 11 of 14; Z, 3 of 6), ‘Contacted a service on my behalf’ was not reported by any women at site X or Z (A, 23 of 86; W, 3 of 13; Y, 5 of 14) and ‘Came with me to a service’ was more common at sites A and Y (A, 23 of 86; W, 1 of 13; X, 2 of 9; Y, 5 of 14; Z, 0 of 6).

Unique aspects of the doula role/what makes the doula role different?

Data collection with key informants identified several aspects of the volunteer doula role that distinguish it from other types of support. These were investigated by asking women how they think of their doula, if and how the relationship would have varied if they had met their doula sooner or later than they had, if and how the relationship was similar with the main and the back-up doula, and views on payment.

Women were asked how they thought of their doulas and were presented with a list of possibilities (Figure 29).

FIGURE 29.

How do you think of your doula?

By far the most common was ‘as a friend’ (74.6%); however, discussion of the relationship’s ‘ending’ (discussed further in Impact of doulas: feelings of loss at the end of support) questioned the possibility of ‘friendship’. One woman, reflecting on the compulsory end to this relationship, characterised it sadly as ‘temporary friendship’ (F337A), a distinction echoed elsewhere.

Focus groups offered more detailed discussion of the ways in which the doula role was similar to and contrasted with support from family, partner and professionals. These discussions illustrated what women valued about the support and offered insights into the ways that it ‘worked’ for them.

Doula’s role is different from partner/family

Thirty-one per cent of women described their doula as ‘like a mother’ and/or ‘like a sister’. Of note, none of the 21 women likening the doula to a mother had a supportive mother around during their pregnancy and no-one with a supportive mother described the role in this way.

Women who had social support from a family or friend nonetheless valued the support of their doula. They pointed out how support from a doula differed from support from family or friends, who are emotionally involved in the situation and may have their own needs and agenda. In particular, family and friends would tell women what to do whereas doulas talked through options and asked what the women wanted to do. They were also better informed than family and friends about pregnancy and birth:

Your Mum, your sisters, even your husband, they are busy, everyone’s busy, and [the doulas], this is their job, and they willingly do the job, they want to help someone. And if you ask for help from your family, they are, like, forced to do that, but the [doulas] do it willingly. They were both like an angel for me.

Woman F369Z

My husband, as much as he wanted the birth that I had, he just wanted the pain to stop, but the doula, was more like, well, no, that’s not what we agreed.

Woman F337A

With your friends and family, you almost feel like you’ve got to look after them . . .

Woman F315A

[The doula is] somebody you can tell things that you can’t tell your partner about, because he’d start panicking. Whereas your doula has been there, she has done that, she has had that dodgy feeling on the boob or whatever.

Woman F337A

Doulas are different from health professionals: continuity and trust

Many women said that they had been cared for in pregnancy and postnatally by teams of midwives and health visitors and saw a different person at every appointment; the midwives at their birth were strangers. By contrast the doula provided continuity during the perinatal period. Women had time to get to know and trust their doula and to feel that they were known and understood by her; they regretted that this relationship did not exist with health professionals. Women saw doulas as a reliable and accessible source of up-to-date information. Some women were reassured to hear the same information from their doulas as from their midwives, although in many cases the doula was filling information gaps that had not been addressed by the midwives. Doulas were also contrasted with midwives in that women felt they could ask them anything, even questions that they were embarrassed to ask busy health professionals, who sometimes gave dismissive or patronising answers.

Obviously, through pregnancy, because this was my first baby, everything is completely new and, you know, you have no idea what’s going on and I had so many questions that I asked that were all silly, in my head, or embarrassing, that you wouldn’t want to ask somebody, but I think you make a relationship with [the doula] and you trust everything that they say to you and you say [it] to them . . . When you’re with your midwife . . . I think that they’ve got an agenda, things they‘ve got to do, they have to check your blood, check the baby, check this, check that and then . . . they say, is there anything else you want to ask? So that kind of [embarrassing] thing, for me, I think I won’t ask that, you know what I mean!

Woman F341Y

They also described how the doula was a person you could rely on when you needed her; this was again in contrast to midwives, who gave out their phone number but did not always answer the phone:

[The doula is] the one constant person that you get to know through it all, because I don’t know about everyone else, but I had different health visitors every time I went, different midwives every time I went, whereas the doula from when I got to know her all the way through . . . for me my doula was a stop gap between what these people should have been in the first place . . . it’s the person you tell it all once to and then sort of carries you through all the other things.

Woman F315A

I called the midwife and I wouldn’t get through and that really, really scared me and I can only imagine if other women, who are going through the same thing in a more difficult situation and the fact that I actually had a [doula] to call, to even make me feel relaxed and calm and give me that confidence, don’t worry, give me that support.

Woman F386Y

She was like a friend to me, whereas my midwife, I could only visit her when there is an appointment, or she could just examine the medical things, but an emotional support and all those things were coming from my doula.

Woman F468W

For the small proportion of women who were critical of the doula service, it was most commonly their doula’s unreliability or inability to provide continuity that was criticised: the doula had not visited or contacted them often enough, for long enough or with her full attention, or had not kept appointments. Some women were very disappointed that their doula had not turned up when expected in labour. This perceived lack of commitment was felt to undermine the very purpose of the doula’s involvement; for some women there was no meaningful emotional support, useful information or birth support:

[The doula] didn’t come regular and seemed distracted with personal matters.

Woman Q346A

The doula contacted me a few times but would have loved to have been consistent up until my labour.

Woman Q458Z

Went into labour and could not get hold of Doula and ended up with emergency C-section.

Woman Q430A

Doulas are different from health professionals: a woman-centred service

Women commented that one of the distinctive features of the doula service was that the doula focused completely on the woman and her needs with no other agenda, and had time to spend with her, unlike busy health professionals for whom you are ‘a number . . . get out, hurry up!’ (F409Y). This was true both during pregnancy and in labour, and some women described how doulas supported and comforted them in labour while the midwives focused on clinical checks. Some women explicitly perceived the doulas as filling in gaps left by overworked health professionals. Others contrasted the doulas’ non-directive information with their experience of information from health professionals. As noted above, women said they could ask the doulas anything, even things they were too shy to ask health professionals, who could make women feel like they were wasting the health professionals’ valuable time:

I think the difference is really that you have somebody who has complete time for you, it’s just that they’re there for you and they haven’t got any other pressure of things that they have to do, or go.

Woman F341Y

Having someone there to support me, who wasn’t a health professional and gave me the time I needed rather than treat me like a number.

Woman Q373Y

Midwives are so stretched [doulas] help when midwives are not present.

Woman Q427A

There are a lot of services that women are not aware of, my doula empowered me to get well informed which in turn gave me control over what I wanted, rather than someone deciding for me, which tends to happen as most health professionals take that right away.

Woman Q373Y

Doulas are different from health professionals: a flexible, personal service

The flexible and responsive nature of doula support was shown by the perception that doulas adopted multiple roles; for example, 24 of 34 women describing the doula as a professional also described her as a friend. This was summed up in a focus group:

We were like friends, she was like a sister, a friend and a mum!

Woman F366Y

Some women expressed their surprise and appreciation for the level of commitment their doulas showed, attending extremely long labours or making themselves available whenever the woman needed them. Doulas had often gone beyond the remit of their role in the support provided: they had checked a diabetic woman’s blood sugar every hour in labour when she could not manage because of drips; helped fill in benefit forms for a woman who could not read or write; taken a woman’s clothes home from the hospital to wash them; and supervised the contact visits between a woman and her older child in care:

I was stay 3 days in labour . . . she is with me all the time, like 8 hours, 7 hours a day she is staying with me and then after when I had the baby, she stay with me as well [. . .] I was 5 days in hospital. She stay with me, she’s helping me how to feed the baby, everything she’s helping me and after that 1 month she is coming at home as well, she visit me and she show me everything about the baby, she is helping me, everything.

Woman F404W

When I was in labour and needed to go to hospital it was . . . nearly 5:00 in the morning. I called [the doula] and told her about that. She only took about 15–20 minutes reaching my home. Then she took me to the hospital and stayed there with me for almost 11–12 hours . . . She helped me and took care of me just like my mother.

Woman Q455A

[The doula said] Or any time if you need me, you’re feeling depressed, give me a call, even if it is midnight, 12 o’clock in the night, you can give me a call, I can come, we can have a chat.

Woman F468W

Another aspect of the doula service that women commented on positively was the way that the doulas followed their lead and tailored the support to what they wanted, although it varied with how proactive the doula was in arranging to meet up:

They, kind of, take control of the situation without being controlling, you know? . . . She reacted to what I asked her, she’d say, what do you want to talk about next week? Is it, you know, and then she’d find the information ready for the week after.

Woman F341Y

However, there were also some women who had been disappointed at the limitations of the doula service and would have appreciated being able to ask their doula for other support, for example to help care for older children while they were in labour if necessary. One woman described how the service was presented to her entirely in terms of boundaries:

I don’t know it was quite peculiar because I didn’t know what to expect from her and when she initially turned up she went through a whole ream of things she wasn’t allowed to do but didn’t really go into the specifics of what she could do. So it was a bit strange at first. And there was a point I nearly said I don’t want to carry on with this.

Woman F385X

Timing of match

The CMO development had identified the timing of onset of support as a potentially important Context; specifically, the intervention may not ‘work’ where the match happens late in the antenatal period and there is not time to establish a relationship (including ‘trust’) before providing intrapartum support.

Understanding the relationship: comparing the woman’s relationship with her main doula and her back-up doula

Fifty-two women (43.7%) reported having a back-up doula; this was most common at sites W and X (A, 31 of 82; W, 9 of 13; X, 5 of 7; Y, 5 of 14; Z, 0 of 1) and among women having intrapartum support (59.4% vs. 22.0%). Frequency of contact varied. Of the 32 (64.0%) who came on joint visits, 12 did so only once. One-third (3 of 9) of those who attended the birth had not previously met the woman. Two women reported having had separate visits with the back-up doula.

Women were asked if they felt as comfortable with their back-up doula as with their main doula and to provide reasons for this. Women who were not born in the UK were significantly more likely to report feeling as comfortable with the back-up doula (18 of 23, 78.3% vs. 9 of 21, 42.9%).

One woman reported feeling as comfortable because she saw both for the same amount of time. All other responses indicated a sense of similar attributes and qualities that meant that this was viewed as the ‘same service’ (Q314A) and part of the woman’s team; for example:

Because she’s got the good qualities as the main doula.

Woman Q359A

Because they’re all friendly and nice.

Woman Q361A

Very similar & supportive.

Woman Q429A

Women who did not feel as comfortable with their back-up, including two who first met the back-up at the time of labour, usually attributed this to not having met her and this meant that there had not been the opportunity to develop a relationship. In the focus groups, most women felt that they did not have a meaningful relationship with their back-up but felt that having an unknown back-up was better than having no birth partner, consistent with the views of a woman who had experienced this. Others felt that the difference was due to the back-up’s background or personality, which was consistent with key informants’ discussion on the importance of ‘matching’:

I feel [that] my [main] doula [was] like mum.

Woman Q306A

[Back-up] was more outgoing and dominating.

Woman Q334A

She was young with no children. I didn’t feel she would be as knowledgeable at the birth. Although she was very nice.

Woman Q423A

Understanding the relationship: views on payment/doulas as volunteers

Women who had received doula support were asked their views on whether or not the ‘volunteer doulas’ should be paid under two conditions: with women still receiving a free service and with women paying for the service. Whereas 75.7% of women (84 of 111) felt that doulas should be paid if the service remained free to women, this reduced to only 25.5% of women (27 of 106) if women were required to pay for the service. There were no statistically significant differences between sites in the proportion of women advocating payment of doulas provided that the service remained free to women. In contrast, when payment of doulas required payment by women, there was a statistically significant association between site and view on payment, with the roll-out sites being more positive about this scenario than site A (A, 22%; W, 40%; X, 50%; Y, 58%).

Women expressed deep appreciation that a stranger was willing to devote so much time to them as a volunteer. On the other hand some women could feel inhibited about bothering a volunteer even when they wanted her support and it was against cultural norms for a relationship to be so completely one-sided (e.g. not being allowed to give the doula even a small gift).

I felt I could not ask for more support when needed.

Woman Q380Y

Request system to contribute to doulas because I heard they do voluntary work. It used to bug me. I felt bad she is helping me for nothing.

Woman Q369Z

Some women saw an element of reciprocity in that doulas were gaining valuable job experience or clearly enjoyed their role. One woman described how she trusted the doula more because she had refused a gift (Q305A).

Women had mixed views about whether or not doulas should be paid: many felt that they deserved to be paid because they work hard supporting women in their free time and some (particularly those who were dissatisfied with the amount of support they had) thought payment would be good as it would increase the amount of time doulas could spend with a woman. Others felt that paying doulas might change their motivation, distort the relationship or make the service more vulnerable to cuts; some noted that doulas want to be volunteers to gain valuable experience and they enjoy volunteering. There was, however, a very strong view that women should not have to pay the doulas themselves, as this would be an insuperable barrier to using the service for those who need it most, or that only women who could afford it should be asked to pay.

There were reservations about the implications if women had to pay for the doula service:

Because people wouldn’t engage in it if it cost money – like paying for somebody to be your friend.

Woman Q341Y

One woman (Q394A) mistakenly believed that ‘This is part of the NHS services and should be free’.

Comparisons with the Picker Institute maternity survey (site A)

Some of the areas of impact were measured using questions from the Picker Institute maternity survey. This allowed us to make comparisons between women who had received the doula service and data collected in the same PCT on some additional outcomes that are not routinely collected, such as methods of pain relief and feeling involved in decision making. We were able to do this only for site A, as numbers were too small in the roll-out sites. Even in site A, analysis had to be limited to the 83 women who had used the service and self-completed the questionnaire, since the relevant questions were not included in the shorter telephone version. Because of small expected values for some response options, cells were combined so that all analyses were comparing the first response option (e.g. ‘yes, always’), with all other combined responses. Tables show data for all trusts in the final column for comparison, but analyses were carried out only with respect to the Hull PCT. Columns for Hull PCT and all trusts show percentages only, not frequencies. This is because the data made available to us had been weighted so that trusts could be standardised to age and parity distributions for the country as a whole. We applied the same weighting formulae to our own data, but it made little difference to the findings, so we have presented unweighted data for site A for clarity. Nevertheless, the parity distribution of our data was significantly different from that of the local Picker survey: only 31% of site A women in our sample were primiparous, in comparison with rates of 54.7% in the local trust and 49% nationally. Our sample may also be atypical in terms of ethnicity, although ethnicity data were not available for the Picker sample.

Impacts on women’s emotional health and well-being

Mental health and well-being were targeted through one question concerned with postnatal mood: ‘Thinking back to the weeks after your baby was born, how did you feel?’ Nearly one in five women (19.2%) reported feeling ‘depressed or low spirited most of the time’, 38.4% felt ‘sometimes cheerful and sometimes depressed or low spirited’ and 42.4% felt ‘reasonably cheerful most of the time’.

The qualitative data highlighted the significance of doula support for mental health and well-being. Some women said explicitly that the doula helped them with feelings of depression; this was not limited to those women who had been referred because of past and current mental health difficulties:

Before I got my doula I was terrified & depressed (because of the health problems) & she helped me change my view around.

Woman Q456A

She turned all my negatives into positive and encouraged me to lead a positive and happy life as this would also have impact on lives of my children . . . In a few weeks her positive attitude really influenced me and I feel I will never forget the support she gave me.

Woman Q453A

This service would have been so appreciated and needed during my 1st pregnancy and I feel if I had had a doula then I probably would not have suffered with post natal depression (as much) and highly recommend it to all 1st time mothers even if they have family support.

Woman Q353A

I could text her at 2:00 in the morning if felt down and she would reply back to me, you know, 24 hours on the phone.

Woman F319A

I think the service should be there for all mothers so won’t feel scared or lonely, or oh my god, that’s the end of life . . . I had depression . . . When I found out about [the doulas], I really needed them and they came straight to see me. That’s when I saw hope.

Woman Q369Z

Impact of doulas: making healthier choices

Impact of doulas: better birth experiences

Impact of doulas: empowerment

Many women commented on how the doula had changed the way they felt about their maternity care. The combination of non-directive information and building up women’s confidence through affirmation led to women developing an awareness that they had choices and feelings of being in control of their maternity care. Women experienced this as very empowering, whether they were nervous first-time mothers or mothers who had experienced a previous difficult birth. In some cases this led to them making different birth choices about pain relief or attempting a vaginal birth after caesarean.

Several mothers said that, even though the doula had not been with them during labour, the way she had given support during pregnancy meant that they could cope better in labour because they felt more in control:

The doula that I had, she was very, she was my second voice, I would tell her things and then she would do, well we could do this, well we could do that . . . She gave me the confidence to say, no, I don’t want to do that, or, yes, I want to do this, or, this is how I’m feeling right now.

Woman F337A

Made me feel in charge of my care in pregnancy and childbirth.

Woman Q444A

Gained confidence and belief in myself to deliver naturally and once my baby was born to get out the house with two babies.

Woman Q334A

Incredibly empowering that she helped me stick to my guns, give birth at home with no pain relief.

Woman Q302A

If I didn’t have a doula I would have opted for an epidural like with my other two but she suggested gas and air and if needed go from there and I’m glad I only had gas and air as I could deal with labour better.

Woman Q352A

One woman described how her doula researched the possibility of having a vaginal birth after an earlier caesarean, accompanied her to all antenatal appointments and coached her on what to say to the midwife. Although this account may raise questions about the boundaries of the doula’s role, the woman herself was extremely positive about this support:

[The doula] did her research, did everything and she was there, sitting there and it was before we went in, she told me what to say and then I had so much confidence to talk to this midwife and I was telling her, in fact, I found my voice and then when we left, she was saying, I don’t like what [the midwife] said, I’m going to find out and, you know, I found out that actually I was empowered, I didn’t have to do what the midwife told me to do, I have to do what I want to do . . . we were prepared before every antenatal, because she went for every single one with me, each time when I call her she was there, she just sits and listens and I say what she’s told me to say . . . I say it because I know it’s right.

Woman F409Y

Impact of doulas: better communication with health professionals

Some women felt a gulf between themselves and the health professionals in hospital. In these situations the doula was an effective advocate for them and they valued her role as a birth partner who could talk to health professionals on their behalf:

[The doulas] give mother second mouth, partner, to talk for her, when she need help.

Woman Q336X

Yeah, I think what my doula did was to introduce herself and say, look, she’s weak, she won’t be able to communicate with you much, so I’m here to communicate, you know, what she wants and I was very weak, I was sleepy most of the time, the only time I wake is when I actually need something and I tell her and she, you know, help pass the information around.

Woman F386Y

A friendship built on trust was formed. So during labour she encouraged me to do as the midwives suggested and I trusted her to make the right judgements.

Woman Q409Y

In some cases the doulas supported women who could not communicate with health professionals, either because of language differences (where no interpreter had been provided, especially at site Z) or because the mother felt overwhelmed and confused by medical terminology:

[The doula] also discussed waiting to clamp and cut the cord while having skin to skin with the healthcare professional – something they had been reluctant to let me do.

Woman Q457A

[Because we had a doula,] after the birth we didn’t feel useless about communicating with the midwives.

Woman Q315A

Things and terminology [the doula] was able to tell me what was happening so I understood.

Woman Q454A

One non-white woman described how being accompanied by a doula to outpatient appointments meant that she was treated more respectfully by staff than if she had been alone:

The hospital, the doctors and the nurses, if they see a [doula or community midwife] with the patient, they don’t do that [behave badly], but if the patient is alone, by themselves, even if you speak English, it’s two different things. They would listen more if you are not alone.

Woman F369Z

Impact of doulas: supporting partners and mothers’ relationships with their partners

Women were asked about potential impact of doula support for the partner through two questions: one concerning intrapartum support and one concerning the woman’s relationship with the partner.

Of the 36 women describing the impact of doula support for their partner, husband or other support person, 22 (61.1%) reported that the doula ‘made sure they were included’, 22 (61.1%) that the doula ‘showed them how to support me’ and 20 (55.6%) that the doula ‘gave them breaks from supporting me’. Occasionally, emotional impacts were also described, for example:

Helped them stay positive.

Woman Q456A

Asked if having a doula made a difference to the woman’s relationship with her partner or husband, the majority (n = 46; 59.0%) said no, 24.4% (n = 19) said yes and 16.7% (n = 13) were unsure.

Most of the comments from women with partners reflected one of five themes: having a doula to confide in meant the woman did not have to bother her partner with her worries; having a doula reduced stress because roles were shared, with the partner able to focus on looking after other children or (for a disabled mother) physical aspects of birth support; the doula supported both parents during birth, which meant the mother did not have to worry about supporting her partner herself; the doula had a direct impact on the partner by explaining things, motivating or reassuring him; or the doula had a direct impact on the relationship by helping the couple at a stressful time.

Women described how the doula service was very well received by their partners who did not want to be involved in the pregnancy or birth:

When I get pregnant first couple of months is fine, but then I feel that he is going emotionally far away from me, he’s kind of fed up, so he felt relief that there was someone else there and I was pleased that I don’t have to beg him every time. Especially when I had to go to appointments with doctors, midwives.

Woman F369Z

Well, he had to stay the first time, all through the caesarean he would stay with me and he just didn’t want to go through that again, so having someone to replace him there was a relief for him and, of course, then he could stay with my daughter.

Woman F341Y

Where the woman’s partner did want to be involved there might be a degree of suspicion or hurt feelings at first, but women described how their partners were won over by the doula and how she became a support for them both:

Yes, when I first said it to my husband he was like, does that mean I was rubbish then? I was like, no, it’s just extra support . . . he could just focus on me, rather than worrying about everything else, but, yes, it was a bit of a, yes, sore point at the beginning, but then he got to know her as well, and it was fine.

Woman F315A

[My baby] wasn’t breathing when his head popped out, so he had to go off in an incubator . . . and [my husband] knew he could call [the doula] and sob his little heart out down the phone, and 20 minutes later, she’s stood next to him, giving him that ultimate needed support.

Woman F337A

I think it’s like the pair of you though, as well, like you’re both looked after by her, not just, like you don’t have to worry about looking after him, because you’re both just sort of looked after as well.

Woman F315A

In times of stress, she helped us to keep communicating with each other and kept things calm.

Woman Q373Y

Because we all got on so well and reduced stress about what to do with our toddler. [My partner] felt much less stressed. She also reassured him about me giving birth at home.

Woman Q302A

In some cases the partner and doula would take turns supporting the mother in labour, giving each other breaks:

My partner . . . he’s coming to visit me to hospital and my doula, she stay with me 8 hours when I was in labour and is coming to visit me and he sent her to go to sleep, they help each other and after 2 hours, I will call her again, because he’s going and she’s coming again.

Woman F404W

The doula also helped couples deal with a less than optimal birth:

Helped him to understand what I had been through.

Woman Q339X

Made us closer after a difficult birth.

Woman Q427A

One woman attributed the fact that she had a partner at all to the doula’s influence:

I met my partner at 5 months pregnant. She gave me the confidence to let a man into my life.

Woman Q448A

However, one mother commented that her partner had not been fully included:

To [invite] my partner in the things and not just me.

Woman Q311A

Impact of doulas: feelings of loss at the end of support

The focus groups with volunteer doulas and the user panel carried out in the first stage of data collection identified the need to examine endings in some detail. Across women’s data sources, the feelings of loss associated with endings were identified as constituting an impact for women.

More than half of women (61.7%) reported that their doula prepared something as part of the ending. Most common was an account of the doula and woman’s time together (27.0%), followed by photographs (26.1%) and a birth story (20.0%).

One-third of women felt that support had ended too soon (33.1%), although this varied across sites (A, 30.1%; W, 14.3%; X, 33.3%; Y, 50.0%; Z, 71.4%). Similarly, when asked what they would change if they could change one thing, the most common response concerned having support for longer after the birth (n = 13), followed by staying in touch with their doula (n = 8).

Several women found that the ending was quite abrupt because they had lost track of the weeks:

I think like they say beforehand and then you get 6 weeks after, but it just flew by, like that first 6 weeks is so, I don’t know, hard and fragile, and just chaos as well, and then it was over, it is quite tough.

Woman F315A

I found it really hard actually, I kept asking if I could keep in touch with her and stuff like that, but we couldn’t . . . At least should have been able to keep their numbers or something like that, so you can have that bond, like, because once a friend they become a friend don’t they and that’s it.

Woman F319A

Yeah, it was so sad for me, because I need to more to help me at that time as well, because the service they send me letter, we stop.

Woman F404W

It happened too soon, I felt I bonded well with my doula and you get used to seeing them and receiving support and then it all stops.

Woman Q332A

I didn’t see Doula for last visit as it was cancelled. So felt it was a little abrupt and did not say goodbye and thank you face to face. Made me feel a little discarded!

Woman Q380Y

We both lost track of visits and ultimately I got a text to say that I needed an evaluation meeting with [the manager]. Wish I could have said thank you.

Woman Q315A

We hypothesised that greater preparation would facilitate endings. Comparing those women whose doula prepared something with those who did not, no statistically significant difference was found concerning feeling that the support ended too soon (31.0% vs. 37.5%) or that they lacked other sources of support when the relationship ended (14.5% vs. 16.7%). However, a greater proportion of those for whom a written account or photograph had been prepared would have liked to have stayed in touch (72.5% vs. 52.4%). This is possibly because the preparation was a reflection of the quality of the relationship, rather than a cause. Women who had received birth stories from their doulas said this was a very positive aspect of closure.

Among those who thought of their doula as a family member, the vast majority would have liked to have stayed in touch (mother, 90.5%; sister, 91.7%). This was greater than those who thought of the doula as a friend (69.0%) or professional (59.5%).

Many women said that they found the ending of doula support very hard, either because they did not yet feel confident in their postnatal role and still needed support, or because they valued the relationship as much as the support and did not want to lose it.

I had a caesarean section, so somewhat depressed at times. Wish the official time to discharge a mother who has undergone caesarean should be longer than a mother who had a natural birth.

Woman Q409Y

It seemed like ending a friendship I cherished.

Woman Q327Y

She made me feel better in myself, everything we did was great together. Felt very depressed when she was gone.

Woman Q444A

And is not fair according to your policy, that when you finish the last day that’s it, not fair at all. She was more than a doula – like family.

Woman Q336X

Women proposed two solutions: that the ending of support should be more timed to the woman’s needs (particularly in situations where she had been ill after birth and was not ready to go out to groups), or that there should be some contact permitted between doulas and women after support ended – a friendly catch-up by phone, over a coffee or at reunion meetings (although some women did have unofficial ongoing contact with their doula). They said specifically they would like to have let the doula know how they were getting on and to show her how their child was flourishing at developmental milestones. Some women framed this in terms of thanking the doula by showing her the long-term impact of her support.

It would have been nice to meet up with her when my son turned one to say she help[ed] bring him into world.

Woman Q419A

So I could show her my perfect family because of her and her help.

Woman Q427A

Just to let her know how I was coping with baby through all her advice.

Woman Q367A

Many women felt that allowing them to keep in touch would acknowledge that they had a real and meaningful relationship with the doula. Women who had formed a close relationship with their doula spoke of their sadness when it had to be broken off.

There was a day she told me that I’m not allowed to get in contact with her, that is not how they do their services, I cried, I said, God, I said, okay, what about in the future? So I just, I said, okay, maybe later I will get her again! But definitely, oh, I miss her, like, I really miss her.

Woman F366Y

What I can say is necessary to stay in touch because when someone do a good thing in life is better to have contact with him/her, like a doula you can even invite her in birthday of your baby for example.

Woman Q440A

By contrast, other women found that the support had ended at the right time, when they felt confident to cope, and some said the support had petered out naturally as they were ready to move on. This was particularly likely for women who primarily wanted information from their doula and was also particularly likely at the one site with an extended postnatal support period of 3 months.

The ending was in the right time, after I felt confident with my baby.

Woman Q408W

It was a good experience of going to coffee shop with my doula and two young girls built up my confidence to go out with my daughters with them both being so young.

Woman Q334A

Felt prepared and confident about the future.

Woman Q341Y

It was very emotional, cos we both know that something beautiful is ending. It was very nice experience.

Woman Q301A

Some women had given their doulas a small gift or card and others were unhappy that they were not allowed to do this or had not been able to; this was also seen as a potential support for feelings of closure:

It was very sad. I wasn’t even allowed to buy my Doula a gift which I didn’t understand and which upset me at the time.

Woman Q312A

Did the service help you in the way you hoped?

While 88.4% of all women reported that the service had helped them in the way they had hoped, the proportion was not consistent across sites; only four of seven women (57.1%) endorsed this at site Z compared with 85.7–92.3% at other sites. Of the 14 women reporting that the doula did not help in the way that she had hoped, 12 provided reasons, the majority of which concerned not having had enough support:

She was of little help because she came too late in my pregnancy and was not at the hospital for my labour and birth at all . . . I was thinking there should be a back-up doula in case of emergency.

Woman Q390W

Doula help[ed] me during labour – not before not after.

Woman Q349A

I felt that the visits were too short and not enough of them.

Woman Q380Y

I needed someone to be with me more often.

Woman Q435A

I needed more support at birth. Doula didn’t get involved much.

Woman Q329Y

Needed it earlier on in the pregnancy.

Woman Q449A

Never saw her really.

Woman Q307A

Was not contacted.

Woman Q414Z

Was only contacted a few times but would have liked more.

Woman Q458Z

For three women, their complaint was about quality rather than quantity.

Didn’t feel supported at all and had a really bad birth.

Woman Q313A

Didn’t give me information that I hoped.

Woman Q322A

I hoped the support would be more professional.

Woman Q362A

Would you recommend the service to a friend or family member?

Only five of the 133 respondents (3.8%) reported that they would not recommend the service, of whom four were from site A and one was from site W. For two of these women (Q307A and Q313A), the reasons for their response reflected not feeling supported and having limited or no face-to-face contact with the doula (discussed in Did the service help you in the way you hoped?). The remaining three women gave little explanation and gave high satisfaction ratings (4 or 5; see Overall, how would you rate your experience of being supported by a doula?). Possibly the question had been misunderstood.

Overall, how would you rate your experience of being supported by a doula?

On a 0–5 rating scale (very poor to very good), 77 women scored 5 (58.3%), 37 scored 4 (28.0%), 15 scored 3 (11.4%), two scored 2 (1.5%) and one scored 1 (0.8%). Scores below 3 were limited to site A. The percentage scoring below 4 at each site was 13.5% (A), 7.1% (W), 11.1% (X), 21.4% (Y) and 16.7% (Z).

There was some agreement between the overall doula rating and whether or not the service had helped in the way that the woman had hoped. Six out of 13 with a rating of 3 and all three women with a rating less than 3 felt it had not helped in the way they had hoped.

What factors are associated with a low rating of the doula experience?

Scores were regrouped into ‘high’ (4 and 5) versus ‘low’ (less than 4) to further investigate the association between lower ratings and various factors that were identified in the CMO configurations. No association was found for the following: self-referral, social complexity, not having whom they wanted at the birth, otherwise birthing alone and having declined any of the support offered.

Lower ratings were more common among women whose birth was not attended by a doula (22.4% vs. 6.8%); however, this appeared to partly reflect the strong association between lower ratings and having had only postnatal support (meeting before 20 weeks 7.7%, meeting at 20–30 weeks 19.0%, meeting after 30 weeks 9.1%, meeting after the birth 44.4%). Two of the five women who were matched late (operationalised as having received intrapartum support without antenatal support) assigned lower ratings.

Analysis of the key informant data and feedback from the user panels identified elements of doula support that women valued. We therefore hypothesised that women who felt that they had not experienced these would be less likely to rate doula support highly. Significant associations were found between lower ratings and not having received doula support with certain characteristics: ‘seeing the same person most of the time’; ‘support shaped around what I wanted to do’; and ‘support from someone who was available when I needed them’. There was no association between lower ratings and the doula not being ‘someone who had given birth herself’ or ‘able to speak my language’.

The association between rating and year of introduction to the service was explored at site A. Lower ratings were most common in 2010 (6 of 24; 25.0%) and 2011 (3 of 17; 17.6%). The following demographic variables were also investigated: age, education, UK born, ethnicity, parity. None of them was significant.

There was a small but significant positive correlation between intensity of support and rating of doula support. However, ratings did not always correspond with the average number of hours of weekly support. For example, the two women who reported an average of 0 hours assigned ratings of 1 and 4.

It is important to note that there appeared to be some systematic bias in ratings not being completed where women felt that they had not received the service as intended and therefore had not completed later sections of the questionnaire.

Administration of the doula services

There were both positive and negative comments about administration. Some women were impressed at how well they were matched with their doulas and how quickly they ‘clicked’. At one site (Y) they particularly commented on how the project staff really took time to get to know them before deciding on a match and on how quickly the doula support began after a referral (this was contrasted with antenatal care).

A few women found the services disorganised or overly formal. One woman with considerable needs had been referred to one of the doula services but had never heard from them and lacked the English language skills to follow this up; she was emphatic that she wished she could have had a doula to advocate for her (site Z). Another described how a member of service staff did not keep follow-up appointments (site X):

There was supposed to be an evaluation of the service shortly after I had finished seeing her and the [staff member] got in touch on a number of occasions and failed to turn up for appointments and then rang up and said ‘oh I am sorry I didn’t make it let’s make it for another time’, made another appointment and still never turned up.

Woman F385X

More awareness of the service as I was not aware until I was nearly 8 months pregnant and unfortunately the service is not readily promoted.

Woman Q373Y

To be less formal.

Woman Q382A

Less formal – tied up in data protection red tape, health + safety, etc.

Woman Q386X

Very attentive, immediate support, well organised.

Woman Q362A

More young people.

Woman Q430A

Disappointed that she didn’t have any contact with a [doula] after they first called.

Woman Q414Z

Women who were not supported by the service

Response rates for doula questionnaires

All sites sent questionnaires to all doulas who had completed training, with the exception of site A, where, by negotiation, nine were omitted who either had moved abroad or had some other reason not to be contacted. Numbers received and sent by each site are shown in Table 18. Between-site differences were not statistically significant.

Of the 89 valid questionnaires received, 71 were from doulas who had experience of supporting women one-to-one. The remaining 18 had trained as doulas but had not (yet) supported a woman. Table 19 shows the distribution of respondents by study site and whether or not they had supported at least one woman. It will be seen that 35 (39%) were from the original site, with sites X and Z being relatively poorly represented in the overall sample. At site W, over half the respondents had not yet supported a woman.

Forty-three per cent of doulas had completed training in 2012 or in the first half of 2013 and a similar number in 2010 or 2011. The 13 who had received training earlier than 2010 were all from the original site.

Ages ranged from 18 to 65 years, with 69% being between 30 and 39 years. Mean age was 36.8 years (SD 9.4 years). Age of leaving full-time education was categorised as less than 16 years (4.5%, n = 4), 16 years (33.0%, n = 29), 17–19 years (31.8%, n = 28) or 20 or more years (30.7%, n = 27). The majority of respondents identified their ethnicity as ‘white’ (80%, n = 71); 11 were ‘Asian’ or ‘British Asian’, five were ‘black’ or ‘black British’, one was ‘mixed’ and one was ‘other’. Forty-one respondents were currently in employment, 16 full time and 25 part time. Of those not in paid employment, 13 were studying or had other major demands on their time in addition to doula activities.

A modified questionnaire was sent to volunteers who were known not to have supported any women. Of the 18 returned, five said that they had stopped volunteering as soon as they completed their training, two at site A, two at site Y and one at W. In three cases this was because a paid job had been obtained. The remaining two both found it too time-consuming, with one additionally feeling that she was not suited to the role and that the role was not what she expected. A further six were still awaiting a match: four at site W and one each at sites X and Z. Five of these six had only recently completed their training but, at site Z, training had been completed in 2010/11. Other reasons offered for not having supported a woman were all personal, for example pregnancy, moving away and the need to support a family member. At site Y two volunteers had moved into outreach work within the organisation, as an alternative to supporting women one-to-one.

Except where otherwise specified, data presented in the rest of this section necessarily refer only to respondents who had supported at least one woman. Of these 71 respondents, 40 (56%) were still active volunteers, 14 (20%) were having a break and 17 (24%) had stopped being a volunteer doula. Of those still actively volunteering, 19 were in paid employment (5 full time and 14 part time) and four were studying.

Nearly half (47%) had taken a break from volunteering at some time during their time as a doula, ranging from a few weeks to a couple of years, generally for health or family reasons. Of those currently active, 76% had never taken a break, including three who had trained before 2010 and 12 who had trained in 2010/11.

There was considerable variation in the number of women supported, ranging from 1 to 19 (including women supported in a back-up role). As would be expected, those trained more recently had supported fewer women (Table 20). In total the 69 doulas responding to this question had supported 260 women as main doula. A further 108 women had been supported in a back-up role by the 66 doulas responding to this question. Doulas at sites X and Z had supported fewer women than those at other sites, even allowing for those trained before 2010.

As might be imagined, not having enough women to support was a source of frustration to some doulas, as one explained in a telephone interview:

Maybe there’s an issue about referrals, maybe we’re not getting enough mums through the doors to support, but whatever that reason is I think I’m a wasted resource, because I’ve been a doula for two years and I’ve only supported four mums, two properly and two as a back-up. But, I’ve had a year just gone like that I know I could have supported 16/17 as many as possible, so it’s a wasted resource I think. . . . I’m passionate about the doula scheme and my main concern is it won’t work if we don’t get consistent work, there’s no proper management of it, you get frustrated.

Doula 067X

Doulas’ views on how their role affects women

Doulas were asked how they thought that doula support had an impact on women in a range of ways. Virtually all respondents (97.1%) thought that it helped women access different services and/or join different groups. Eighty-one per cent thought that it helped women to meet other mothers. Those not agreeing with these statements thought that doula support had no effect on these behaviours; none thought that it had a deleterious effect. All respondents agreed that doula support increased women’s knowledge and skills around pregnancy, childbirth and looking after a baby. This was explained in two ways. A frequent response referred to direct transmission of information from the doula to the woman. However, another prevalent response, particularly at sites Y and A, identified the woman as an active agent and the doula as an enabler, thus, for example:

Gives them choices, information and confidence.

Doula 049A

A doula encourages women to have an informed choice and they seek info together.

Doula 003X

Able to feel more confident and ask more questions.

Doula 008Y

Awareness + confidence = empowerment.

Doula 016Y

This support allows them to get better understanding of coping and knowing support sources around them.

Doula 045Z

The majority (88.4%) also thought that having doula support helps women to make lifestyle changes for the health of themselves and their babies, most frequently citing breastfeeding, smoking cessation, diet and cutting down on drugs. A similar distinction was seen in the explanations of how this was achieved. Most referred to women now having knowledge that they did not previously have, for example about how to breastfeed. Others, though, referred to women making informed choices and having confidence to make different choices.

Finally, doulas were asked to tell us about any other positive or negative effects that doula support has had for women. Most responses reinforced and summed up earlier answers.

Doula support has a positive effect by empowering woman to make choices and signposting them to other services to receive support.

Doula 014A

Women feel empowered because they’ve chosen how, when, where, instead of being told what to do.

Doula 017Y

If we have enough time to make a close relationship with the woman, our input can be amazing.

Doula 054A

Only two responses mentioned any negative effects:

Some feel a loss of friendship at the disengagement time.

Doula 028A

One lady became reliant on me, had to gradually decrease time with her.

Doula 029A

Should doulas have had a child themselves?

Doulas were also asked how important they thought it was that a doula should have had a child and/or parenting experience herself. This question was also asked of those who had not yet supported a woman. One-third of respondents (33.7%, n = 30) said it was not important, 43.8% (n = 39) ‘quite important’ and 22.5% (n = 20) ‘very important’. This varied by site, with only one doula from site A (2.9%) saying that it was ‘very important’, compared with 38.5% at W and 45.8% at Y. The reasons given for why it was not important included:

’Cause it’s not about our birth experiences.

Doula 015A

A doula can support and advocate for a woman if she has no children – that’s like saying a nurse cannot look after a patient who has had a cardiac arrest if she hasn’t had one!

Doula 049A

Be[ing] qualified and properly trained is enough to conduct and support a mother.

Doula 047Z

Those who thought that having your own experience was important mainly felt that this gave them insights and empathy that someone who had not given birth would not have:

Because if you have been through the same situation you have more understanding of what they are going through.

Doula 037A

I feel there are experiences that you cannot learn and you are able to empathise – relate to the woman better.

Doula 003X

Experience and knowledge together is a fuller package.

Doula 022Y

Two commented that they had been a doula before they had children of their own, but now felt that their experience had added to what they had to offer. Some stated their position in quite extreme terms:

Can’t preach what you haven’t practiced.

Doula 062Y

How can you give good advice, if you’ve never needed it yourself!

Doula 052Z

Some also made the point that:

The client may feel more confident in her doula if she has experienced childbirth herself.

Doula 014A

Those rating personal childbearing experience as ‘quite important’ were more likely to feel that it was ‘more of an advantage than importance’ (007A):

I don’t think it is essential for all women but some women may find it more reassuring to have someone supporting them who has personal experience of labour and pregnancy. If the doula is knowledgeable and confident this should not matter.

Doula 048A

Knowledge, empathy and professionalism matter most but the mothers I worked with were reassured that I had also given birth so knew what they were experiencing.

Doula 051Z

Matches that were ‘not the best fit’

Fifteen doulas felt that they had had matches that were ‘not the best fit’; these represented all five sites. In some cases this was put down to the doula’s own inexperience at the time or where the woman wanted something from the service (e.g. child care) which was not on offer. There were other examples where the woman was not really engaged:

The woman was hardly in contact – think she felt she had to have a doula – but didn’t want one.

Doula 039A

Only a minority of examples did specifically relate to matching, that is it was implied that a different doula–woman combination would have been more successful:

Some people you get on better with than others.

Doula 018A

Young woman who did not want support from an older doula.

Doula 027S

Could just tell the teenager wasn’t happy, but her main doula was ok for her.

Doula 017S

The woman thought I had more experience than I did and she wasn’t very keen when she found out.

Doula 080

Timing of when the mother meets the doula

In addition to the ‘who’ of matching, early key informant interviews had also talked about the ‘when’. Doulas were therefore asked ‘Do you feel that your relationship with a woman is affected by the timing of when you meet?’ Response options were ‘yes’, ‘no’ and ‘depends on the woman’s needs’. Perhaps surprisingly, just over half (5.2%, n = 37) said no, with 13.4% (n = 9) saying yes and 31.3% (n = 21) saying that it depended on the women’s needs.

All the doula services target ‘disadvantaged’ women; two are restricted to women from minority ethnic groups and a third serves an area with a very large minority ethnic population. Thus, it is to be expected that many women will not speak English, and this was borne out during data collection from women. Fifty-one per cent of the doulas (n = 36) said that they had supported a woman who did not speak English, and only seven of these (18.9%) said that they and the woman spoke the same language; one of them had also supported another non-English speaker with whom she had no common language. Thus, a substantial proportion of the volunteer doulas (30 of 71, 42.3%) had supported a woman without the benefit of a shared language. Somewhat surprisingly, the site with the highest percentage of doulas who had supported a non-English-speaking woman was site A (71%). This may reflect the greater experience of site A doulas rather than a higher proportion of non-English-speaking women.

A telephone interview with a site A doula who had worked with interpreters added to the questionnaire data:

I had had a few problems with the interpreter in that instance as well, where the interpreter sort of took over really, and didn’t interpret but actually did my role as well, in some, a little bit . . . she actually acknowledged that she had done that, and the next time I thought she was entirely different.

Doula 069A

However, she went on to make the important point that:

most of the family seem to know them [interpreters] quite well so they must come across them quite regularly, and they obviously have a really good bond with them, it’s just a little bit of an obstacle I guess, to just get the response that you would have with someone who speaks your language.

Doula 069A

A telephone interview with a bilingual doula raised the situation where one doula can communicate with the woman in her own language but the other cannot:

I’ve been interpreting, I don’t mind doing it, but I personally don’t think it’s fair on the other doula. Because I’m developing my relationship with the mum and the other doula can’t because there’s this big barrier, so I’ve had to do all the interpreting, which I don’t mind doing, but I just don’t think it’s the essence of what the doula scheme is all about.

Doula 067X

This doula also raised another issue about second language use:

with the doula scheme in X we’re allowed to use a second language in the homes but not in the hospitals.

Doula 067X

Have you ever felt out of your depth?

Doulas were asked if they had ever felt out of their depth. Two (2.9%), from two different sites, answered ‘often’; 18 (26.1%), from four different sites, ‘occasionally’; and 49 (71%) ‘never’. The examples given from site A reflected the increasing emphasis on women with complex problems: drug abuse; domestic violence; and babies being taken into care. However, there were also references to language problems and to inappropriate expectations.

Complex cases were also mentioned by one respondent at site Y. Comments from other sites referred to communication issues that will be explored further below, such as with a rude and non-communicating midwife (064Y) or a non-participating back-up doula (047Z). Two further comments raised training issues: one who had needed more knowledge of the benefits system (060Y) and one who felt she had not had enough practice with massage and birth positions, although ‘I revised with a fellow doula which helped’ (059X).

Do you feel that a professional has ever misunderstood your role?

The majority of the sample (64.3%, n = 45) had never felt that a professional had misunderstood their role, which is very heartening. However, 12.9% (n = 9) said that it had happened once and 22.9% (n = 16) said that it had happened more than once. Examples given provide interesting insights into doulas’ clear perceptions of their role and its boundaries:

Midwife/social worker. They sometimes see me as a support worker rather than a volunteer.

Doula 026Y

[Thought I] was here ‘to mop her brow’. Treated like a lady’s mum. But once discussed, my role is understood.

Doula 039A

Social worker contacted me for information. Directed her to Service Manager.

Doula 054A

Others suggested that they made professionals uncomfortable:

Expecting more than we can offer, feeling threatened by our role.

Doula 003X

They don’t know the [parent organisation] and not sure why I’m there if I’m not a friend or family member or health professional.

Doula 051Z

A small number alluded to ‘teething problems’ with midwives when the service was new, with the implication that these had been overcome, and others that certain professionals were simply unaware of their role or had never heard of a doula:

Early days professionals were dubious of doula service – but through hard work and marketing we have built a good rep[utation?].

Doula 053X

In a telephone interview, one doula reinforced that these were isolated occurrences and that they could be dealt with:

It’s usually particular individuals, it’s not the midwifery team as a whole; most of them are absolutely wonderful . . . I have had three instances where the midwives have been difficult, but I have had a lot more where they have been wonderful . . . Well I actually spoke to somebody at the doula office about the recent situation . . . when I was asked if I wanted to take it further I said no, what I am going to do is try and build a relationship with that person because I will be seeing her quite regularly so I would rather change her views through building that relationship, and if I can’t I can’t.

Doula 069A

Do you feel the role of doula is clear in relation to the role of other services?

Eighty-five per cent (n = 57) felt that it was clearly a different role and, perhaps surprisingly, only 14.9% (n = 10) thought that there was overlap with other roles (four failed to answer the question). The most obvious overlap was with the midwife’s role, both the support function and that of signposting to other services:

Sign-posting – baby cares etc. HV [health visitor] and MW can also do these. I think we can free up professionals time by helping in this.

Doula 039A

More than one respondent at site Z implied that midwives took advantage of this overlap:

Midwife leaves you to it!

Doula 052Z

A small number of respondents also mentioned overlap with other voluntary organisations and one suggested that areas of overlap between different services could cause confusion to the mothers.

How well do you feel that you and midwives work together?

Six respondents answered ‘not applicable’, as their role had not yet brought them into direct contact with midwives, and one omitted the question. Of the remainder, 67.2% (n = 43) answered ‘well most of the time’, 31.3% (n = 20) ‘well some of the time’ and one person (at site A) answered ‘not well at all’. This seemed to be mainly for the reasons outlined in the earlier questions, with an additional couple of examples of particularly poor communication being given from site X and site Y:

Midwife was very rude, did not show respect for anyone, would not communicate.

Doula 064Y

Our CMOs had led to the hypothesis that role ambiguities and working relationships would be better in the sites where the service had closer NHS links, but no significant between-site differences were found for any of the three questions just reported and, as indicated, examples of friction were isolated.

How well do you feel that you and other health and social care professionals work together?

Five respondents answered ‘not applicable’, as their role had not yet brought them into direct contact with other health and social care professionals, and one omitted the question. Of the remainder, 57.1% (n = 40) said ‘well most of the time’, 32.9% (n = 23) ‘well some of the time’ and two answered ‘not well at all’. Here a significant between-site difference was found, with more problems being reported at sites W and Z. However, numbers at both of these sites were small.

Have you ever found any barriers to signposting women to other services?

Twenty-four per cent (n = 17) of respondents said that they had experienced difficulties with signposting. The reasons for them were ‘I did not know what services were available’ (four); ‘A lack of available services’ (eight); ‘Waiting times for services’ (10); and ‘Women not wanting to use the services’ (nine). Two ‘other’ reasons were given:

Frustration at services passing the buck.

Doula 001Y

Sometimes it’s too early for the mums to know what they might want.

Doula 037A

Have you ever felt there were too many other services involved in supporting a woman?

Only eight respondents answered yes to this question (11.6%). The effects of this were felt to be ‘Disrupt or interfere with other support work going on’ (four); ‘Repetition of support by different people’ (seven); ‘Contradictory support given by different people’ (six); and ‘Woman does not appreciate the support because receives so much’ (seven). One other response was:

Woman gets confused about each person’s role and what they are there for.

Doula 037A

What makes for a difficult ending?

At all sites the main thing that doulas said made an ending difficult was when they had formed a bond of friendship with the woman and were not allowed to have further contact:

When you have a real friendship and both want it to continue.

Doula 028A

When you have bonded well and really helped mum through difficult times.

Doula 001Y

Never being able to see or speak to that family again.

Doula 086A

A couple of doulas said it was difficult if this relationship was one-sided:

When they want you to still be their friend.

Doula 018A

Some doulas identified a woman’s overdependence on the doula as a problem for endings. One described this as ‘giving too much support’ (034Y). Some saw this issue from the opposite point of view, in terms of unmet needs: if the woman was very isolated and not yet fully ready to cope, it was difficult for the doula to have to withdraw.

If women feel they still need you, they are lonely, low confidence.

Doula 003X

When a mum has really needed the support and doesn’t think she could cope alone . . . It’s an emotional time for both doula & woman and they sometimes feel abandoned.

Doula 066A

If you both don’t want to end the support especially if all issues not resolved.

Doula 012A

Some doulas were not confident that there was other adequate support available. In considering how endings could be done differently, many doulas responded that they did not know or that there was no way to do endings differently. Others made various suggestions, most commonly that the support should be more flexible around the mother’s needs:

Maybe offer extra time for 1st time mums – until they feel able to cope alone.

Doula 090W

6 weeks should be a guide. Sometimes a little more is needed e.g. hospital appointments for mum or baby.

Doula 039A

It was also suggested that support should be phased out rather than cut off, or that the ending of contact should not be absolute but supplemented with catch-up sessions or informal contact if the doula and woman so chose. On the other hand, a few doulas thought that ongoing contact would carry its own risks:

It’s best not to follow-up otherwise the emotional connection will be difficult to break off.

Doula 047Z

If we could stay in touch with the mother it may put the volunteer in difficult positions e.g. being called to a problem she is having or constant phone calls.

Doula 042Y

One doula suggested that articulating the reality and importance of the relationship would support feelings of closure:

You always need to remember the rules and the reasons for them, about professionalism given by the doula project, but say to the mum and family also you will never forget the time you spent together with them and that you will cherish it in your heart always.

Doula 065A

Another suggested that the doula could turn to project staff for emotional support around the ending:

Bad endings can be turned to good endings by talking to management team – getting there early for support rather than leaving it ‘til too late.

Doula 053X

Another doula said that it was easier to overcome feelings of loss if the doula went on to support another woman quickly:

Hardest part of being a doula. It is easier if meeting a new mum that week.

Doula 069A

Several of the doulas reflected that accepting endings was an inherent part of their role:

It’s a service with a natural time line. Bit like teaching a bird to fly.

Doula 088Z

Preparing something for the woman as part of the ending was the norm: 71.4% (n = 45) of those answering said that they did and 28.6% (n = 18) that they did not. (Eight did not respond.) Of these 45, 28.9% (n = 13) said that they prepared an account of the time spent together, 66.7% (n = 30) a birth story and 62.2% (n = 28) photographs. Giving a birth story was much more likely at site A (85.2%) than at other sites, and similarly giving a photograph (81.5%), but numbers responding at the other sites were all low.

Ten respondents (22.2%) indicated that they gave something else. Three, at three different sites, said that they gave a small card or gift for the mother, the baby or both. The other seven, representing all sites except W, referred to less concrete things that they gave the mother, for example:

Ensuring they are confident in accessing services.

Doula 047Z

Sum up the (positive) changes I have seen.

Doula 026Y

The majority of respondents (64.1%, n = 41) said that they had been offered a token of appreciation by women, while 23.9% (n = 23) had not (seven failed to answer this question). This was least likely at sites X and Z, although the numbers of respondents from these sites are small. Since the protocol forbids doulas to accept gifts from women, respondents were asked if an offer of a token of appreciation had ever caused any problems. Five said yes:

Mum became very upset when she was told by the doula office that I could not keep the gift.

Doula 069A

One woman wanted to prepare a meal for me and for me to bring my daughter – she was upset and possibly offended when told this was not allowed.

Doula 048A

Twenty-one of the 64 doulas answering the question (32.8%) said that they had contact or received updates beyond the official ending. This proportion was approximately the same at each site, except at site X, where it was zero. Most respondents (77.3%, n = 51) would like to have more update information on women whom they have supported. This proportion was approximately the same at each site, except at site Z, where it was zero. Fifty per cent (n = 32) said that there had been women with whom they would have liked to have stayed in touch after the official ending. Reasons given were concern about the woman (n = 11); concern about the children (n = 3); feeling that the woman needed continued contact (n = 13); and that she had no-one else (n = 7). Eighteen respondents gave additional reasons, 15 of whom alluded to friendship which either had developed or it was felt could have developed. The remaining three focused more on confirmation of progress and outcomes:

It’s nice to observe the parent parenting confidently and affirms the time spent working with her.

Doula 026Y

In a telephone interview, one doula said that her strategy was to move on immediately to a new match:

your focus has to immediately change to somebody else and that’s the only way I find to deal with that ending,

Doula 069A

Emotional impact

Impact on health and well-being

Social impact

How has being a doula affected you socially?

As Figure 37 shows, each of the potential social benefits of being a doula was strongly endorsed by over half of the sample.

FIGURE 37.

How has being a doula affected you socially?

Nine respondents offered other ways in which they had been affected socially mainly becoming more confident and:

Realising that I lead a privileged life!

Doula 013A

Two alluded to negative impacts on their social lives when they were on call.

Work-related impacts

How has being a doula affected you in work-related ways?

As we have seen, many doulas were not motivated by a desire to undertake further training or employment, but most agreed ‘somewhat’ or ‘a great deal’ that being a doula had had employment-related benefits, as shown in Figure 38.

FIGURE 38.

How has being a doula affected you in work-related ways?

Impacts on doula’s family

How has being a doula affected your family?

Perhaps surprisingly, there was little indication of negative impact on the families (Figure 39). Being ‘less available to family’ and the ‘car being less available to family’ were each strongly agreed with by only four respondents. The most strongly endorsed impact on family (52.9%) was ‘Increased knowledge and accessing of services’.

FIGURE 39.

How has being a doula affected your family?

Four respondents added additional ways in which their families were affected:

Affected social life.

Doula 029A

Hope it has given them a broader understanding of people’s different experiences and encouraged them to think about what they can do to support others.

Doula 039A

My children are proud of me and have an understanding of my role and how important volunteering is.

Doula 048A

My husband has been 100% supportive. He likes me being a doula.

Doula 054A

Other impacts

An optional question invited respondents to describe any other ways in which they had been affected by being a doula and 50 of 71 (70%) responded. Most comments were positive and included a number saying that they had become less judgemental:

It has opened my eyes to how sheltered my life was. Also a lot of my opinions were misguided and ill-informed. I am a lot more open to many things and I do genuinely feel less judgemental in a positive way.

Doula 033A

A number also commented on the transferability of the skills that they had learned:

I love it. I don’t know what I would do if the service ceased to operate. It has made me a better friend and given me skills to support anyone in my life.

Doula 069A

The third category of positive comments referred to increased confidence and fulfilment:

It has made me more pro-active and has encouraged me to engage with people.

Doula 047Z

There were also references to helping with career choices: 023A was ‘now a newly qualified midwife’ while for 051Z ‘[it] helped me to decide NOT to pursue a career in midwifery.’

There were a small number of more negative comments:

It can be hard seeing how some people live and the conditions.

Doula 018A

As a grandma, there are expectations that I am available for child care from time to time. Sometimes this can be difficult.

Doula 054A

Decreased my confidence in ‘health services’, seems like a culture of ‘bullying’ in the work environment.

Doula 067X

Volunteers who had supported women: why did they stop?

Seventeen respondents who had supported women said that they had now stopped volunteering and all answered the subsequent question explaining why. No respondents from site X had stopped volunteering. The reasons given by those from sites A, W and Y were very similar, citing family commitments, studying, having moved away from the area, not having a car or own health problems. In contrast, at site Z the reasons given were all about the organisation and the role rather than about the individual.

The organisation was unprofessional.

Doula 051Z

Finances, lack of communication, felt uncomfortable in office and meetings, didn’t feel like I was accepted.

Doula 052Z

Am going to stop because I don’t enjoy it.

Doula 079Z

I think I have stopped, may give it one more try if they call.

Doula 089Z

It is clearly in the services’ interests to minimise attrition of trained volunteers. We therefore asked all responders (i.e. whether or not still volunteering and whether or not they had yet supported a woman) ‘What would (have) encourage(d) you to keep volunteering?’ Responses are shown in Figure 40. It will be seen that very few of these potential motivators were rated as important by the volunteers, the modal response being ‘Not at all’ for all except ‘Fewer commitments in personal life’. There were few noticeable differences between sites except that volunteers at site Z were more likely to endorse ‘More support from service/project staff’. There were also no significant differences between those still volunteering and those who had stopped, except that those who had stopped were likely to rate ‘Fewer commitments in personal life’ as very important.

FIGURE 40.

What would (have) encourage(d) you to keep volunteering? Note: Includes doulas who had not yet supported women.

A small number of respondents offered additional motivators, with some pointing out that their reasons for stopping were intrinsic (e.g. having moved away) and not anything within the service’s control. Among those that the service could potentially act on were to have variety in the women they support; not to have to wait too long to support a woman; small group supervision; money towards childcare costs when needed; and further training. Two rather more negative comments came from site Z from doulas wanting better support from service staff and a further one from site Y from a doula who felt that the service had not been honest with her about the time commitment required.

What is the best thing about being a doula?

The overwhelming majority of doulas who answered the question ‘What is the best thing about being a doula?’ said that it was about being able to help and support women and making a difference to women and their babies:

The greatest feeling you get knowing you have made a big difference in a person’s/family’s life.

Doula 061A

Being able to share our expertise with vulnerable mums, assist and support them at a very important time in their lives.

Doula 059X

Many described it as a privilege to be allowed to share in a woman’s pregnancy and birth:

The privilege of being part of a woman’s journey from pregnancy to birth and looking after her new baby. Seeing a family blossom and a new baby thrive in that family.

Doula 027Y

Some talked in terms of ‘feeling useful to the community’ (059X) or ‘giving summat back’ (036A).

Some doulas, especially at site Y, particularly referred to the pleasure of seeing women’s confidence grow through their support:

Seeing a woman gaining confidence in her own abilities to birth her baby and care for her child.

Doula 056Y

The best thing is supporting a really vulnerable mum and within 6 weeks she becomes independent and in control especially at the birth – and to watch her grow strong with her baby and family.

Doula 053X

Some doulas framed ‘the best thing’ in terms of their own development: they had gained confidence, enjoyed meeting new people from diverse backgrounds and liked feeling they were part of a team of professionals.

I love meeting new people and I love supporting women from all ethnicities. I enjoy working with other health professionals as part of a team.

Doula 020W

A couple of doulas referred to the camaraderie of the doula service:

Contact with other doulas was essential – part of big, lovely team!

Doula 023A

Meeting the other doulas and wonderful support staff. The positivity around the service.

Doula 053X

Some doulas described the personal satisfaction of feeling needed or appreciated and pride in accomplishing what they had set out to do:

Giving help and support to someone and get the appreciation from her, especially if this woman is new in this country and doesn’t have family members or friends.

Doula 078W

When you really feel you have made a difference to someone. Some women expect to be given help, almost as a right. Others really need support and say thank you. That makes my day!

Doula 054A

What is the one thing you would change about the doula role?

There was considerable variation in answers to the question ‘What is the one thing you would change about the doula role?’ Many did not answer or said ‘nothing’. Some doulas commented on aspects of the service organisation: they would like to be offered more doula work or offered it sooner after finishing training; less paperwork; not to have to phone in every hour; or to have feedback from the women they supported. One doula said that the service should work around and not overlook family commitments of volunteers:

The fact that we are mothers with young children – we feel that sometimes late evening and night support could be a problem – so reserve two back-up doulas in order to maintain the service and not let the mother down.

Doula 047Z

At one site (A), two doulas commented on the postcode restriction of the service:

There are many women that are missing out on extra support as they live out of the area.

Doula 041A

At site W it was similarly suggested that the service should be available to young indigenous women, not only women from the black and minority ethnic (BME) community. Some doulas said they would like more support from the service: updated information, emergency back-up, more contact with other doulas. As we had been led to expect from our initial doula focus group, many doulas (55.2%) would like more contact with other doulas and this was similar across all the sites. It was also suggested that a doula co-ordinator would help to maintain motivation:

I know it isn’t possible any more but when I started in 2007 the locality workers were available on call 24/7. Very, very occasionally it would be good to have this support. The middle of the night can be lonely! If things are going wrong.

Doula 054A

Think a co-ordinator doula is needed. Someone to organise meetings, support and keep doulas motivated. Somebody just for them (to contact anytime).

Doula 071W

Overall, 34.8% of respondents had said that they thought that doulas should be paid and this was mentioned by a small number as something that they would like to see changed:

It would be nice to have some sort of payment for the time and effort that we all put in and it would be nice to be more recognised for the work we do and the differences we make.

Doula 009A

Some doulas said they would like to change some of the restrictions imposed by the services at some sites; in particular, rules about not giving women gifts and not sharing personal experiences:

That we could help them more in ways of donations, i.e. be able to give them small items or buy them a small gift when baby is born.

Doula 007A

Perhaps being able to ‘advise’ and ‘influence’ and share experiences.

Doula 067X

Several referred to the rule that doulas could not stay in touch with the women they supported:

Would like the option of some follow-up (after the 6 weeks postnatally) to be available so women don’t feel you have just moved on and forgotten them.

Doula 048A

One explained her mixed feelings about this issue:

A lot of us doulas say we don’t like the no contact with mums after the match has finished. We all know and appreciate the reasons behind it. We understand but find this the most difficult sometimes in our role as a doula. Is it something we would like to change – yes. Is it something that should change – really no.

Doula 033A

In a telephone interview, one doula pointed out:

the other thing I find really difficult particularly is trying to get mums out of the house, most of them, a lot of them don’t want to leave the house in the first month, and trying to get them to go places with you, to go to children centres and stuff is really difficult and also things like baby massage and that they don’t really want you to take your babies until they are at least 6 or 7 weeks old anyway, so that’s the bit I find most challenging about only having 6 weeks, and women in certain cultures don’t leave the house for 6 weeks anyway.

Doula 069A

A final theme was the need for the doula concept to be better known and understood:

The constant need to explain to health professionals the role we play and the importance of that role.

Doula 026Y

I would like more recognition. It would be nice when I tell someone I’m a doula for them to know that is!!

Doula 056Y

In a telephone interview, one doula said:

I’ve got this policy of telling one new person every day about what a doula is.

Doula 067X

Would you recommend being a volunteer doula to other people?

Only 2 of the 71 respondents who had supported at least one woman answered ‘no’ to this question. Both were from site Z.

Overall, how would you rate your experience of being a doula?

On a 0–5 rating scale, 49 doulas scored 5 (69%). The percentages by site were 81.3% (A), 83.3% (W), 71.4% (X), 61.1% (Y) and 25% (Z). A further 16 scored 4 (22.5%). Thus, over 90% rated their experience as either 4 or 5, including all respondents from sites A and Y. Two respondents gave a rating of 0, both from site Z.

What factors are associated with a low rating of the doula experience?

Scores were regrouped into ‘high’ (4 and 5) versus ‘low’ (< 4) to further investigate the factors associated with low scores. Based on the CMOcs developed earlier, variables investigated were degree of experience as a doula, measured in terms of both length of time since training and numbers of women supported; whether or not the doula had ever taken a break; whether or not she had ever felt out of her depth; whom the doula could talk to in an emergency; whom she could talk to for moral support; type of contact with doula service staff; satisfaction with availability of doula service staff, both during and outside office hours; satisfaction with availability of resources; whether or not more contact was wanted with other doulas; whether or not a professional had ever misunderstood her role; whether or not the doula role is distinct from other roles; and whether or not ‘volunteer’ doulas should be paid. Demographic variables were also examined: age, education and whether or not in paid employment. Ethnicity could not be considered because it was strongly correlated with site.

The very small number of low scorers (n = 6) meant that Fisher’s exact test needed to be applied, rather than a standard chi-squared analysis. Very few of the analyses were found to be statistically significant, although most showed figures in the expected directions. Those that were significant were exclusively to do with their relationships with the doula service. Low scorers were less likely to say that they were able to speak to doula service staff in an emergency or for moral support. Those who specified some other person that they could talk to for moral support were more likely to be low scorers, perhaps reflecting the fact that the service staff were not meeting this need. Respondents who said that they did not have scheduled supervision meetings were much more likely to be low scorers, as were those who did not have scheduled telephone conversations, for example before or after an initial meeting with a new woman. In the sample as a whole, five respondents said that their contacts with the doula service staff did not include telephone conversations when they needed them, and four of these were in the low-scoring group. Low scorers were also significantly less likely to say that they were able to have face-to-face conversations with staff when they needed them and more likely to say ‘I usually wait until they contact me’. All who were ‘very satisfied’ with the availability of service staff inside office hours were high scorers, whereas all who were ‘not at all satisfied’ were low scorers. The pattern was similar for availability outside office hours except that a small number of high scorers also expressed themselves as ‘not at all satisfied’. Low scorers were also less likely to feel that there were enough resources, such as interpreters and telephones, to support their work.

The difference between ‘good’ and ‘very good’

In previous work47 we have found it informative to examine differences between respondents who give maximum ratings and those who, while generally positive, still have some reservations. It was felt that comparing those scoring 5 (n = 49) with those scoring 4 (n = 16) for their doula experience may give us some insights into what it is that makes the doula experience less than optimal. We therefore re-examined the data omitting the six respondents scoring less than 4 from the analysis and comparing those scoring 5 with those scoring 4. The same list of independent variables was used as reported above for high versus low scorers. The only analysis to show a statistically significant result was that experienced doulas (those who had supported more than two women as a main doula or more than three as a back-up) were more likely to give a rating of 5 than of 4. This might be thought to be because those who are less satisfied do not stay long enough to become experienced. This was partially supported by the finding that 90.9% of those training before 2010 scored 5 compared with 78.6% of those trained in 2010/11 and 65.2% of those trained in 2012/13. However, that chi-squared analysis was not significant (p = 0.24).

Demographic variables

Age, ethnicity and education were compared. Neither age nor ethnicity distinguished satisfied (scoring 4 or 5) from dissatisfied (scoring 0–3) doulas. All the dissatisfied doulas, including the one who had never been offered a woman to support, had had more than the minimum years of education, whereas this was only true for one of those who were satisfied. There were also suggestions that at least two of the dissatisfied doulas had relevant professional experience which gave them particular insights into the shortcomings of the organisation. Possibly this may have been seen as threatening by some individuals within the organisation, thus amplifying the friction. It was hypothesised that, if the dissatisfied doulas were more likely to have a professional background, then they might be more likely to have cited ‘to use the skills I already had’ as a motivation for training than those who were satisfied. However, the hypothesis was not supported; seven of the nine respondents rated that motivator as ‘very important’ and the other two (both dissatisfied) rated it as ‘quite important’. No other demographic variables were available for comparison.

Training cohort

The doula who was interviewed had reported particular problems around the training of her cohort. All four dissatisfied doulas had trained in 2012, whereas three of the four satisfied doulas had trained in 2010. However 088Z, the positive doula quoted above, had also been part of the 2012 cohort. It was known that there had been some disruption to the normal staffing at site Z during that period, so these doulas’ experiences may not be representative of other cohorts. The doula who was interviewed had specifically reported problems in the group dynamic of her training cohort, with one individual described as constantly dominating the conversation, talking about herself, being opinionated and intimidating and with the tutors not challenging her behaviour. ‘[It’s] not useful to have one member of the group do 90% of the talking’ (doula telephone interview). A formal complaint had been made about the quality of the training and it was implied that there was some polarisation of the group thereafter into those critical of the organisation and those supportive.

There were I think 13 of us on the course and the meeting where people were airing their grievances as it were, stuff they weren’t happy about, 10 of them were saying exactly the same thing that I was saying.

Doula telephone interview

Thus, it seems likely that 088Z was not one of those 10 who were critical of the organisation, but she will almost certainly have been aware that such views were being expressed. This may account for particularly strong endorsement of the staff in her final comments: ‘fantastic’, ‘wonderful’.

Number of women supported

Of the four dissatisfied doulas, two had only supported one or two women as a back-up and had supported none as a main doula. A third had had two matches, one as main and one as back-up, but neither had developed, since one was out of area and the other wanted child care rather than a doula. Thus, neither actually met the organisation’s criteria and should never have been matched at all. She felt that this reflected poorly on the organisation. The fourth had supported two as a main doula and one as back-up. In contrast, the very positive doula, 088Z, had supported four women as a main doula since completing her training, which was more than any other doula respondent from this site. One of the nine respondents had completed training in 2011 and was still waiting to be offered a woman to support. As we saw above, volunteers at site Z had tended to support fewer women on average than those at other sites and a shortage of clients was recognised as a problem by the organisation. However, we also spoke to one woman at that site who had wanted a doula and been referred to the organisation but had never been contacted by them. Furthermore, there was a perception that such women as were available to be supported were not being shared out equally.

It did seem like there were several women [i.e. volunteers] on the [training] course who were not even getting the opportunity to work with anyone . . . it seemed to me like it was operated on the basis of who [staff member] liked the best.

Doula telephone interview

However, we were also aware from our other data sources that there were challenges at this site particularly in matching women to volunteers who could be available at the times that women wanted them, because of doulas’ family and other commitments. It may thus be that those not offered matches could not be available at the times needed, for example potentially staying with a woman during labour at night, and that this was why no matches had been offered.

Thus, it seems that the answer to ‘why did not all doulas at site Z feel the same way?’ is probably that ‘not all volunteers were treated the same way’ and that perceptions of unfair treatment have led to divided camps. A CMO can thus be hypothesised that, when staff are perceived as being unprofessional (Context), volunteers feel frustrated and marginalised (Mechanism), which leads them to be dissatisfied and to leave the organisation (Outcome). It is also interesting, that, as far as we can tell from our limited sample, negative views of the training and the organisation are restricted to one training cohort during an atypical period. This could suggest another speculative CMO, which is that a poor training experience (C) leads to a lower threshold for negative perceptions of the organisation (M) and lower role satisfaction/leaving the organisation.

Establishing the original service

Based on interviews with eight operational and strategic staff in the first phase of data collection, we identified 11 facilitators or ‘Mechanisms’ considered important for implementing and sustaining the site A service. In Table 28 below, these Mechanisms are listed and illustrated with quotations from the interview data.

Maintaining the original service

We conducted follow-up interviews with operational and strategic staff at all five sites, approximately 1 year after the initial interviews. We revisited questions from the initial interviews, to further explore the Mechanisms identified for implementing and sustaining the services, this time specifically to explore their relevance for sustainability. Four of the Mechanisms identified in the initial interview data were identified as important for the sustaining of the site A service:

1. strong support for volunteers

2. widening criteria for referrals

3. shaping the service to fit with local service drivers

4. funding and goodwill work and costs absorbed.

Challenges and barriers at the original site

Analyses of interviews with service staff at site A highlighted challenges that had been encountered in implementing the service. The first of these concerned a legacy of tensions between paid doulas and maternity service staff. A senior staff member said that:

In the context of, you know, working with, within local services and linking with other agencies is, the, the only area of tension is this business of the paid doulas going in and working with midwives and it not going so well and then volunteer doulas going in and it makes it harder for you.

Site A, doula service staff, initial interview

It was also reported that maternity services were concerned about role boundaries:

the question that I get asked most when I go is where, what is the doulas, what are the boundaries, first off. What are the doulas, what are the volunteer doulas going to be doing actually in here? Are they going to be taking the jobs of maternity assistants?

Site A, doula service staff, initial interview

Secondly, staff talked about the resource implications of the increasing number of referrals of women with Social Services involvement. These referrals require a lot more paperwork and attendance at many more meetings, including multiagency risk assessment conferences (MARACs) and case conferences. This can be quite challenging for the volunteer doulas and thus for project staff, as increased supervision and debriefings are required as well as extra training on how to record information, serious case reviews and court reports:

When you look at the referrals, before it was either self-referrals or referrals from the hospital, usually, or children’s centres. When you look now, Social Services are recommending us at MARACs. We’re getting recommended in, in childbirth, you know, written plans, case conferences, because we do make a difference for that woman engaging with other, in a multiagency approach. So it’s very different from what it [was].

Site A, doula service staff, initial interview

Smoking cessation work was also proving very challenging. Doulas felt that getting women to stop smoking was too much to aim for, and feared that being too insistent about the issue might mean women disengaging from the service. A partial solution had been found through local health champions:

We’re training up six health champions. So when the Locality Workers go in and talk to mum and say, have you, ‘who smokes in the family?’, ’cause it’s one of the questions. ‘Who comes into house and smoking?’ ‘Would you like some help asking Auntie Matilda to smoke outside?’, for example. Then they would link in with a health champion who’d go on a joint visit with the doula. So it’s them that’s actually discussing the public health outcomes. But then the doula continues to ask them at each visit, ‘and how is it going?’

Site A, doula service staff, initial interview

We revisited these problems raised in the initial interviews in the follow up interviews, to find out the extent to which they have been challenges or barriers in the sustaining of the service. The relationship with maternity services in site A was reported to be a very happy one. Former problems based on paid doulas’ interactions with midwives were no longer an issue:

I certainly think our relationship with maternity services in [name] is excellent now, we don’t have half the challenges we had when we first started, we have been accepted and recognised as a highly professional volunteers service.

Site A, doula service staff, return interview

Tensions were identified with other services and organisations. These were reported as ‘potential’ rather than ‘actual’ and included other third-sector organisations, because of competition for funding:

I think there is potential for a bit of conflict with the [another third-sector organisation] because [name] have now received funding through the Department of Health to deliver a similar type of service. So they are building that doula-ing type approach into the [name] work . . . Now in theory we should all be able to share and work together but I don’t know whether that will be a risk or challenge.

Site A, doula service staff, return interview

The resource implications in supporting an increasing number of women with Social Services involvement continued to be an issue, with paperwork being a particular challenge:

The [staff] have to be a lot more specific in how they do their report writing because they produce a day-to-day report from the doulas’ information and their information which comes to [senior staff] ready for it to go to the professionals’ meetings. So they are a lot more in-depth.

Site A, Doula service staff, return interview

Supporting women with Social Services involvement means that the service can support fewer women in total:

I think sometimes that’s where we struggle if we are looking at how many doulas we’ve got because what you are doing is saying when you’ve got eight or nine safeguarding cases, the time that you take to support them, you could quite easily have supported 12 cases that were not safeguarding.

Site A, Doula service staff, return interview

Smoking cessation work had also continued to be challenging. This seemed to be in part because it represents a culture change for the service:

It has been interesting; because it has been a culture change for the organisation, the volunteers were reluctant to take on that side of work. It wasn’t what they thought was the priority for the service. I’m sure that is what came across when you did those interviews before but I think and the managers weren’t that keen either because really, when the volunteer doula project started, it was really focusing on a natural childbirth and the impact of a natural and enjoyable birth had on parent and baby opportunity to bond. And I think and to breastfeed. And I think that then became a bit tricky because they suddenly had to start challenging women about behaviour and that is kind of changing the relationship a bit. And I think that is what the challenge was for managers in terms of how would they pass that information through to volunteers. It has taken about a year or so to for them to really realise that has to happen and that the mother’s health and behaviour during her pregnancy has a huge impact on the health of the newborn baby.

Site A, Doula service staff, return interview

Whilst progress had been made, Smoke Free Zones work was embedded into the service level agreement in 2013 and remained more successful than smoking cessation work:

definitely in terms of the smoke free stuff we are finding that women are very receptive to that and recognise it is not a good thing. But there is still a number, our stats around women smoking and being referred to the smoking service are still very low. And I don’t really know why that is to be honest, I don’t know if it is because they are lying to us or because the volunteers aren’t pushing the issues or because they genuinely aren’t smoking.

Site A, doula service staff, return interview

The return interviews with the staff also drew attention to the rising costs associated with translation and interpretation:

I think the cost that is the one that fluctuates the most is translation costs. Translation and interpretation ultimately depending on the numbers of women that are referred who don’t have English as a first language and need quite a lot of translation and interpretation support. So that fluctuates, I mean I say fluctuates it tends to go up rather than down.

Site A, doula service staff, return interview

The service had experienced some problems relating to the balancing of referrals and volunteers. It was suggested that criteria needed to be tightened in order to keep the number of referrals manageable but ‘I don’t think they like to say no to anyone actually’ (site A, doula service staff, return interview).

Alongside increasing numbers of referrals, the fluctuating availability of volunteers has to be accommodated. Service staff spoke of the stress that can arise when doulas become unavailable because of illness, family commitments or other personal reasons. No woman has ever been left unsupported, but there clearly have been ‘tense times’.

Thirdly, staff explained that it can be a problem securing the references needed for volunteers, which slows down the process of recruitment. To speed the process up, the service has begun taking verbal references and then asking for a signature on a standard reference form:

We reflect, every time we do an intake, we reflect on how that pathway has worked, have there been any difficulties and I think one of the things that we struggle with when we get the volunteers in is having references. You know, getting the written references. So what we’ve decided to do, once they have been successful, is we now take verbal references on the telephone when we’ve offered a place . . . We get a verbal reference and then we send out our standard form for them to confirm it in writing.

Site A, doula service staff, return interview

In summary, challenges and barriers to the sustainability of the original doula service include the balancing of referrals and active volunteers, which is affected in part by the increasing number of women supported who have Social Services involvement, and the changing focus or direction of the service from natural birth to public health issues, which some volunteers have found hard.

Commissioners

The original site and three of the four roll-out sites (all except site X) had been commissioned at the time of the evaluation (site W had received seed funding only). Site A was initially funded jointly by the local authority and the PCT, then by the local authority only and finally by the PCT only; site W received PCT funding; site Y was jointly funded by the local authority and the PCT; and site Z was funded by the PCT.

A total of six commissioners were interviewed whose fields covered a range of specialities within health, children, families, young people, maternity and early years.

All six of the services’ commissioners knew about the services prior to being asked to commission them. All three commissioners of the original service had prior connections with the host organisation and at site Y the commissioner had previously worked with the director.

Local champions

Four local champions were interviewed and those at sites Y and Z were involved in the commissioning process. A local champion at site Z explained that helping to secure funding had been their key role within the service:

With my commissioning hat on, with my sort of CCG board member hat on, my role was partly sort of trying to secure ongoing funding, and we had a, a sticky few months where people were saying, yes, you want to fund it but we think it might have to go out to any willing provider and go out to open tender. Well they all know this is (laughs) going to be a disaster. And we’ve managed to find a way to carry on funding it for the rest of this financial year, and we’re putting it in the commissioning, the strategic plan . . . for the next year.

Site Z, local champion

The local champion at site Y had been working as a commissioner at a point when services in the antenatal period were being commissioned, and had invited the social enterprise director to apply; that application became the basis of the local doula service. This local champion was also instrumental in protecting the service’s funding when it was under threat, by explaining to senior commissioners what the service actually comprised. This would not have been possible had this individual been less involved with the service.

Why the service was funded

The commissioners of two of the services – at sites A and Y – said that they had been able to give the services the amount of funding requested and for the time period required. Site W received only transformational seed funding, because of the rules associated with its location within a hospital trust. At site Z, 1 year’s funding had been sought, but only 8 months’ was awarded.

The commissioners of the four services were asked about how and why they had funded the services. First, what information had they needed when making a decision about funding such a service? Responses alluded to a wide range of information including:

• epidemiological data

• information about maternity services, including national guidance on how maternity services are working; booking targets; breastfeeding targets; pressure points identified via needs assessment; equity of access to maternity services; experiences of maternity services of disadvantaged groups

• information about the local population: number of birth; number of births requiring interventions; breastfeeding rate; C-section rate; and area differences in these

• information about the doula service including the number of women supported; how it supports vulnerable families; costs; and outcomes.

Second, commissioners were asked what additional considerations apply when deciding whether or not to fund a third-sector service. Responses included:

• experience; reputation; ability to deliver/do the work

• previous experiences with the host organisation

• whether or not the service fits with wider remit of commissioning work

• whether or not the service fills a gap in statutory service provision

• whether or not the host organisation is able to take a community development approach

• whether or not the host organisation is able to work with relevant sectors of the community

• competitive costing.

Third, commissioners were asked specifically what had led to a positive funding decision in this case and what they had particularly liked about the services. All six commissioners said that they had chosen to fund the service because of the likely impact on two groups: the service users and the volunteers. They also alluded to wider strategic, public health and service issues. The reasons given thus fall into three categories: reaching the families most in need – the most vulnerable and deprived families and/or those least likely to access statutory maternity services; helping local women in areas of low employment into training and paid work; and fitting with wider health/public health strategies and/or filling gaps in statutory services provision. These reasons are illustrated below.

Processes involved in the commissioning of the services

The commissioners were also asked about challenges in the commissioning process, and what advice they would give colleagues on the commissioning of services such as the volunteer doula services. The local champions were also asked about challenges faced in their roles championing the services. When interviewed at the start of the evaluation, service staff were asked about securing initial funding. The answers to all of these questions highlight some of the processes that have been or may be important in the commissioning of such services, according to some of those involved.

Facilitators at the roll-out sites

We revisited questions from the initial interviews with staff at the four roll-out sites, to further explore the Mechanisms identified for sustaining the services. All of the Mechanisms identified in the initial site were identified as important for sustaining the roll-out sites, together with an additional Mechanism – modifying the training to aid retention of volunteers – which emerged from the return interviews data:

1. strong support for volunteers

2. modifying the training

3. flexibility in eligibility criteria

4. marketing to gain referrals

5. shaping the service to the local context

6. responding to funding availability

7. goodwill work and costs absorbed.

Challenges and barriers at roll-out sites

Three problem areas were identified in the initial interviews with staff at the roll-out sites. First, all four of the roll-out sites were experiencing funding challenges. Site X was in need of funding to train a new cohort of volunteers, which was described as a big expense, in part because of the childcare costs involved. The overall goal was to grow the service, become an established service and secure PCT funding. At site W a number of problems were mentioned that relate to funding: an ongoing lack of administrative support for the service; a plan to reduce volunteers’ supervisions from three to two over the course of the minimum post-training support work; and stopping using the site A service database because of the high costs involved in doing so. Site Y had experienced 25% funding cuts and because of this had postponed training of the next cohort. At site Z initial funding was coming to an end and it had been rather overwhelmed to discover how much work was involved in running the service:

We were really running to keep up, weren’t we, to start with, because there was so much to do, and we didn’t have the staff that [site A] have.

Site Z, doula service staff, early interview

Second, all four of the roll-out sites were experiencing some sort of tension with another service. Both site X and site W had experienced some tensions with midwifery. At site X these included midwives’ suspicion about doulas at births doing some of their work; doulas forgetting some of the boundaries addressed in the training; and doulas and midwives both writing birth plans with women and then clashing over this. At site W a history of mistrust of paid doulas by midwives on the delivery suite was cited and a site W delivery suite midwife also talked about this legacy. Sites Y and Z had both experienced tensions within the local voluntary sector over competition for funding.

Third, problems with volunteers were cited at three of the roll-out sites. At site X, these related to failing to follow procedures for contacting the service, using the designated mobile phones – a health and safety matter – and volunteers not adhering to the rules regarding expenses. At site Y, there was concern that some volunteers had unrealistic expectations of the role. At site Z, 50% of women referred wanted a doula with them for the birth, which was a much higher proportion than expected. While some volunteers were excited about this, others were not able to offer support at night. Some volunteers at site Z were also not happy with the current level of security and wanted ID badges.

In the follow-up interviews with service staff, we revisited these three problem areas raised in the initial interviews, to find out the extent to which they have been challenges or barriers in the sustaining of the services. Funding problems, experienced at all four roll-out sites, were ongoing. The site X service was still not commissioned and continued to be funded by the social enterprise host organisation. They were still unable to afford to train another cohort of volunteers, which was becoming a more pressing problem as volunteers moved on.

At the time of the follow-up interviews, site W still lacked administrative support and a business case had been written for additional support for supervision work. At site Y the effects of the changes to the commissioning system were being felt, with only short-term funding being made available:

The issue for us is that the commissioning structures that were in place with our previous contract are now not there any more so it creates instability for everybody in the NHS in that it’s changing over to now a clinical commissioning group so now it’s obviously a completely different ball game. They can’t commit to a long-term contract at the moment because of the changeover so we know . . . we’ve had it confirmed that our funding for this year but clearly it’s much more stable if you have a 3- to 5-year contract but it’s not possible because of the government changes. I mean, that’s the case for everybody at the moment.

Site Y, doula service staff, return interview

Site Z was awaiting confirmation of 1 year of funding, at the time of the return interviews, with training work on hold until then:

The funding issue has made it – the insecurity that brings and not knowing whether and when we are doing the next training. That has been difficult to manage.

Site Z, doula service staff, return interview

While funding problems were ongoing, tensions with midwives had apparently resolved at the two roll-out sites which had experienced them: site X and site W. At site X, this was ascribed to increased visibility and marketing. Doulas had attended births and supported mothers at the new birth centre so they were becoming more known and midwives were recognising the work that the service does in supporting vulnerable women:

I have to say I’ve heard a lot of good feedback from the doulas themselves, midwives are praising them, giving them really good feedback and they’ll send that back to me via email and we’ll make sure we keep a record of it. So it’s been very positive.

Site X, doula service staff, return interview

Marketing materials had been produced and circulated, which addressed the doula role and boundaries:

We’ve had good marketing material as well, I think that’s helped because on that we’ve really tried to highlight exactly what the role is, what the boundaries are, the little sections of the bottom of the leaflet saying that we don’t provide medical or clinical advice. So very much pointing out exactly what we do and what we’re here for. So I think that’s helped a lot as well.

Site X, doula service staff, return interview

At site W, relationships with midwives were similarly reported to have improved:

I think it took them a bit of a time for them to get used to it, thinking are these are private doulas, but I think once they got used to it they know that they are part of the team, part of [host organisation], and they are volunteers and as they have worked with them on delivery suite, things are changing, positive working relationships. I think we used to have this thinking ‘oh private doulas’. We have always highlighted that they have responsibilities, their role is a non-clinical role and the midwives are there clinically so knowing that they are not going to step on each other’s toes really.

Site W, doula service staff, return interview

The tensions within the voluntary sector that had been reported in initial interviews, were, however, persisting in both sites where it had been raised. In both cases the issue was concern that a large charity would take up their ideas and that they would not be able to compete for funding.

Some of the volunteer-related challenges had been successfully addressed at the time of the return interviews, while others were ongoing. Security issues in both sites had been successfully resolved by using external security companies. At site X, volunteers’ safety when going into family homes has been managed via a security company, already used by the host organisation, that volunteers phone before entering and after exiting homes.

This has relieved a lot of pressure on project staff:

in the early days it was quite a lot of a burden where doulas would ring me to say this is where I am, this is my location. I would have to call them back after an hour later and literally throughout the day and night. So it was a bit too much for me to take on. So luckily in the last year we’ve had [security company] who’ve taken over that, and they’ve been . . . really, really good.

Site X, doula service staff, return interview

At site Z an external company has similarly been contracted for nights and weekends, but the project workers are still responsible for checking doulas’ whereabouts during office hours:

because it is expensive and we don’t have funding to pay for a full system. I think [another site] are doing the same work and so I am following their advice, in daytime staff take the call, and evenings and weekends they go to the security.

Site Z, doula service staff, return interview

At site Y, staff felt that the problem of volunteers’ unrealistic expectations of the role had been successfully addressed by the time of the return interviews. They felt better able to tell volunteers the commitment expected, including support work, supervisions and training. This was ascribed to experience they had gained in running the service:

I think we recruit with a little more information as well. I think everything we do has been more informed by the fact that we are now, you know, a lot more experienced in what we are doing and so those that come on board I think are clearer as to what they will be doing.

Site Y, doula service staff, return interview

At site Z, the problem of volunteers not being able to work nights continued, with approximately 70% of the volunteers being unavailable to work at night. This clearly is a major challenge to allowing the service to meet women’s needs:

The mums who need labour support means you can be called any time 24 hours but only a few volunteers are available to support during night-time, most only in daytime 9–5, 9–6, 9–3, so it is really hard, whoever the team is that are available it is really difficult. I am going to be a bit careful who I approve next time . . . In the induction day we are making them sign an agreement that they have to be available and when I make the presentation on the open day I will make it all clearer, repeating myself. I did it all the time. But when I do the interviews I will ask the question specifically about availability and who will look after their kids and all that stuff. We can select on the recruitment.

Site Z, doula service staff, return interview

Staff at sites W and Z expressed concerns that not all volunteers were fully engaging with the role:

there are volunteers who for various personal reason just can’t offer time. And my feeling is that for whatever reason they just either have never been or are just not really engaged in the project.

Site Z, doula service staff, return interview

The return interviews with the staff at three of the roll-out sites also highlighted problems not identified at the initial interviews. A senior staff member at site X (who was not in post at the beginning of the service) talked about the service not being sufficiently linked in to the wider health/social care infrastructure. At site W there was concern that women were confused between the doula service and a new voluntary sector project with overlapping aims.

Site Z had also experienced challenges due to staffing discontinuities. A key person in the organisation had had a period of absence. At the time of the return interview another key person was just beginning a period of leave and staff were very aware of the difficulties that this created.

In summary, challenges and barriers to the sustainability of the roll-out services include funding problems; volunteer related problems; relations with other services; and staffing discontinuities. These differ somewhat from those in the original site, which, while including funding, also included balancing of referrals and active volunteers, and the changing focus or direction of the service.

Table 30 provides a summary of the barriers and challenges to sustainability of the original doula services and the four roll-out services and Mechanisms that addressed some of the initial barriers.

Service documents

Documents developed and used by the original site included those related to the recruitment and selection of doulas; service policies; information about the role for doulas and for women; equipment use and reclaiming expenditure; and documentation included in each woman’s file. All roll-out sites had also developed additional documentation of their own; some were novel and site-specific, and others were modifications of those developed by site A.

Original site database

Variations from the replication package with volunteers

The four roll-out sites deviated from the original site’s replication package in a range of different ways, including supervision, training and roles of volunteers. The quotes in the following sections illustrate these variations.

Supervisions

Three of the four roll-out sites have made changes to the replication package’s model of supervision of volunteers (which is three supervisions per woman supported). At site W, at the time of the initial interviews, the three supervisions had become two. However, by the time of the return interview, a need to increase the number of supervisions was expected as the service grows. In the meantime, experienced volunteers were providing additional non-supervision support to other volunteers:

With regard to supervision I think it has always been my concern that as the project grows the amount of supervisions that is required grows as well. I have put in a business case, put it forward to the managers, for the need for having additional support for supervision. At the moment she [the Locality Development Worker] is doing it using some of the more experienced volunteers as sort of, using them to be not necessarily providing direct supervision but providing the support side of the supervision. She has set up little teams which I think is a very creative way of doing it without having that additional person in post. But we have just another nine who have qualified and are now being matched, and there is another cohort finishing who have just had nine there, and I am very aware that the amount of supervision that is required is increasing. So I have written a business case to try to recruit some additional support, to support and provide supervision.

Site W, doula service staff, return interview

At site Z, group supervisions had been added to the individual supervisions that are part of the replication package, at the time of the return interviews:

Once a month there is group supervision, so they can come as a group and request if they particularly want an input from a particular health professional then that can come from them.

Site Z, doula service staff

A new supervision policy, one more empowering for volunteers, had been introduced at the site Y, at the time of the return interviews.

Training

Two of the four roll-out sites have made changes to the model of training of volunteers. Staff at site W in the initial interview reported both adding and removing training topics and, in particular, added more training around different cultures and faiths. They were aware that site A modified the training package as needed.

Site Y combined the original site’s OCN accredited training with its existing City & Guilds accredited training, thus creating a whole new training programme:

We’ve really sort of built a whole new training programme haven’t we?

Site Y, doula service staff, interview

Site Y also runs a taster training course as a precursor to the full training course, to help avoid volunteers dropping out after expensive training without supporting any women:

something else a bit different to the way things are done in [original site], because when I came into all of this, the background with community parents, what had shown as particularly worked well was having initial taster courses, rather than taking volunteers straight in. And you can invest a lot of time in quite expensive training, and people can drop out or you find out they’re not really suitable and all that sort of thing, that we’d, we had trialled taster courses and, where people could test the water and we could find out whether they’d be suitable during that process, and also whether they’re going to come to each of the sessions.

Site Y, doula service staff, interview

Volunteer roles

At the time of the initial interviews, site Y had also made changes to the role of the volunteers, with some carrying out outreach work:

the thing that we have put in place that has generated referrals is volunteers doing regular sessions in places where pregnant, where parent, where parent, parents-to-be will be. So we’ve got volunteers doing outreach sessions every week or every 2 weeks. So we’ve got say a, a midwife clinic who’s, the midwife who’s running her clinic at a children’s centre, the volunteer will be in the waiting room talking to those waiting to go in to see the midwife, or as they’re coming out. So have sort of a bit of a display, sharing information, trying to identify those that might need extra support. And what we found is, because they know the programme really well, they can talk about it in the best way to, to parents who might need it.

Site Y, doula service staff, initial interview

Service costs

The costs of providing the doula service were calculated from information collected from service managers of the Hull Goodwin doula project and the roll-out sites. Managers were asked to provide information relating to the costs of doula recruitment and training; doula costs (travel and telephone); and the costs of running the service including premises, equipment and consumables. This information was requested for three financial years: 2010/11, 2011/12 and 2012/13. The questionnaire issued to service managers appears in Appendix 13. The roll-out sites had some difficulties in providing information with several items missing or estimated. In some cases, costs could not be disaggregated from those of host services.

Health and clinical outcomes

Doula service costs

Health and clinical outcomes

The following section draws on outcome data available from service A database, as described in Chapters 2 and 3.

Costing summary

In the following section, service site cost aggregates are presented in Table 46 for the Goodwin volunteer doula service so that potential cost savings can be contextualised by the costs associated with running a mature volunteer doula service.

Doula payment

The doulas in this intervention provide their labour voluntarily, so from the NHS + PSS perspective the time spent by the doula in client support is a free resource and therefore imposes no cost burden.

On the other hand, if the perspective is (momentarily) widened such that the doula’s opportunity cost of time forgone on other activities could now be included in the cost schedule, then some portion of that time may involve productive activity involving paid employment and so could be costed at that level. However, when we surveyed the extent to which their role as doula affected their ability to take on paid work or devote additional hours to their existing work, 77.9% (53 of 68) responded that there had never been any effect on this. Only 4.4% (3 of 68) recorded a frequent impact; the remaining 12 doulas reported that they had sometimes been affected.

Departing from the volunteer model to one in which doulas are paid as employees by the NHS, the previous result (showing little hindrance if seeking additional time for productive labour) suggests that the reservation wage to work as a doula may be in the region of the minimum wage. This is also in line with responses to questionnaires completed by doula-supported mothers, in which exactly one-half favoured doulas being paid an hourly wage similar to that of a maternity support worker (£7–10 per hour), with the next most popular response (31%, 24 of 78) suggesting pay be at the minimum wage, £6 per hour. The remaining 13% (10 of 78) suggested wages be similar to those of a newly qualified midwife (£11–14 per hour).

The implications for service costs of an NHS-paid doula model are illustrated with one example in the context of service A. Assuming a wage rate of £8 per hour with 30% add-on costs, an active list averaging 34.25 doulas at any one time and contact with client and service totalling 4 hours per week on average [from responses to questions 5(a), 5(b) and 6(a) in the doula questionnaire] would add approximately a further £74,000 to total costs per year.

Service costs

The per birth cost to the NHS of the volunteer doula service for disadvantaged childbearing women was estimated to be £2349. This figure was arrived at using Hull Goodwin expenditure data, where this, the original service, is based within the third sector and is a mature service, relative to the roll-out sites.

Health and clinical outcomes

The per birth contribution to NHS savings resulting from improvements in health and clinical outcomes resulting from doula assistance was estimated to be £487. This figure was arrived at by aggregation across a number of health and clinical outcomes hypothesised to be influenced by the doula intervention. It is worth noting that those outcomes may not necessarily be exhaustive, for there may be benefits from doula support that exist beyond the maternity episode that may be of wider societal benefit for women, and indeed for doulas.

The major contributor to the estimate was NICU use and this was shown to be sensitive to changing assumptions. For example, one additional NICU admission per year (see the ‘+ 1’ column in Table 45) reduces to £282 per birth the estimate of the net contribution to savings versus standard practice in NHS Hull.

Births by C-section among women who received doula support from the Hull Goodwin service were fewer than both comparators, but only marginally so: 21.2% versus 25% (NHS England) and 23% (NHS Hull). Despite the modest improvement, cost savings due to the doula service were evidenced even in the sensitivity analysis scenario in which all caesareans were assumed to be performed as emergency procedures. Contrasting the marginal difference observed here are two US studies, by Kennell et al. 15 and Kozhimannil et al. ,62 in which the differences between the two caesarean rates were of the order of 10%, in favour of doula support. If this level of improvement were translated into the UK context, then the contribution to cost savings due to the doula service would be substantial.

Epidural use among doula-supported women in Hull for normal or assisted birth averaged 23.3%. This was a higher rate of use than that observed throughout NHS England (16%) but does represent a one-third reduction on the rate of use seen in the local region NHS Hull (35%). When the doula service was compared against the local comparator, NHS Hull, less epidural use resulted in cost savings to the NHS. However, this reversed when the comparison was with countrywide practice. There was, of course, statistical uncertainty about both estimates due to small sample size. Finally, we note a similar epidural rate of 27.9% (n = 1079) reported in a recent study by Kozhimannil et al. 62 for doula-supported births to women of lower socioeconomic standing, where the cohort in that study was of Medicaid recipients in Minneapolis, MN, USA.

Costing summary

The net per birth cost to the NHS from doula support to disadvantaged childbearing women was estimated to be £1862. The caveats already noted on the relevance, weighting and uncertainty affecting each component in this calculation will be inherited by the net estimate.

Disadvantage

As noted in Chapter 1, the term ‘disadvantage’ covers a range of social, economic and psychological states which frequently coexist. Our evidence suggests that the doula services are indeed successfully engaging with women who are disadvantaged, as can be seen by data from a range of sources. Referrals to the original service were in areas with higher than average material deprivation as categorised by the IMD. The IMD comprises seven dimensions of deprivation: income; employment; health and disability; education, skills and training; barriers to housing and services; living environment; and crime. In several of the sites we included, other initiatives that target women in disadvantage were evident, for example Family Nurse Partnership programmes.

Referrals to the original doula service reflected higher rates for younger women and lower proportions of white British women than in the local population of childbearing women. Midwives described the positive features of doula support for isolated women without family networks or other support; they perceived that doulas supported access to and uptake of services, for women who may have otherwise found it difficult. Through women’s questionnaire responses, it was evident that a higher proportion of women who engaged with doula services were born outside the UK than national rates for women giving birth in England. A large proportion of women referred for doula support lived alone or had no supportive friends or family available to them. Women also reported their difficulties and thus disadvantage in trying to access services due to language or cultural barriers. Doulas also spoke of the difficult lives of some of the women they supported.

Ethics and governance issues

As with any research of this scale, considerable effort was invested in the ethics and research governance processes. The particular nature of the service being researched, that is one that comprises volunteer doulas working in the main from third-sector agencies, provided challenges in ensuring the appropriate governance mechanisms were pursued, as NHS governance staff are generally unfamiliar with the role of doulas and do not frequently encounter research that evaluates services provided by trained volunteers, based in non-NHS agencies. Some of the third-sector agencies were unclear of their status and governance requirements, if they received some funding contribution from NHS organisations. Finding answers to those queries required them to seek advice which added to the time required for governance approvals.

We encountered delays while trying to achieve portfolio adoption, until our research was reallocated within a team. We also encountered situations where NHS R&D departments would not initiate review, if, for example, staff were due for annual leave, because of the time targets they are required to meet.

We are aware that the NIHR is familiar with these issues but report them here because governance processes continue to result in serious delays and threats to timetables.

Patient and public involvement

We described our convening of two user panels in Chapter 2, section Advisory group and public involvement mechanisms, and their contributions to identifying important topics to be explored: in advising on approaches to data collection and in testing out data collection instruments for feasibility and acceptability. Public involvement in our research was further strengthened by a multidisciplinary advisory group that comprised a range of stakeholders including staff of doula services, women who had received doula support and doulas, in addition to maternity service commissioners, NHS staff and academic experts. These individuals provided advice both in and between meetings and also supported our research in their local constituencies, where applicable.

Measurement of clinical and public health outcomes

Measurement of public health outcomes was possible only for the original site, as numbers were too small at the roll-out sites. Even so, the sample size available in the Goodwin database was rather lower than we had expected and thus, for some clinical and public health outcomes, estimates of effect are very wide. The data held in the Hull doula database had a substantial data cleanse, routine checks were carried out for inconsistencies and data were checked against manual records. However, the limitations of the relatively small sample of doula-supported mothers must be noted in interpreting outcomes. We were constrained by what was available and inevitably use of comparisons is less robust than determining effects through prospective randomised controlled trials. In our application, we were always clear that such a trial was impossible for a range of reasons including the provision of DH funding to support roll-out and the fact that doula services were not in equipoise. Given the impossibility of a randomised controlled trial, the use of reference data sets of routinely collected data, available through the public health observatories, provided the most robust alternative. The data used were produced at different geographies, hence the use of PCT, local authority and trust data to strengthen comparisons.

Hull PCT and Kingston upon Hull local authority were considered coterminous and were used, as comparisons are often made in these situations. Other data were derived specifically from the maternity database at HEY. When this data source was used, the data were filtered to include only mothers who were resident in Kingston upon Hull. Analysis of deliveries in HES shows that 99.2% of deliveries to mothers who were resident in Hull PCT took place in HEY. No data quality issues were identified in the published data used for comparison. Data published by the Office for National Statistics, the DH and the Health and Social Care Information Centre have undergone quality assurance before publication. While some PCTs have known data quality issues relating to breastfeeding, this was not the case for the Hull PCT or its statistical neighbours. Data from HES are known to have some data quality issues, some fields more than others. Areas of known concern were avoided.

Data from doulas and women

We were surprised by the lower than expected response to questionnaires among doulas, especially in the context of the demonstrated commitment to the role. This constitutes a limitation to the interpretation of findings, although we were able to identify and offer explanations for both positive and negative experiences of the doula role. We can only hypothesise that our distance as researchers with information brokered through a third party may have contributed to the low response rate. It may also have been that doulas perceived our research as a potential threat to their organisation and thus did not engage.

We had expected a low response rate from women offered doula support and this was our experience. The implications of this low response rate must be noted. Despite multiple options for providing their views, response rates to our questionnaires were low, with 23.6% of women completing questionnaires out of the 84% of women who had been introduced to the service and sent a questionnaire. The proportion of multiparous women was higher than expected; unfortunately we do not have information about their previous childbearing experience. We suggest that reasons for the low response include the fact that women who were eligible to receive doula support were in groups traditionally hard to engage in research. They may well be suspicious of strangers asking questions and may worry about the implications of answers or even of responding, for example with regard to benefit entitlement or applications to stay in the UK. We had hoped that approaching women via an organisation previously known to them and reassuring them of confidentiality would help to maximise response rates. We also tried to ensure that potential participants understood that their responses would not be fed back to the doula services. We knew that some of the services focused their support on women from BME communities that would include some women unable to read and write in either English or their own language. That was part of the rationale for offering a variety of different ways in which women could take part. We also appreciate that some topics, for example emotional well-being, are sensitive ones about which women from some cultures may feel diffident about providing information. We realised that some of our potential participants had accessed the service several years previously and may well have moved from the address recorded while they were in receipt of doula services. That was borne out by the proportion who could not be contacted via the information held by the doula service. We do, however, feel that the number of women included who gave their first language as other than English (45.2%) is a strength of this research.

Low response generally means that those who are less articulate and those without strong feelings, in either direction, are less likely to offer a view. Since so many of the women were so positive about the service, it may be that we have failed to get responses from those who did not rate it so positively. It was evident, in at least one focus group, that women saw their attendance as a way to repay to the service the benefit that that they had received. Those who did not think that they owed the service anything will not have been motivated in this way. Thus, although we were as facilitative as we could be to enable women to attend the focus groups (e.g. providing transport, lunch, crèche and financial acknowledgement), it was evident that for many women the effort of getting themselves and their children to a specific place at a specific time was quite substantial.

In the absence of randomly derived groups, we had given considerable thought and taken advice on the identification of potential comparison groups against which, for example, women’s experiences of labour could be compared and from which we could identify questions acceptable to women that could be incorporated into our questionnaire survey. In using the Picker survey41 as our comparison, we were aware of some differences between the Picker sample and our own. These included the different years over which experience of the maternity service was compared (i.e. women responding to Picker gave birth in January and February 2010) and that Picker outcomes relate to a sample rather than all women giving birth at the trust; our sample was drawn from women who had experienced maternity care during all of the years in which the original doula service was available. Picker respondents are generally more likely to be older, white mothers, who may rate maternity services more positively than younger and non-white women (Anderson, Public Health England, 2013, personal communication). However, there were also similarities in that both were postal questionnaires (generally associated with a low response rate) and the means of maximising response rates for both questionnaires was informed by consultation with service users. In addition, the questions from the Picker survey included in our questionnaire have been utilised in several national surveys of maternity care.

Language support

Interestingly, the possession of a common language for communication is not considered essential by a significant proportion of women and doulas. Some midwives appeared to see the doula’s main purpose as language support and were thus mystified when a doula did not share a common language with the woman that she was supporting. There were also suggestions that some midwives might be more comfortable with the idea of a doula fulfilling a role as an intermediary which complemented their own role rather than challenging it.

A particular challenge in this research was involving women who were not fluent in English. Over half of the doulas had supported women who did not speak English. Of the women on the database of the original site, 37.1% were not ‘white British’ and, across all the sites, 45.2% of women completing questionnaires identified themselves as having a first language other than English. To identify these women, we were dependent on service records to tell us that the woman required an interpreter. We were surprised that smaller numbers of women than expected apparently required language support for their participation. Doula service documentation concerning language was largely limited to the need for an interpreter, rather than documenting the main language(s) spoken, which led to under-reporting of languages that could be pursued for the non-English-language focus groups. It is also possible that the need for language support was under-reported for pragmatic reasons: the financial implications of interpreter support; the greater complexity of involving an interpreter; and possibly as a strategy to encourage communication in English language and social integration. It is also the case that participation in research may require a different vocabulary than is required to engage with services.

Working with third-sector agencies

The organisations that host the doula service were generally of small to modest size with insecure funding in the longer term. Staffing levels in all of the doula services are low and administrative support limited. Priorities for these services include securing their own funding, training and running the service on a day-to-day basis.

We were aware that, in seeking the doula services’ support in steps towards data collection, we were placing additional burdens on them. Despite the provision of some funding for administrative support, it appeared difficult for services to put this to positive use within a short time scale. We identify these as issues that researchers may wish to consider in future research with such services.

Prior to commencing the research, we had been given to understand that the database at the original site would yield data for approximately 600 women and that data quality was considered good. The database and reports available from that system were, of course, designed for service provision rather than for evaluation; thus we identified the need for work to quality assure the data set before extraction could begin. Following completion of data cleaning and other aspects of quality assurance, it became apparent that outcome data were not available for as many doula service recipients as we had been given to understand. Further reflections, which may be of interest to researchers, on conducting research into volunteer roles and with third-sector agencies are included in Appendix 14.

Challenges with realistic evaluation

Realistic evaluation emphasises that interventions are intrinsically theory based: a given intervention is implicitly meant to tackle an underlying ‘social problem’ and people have, often unspoken, assumptions (theory) about how and why change happens (Mechanisms). The initial work in a realistic evaluation is to make the theory explicit. Identifying what different people see as the problem that is being addressed by the intervention was an important first step, achieved through our early interviews with doula service staff and focus groups with women’s and doulas’ user panels. Through initial data collection we identified each service’s key features and how they operationalised doula support. By defining the interventions we hypothesised the implicit CMOs at each site. These then had to be rendered accessible to key informants and we adopted the approach of working backwards from Outcomes to elicit informants’ beliefs about Mechanisms that would achieve those from a range of starting points or Contexts.

Published accounts of the use of realistic evaluation most commonly investigate interventions that occur at a system level. Furthermore, these are commonly investigated using a more discursive (qualitative) approach. The study of the Scottish policy initiative Keeping Childbirth Natural and Dynamic64 provides an example of this approach to programme evaluation. In our research, CMOcs were constructed at different levels of Context: woman level, doula level and doula service level. There is rather less literature on the application of realistic evaluation where an intervention is targeted at the individual level, as is the case with volunteer doulas.

Some may question the extent to which realistic evaluation is consistent with a quantitative paradigm unless considerably larger data sets are used. Further, how ‘testable’ are CMO configurations within the quantitative paradigm? The characteristics (Contexts) that have been identified as ‘for whom’ and ‘in what circumstances’ occur in combination, so bivariate analyses are unable to tease out the possible interactions that a realistic evaluation approach seeks to disentangle.

Context, Mechanism and Outcome configurations can be challenging to create or propose for several reasons. The CMO can ‘slide’; that is, a Context may sometimes be viewed as a Mechanism (e.g. provision of support), a Mechanism may sometimes be viewed as an Outcome (e.g. self-efficacy) and an Outcome may, itself, become a Mechanism for a longer-term Outcome.

A further challenge is discriminating between the description of the active intervention and the Mechanism; at times these appear to be the same, but the description of the intervention concerns its components, whereas the Mechanism describes the process by which the intervention (or its component parts) works. There are suggestions that the Mechanism must relate to how the intervention impacts on individual’s reasoning,33 but for some interventions this may not be achievable. An example from our research may be where accompanying a woman to appointments is the simplistic view of the Mechanism (and describes the intervention) but the Mechanism is reducing barriers to access, whether they are linked to lack of knowledge of the system, lack of transport for access or anxiety, but they do not relate to reasoning.

Another challenge is accessing the elements of the CMOc so that they are testable. This is particularly challenging when the intervention occurs at an individual level: the Context can be highly sensitive and the Mechanism can be subconscious. Participants are more likely to identify reasons for the intervention not ‘working’ that are at the doula or site level, rather than the woman level, for example engagement or motivation. However, the approach to data collection is also important: focus groups led to greater discussion of woman-level factors (Context) than obtained in the questionnaire and this reflects both the richness of data gathered in conversation and also the positive contribution of different but complementary approaches to data collection in our research.

Women’s perspectives

Doulas’ perspectives

A response rate of 34.5% was achieved. It included doulas who had experience of supporting women and a smaller number who had not. Doulas came from a wide age range and different ethnic groups and included women in employment and education. Among doulas who had not supported women, some had discontinued because paid work became available; others had not yet had an opportunity to provide doula support.

Almost all doulas felt that their role helped women to access other services and increased women’s knowledge and skills related to childbearing and looking after their baby. When considered against a framework of social capital theory,66 both bonds and bridges were formed within and between groups and also among individuals. The most important factor that allowed them to achieve that was that they ‘got on well’ with women they were matched with, rather than having a shared language, being of a similar background or living in the same geographical location. Where matches had not been optimal, this was ascribed to the doula’s inexperience at that time or women hoping for support in areas that the doula service could not provide. In a small number of situations, there were suggestions that, with other partnerships, things might have been better. Doulas did not feel that their relationship with a woman was affected by when they met.

Some doulas had experienced communication problems with health and social care professionals, especially in the early days of the service, but these were rare. Doulas were mostly of the view that the role was distinct from other roles; just under 15% felt there were some overlaps, predominantly with that of midwives.

Reasons why doulas stopped volunteering varied. In three sites, volunteering had been discontinued for personal reasons; at site X there were no discontinuations of volunteering; at site Z, reasons for discontinuation related to the organisation of the service and the role.

With two exceptions, all would recommend being a doula to other people, and over 90% rated their experience very highly. Low ratings were associated with perceptions of the doula service, including the staff and their availability.

There are parallels between women and doulas for positive experiences and outcomes. Our findings support existing evidence67 that experience as a volunteer doula has positive impacts for the majority of volunteers, across a number of domains including the doula’s own health and well-being, and for their wider social context. Training appears to prepare them well for the role, as reported elsewhere,63 and to be a Mechanism that supports satisfaction with the role, together with positive experiences. Satisfaction with, and being supported by, their doula service are also Mechanisms that support positive experiences (Outcomes).

Doula services

NHS maternity services

Other schemes that support disadvantaged women

We acknowledged, in Chapter 8, Women’s perspectives, the availability of the Family Nurse Partnership programme, which provides support to young women who meet specific criteria; this programme is available in some but not all areas. In that programme support is provided by salaried nurses specifically trained for that role. 75

There are a number of other peer-based approaches. We note some similarities between the volunteer doula role and other community-based support programmes for women who may find it difficult to access or engage with services. 74 A randomised controlled trial of pregnancy outreach workers, working from a base in the maternity services, found that support has a modest effect on maternal mental health for women with identified social risk factors. 76 In that trial, all women accessing maternity services are assessed for social risk in addition to medical and obstetric risk. Women who have been assessed and identified as being at social risk and who have been randomly allocated to the intervention arm of the trial are referred to a pregnancy outreach worker for a programme of support that includes home visits and that aims to support changes in behaviours such as smoking and breastfeeding and to ensure that available benefits are accessed. Workers are trained to a National Vocational Training Level 3 standard and have access to ongoing support.

A befriending scheme that supports pregnant asylum-seeking and refugee women comprises a package of support from a befriender that lasts from early pregnancy to 2 months after birth. 77 When applying to become befrienders, women who are themselves often asylum-seekers or refugee mothers receive a 4-half-day training programme. Befrienders are matched with their clients as closely as possible for cultural background and shared language and are expected to meet clients for up to 3 hours per week. Referrals to the service are received from community midwives and early years services. Early findings suggest that clients value befrienders’ support for a range of reasons and activities reported include help for clients to access a range of services and being accompanied to appointments. This scheme, which commenced in 2011, has exceeded some of its initial targets for volunteers and clients, and volunteers are also experiencing benefits from their experiences in the role.

Clinical and public health outcomes for women and their babies

Our findings suggest some improvements in clinical and public health outcomes, but these do not always reach statistical significance and must be considered in the light of the limitations described above. Information available from the original site prior to the start of this research suggested positive impacts on C-section and epidural rates. Reduced C-section rates were reported in studies of doula support in non-UK settings, for example by Kennell et al. 15 and Madi et al. 78 We have not found significant differences but, it must also be noted, we have not observed any worrying trends in any of the outcomes examined that would seriously call into question the safety of doula support. In the context of concerns about C-section rates in the UK, where significant reductions in rate are unlikely to be achieved, it is worth considering practices that may help to avoid increases. Furthermore, we are aware that some women supported by the doula service, especially at the original site, are not typical and that some multiparous women have had previous difficult birth experiences which may increase their chances of a C-section. Unfortunately, we do not have the data available to investigate this hypothesis further but it merits inclusion in further research.

For both epidural use and the prevalence of smoking during pregnancy, Hull is an area with higher rates than the comparators used. Interpretation of non-statistically significant reductions in epidural use is compromised by the necessary exclusion from the analysis of women who had a C-section. Apart from planned C-section, the decision to use an epidural usually precedes that of mode of birth and cannot be considered independent. It is known that women who have an epidural are significantly less likely to have a normal vaginal birth. 79 It is quite plausible that doula support may improve the likelihood that a woman with an epidural will go on to have a vaginal birth and this hypothesis could usefully be examined in future research.

It is difficult to attribute smoking cessation findings to doula support when so few women reported doula involvement in conjunction with changes to their smoking behaviour. In some sites, it appears that involvement in smoking cessation work has not always been a prominent or easy part of the doula role; thus the relatively low level of doulas’ involvement may be attributed in part to that.

The impacts on breastfeeding appear more convincing, although breastfeeding rates, while generally lower among disadvantaged women,4 are higher among women from some minority ethnic communities. Nevertheless, breastfeeding rates are encouraging and compare well with those for similar groups of women reported in UK national infant feeding surveys. 4 Doula activity is reported that includes direct support, signposting to other sources of support and obtaining specialist help for women in situations of difficulty. It would appear that the doulas’ non-directive approach may have been an enabling factor, as is their practical support for skin-to-skin contact between mothers and babies at birth. Doula-supported women at the original site were significantly more likely to hold their babies skin to skin at birth than the Picker survey comparison group. Peer support, generally provided on a voluntary basis, has been found to support the initiation of breastfeeding. 80 The impact of peer support is less clear for continuing breastfeeding. 81 A randomised trial of peer support (mother to mother) provided by telephone was acceptable and associated with significant increases in breastfeeding rates at 3 months postpartum, compared with standard support provided in the main by hospital-based health-care professionals. 82 In a trial where doula support in a non-UK setting was restricted to labour only, increases in breastfeeding at 6 weeks were hypothesised to be due to increased empowerment by the doula during labour that increased women’s sense of confidence and thus ability to breastfeed. 11

Costs to the NHS

We utilised a predominantly cost–consequence approach when evaluating the impact of doula support across key health and clinical outcomes. In the absence of a primary outcome, for example the quality-adjusted life-year, a fully aggregated incremental analysis was not possible. In this approach, we deliberately did not assign a priority to one health or clinical outcome compared with another, as we were aware that priorities and challenges may vary between geographical areas.

Our findings suggest some improved outcomes associated with doula support. It is noteworthy that these improvements are achieved through relatively small amounts (in numbers of hours) of volunteer support. These improvements have the potential to reduce NHS costs if doula services do not receive NHS funding. If, however, NHS funding into the original doula service is considered, then the net per birth cost to the NHS was estimated to be £1862. We do, however, have some reservations about this calculation. Historically, volunteer doula services have not always been NHS commissioned and funded, and the context of a third-sector agency is very different from being a part of mainstream NHS services. As our study has demonstrated, doula services have needed to expend considerable energy and staff time in securing their funding, often from multiple sources. Funding insecurity adds costs both directly through staff time and indirectly through inhibiting planning and the efficiencies that come with that. Furthermore, the figures presented are based on only one service, since the roll-out sites were not yet in a position to provide comparable data. A final caveat is that the costing of benefits to the NHS is necessarily based only on outcomes that can be measured in the short term. Such estimates exclude the positive impacts of psychosocial outcomes and other benefits to women that may persist beyond the maternity episode, and the wider societal benefits for women, families and doulas.

Doula services

NHS maternity services

Midwives need to be aware of their role as a key route of referral to the service and to be familiar with the criteria for referral to their local doula service. Where doula services are being considered for a locality, midwifery staff should accept opportunities for involvement from an early stage, including the opportunity to contribute to training.

Ensuring that communication systems are established to support the two-way flow of information is in the best interests of both doula and maternity services. These systems need to be continued, to cope with changes in personnel of both services.

While efforts to accommodate and integrate volunteer doulas within maternity services are generally to be encouraged, it is important that doulas are not utilised inappropriately to fill gaps in pressurised services. Neither does current evidence support the assimilation of volunteer doulas as (paid) members of the NHS workforce, as their independence is highly valued. The challenging experiences that volunteer breastfeeding counsellors encountered in a transition to their incorporation into statutory services offer important warnings in this regard. 83

MEDLINE and MEDLINE In-Process & Other Non-Indexed Citations (Ovid)

1948 to October week 4 2011.

Date of search: 8 November 2011.

Number of records found: 1347.

1. (doula or doulas).ti,ab. (133)

2. (labour attendant$ or labor attendant$).ti,ab. (6)

3. (labour companion$ or labor companion$).ti,ab. (10)

4. (labour assistant$ or labor assistant$).ti,ab. (2)

5. (birth companion$ or childbirth companion$).ti,ab. (10)

6. (birth assistant$ or childbirth assistant$).ti,ab. (15)

7. birth partner$.ti,ab. (11)

8. supportive companion$.ti,ab. (17)

9. monitrice$.ti,ab. (7)

10. childbirth support$.ti,ab. (5)

11. ((birth or birthing) adj1 support$).ti,ab. (46)

12. (labo?r adj1 support$).ti,ab. (109)

13. ((childbearing or child-bearing) adj1 support$).ti,ab. (7)

14. ((intrapartum or perinatal$) adj1 support$).ti,ab. (44)

15. ((prenatal$ or antenatal$) adj1 support$).ti,ab. (59)

16. (pregnan$ adj1 support$).ti,ab. (174)

17. (pregnan$ adj1 outreach).ti,ab. (5)

18. labo?r coach$.ti,ab. (12)

19. (((continuous or continual) adj1 support$) and (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (55)

20. (advocacy adj5 (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (38)

21. (befriend$ adj5 (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (1)

22. (human support adj5 (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (1)

23. (emotional$ support$ adj5 (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (32)

24. ((paraprofessional? or allied health personnel or allied health worker? or support worker? or home health aide?) adj5 (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (15)

25. (trained adj3 (volunteer? or health worker? or mother?) adj5 (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (21)

26. ((lay or voluntary or volunteer? or untrained or unlicensed or nonprofessional? or non professional?) adj5 (worker? or attendant? or aide or aides or support$ or person or people or helper? or carer? or caregiver? or care giver? or assistant?) adj5 (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (105)

27. maternity assistant?.ti,ab. (2)

28. ((peer or community) adj2 (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (330)

29. (((peer or community) adj1 support$) and (labo?r or birth or childbirth or prenatal$ or antenatal$ or perinatal$ or intrapartum or birthing or pregnan$)).ti,ab. (193)

30. community parent$.ti,ab. (46)

31. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 (1403)

[All terms relating to doulas]

1. exp animals/ not humans/ (3707268)

[Excludes animal-only studies]

1. 31 not 32 (1347)

[Final results set]