NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women's and Children's Health (UK). Intrapartum Care: Care of Healthy Women and Their Babies During Childbirth. London: National Institute for Health and Care Excellence (UK); 2014 Dec. (NICE Clinical Guidelines, No. 190.)

  • Update information February 2017: Sections that have been updated (see addendum files) have been marked with dark grey shading.

Update information February 2017: Sections that have been updated (see addendum files) have been marked with dark grey shading.

Cover of Intrapartum Care

Intrapartum Care: Care of Healthy Women and Their Babies During Childbirth.

Show details

4Care throughout labour

4.1. Communication between women and healthcare professionals

4.1.1. Introduction

Effective communication in all its forms is a fundamental aspect of today's maternity services. The overall aim of caring for women during labour and birth is to engender a positive experience for the woman and her family, while maintaining their physical and emotional health, preventing complications and responding to emergencies. To successfully achieve this aim, good communication between all those involved in the care of women during the process of childbearing is crucial. Developing a rapport, trust and effective communication between healthcare providers and women is important to a woman's positive childbirth experience. Other factors include involvement in decision-making, informed explanations and meeting personal expectations. All these elements have a powerful impact upon women and their childbirth experience. Their influence on whether the experience of a woman and her family is good or bad cannot be overestimated.

The views, beliefs and values of the woman, her partner and her family in relation to her care and that of her baby should be sought and respected at all times. Women should be fully involved so that care is flexible and tailored to meet her and her baby's individual needs. Women should have the opportunity to make informed decisions about every aspect of their labour and birth. Women sometimes decline the offer of interventions for numerous reasons including previous unpleasant experiences. Individualised care should be supported by giving evidence-based information and active informed consent should be sought from women before all monitoring procedures, examinations and treatments.

4.1.2. Review question

What effect does communication have on a woman's perception of her birth experience?

  • Interventions include the effect of control, choice and decision-making on psychosocial wellbeing in the medium and long term.
  • Outcomes include postnatal depression and post-traumatic stress disorder.

For further details on the evidence review protocol, please see appendix E.

4.1.3. Description of included studies

The search yielded 2615 titles, 182 of which were selected for retrieval. The search did not impose geographical limits, but papers were not included if it was felt that the cultural setting of the research would be unlikely to generalise to women in the UK. Papers were also rejected if they did not have information on caregiver behaviour linked to psychosocial outcomes for women. Within the remaining papers, 19 were selected as key, either because they were methodologically sound empirical studies specifically designed to address the link between caregiver behaviour and psychosocial outcomes for women (n = 18) or because they were reviews that highlighted this link (n = 1).6785

4.1.4. Review findings

A systematic review of 137 reports of factors influencing women's evaluation of their childbirth experiences was included.67 [EL = 3] The review identified four factors that were seen as key in shaping women's experience of labour: personal expectations; the amount of support from caregivers; the quality of the caregiver–patient relationship; and the involvement in decision making. It is concluded that the influences of pain, pain relief, and intrapartum interventions on subsequent satisfaction are important but not as powerful as the influences of the attitudes and behaviours of the caregivers.

A Swedish longitudinal cohort study of 2541 women measured women's global experience of labour and birth and obtained information on the possible risk factors during pregnancy and 2 months after birth.68 [EL = 2+] The following categories of risk factors were identified that were associated with women's experience of labour and birth:

  • factors related to unexpected medical problems
  • social factors
  • factors related to the woman's feelings during labour, such as pain and lack of control
  • factors that may be easier for caregivers to influence, such as lack of support in labour and administration of analgesia.

A UK prospective study sent questionnaires to women 1 month before the birth to assess their preferences and expectations, and at 6 weeks after birth to discover their experiences and assess psychological outcomes.69 [EL = 2+] Findings are based upon data from 1146 women. Parity was found to be strongly associated with feeling in control, with multiparous women feeling more in control than nulliparous women in all cases. In logistic regression analyses, the feeling of being in control associated with staff behaviour was found to relate primarily to being able to get comfortable, the feeling of being treated with respect and as an individual and perceiving staff to be considerate.

As part of a large randomised trial in the UK, which assessed the timing of intervention in prolonged labour, women's views were explored using a specifically designed questionnaire.70 [EL = 3] Analysis of findings from 412 nulliparous women in response to an open-ended question revealed the following main themes: support, information, intervention, decision making and control, and pain relief. One hundred and eight women said they wanted to participate in decision making but the degree of involvement varied among women.

Secondary analysis of questionnaire survey data, also collected during an RCT, was carried out to explore factors relating to women's experience of birth. Data were collected from women receiving either care in an alongside midwife-led unit or standard hospital care.71 [EL = 3] The two groups were combined for the purposes of this analysis (n = 1111). Logistic regression analysis identified five explanatory variables: involvement in the birth process (perceived control) and midwifery support were predictive of a positive experience; anxiety, pain and having a first baby were predictive of a negative experience.

Findings from a questionnaire survey (Sweden) distributed to women 1 day after giving birth (n = 295; response rate = 91%) showed that women usually experienced severe pain and various degrees of anxiety, and most were seized with panic for a short time or for some part of their labour.72 [EL = 3] Despite these negative feelings, most women felt greatly involved in the birth process, were satisfied with their own achievement and thought they had coped better than expected. Of the 38 variables tested in regression analysis, the six that contributed to explaining women's overall birth experience were: support from the midwife (sensitivity to needs); duration of labour; pain; expectations of the birth; involvement and participation in the birth process; and surgical procedures (emergency caesarean section, vacuum extraction, forceps, episiotomy).

Another questionnaire survey was sent to women 8–9 months after they had given birth (Australia) (n = 790; response rate = 71%).73 [EL = 3] Findings revealed that not having an active say in decisions was associated with a six-fold increase in dissatisfaction among nulliparous women and a 15-fold increase among multiparous women. When adjusted for parity in a logistic regression model, the following factors were highly related to dissatisfaction with intrapartum care: lack of involvement in decision making (P < 0.001); insufficient information (P < 0.001); a higher score for obstetric interventions (P < 0.015); and the perception that caregivers were unhelpful (P < 0.04).

A second Australian cross-sectional questionnaire survey returned by 1336 women (response rate = 62.5%) 6–7 months after they had given birth found that, after adjusting for parity, social factors and obstetric care, caregivers perceived as unhelpful and not having an active say in decisions about their care had the greatest impact on women's experience of birth.74 [EL = 3]

A third Australian prospective descriptive study employed telephone interviews conducted 4–6 weeks after birth to investigate women's experiences (n = 499 women).75 [EL = 3] One in three women identified a traumatic birthing event and reported the presence of at least three trauma signs. Twenty-eight women (5.6%) met DSM-IV criteria for acute post-traumatic stress disorder. The level of obstetric intervention experienced during childbirth together with the perception of inadequate intrapartum care during labour was consistently associated with the development of acute trauma symptoms.

A questionnaire survey of first-time mothers in Finland (n = 271; response rate = 83%) investigated women's perceptions of labour and birth.76 [EL = 3] Regression analysis showed that positive childbirth experiences were associated with the positive characteristics and professional skills of the attending midwife, the positive attitude of the child's father towards the pregnancy and a short labour.

In the USA (early 1990s), there was a convenience sample of 15 women (eight first-time mothers) who told 33 birth stories.77 [EL = 3] From the findings, the researchers concluded that when decision making was increasingly shared between the women and the caregivers, the women expressed more positive emotions. Professional knowledge and power needs to be supportive, not directive, of the birthing processes.

A Swedish qualitative study using interviews with 18 women (six primiparous) who were 2–4 days post birth investigated women's experiences of labour and birth. The study took place in Sweden in 1994.78 [EL = 3] Three main themes emerged: the need to be seen as an individual; to have a trusting relationship; and to be supported and guided on one's own terms. These themes were associated with a positive birth experience.

Another small-scale (n = 14) interview-based study conducted in Iceland also explored women's experience of giving birth.79 [EL = 3] Analysis of the data showed that women have a need for a sense of control as well as a need for caring and understanding. Additionally there was a need for a good relationship with the midwife, which included the women feeling safe and secure. An explanation of events and reassurance regarding progress were also important to women.

A second Icelandic qualitative study sought views and experiences from a purposive sample of ten women who had experienced both caring and uncaring encounters during childbirth in Iceland.80 [EL = 3] The authors summarised three traits of the caring midwife which were defined as follows:

  • competence – has the necessary knowledge and skills needed to coach a woman through the journey of labour and giving birth; is responsible, attentive, deliberate and communicates effectively
  • genuine concern and respect for the woman – gives of her or himself, shows solidarity and sharing, is encouraging and supportive, respectful and benevolent
  • positive mental attitude – is cheerful and positive, reliable and trustworthy, considerate and understanding.

Similarly the authors summarised three traits of the uncaring midwife:

  • lack of competence – being rough when giving care to women, ineffective communication, not taking the initiative when needed and lack of understanding and flexibility
  • lack of genuine concern and respect for the woman as a person – being thoughtless, strict on routines and rules, not taking notice of the woman and lacking in cooperation; being indifferent and untouched by the event as such, lack of interest and understanding in general, being non-supportive and insensitive, being hurried and in a rush
  • negative character traits – being gloomy and brusque, cold, unkind or harsh.

An interesting US study showed a sample of 20 women videotapes of their births while simultaneously interviewing them.81 [EL = 3] In separate interviews, the 25 caregivers were also shown the videotapes and interviewed. Although women and caregivers appeared to agree about what information women required and how it should be given, caregiver perceptions were more positive than those of the women. Many women wanted more information and valued detailed information to explain what was happening.

A discussion paper based on a previous paper82 puts forward an idea that women are of less interest to the caregivers than the equipment, and that lack of information disempowers women. [EL = 3] Caregivers were seen to block women's worries or concerns by silence, changing the subject or by neutral statements such as ‘let's see how we go’.

Participant observation of a convenience sample of 12 primiparous women in the second stage of labour examined communication between midwives, student midwives, labouring women and their partners, by analysing videotaped recordings.83 [EL = 3] Communication was categorised using one of the following: innovation, encouragement, directing, educating, questioning, social and professional. Findings revealed that most communication was categorised as being directing, encouraging or educational, with the latter two categories showing a degree of overlap. Midwives were found to fall into one of two groups: those that tend to be directing or those that tend to be encouraging and educating. Women preferred the latter type of communication.

The Caring Behaviour Assessment tool has been used on a convenience sample of 31 women following normal birth (USA) to look at women's perceptions of caring behaviour from nurses during childbirth.84 [EL = 3] Findings showed that the behaviours perceived by women to be most indicative of caring focused on professional competence and monitoring of the woman's condition. The most caring behaviours included knowing what they were doing, treating the woman with respect and as an individual, being kind and considerate and reassuring the woman.

A cross-cultural qualitative study compared responses from semi-structured interviews conducted with ten Chinese women and ten Scottish women (giving birth in Scotland).85 [EL = 3] In addition, 45 unstructured interviews were undertaken with health workers, relatives and friends. Responses to the birth experience were partly related to the woman's culture, with Chinese women being more accepting of care given, but there were issues that were common across all the women irrespective of cultural background, notably that the feeling of being in control was linked to a better emotional outcome. Caregivers' failure to engage with the woman as a human being was experienced as very traumatic.

4.1.5. Evidence statement

The studies included in this review varied in the methodology that they used as well as the method of analysis undertaken. Nevertheless, a number of strong common themes emerge and it is apparent that the way caregivers relate with the labouring women is hugely influential upon the woman's experience of birth. The first theme highlights that women value being treated as an individual, with respect and care. Secondly, most women need information and interpretation of that information in order to feel guided and supported throughout the birth.

These findings are usefully summarised by the words women use to describe both the midwife and the feelings involved in a positive birth experience. These words include: caring, considerate, understanding, competent, trustworthy, empathic, tender, kind, friendly, calm, alert, peaceful, having professional expertise, unhurried.

Women want to receive information and assistance, to be involved, to feel safe and secure, to feel at ease and to be able to be themselves.

4.1.6. Recommendations on communication


Treat all women in labour with respect. Ensure that the woman is in control of and involved in what is happening to her, and recognise that the way in which care is given is key to this. To facilitate this, establish a rapport with the woman, ask her about her wants and expectations for labour, and be aware of the importance of tone and demeanour, and of the actual words used. Use this information to support and guide her through her labour. [2007]


To establish communication with the woman:

  • Greet the woman with a smile and a personal welcome, establish her language needs, introduce yourself and explain your role in her care.
  • Maintain a calm and confident approach so that your demeanour reassures the woman that all is going well.
  • Knock and wait before entering the woman's room, respecting it as her personal space, and ask others to do the same.
  • Ask how the woman is feeling and whether there is anything in particular she is worried about.
  • If the woman has a written birth plan, read and discuss it with her.
  • Assess the woman's knowledge of strategies for coping with pain and provide balanced information to find out which available approaches are acceptable to her.
  • Encourage the woman to adapt the environment to meet her individual needs.
  • Ask her permission before all procedures and observations, focusing on the woman rather than the technology or the documentation.
  • Show the woman and her birth companion(s) how to summon help and reassure her that she may do so whenever and as often as she needs to. When leaving the room, let her know when you will return.
  • Involve the woman in any handover of care to another professional, either when additional expertise has been brought in or at the end of a shift. [2007]

4.2. Mobilisation and position

4.2.1. Review question

What is the effectiveness of the following interventions or techniques in labour on outcomes?

  • mobilisation
  • positions including: ‘freedom to choose’ option; standing; squatting; kneeling; semi-recumbent; lying on back; left lateral; birth stool, etc.

4.2.2. Previous guideline

Mobilisation during labour was reviewed in the Caesarean Section guideline.6 Two RCTs were included. The guideline recommended that women should be informed that walking during labour has not been shown to influence the likelihood of CS.

4.2.3. Description of included studies

Evidence for the effect of different positions and mobilisation during the first stage of labour on labour outcomes is drawn from one systematic review of RCTs86 and five RCTs.8791

4.2.4. Review findings

A systematic review of maternal positions during the first stage of labour was identified which included 14 RCTs (seven of which used women as their own controls).86 [EL = 1−] Most trials where women acted as their own controls were small-scale (n = 23 or fewer in six of the trials). In the other trials, sample sizes ranged from 40 to 1067, with four of the trials involving over 200 women. The trials of positioning during the first stage of labour compared mobilisation or upright positions with one or more horizontal positions in bed. Outcome measures included pain, comfort, uterine activity and labour progress. In trials where women acted as their own controls, they were requested to alternate between two different positions (e.g. standing, walking or sitting up versus side-lying or supine) during labour for equal periods (usually 30 minutes). Measures were made after each period of reported location and intensity of pain, uterine activity and labour progress. Other trials assigned women to an upright group or a recumbent group for a longer period of time e.g. active first stage, the whole of the first stage or the duration of labour. The differences in study design, the lack of detail in most papers regarding measures taken to prevent bias, difficulties of compliance and different pain assessment methods undermine the reliability of the findings and prevent pooling of data. The one consistent finding was that none of the women in any of the studies reported greater comfort in the supine position. In addition, it was found that alternating between different pairs of positions has different effects on uterine efficiency. Alternating between supine and sitting seems to reduce the efficiency of uterine activity compared with alternating between supine and standing or side-lying. It was also noted that many women had difficulty in remaining upright and/or mobilising during labour, especially towards the end of the first stage of labour and during the second stage. No conclusion could be drawn regarding the effects of position and mobilisation on reported pain or duration of labour.

A fairly large US randomised trial compared walking in the first stage of labour (n = 536) with no walking (usual care) (n = 531).87 [EL = 1+] Women in spontaneous labour following uncomplicated pregnancies were randomised once labour had been established (cervical dilatation of 3–5 cm). Neither group underwent continuous electronic fetal monitoring (EFM) unless a fetal heart rate abnormality was detected using intermittent monitoring, epidural anaesthesia was requested or oxytocin augmentation was required. This then excluded any further ambulation. The amount of time spent walking undertaken by both groups of women was recorded by the attending nurse, and the distance walked was recorded using a pedometer (how the use of this instrument may have impacted upon the comfort of the labouring women is not discussed). Of the women assigned to the walking group, 22% chose not to walk. Of the 420 women who actually walked during labour, the mean walking time was 56 minutes (SD = 46 minutes). The degree of ambulation in the non-walking group was minimal. There were no significant differences between the characteristics of women in the two trial groups. Analysis was on an intention-to-treat basis. No significant differences were found between the two groups in terms of labour outcomes (e.g. length, use of oxytocin for augmentation, use of analgesia), mode of giving birth, maternal or neonatal outcomes. Of those women who walked during labour, 278 were asked if they would do so in a future labour: 99% said that they would.

A prospective Australian RCT was carried out to determine whether there was any advantage or disadvantage to giving women the option to ambulate during labour compared with labouring in the recumbent position.88 [EL = 1+] All women entering the trial (n = 196) underwent continuous EFM using a scalp electrode. This was carried out via telemetry for women in the ambulant group. The demographic and obstetric characteristics of the two trial groups were similar. Analysis was carried out on an intention-to-treat basis. No significant differences were found between the groups in terms of labour outcomes, mode of giving birth, maternal or neonatal outcome. Only 37 of the 96 women allocated to the ambulant group (39%) actually chose to ambulate for 30 minutes or longer. Of those who did ambulate, the mean time spent in an upright position was 1.5 hours (SD 0.8 hours). During the time of recruitment of women into the trial, 389 declined to participate, 46% for fear of losing the option to ambulate during labour.

In a small, older, UK prospective RCT, 68 women in spontaneous labour were allocated to either an ambulant or recumbent group for the first stage of labour.89 [EL = 1−] Trial participants were recruited from a group of women who had expressed antenatally a desire to be ambulant. Each group comprised 17 nulliparous women and 17 parous women. Continuous EFM was performed for all women with the use of a fetal scalp electrode (via telemetry for the ambulant group) and contractions were monitored using an intrauterine pressure catheter. A number of significant differences were noted between the two groups, all in favour of the ambulant group. Ambulant women were given less analgesia, contractions were less frequent and were of greater amplitude, duration of labour was shorter, there were more normal births and babies' Apgar scores were also higher in the ambulant group. For women in the ambulant group, the mean time spent mobilising was 2.2 hours [range 0.8 to 8.3 hours]. The selection bias inherent in this study needs to be taken into consideration when interpreting these findings.

An RCT conducted in Argentina compared the pain perceptions of two groups of 50 women allocated to adopt alternately a vertical (sitting, standing or walking) or horizontal (lie on side or back) position for periods of 15 minutes throughout the first stage of labour.90 [EL = 1+] Each woman thus acted as her own control and was asked to adopt a position of her own choosing between the assigned position periods in order to reduce ‘carry-over’ effects. The participants were all staff connected with the public education sector. Pain levels were measured during each 15 minutes horizontal or vertical position period using two validated pain scales (a Likert-type scale and a 10 cm visual analogue scale (VAS)). Pain scores were reported for each dilatation interval (2–3 cm, 4–5 cm, 6–7 cm and 8–9 cm). During the first half of the first stage (i.e. from 2 to 5 cm cervical dilatation) there was no difference noted in reported pain between the two positions. As labour progressed however, there was a statistically significant difference noted in measured pain levels, both abdominal contraction pain and lumbar pain, with higher levels of pain being associated with horizontal positions.

A small US trial randomly allocated nulliparous women in spontaneous labour to upright (n = 20) or recumbent (n = 20) groups.91 [EL = 1+] The recumbent group included the options of supine, lateral or all fours. The upright group included standing, walking, kneeling, sitting or squatting. Outcome measures included the duration of the active phase of labour (defined as 4–9 cm dilatation), uterine contraction pattern and maternal comfort (as measured by a researcher using a standardised tool). Women allocated to the upright group had a significantly shorter active phase of labour (mean difference 90.25 minutes, P = 0.003) and had contractions that were longer lasting and more frequent than women in the recumbent group. There was no significant difference in reports of women's physical comfort.

4.2.5. Evidence statement

Surprisingly, there are no trials examining the effect of freedom of movement throughout labour compared with restriction of movement on outcomes such as comfort, labour progress and fetal wellbeing. There is a lack of high-level evidence to suggest that either mobilisation or any particular position in the first stage of labour affects outcomes.

4.2.6. Recommendation on mobilisation and position


Encourage and help the woman to move and adopt whatever positions she finds most comfortable throughout labour. [2007]

4.3. Support in labour

4.3.1. Review question

Is there evidence that support in labour for women improves outcomes? Interventions include:

  • any support from partners
  • other birth supporters
  • health professionals
  • continuity of care

4.3.2. One-to-one care Introduction

Traditionally, women have been attended and supported by other women during labour and birth. However, with the movement of the majority of births from home to hospital since the middle of the 20th century, continuous support has become the exception rather than standard care. Women's support needs in labour have been shown to have four dimensions: emotional, information support, physical support and advocacy. Women in the UK today usually labour with their partners present, providing them with physical and emotional commitment, but for some women this may be insufficient to provide them with the level and type of support that they need in the context of a modern institutional birth environment. Previous guideline

One-to-one care is defined as continuous presence and support either by husband/partners, midwives or other birth supporters during labour and childbirth. One-to-one care was reviewed in the NICE Caesarean Section guideline.6 The guideline reviewed one systematic review and recommended that women should be informed that continuous support during labour from women with or without training reduces the likelihood of CS. Description of included studies

The updated systematic review was identified during the search for this guideline.92 The systematic review examined 15 trials including 12791 women in both high income and low income countries (Australia, Belgium, Botswana, Canada, Finland, France, Greece, Guatemala, Mexico, South Africa and the USA). The impact of one-to-one care was considered different by status of caregivers, so that the review was stratified by the care providers. In eight trials, the support was provided by a member of the hospital staff, e.g. a midwife, student midwife or nurse. In the remaining seven trials, the providers were not members of the hospital staff; they were women with or without special training, a childbirth educator, a retired nurse, or a close female relative, usually the woman's mother. There is no identified trial that investigated the effectiveness of continuous support by husbands or partners. In nine of the trials, hospital policy permitted women to be accompanied by their husbands/partners or other family members during labour, while in the other six trials, no additional support people were allowed. Presence of husbands or partners was considered as usual practice in the UK. [EL = 1+] Review findings Labour events
  1. Stratified analysis by care-providers
    Women supported by a member of the hospital staff were less likely to have analgesia than women receiving standard care (RR 0.97 [95% CI 0.95 to 0.99]). This difference was also apparent if the support was provided by birth attendants other than professionally trained staff (RR 0.83 [95% CI 0.77 to 0.89]).
  2. Meta-analysis of all trials
    Meta-analysis of findings from nine trials without stratification, which included 10 322 women, showed no significant difference in length of labour (WMD (random) −0.28 hours [95% CI −0.64 to 0.08 hours]). Birth events
  1. Stratified analysis by care-providers
    Women supported by a hospital staff member were more likely to have a spontaneous vaginal birth (RR 1.03 [95% CI 1.01 to 1.06]), less likely to have an instrumental vaginal birth (RR 0.92 [95% CI 0.85 to 0.99]) or caesarean section (CS) birth (RR 0.92 [95% CI 0.85 to 0.99]). If support was given by non-hospital staff, the positive impact on spontaneous vaginal birth, instrumental vaginal birth and caesarean birth remained, with RR of 1.12 [95% CI 1.07 to 1.18], 0.59 [95% CI 0.42 to 0.81] and 0.74 [95% CI 0.61 to 0.90], respectively.
  2. Meta-analysis of all trials
    There appeared to be no difference in the rates of perineal trauma. One trial, which investigated the rate of episiotomy when support was provided from a specially trained nurse, found no significant difference between supported women versus those with standard care (RR 0.97 [95% CI 0.90 to 1.05]). Meta-analysis of two trials, both of which investigated support by a member of hospital staff, showed no significant difference in perineal trauma (RR 0.99 [95% CI 0.95 to 1.03]). Newborn events

Meta-analysis of trials showed no significant difference in low 5 minute Apgar scores (seven trials, total RR 0.81 [95% CI 0.56 to 1.16]; with support by a member of hospital staff RR 0.83 [95% CI 0.56 to 1.22] and with support by non-hospital staff RR 0.64 [95% CI 0.22 to 1.92]); and admission to neonatal units (four trials RR 0.94 [95% CI 0.82 to 1.09]). Women's satisfaction and experience of childbirth

Meta-analysis of eight trials showed that there was no significant difference in dissatisfaction and negative experience of childbirth between women supported by a hospital staff member (RR 0.83 [95% CI 0.67 to 1.02]) and women receiving standard care, but there was a significant difference if support was provided by a non-hospital staff member (RR 0.64 [95% CI 0.58 to 0.78]). Women's mental and psychological health

There was one trial that investigated the incidence of postpartum depression in women given support by a specially trained nurse.93 There were fewer supported women who reported post-partum depression than those receiving standard care, but this difference was not statistically significant (RR 0.89 [95% CI 0.75 to 1.05]). Another trial investigated the impact of postpartum self-esteem on women given support by a retired nurse.94 There was no evidence of a difference in the number of women with low postpartum esteem, between supported care and standard care (RR 1.07 [95% CI 0.82 to 1.40]). Long-term outcomes

One trial investigated the long-term outcomes of support by a specially trained nurse for women in labour. There were no significant differences between the trial groups for poor relationship with partner postpartum (RR 1.00 [95% CI 0.80 to 1.23]), postpartum urinary incontinence (RR 0.93 [95% CI 0.81 to 1.06]) or postpartum faecal incontinence (RR 0.89 [95% CI 0.64 to 1.24]). Evidence statement

In general, the included studies were of good quality. A range of professionals providing one-to-one care, including obstetric nurses, was identified within the studies. There is evidence to suggest that women with one-to-one care throughout their labour are significantly less likely to have caesarean section or instrumental vaginal birth, will be more satisfied and will have a positive experience of childbirth. This impact becomes more apparent when non-professional staff members, rather than professional staff members, care for them. The non-professional person providing one-to-one care in labour within these studies varied in their level of training, background and in the context of care.

There is little evidence on perinatal mortality and the long-term wellbeing of women and their children.

There is also a lack of high-level evidence to suggest that support by partners, other family members or friends affects clinical outcomes. GDG interpretation of evidence

Although in the UK midwives usually provide the majority of care during labour and childbirth, there were no studies identified that compared one-to-one support from a midwife with that provided by another professional. The reviewed studies are from a range of countries, some of which are not representative of the UK setting, especially in that partners/support persons were not usually allowed to accompany women during labour. This means it is not possible to extrapolate all these findings regarding support from a non-professional person to the UK. The role of maternity care support workers remains unevaluated in the UK. Recommendations on one-to-one care


Provide a woman in established labour with supportive one-to-one care. [2007]


Do not leave a woman in established labour on her own except for short periods or at the woman's request. [2007]


Encourage the woman to have support from birth companion(s) of her choice. [2007] Research recommendations on one-to-one care


Studies should evaluate the impact of a standardised training programme for maternity care support workers in the intrapartum period. Outcomes should include: maternal and neonatal mortality, adverse outcomes, long-term outcomes, women's satisfaction and costs as outcomes.

4.3.3. Appropriate staffing configuration of midwives and healthcare support staff on labour wards Review question

What is the appropriate staffing configuration of midwives and healthcare support staff on labour wards to support one-to-one continuous care during labour? Description of included studies

There is little evidence linking midwifery and healthcare support staffing levels to provision of one-to-one continuous care during labour. In particular, no evidence was found evaluating the staffing configuration of healthcare support staff (doulas, birth attendants, healthcare support workers, healthcare assistants, obstetric nurses and auxiliary nurses) on labour ward or in midwifery-led units. Although much work has been done to develop scoring systems for quality appraisal of a range of study designs, the NICE interim methods guide for developing service guidance (February 2013) does not contain guidance for quality assessing tool development and validation. Therefore, for the purposes of this review, a pragmatic approach was taken based on the quality assessment scoring outlined for surveys. This meant that reports of tool development were rated as very low.

Two descriptive reports (Ball, 1992; Ball and Washbrook, 1996) and two studies (Ball, 2010; Allen and Thornton, 2012) are included in this review. All included studies assessed the configuration of midwives based in UK hospitals. Two additional UK surveys of midwifery staffing levels are presented as supporting information (Ball et al., 2003; BBC Panorama survey, 2011).

Ball (1992; based on Ball, 1988) introduced a method to assess workload in the UK's maternity units known as Birthrate. Birthrate is a model for calculating midwifery staffing levels based on a retrospective assessment of events and factors arising during labour and birth. It was developed using evidence from previous nursing and midwifery workforce planning studies which used non-participant observation, qualitative methods and expert consensus to estimate the time taken to provide care to women across a range of dependency levels. Its objective was to provide a tool that could be used to calculate midwifery staffing levels needed to provide continuous one-to-one care during labour (chosen as an agreed quality standard representing high quality care). This was further developed to Birthrate Plus (Ball and Washbrook, 1996). Birthrate Plus provided a method to support workforce planning and strategic decision-making in maternity services.

Birthrate and Birthrate Plus have been approved by the Royal College of Midwives (1994, 1999) and the Royal College of Obstetricians and Gynaecologists (1994) as a planning tool to predict the number of midwives required on labour wards in order for women to receive continuous one-to-one care during labour.

Birthrate Plus has 3 main components: a scoring system; actual midwife time (based on a staff survey); and a staffing formula. The scoring system uses 5 categories of labour and birth complexity which reflect the process and outcome of birth for both the woman and the baby. The midwife time required for each category is recorded based on the provision of one-to-one care throughout labour. The amount of midwife time increases for more complicated cases (that is, those in higher categories). All staffing calculations take into consideration variability of work load, holiday, sickness and study leave. Then, using a staffing formula, this data are converted into the number of midwives required to meet the workload.

Both Birthrate and Birthrate Plus have been well described and are evidence based. However, although the tools have been developed in a robust way and have face validity, they have not been externally validated. For this reason they have been graded as very low quality in relation to the key outcome for this review, namely the provision of one-to-one midwifery care throughout labour.

Birthrate Plus methodology was further developed to prospectively measure the ‘acuity’ workload on a labour ward (Ball and Washbrook, 2010). Acuity is defined as “a measure of the intensity of need within a delivery suite” (Ball and Washbrook, 2010). This was developed to enable managers to assess and record fluctuating workload with midwife availability in ‘real time’. The tool enables the Birthrate Plus classification system to become a predictive/prospective tool rather than the retrospective assessment of process and outcome of labour used in the standard Birthrate Plus.

Allen & Thornton (2012) examined the ability of midwifery staffing levels calculated using Birthrate Plus to enable the provision of one-to-one care during labour. Variation in workload in maternity units and how much reserve resource was needed in times of high workload was also assessed. Computer simulation of labour wards was used to investigate the expected effect of changing midwife numbers throughout the day and across the week, and also was used to see how the size of unit affects the unit's ability to cope with different levels of complexity and workload. This validation study is graded as being of low quality. Evidence profile

Due to the descriptive nature of the evidence for appropriate staffing configuration of midwives on a labour ward to support one-to-one continuous care during labour, the findings are presented in a simple summary table below. It should be noted that although they are based on evidence, Birthrate and Birthrate Plus do not in themselves constitute evidence but rather a method of calculating midwifery staffing needs in order to provide one-to-one care in labour.

Table 43. Appropriate staffing configuration of midwives on labour ward to support one-to-one care.

Table 43

Appropriate staffing configuration of midwives on labour ward to support one-to-one care. Supporting information

In Ball (2003) an outline of the Birthrate Plus programme during 2001–2002 in England and Wales was reported. The project was funded by Department of Health to help maternity units in England and Wales to implement Birthrate Plus and it required units to complete the Birthrate Plus data collection tool. Results from 44 maternity units in England and Wales were reported, representing a variety of unit sizes and locations. All units collected a minimum of 6 months of data.

Further, a Freedom of Information request by the BBC's Panorama television programme (2011) requested midwife-to-birth ratios from 171 trusts in the UK. Data were submitted by 149 trusts. The number of births in 2010 and the whole time equivalent midwives in post on 1 May 2011 was reported for each trust in this survey.

Both sources provide data for number of births per midwife ratios and are presented in the summary table below:

Table 44. Number of births per midwife ratios reported by maternity units in the UK.

Table 44

Number of births per midwife ratios reported by maternity units in the UK. Evidence statements

There is very little evidence to inform the staffing levels needed to provide one-to-one continuous care in labour.

Birthrate and Birthrate Plus were developed from evidence and field tested for usability and reliability, but they have not been validated in terms of whether midwifery staffing levels calculated using their formula results in midwifery staffing levels that allow provision of one-to-one continuous care during labour. The evidence was of very low quality.

The Birthrate Plus Acuity tool has been developed to allow real time assessment, prediction and planning of midwifery staffing workload on a labour ward taking into account variations in workload across time of day and days of the week. In pilot testing this tool has been found to be useable and reliable but its validity has not been established. The evidence was of very low quality. The Birthrate Plus formula for calculating number of midwives allows for 15% extra resource (above the average Workload Index) for coping with variation in workload, but 1 computer modelling study indicated that when it was applied in practice, the Workload Index exceeded planned resource 36% of the time and the number of women exceeded the number of midwives 13% of the time. The simulation model used in the study suggested that overload can be reduced by 15–25% if the resource were matched to known pattern of workload, taking into account variations between weekdays and weekends and allowing for peak hours. It also suggested smaller units are more prone to work overload than larger units. The evidence was of low quality. Health economic profile

No published economic evaluations were identified for this question. Evidence to recommendations Relative value placed on the outcomes considered

It is recommended that all women should receive supportive one-to-one care during labour (recommendation 23). The intention of this review was to build on that recommendation and to identify whether there are particular staffing models which can be applied to enable this provision. Receipt of one-to-one care was therefore the key outcome for this review. Unfortunately, no comparative studies were identified which evaluated the effectiveness of any particular staffing model, although a computer modelling study did attempt to evaluate the ability of Birthrate Plus to provide appropriate staffing levels to allow one-to-one care for all women on the delivery suite. Three other reports that were included in the review described the development and usability of 3 tools for calculating midwifery staffing levels (Birthrate, Birthrate Plus and the Birthrate Plus Acuity Tool). Consideration of benefits and harms

Because of the nature of the available evidence, it was not possible for the group to adequately consider the benefits and harms of the different staffing models. No outcome data were provided and so it was not possible to determine if use of the models in practice would lead to an improved level of one-to-one care. Similarly, it was not possible to determine if there were any positive or negative outcomes associated with changing midwifery staffing levels. Consideration of health benefits and resource uses

Again, owing to the lack of data available from the studies, the guideline development group did not feel that it was possible to determine the cost effectiveness of the different staffing models. They recognised that even if the different models achieved their aim of ensuring that a greater proportion of women were able to receive one-to-one care, without proper evaluation and validation it was not possible to determine whether the models would achieve this in a cost-effective way. None of the included papers evaluated comparisons between different models for service provision. So, for example, the group noted that Birthrate and Birthrate Plus calculate midwifery staffing levels based upon the underlying assumption that support staff, such as healthcare assistants, are also available on the labour ward or midwifery-led unit. The tools do not include calculations for the number of support staff required. Potentially, staffing models using different configurations of midwives and other support staff might be more cost effective than that proposed by Birthrate or Birthrate Plus, given the difference in staff costs. The group did not necessarily believe that such a model would be more effective, but without any formal evaluation of the models, and without any comparative data available, the guideline development group did not feel that it was possible to advocate any particular staffing model or staffing calculation tool over another. Quality of evidence

Although the data provided in the studies were clear regarding the development and usability of the tools, they did not provide good quality evidence to answer the review question, all being graded as very low. Without formal validation being conducted, it was not possible to determine if the tools are able to achieve their stated aims of calculating appropriate midwifery staffing levels for provision of one-to-one care. Other considerations

Although the guideline development group did not feel that they could advocate a particular calculation tool or staffing model for providing one-to-one care, they agreed that this should be an aim for all maternity services. They agreed that it was appropriate to use workforce planning tools such as Birthrate Plus as this would then allow benchmarking, so would encourage services to identify the level of care they are currently providing and then assess the difference made by adopting the new staffing model. They agreed that these data would be useful both for the individual unit in identifying the degree to which they were achieving the provision of one-to-one care during labour and their levels of under- and over-staffing, and also at a national level to determine the effectiveness of the workforce planning models used. The group also drafted a research recommendation to ensure that currently available workforce planning tools are evaluated and validated.

The group noted that the BBC Panorama findings showed a large variation in practice across the UK in terms of the ratio of births to midwives, with both England and Wales above the 28 births to 1 midwife ratio that the Royal College of Midwives recommends. However, the group was not aware of any published evidence which shows that this particular ratio allows for one-to-one care. Given this, they did not feel that it was appropriate to refer to this proposed ratio in the recommendations. Recommendations


Maternity services should

  • provide a model of care that supports one-to-one care in labour for all women and
  • benchmark services and identify overstaffing or understaffing by using workforce planning models and/or woman-to-midwife ratios. [new 2014]

4.3.4. Continuity of care Introduction

Continuity of care in maternity services refers to both continuity of carer and consistency of care. The former has received most attention both in terms of policy and in research where continuity of care is defined in terms of continuity of carer and describes care provided by a midwife or a small group of midwives, from early pregnancy to the postnatal period. Continuity of carer was highlighted as a key component of good maternity care in the Health Committee Second Report: Maternity Services, vol. 1 (1992) (the Winterton Report),95 and further endorsed by the Report of the Expert Maternity Group at the Department of Health (the Changing Childbirth Report) (1993),96 which identified among its ten key indicators of success (page 70) that:

  • every woman should know one midwife who ensures continuity of her midwifery care – the named midwife
  • every woman should know the lead professional who has a key role in the planning and provision of her care
  • at least 75% of women should know the person who cares for them during their birth.

Two main models of midwifery care have evolved as a way of organising services so as to provide continuity of carer in a way that is sustainable within the existing NHS structure, namely team midwifery and caseload midwifery. Team midwifery is a team of midwives looking after a group of women and caseload midwifery aims for a more personal relationship with the woman and involves a small group of midwives. Sizes of team midwifery teams vary greatly, ranging from four midwives to ten or more, with hospital-based teams tending to be larger than community-based teams. The aim of most team midwifery schemes is to increase the chance that women will be cared for in labour by a midwife they have met antenatally, with the focus on intrapartum continuity often taking precedence over antenatal and postnatal continuity. Caseload midwifery describes a system of care whereby one midwife (sometimes referred to as the ‘named midwife’) is responsible, and provides the majority of the care, for a group of women backed up by a small group of associate midwives (usually two or three). When there is one midwife backing up a named midwife this system is also known as ‘one-to-one’ care.97 Team midwifery schemes have usually been hospital based, or integrated across hospital and community settings. Caseload midwifery schemes tend to be community based. These two systems of care will be reviewed separately below. Some studies investigated a package of care which included both care in midwife-led units and continuity of care. This review includes schemes which provide care in a variety of settings, including traditional delivery suite, birthing rooms within a traditional midwifery suite and separate birth units. For the purposes of this review where one midwife has taken responsibility for a group of women this has been categorised as caseload midwifery. Where there has been shared responsibility between a group of midwives this has been categorised as team midwifery.

While much research confirmed that continuity of carer was highly valued by many women, concern has been raised about the effects on midwives of working in systems designed to provide continuity of care, particularly hospital-based team midwifery schemes.98 Previous guideline

Continuity of care was reviewed in the NICE ‘Antenatal Care’ clinical guideline.99 Two systematic reviews were appraised in the guideline. It was recommended that antenatal care should be provided by a small group of carers with whom the woman feels comfortable and there should be continuity of care throughout the antenatal period. They also recommended that a system of clear referral paths should be established so that pregnant women who require additional care are managed and treated by the appropriate specialist teams when problems are identified. Team midwifery Description of included studies

There were two systematic reviews100,101 and four RCTs102108 identified. One systematic review included two trials,100 and another included seven trials.101 The trials that were included in the former systematic review were all included in the latter systematic review. A new meta-analysis was conducted by using a total of ten trials.100108 [EL = 1+]

Among the ten trials, three were conducted in England, five in Australia, one in Canada, and one in Sweden. A total of 1229 women were involved. The ten trials were all evaluations of team midwifery, with teams ranging in size from four to ten midwives. Six of the ten studies were of community-based teams coming into the hospital or midwife-led unit to provide care during labour and the postnatal period. The review here relates to team midwifery rather than continuity of carer per se.

A cross-sectional study98,109 with a 5% random sample of midwives in England (n = 1166) measured occupational stress, especially burnout, in midwives, comparing those in midwifery teams (hospital-based and community-based) with traditional hospital-based midwives and GP-attached midwives. Review findings

Details of the included trials on team midwifery care are summarised in Table 45.

Table 45. Details of trials of team midwifery care.

Table 45

Details of trials of team midwifery care. Labour events

It was not possible to conduct a meta-analysis on the length of labour owing to the different measures used. There were no consistent findings in duration of either the first, second or third stage of labour. Meta-analysis was conducted for interventions related to labour as follows. Induction (nine trials, n = 10 341): RR 0.96 [95% CI 0.88 to 1.05] (test for heterogeneity P = 0.11); augmentation (nine trials, n = 10 201): RR 0.83 [95% CI 0.78 to 0.90] (test for heterogeneity P < 0.0001); epidural (ten trials, n = 10 399): RR 0.80 [95% CI 0.74 to 0.86] (test for heterogeneity P = 0.04); opioid analgesia (nine trials, n = 10 146): RR 0.75 [95% CI 0.75 to 0.84], P < 0.00001 (test for heterogeneity P < 0.00001). Overall, women receiving care from a team of midwives were less likely to have interventions than women receiving standard maternity care, although there was a significant level of heterogeneity among these trials. Birth events

Meta-analysis was conducted for interventions related to birth, with findings as follows. CS (ten trials, n = 10 622): RR 0.90 [95% CI 0.80 to 1.00] (test for heterogeneity P = 0.31); instrumental vaginal birth (nine trials, n = 10 449): RR 0.85 [95% CI 0.76 to 0.95] (test for heterogeneity P = 0.52); episiotomy (ten trials, n = 9810): RR 0.79 [95% CI 0.74 to 0.85] (test for heterogeneity P = 0.02). Overall, women receiving care from a team of midwives were significantly less likely to have these interventions.

Six trials reported no significant difference in postpartum haemorrhage (PPH) and five trials reported no significant difference in either manual removal of placenta or retained placenta. Newborn outcomes

Meta-analysis was conducted for interventions related to newborn events with results as follows. Condition at birth (Apgar score less than 7 at 5 minutes) (seven trials, n = 6135): RR 1.17 [95% CI 0.81 to 1.680] (test for heterogeneity P = 0.68); admission to neonatal units (nine trials, n = 10 404): RR 0.90 [95% CI 0.79 to 1.03] (test for heterogeneity P = 0.05); perinatal mortality (nine trials, n = 10 423): RR 1.63 [95% CI 1.04 to 2.56], P = 0.03 (test for heterogeneity P = 0.69). Although there were no differences between groups regarding Apgar score at 5 minutes or admission to neonatal intensive care, there was a significantly higher perinatal mortality noted for babies born to women cared for within the team midwifery model. Women's satisfaction and experience of childbirth

Virtually all the trials reported on women's satisfaction and their assessment of the childbirth experience. This was measured using various qualitative methods. All the trials reported that team midwifery systems of care designed to provide intrapartum care by a midwife met antenatally increased women's satisfaction and resulted in more positive experiences of childbirth compared with standard maternity care. Women's mental and psychological health

One trial reported on the emotional wellbeing of women who were given continual support from a team of midwives. Responses to the Edinburgh Postnatal Depression Scale (EPDS) 2 months after the birth showed that 16% of women in the team midwifery care group and 12% in the standard care group were depressed (EPDS score > 12) – a non-statistically significant difference (P = 0.19). Long-term outcomes

There were no long-term outcomes reported in the relevant articles. Wellbeing of healthcare professionals

The cross-sectional study98,109 (n = 1166) measured occupational stress, especially burnout, in midwives, comparing those in midwifery teams (hospital-based and community-based) with traditional hospital-based midwives and GP-attached midwives. Burnout was measured using an adaptation of the Maslach Burnout Inventory (MBI). The study found that burnout was associated with a lack of freedom to make decisions at work, longer contracted hours and low control over work pattern. Findings showed that midwives working in hospital-based teams had the highest reported levels of burnout, followed by traditional hospital-based midwives. No relationship was found between higher levels of burnout and continuity rate, number of nights worked on-call and type of caseload. It would appear, however, that this association is strongly linked with working within the constraints of a hospital-based system where midwives tend to have less autonomy over working pattern and decision making compared with community-based midwives. Caseload midwifery Description of included studies

One UK RCT and one UK cluster RCT were identified for inclusion in this review. The RCT compared women cared for by a named midwife with three associate midwives (n = 648) in a hospital-based midwifery development unit (MDU) with women receiving shared care (n = 651) (majority of care provided by GP with three or four visits to the obstetrician at the hospital).110 [EL = 1+] The cluster RCT was randomised on the basis of geographical area, with three areas in each cluster.97 [EL = 1-] In three areas caseload midwifery care was provided to all low-risk women booked for maternity care (n = 770). The caseload model involved each named midwife being allocated 35–40 women to care for, with back-up provided by one or two associate midwives. In the remaining three areas shared care was provided to women (n = 735) by the GP and community midwife in the established way, with occasional visits to the hospital to see an obstetrician. Details for each study are presented in Table 46.

Table 46. Details of included studies of caseload midwifery model.

Table 46

Details of included studies of caseload midwifery model. Review findings

It was not possible to perform a meta-analysis owing to the methodological differences between the two studies. Labour events

Findings from the (non-cluster) RCT showed that women cared for within the caseload midwifery model had fewer inductions of labour: 199 (33.3%) versus 146 (23.9%); difference 9.4% [95% CI 4.4% to 14.5%]. There was no significant difference found for other labour events including augmentation of labour (difference −3.4%), opioid analgesia (difference 2.5%) and epidural (difference 1.4%). The lower use of epidural analgesia (10% versus 15%) and oxytocin augmentation of labour (46% versus 53%) was also evident in the cluster RCT. No differences in induction of labour were noted, however. Birth events

Findings from the RCT showed that significantly more women in the caseload midwifery group had an intact perineum following birth: 120 (23.6%) versus 160 30.5%, while fewer had an episiotomy: 173 (34.0%) versus 147 (28.0%) or a first- or second-degree perineal tear: 216 (42.4%) versus 218 (41.5%); test for overall difference P = 0.02 (χ2). No significant differences were found between groups for mode of birth, with the incidence of spontaneous vaginal birth being 73.7% in the shared care group compared with 73.5% in the caseload midwifery group. Findings from the cluster RCT showed no differences between groups for perineal trauma or mode of birth. Newborn outcomes

Findings from the RCT showed no difference for newborn outcomes between groups, Apgar score 8–10 at 5 minutes: 565 (96.6%) versus 589 (97.8%), difference −1.2% [95% CI −3.1% to 0.6%]; admission to special care baby unit (SCBU) 40 (6.6%) versus 33 (5.4%), difference 1.2% [95% CI −1.4% to 3.9%]. There were nine stillbirths plus neonatal deaths in the shared care group compared with four in the caseload midwifery group (difference 0.4% [95% CI −0.4% to 1.2%]. Findings from the cluster RCT also showed no differences between the groups in newborn outcomes. There were a total of 11 stillbirths plus neonatal deaths (1.5%) in the shared care group and six (0.7%) in the caseload midwifery group (difference 0.8% [95% CI −0.2% to 1.8%]). Women's satisfaction and experience of childbirth

In the RCT women were found to be significantly more satisfied with their maternity care; antenatal care: difference in mean scores 0.48 [95% CI 0.41 to 0.55]; intrapartum care: 0.28 [95% CI 0.18 to 0.37]; hospital-based postnatal care: 0.57 [95% CI 0.45 to 0.70]; home-based postnatal care: 0.33 [95% CI 0.25 to 0.42]. A basic cost comparison of team midwifery versus conventional midwifery Rationale

The evidence does not suggest that team midwifery leads to significantly better outcomes. Indeed, a meta-analysis undertaken as part of this guideline suggested that team midwifery resulted in statistically significant increases in perinatal mortality compared with the standard model: RR 1.64 [95% CI 1.04 to 2.58], P = 0.03.

Anecdotally, a number of providers appear to have ceased providing a team midwifery service on the grounds of cost. Similarly, one reason team midwifery did not become more widely established was because additional funding was not made available for it. This would seem to indicate, at least from the perspective of service providers, that team midwifery is a more costly service than the conventional model. If it is both more expensive and less effective we can say unambiguously that it is not cost-effective, being ‘dominated’ by the conventional model.

At this stage we do not have the detailed cost data to do a full cost comparison of the two models of care. The only quantitative information we have at this stage comes from a maternity unit in the north of England currently offering a form of team midwifery care. They state that they have an annual midwife to birth ratio of 1:26 against a national average of 1:33. At this stage we do not know how representative this unit's ratio is of team midwifery models in general but it does at least seem consistent with the perception that team midwifery (TM) is a more resource-intensive service. If we assume that this service was typical then we could estimate the additional midwife staffing cost per birth as follows:

Annual cost of midwife= £40,000 approximately
Hospital births= n
Additional midwifery staffing in TM model= (1/26 − 1/33) × n = 0.008 × n
Additional midwifery staffing cost= £40,000 × 0.008 × n = £326 × n
Additional midwifery cost per birth= £326

Clearly, a full cost comparison would also have to include ‘downstream’ cost differentials between the two models of care, especially as the meta-analysis undertaken for the guideline found the following intervention differences for team midwifery:

Induction:pooled OR 0.88 [95% CI 0.80 to 0.98], P = 0.02
Augmentation:pooled OR 0.83 [95% CI 0.76 to 0.91], P < 0.001
EFM:pooled OR 0.30 [95% CI 0.27 to 0.33], P < 0.001
Epidural:pooled OR 0.77 [95% CI 0.71 to 0.85], P < 0.001
Narcotics:pooled OR 0.72 [95% CI 0.66 to 0.78], P < 0.001
Caesarean section:pooled OR 0.91 [95% CI 0.81 to 1.02], P = 0.12; NS
Instrumental birth:pooled OR 0.84 [95% CI 0.75 to 0.95], P = 0.005
Episiotomy:pooled OR 0.73 [95% CI 0.67 to 0.80], P < 0.001

This meta-analysis suggests that women receiving care from team midwifery have less intervention and therefore ‘downstream’ costs may, to some extent, offset higher staffing costs of service provision. The most important of these ‘downstream’ savings is likely to relate to a lower rate of instrumental vaginal birth and the saving per birth that this might be expect to produce is calculated below.

From NHS Reference Costs (2004) finished consultant episode data:

Normal birth= 382 669
Instrumental births= 64 995
Caesarean sections= 130 353
Total births= 578 017
Odds of instrumental conventional birth= 64 995/513 022 = 0.127
Odds of instrumental TM birth= 0.84 OR from meta-analysis × 0.127 = 0.107
Number of instrumental TM births= 55 870
Reduction in instrumental births due to TM= 64 995 − 55 870 = 9125
Cost of instrumental vaginal birth= £1,263
Cost of normal vaginal birth= £863
Cost saving of reduction in instrumental births due to TM= 9125 × (£1,263 - £863) = £3.65 million
Cost saving per birth= £3,650,000/578 017 = £6.30

While this is a substantial saving it falls a long way short of what would be required to offset the additional staffing costs of providing a team midwifery service.

This analysis does not constitute a proper costing of the two alternative models of care. However, if its assumptions are accepted it would suggest that a team midwifery model is more expensive than a conventional model of midwifery care. When taken together with some evidence of higher perinatal mortality it could not be recommended on cost-effectiveness grounds.

4.3.5. Evidence statement Team midwifery

In general, the studies included were of good quality. There was heterogeneity between the studies, particularly in both the settings for intrapartum care and the size of the team, which makes interpretation difficult. There is evidence to support that women cared for by a team of midwives throughout their pregnancy, intrapartum and postnatal period are less likely to have interventions during labour, and that such care is highly valued by women. However, there is an increased perinatal mortality associated with team midwifery care. There was no indication as to which component of care, or combination of components of care, might have contributed to this.

There is some evidence that midwives working in hospital-based teams experience higher levels of burnout than those working in community-based teams. Caseload midwifery

Findings from two trials show that women cared for in a caseload midwifery system are less likely to receive interventions during labour and that women prefer this system of care compared with traditional shared care. No evidence of difference in other maternal or neonatal outcomes was found.

There is no evidence about its cost-effectiveness. Recommendation on continuity of care


Team midwifery (defined as a group of midwives providing care and taking shared responsibility for a group of women from the antenatal, through intrapartum to the postnatal period) is not recommended. [2007] Research recommendations on continuity of care


Studies are needed that investigate the components affecting a woman's satisfaction with her birth experience, including her emotional and psychological wellbeing. A robust method of assessing a woman's satisfaction is also needed.


There should be studies carried out to investigate the effects of caseload midwifery (defined as one midwife providing care and taking responsibility for a group of women from the antenatal, through intrapartum to the postnatal period) on women, babies and healthcare professionals, including cost-effectiveness and long-term outcomes.

4.4. Eating and drinking in labour

4.4.1. Review question

What is the effectiveness of the following interventions or techniques in labour on outcomes?

  • restricting fluids and nutrition.

4.4.2. Reduction gastric aspiration in labour Routine prophylactic drugs in normal labour for reducing gastric aspiration Description of included studies

A systematic review identified three randomised controlled trials.111 [EL = 1+] The intervention was any drug, with any route of administration, in any dosage. The drug categories were particulate and non-particulate antacids, H2-receptor antagonists, dopamine antagonists and proton pump inhibitors, although no trials were identified on proton pump inhibitors. The primary outcome measure was the incidence of gastric aspiration in the woman. The review found none of the trials to be of good quality. Review findings

There was limited evidence to suggest that antacids may reduce the chance of vomiting in labour when compared with no intervention (one trial, n = 578; RR 0.46 [95% CI 0.27 to 0.77]). When individual antacids were compared with each other, when tested in one study, there was no significant difference in incidence of vomiting (Gelusil® versus Maalox® (n = 300): RR 0.83 [95% CI 0.39 to 1.75]; Gelusil versus Mylanta II® (n = 325): RR 1.32 [95% CI 0.58 to 2.99]); Maalox versus Mylanta II (n = 285): RR 1.59 [95% CI 0.69 to 3.65]). There was no significant difference in vomiting (one trial, n = 1287; RR 0.96 [95% CI 0.73 to 1.27]); CS (one trial, n = 1287; RR 0.93 [95% CI 0.59 to 1.47]); emergency general anaesthetic (one trial, n = 1287; RR 0.92 [95% CI 0.62 to 1.35]); PPH (one trial, n = 1287; RR 0.83 [95% CI 0.08 to 9.14]) and stillbirth (one trial, n = 1287; RR 0.69 [95% CI 0.17 to 2.89]) when H2-receptor antagonists were compared with antacids. Again, the number of participants was too small for the results to be conclusive.

Dopamine antagonists given alongside pethidine may reduce vomiting in labour (one trial, n = 584; RR 0.40 [95% CI 0.23 to 0.68]) when compared with placebo or no treatment given alongside pethidine, but the subgroups from the study population were too small to make an assured comment. The trial showed no significant difference in Apgar scores < 7 at 1 minute (RR 1.02 [95% CI 0.62 to 1.69]) or perinatal deaths (RR 1.22 [95% CI 0.24 to 6.21]). When two different dopamine antagonists were compared (metoclopramide versus perphenazine; n = 393) there was no significant difference in vomiting (RR 1.45 [95% CI 0.64 to 3.32]), Apgar score < 7 at 1 minute (RR 0.83 [95% CI 0.47 to 1.47]) or perinatal death (RR 0.25 [95% CI 0.03 to 2.23]). Evidence statement

The studies were too small to assess the incidence of gastric aspiration, Mendelson syndrome and its consequences. There is limited evidence that antacids or dopamine antagonists given alongside pethidine reduce the chance of vomiting in labour. There is also limited evidence that H2-receptor antagonists have no impact on vomiting and other outcomes when compared with antacids.

There were no trials identified on proton pump inhibitors. Recommendations on reducing gastric aspiration


Do not offer either H2-receptor antagonists or antacids routinely to low-risk women. [2007]


Either H2-receptor antagonists or antacids should be considered for women who receive opioids or who have or develop risk factors that make a general anaesthetic more likely. [2007] Research Recommendations on reducing gastric aspiration


Use of either H2-receptor antagonists or antacids in labour should be evaluated for women who have or develop risk factors, who have opioids or who may need a general anaesthetic.

4.4.3. Eating and drinking in labour Description of included studies

One randomised controlled trial, published in 1999, was identified (eating group = 45; starved group = 43). The study population comprised women in labour at 37 weeks of gestation or greater who had one baby with cephalic presentation. The intervention was a low-residue diet compared with water only.112 [EL = 1+] Review findings

The results showed that restriction of food throughout the course of labour results in a significant increase in plasma β-hydroxybutyrate (mean difference (MD) 0.38 mmol/l [95% CI 0.21 to 0.55 mmol/l], P < 0.001) and non-esterified fatty acids (MD 0.35 mmol/l [95% CI 0.22 to 0.48 mmol/l], P < 0.001) when compared with eating a low-residue diet. There was a significant increase in plasma glucose (MD 0.62 mmol/l [95% CI 0.22 to 1.01 mmol/l], P = 0.003) and insulin (MD 15.6 mmol/l [95% CI 2.9 to 28.3 mmol/l], P = 0.017) in the eating group when compared with the starved group. Gastric antral cross-sectional areas measured within 1 hour of labour were significantly higher in the eating group (MD 1.85 cm2 [95% CI 0.81 to 2.88 cm2], P = 0.001) and these women were also twice as likely to vomit at or around giving birth (MD 19% [95% CI 0.8% to 38%], P = 0.046). The volumes vomited by the women in the eating group were significantly larger (MD 205 ml [95% CI 99 to 311 ml], P = 0.001) than the volumes vomited by women in the starved group. Lactic changes remained similar in both groups (MD −0.29 mmol/l [95% CI −0.71 to 0.12 mmol/l], P = 0.167). However, the study showed no significant differences in maternal outcomes (duration of first and second stage of labour, oxytocin requirements, mode of birth) or neonatal outcomes (Apgar scores, umbilical artery and venous blood gases) between the two groups of women (only means reported). Evidence statement

The limited evidence suggests that a light diet significantly reduces the rise of plasma β-hydroxy-butyrate and the non-esterified fatty acids from which it is derived, while significantly increasing plasma glucose and insulin. However, the significant increase in volumes vomited must be considered, given that there were no significant differences in maternal or neonatal outcomes.

4.4.4. Intervention to prevent ketosis Carbohydrate solution versus placebo Description of included studies

Three randomised controlled trials, conducted by the same researchers at the Leyenburg Hospital in the Netherlands, were identified for review. The first study involved 201 nulliparous women randomised at 2–4 cm cervical dilatation (carbohydrate solution n = 102; placebo n = 99).115 [EL = 1+] Women were able to consume small standardised amounts of food or drink on specific demand, with total amount of intake of kilojoules calculated for each woman at the end of the study. The second trial involved 202 nulliparous women randomised at 8–10 cm cervical dilatation, (carbohydrate solution n = 100; placebo n = 102).113 [EL = 1+] Women were not allowed any other solutions. The final study involved 100 nulliparous women randomised at 8–10 cm cervical dilatation (carbohydrate solution n = 50; placebo n = 50).114 Women were only allowed water in addition to the study solutions. [EL = 1+] Review findings

In the first study, the median intake of study solution was 300 ml [range 17 to 1600 ml] in the placebo group and 400 ml [range 0 to 1600 ml] in the carbohydrate group (P = 0.04).115 Similar proportions of women in both groups had a small additional intake (32% placebo group; 32.5% carbohydrate group). The median total calorific intake by the placebo group during the study was 0 kJ [range 0 to 1086 kJ] and 802 kJ [range 140 to 3618 kJ] for the carbohydrate group (P < 0.001). There was no statistically significant difference in the need for augmentation (RR 0.83 [95% CI 0.55 to 1.26]) or in the need for pain-relieving medication (opiates: RR 0.96 [95% CI 0.44 to 2.11]; epidural: RR 1.56 [95% CI 0.89 to 2.73]; Entonox: RR 3.64 [95% CI 0.72 to 15.8]), when women in the carbohydrate group were compared with women in the placebo group. While there was no significant difference between the carbohydrate and placebo groups for spontaneous birth (RR 0.90 [95% CI 0.68 to 1.17]) or for instrumental birth (RR 0.78 [95% CI 0.52 to 1.17]), the number of caesarean sections was significantly higher in the carbohydrate group (RR 2.9 [95% CI 1.29 to 6.54]). There were no significant differences in Apgar scores at 1 minute (P = 0.17), Apgar scores at 5 minutes (P = 0.18) or the arterial umbilical cord pH (P = 0.07) between the carbohydrate and placebo groups.

In the second study, the median intake of study solution was 200 ml [range 15 to 200 ml] in the placebo group and 200 ml (10 ml to 200 ml) in the carbohydrate group (P = 0.42).113 There were no significant differences in spontaneous birth (RR 1.07 [95% CI 0.88 to 1.30]), instrumental birth (RR 1.05 [95% CI 0.69 to 1.60]) or CS (RR 0.15 [95% CI 0.02 to 1.16]) when the carbohydrate group was compared with the placebo group. No significant differences were observed in neonatal outcome: Apgar scores at 1 minute (P = 0.22), Apgar scores at 5 minutes (P = 0.32) or the arterial umbilical cord pH (P = 0.80), when the carbohydrate group was compared with the placebo group. In addition, when the carbohydrate and placebo groups were compared, there were no significant differences in changes in glucose (P = 1.00), lactate (P = 0.07) or plasma β-hydroxybutyrate (P = 0.21). There was a significant decrease in free fatty acid levels (P = 0.02), with the carbohydrate group tending to decrease to a higher degree.

In the third study, there were no significant differences in spontaneous birth (P = 0.30) or vaginal instrumental birth (P = 0.84) when the groups were compared.114 However, the cohort was too small to draw conclusions. There were four caesarean sections in the placebo group and none in the carbohydrate group, but no statistical calculations were made.

Arterial umbilical cord pH, pCO2, pO2, HCO3 and base excess were similar in both groups, as were venous umbilical cord results. However, no statistical data were presented. Evidence statement

There is no evidence of difference in mode of birth, or fetal and neonatal acid–base balance between taking carbohydrate solution and placebo during labour. Isotonic sports drink versus water Description of included studies

One randomised controlled trial conducted in the UK and published in 2002 was identified.116 The study involved 60 women at 37 weeks of gestation or greater, with a singleton fetus having cephalic presentation (sports drink group n = 30; water group n = 30). [EL = 1+] Review findings

In the sports drink group there was a significant decrease in plasma β-hydroxybutyrate (MD −0.63 [95% CI −0.85 to −0.42]) and non-esterified fatty acids (MD −0.36 [95% CI −0.46 to −0.25]) when compared with the water-only group. Mean plasma glucose remained unchanged in the sports drink group, but decreased significantly in the water-only group (MD 0.76 mmol/l [95% CI 0.22 to 1.3 mmol/l]). The total quantity of liquid consumed was significantly higher (P = 0.001) in the sports drink group. The mean calorific intake was also higher for the sports drink group (47 kcal/hour (SD 16 kcal/hour) compared with the water-only group (0 kcal/hour). However, there was no significant difference in gastric antral cross-sectional area (MD −0.63 cm2 [95% CI −1.12 to 0.70 cm2]), volume vomited within 1 hour of giving birth (MD 65 ml [95% CI −141 to 271 ml]) or volume vomited throughout labour (MD 66 ml [95% CI −115 to 246 ml]), when the two groups were compared. There was no difference between the two groups with respect to duration of labour, use of oxytocin, mode of giving birth or use of epidural analgesia. The study authors only presented the data as mean (SD) or proportion (%), but noted that all results were non-significant. Evidence statement

There is a small amount of evidence to demonstrate that ketosis is prevented by relatively small calorific intake provided by isotonic drinks and that these provide an alternative source of nutrition that is rapidly emptied from the stomach and rapidly absorbed by the gastrointestinal tract.

There is limited evidence that labour outcomes were not compromised in either the sports drink group or the water-only group. GDG interpretation of the evidence (eating and drinking in labour)

The development of ketosis in labour may be associated with nausea, vomiting and headache and may be a feature of exhaustion. Limited evidence suggests that a light diet or fluid carbohydrate intake in labour may reduce ketone body production while maintaining or increasing glucose and insulin. However, the volume of stomach contents may increase, increasing the chances of the woman being sick. There are no differences in any measured outcomes. Recommendations on eating and drinking labour


Inform the woman that she may drink during established labour and that isotonic drinks may be more beneficial than water. [2007]


Inform the woman that she may eat a light diet in established labour unless she has received opioids or she develops risk factors that make a general anaesthetic more likely. [2007]

4.5. Hygiene measures during labour

4.5.1. Introduction

Puerperal sepsis was the leading cause of maternal mortality in the UK up until the early 20th century. Deaths due to sepsis fell dramatically following the widespread availability of antibiotics in the 1940s and the passing of the Abortion Act in 1967, with no deaths from sepsis being reported in the triennium 1982–84. Unfortunately, deaths from sepsis have been reported in each of the subsequent triennial reports, with 13 maternal deaths being directly attributed to sepsis in the 2000–02 report: five women following a vaginal birth, two after giving birth at home. The continued deaths of previously healthy women by overwhelming infection following childbirth and the spread of blood-borne diseases such as HIV highlight the importance of adequate hygiene measures during labour, to protect both the woman and her caregivers. Many women are exposed to invasive procedures during labour, all of which have potential to introduce pathogens into the genital tract. While the rituals of perineal shaving and the administration of enemas, previously performed to reduce contamination of the genital tract during birth, have been discredited, well-established practices of cleansing and draping the vulva prior to vaginal examinations and birth are still commonly practised.

4.5.2. General points

General points in infection control were reviewed in the NICE clinical guideline Infection Control, published in June 2003.7 The guideline reviewed 169 articles for the section relating to general principles of infection control. Twenty-six recommendations were provided for areas of hand hygiene, use of personal protective equipment, and safe use and disposal of sharps. The recommendations below are specific to women in labour; however, they do not override the recommendations in the Infection Control guideline.

4.5.3. Review question

Are there effective hygiene strategies for vaginal birth out of water to protect both women and babies, and healthcare professionals?

  • Strategies include vaginal examination and antisepsis.
  • Outcomes include infection control and rates of infection.

4.5.4. Chlorhexidine vaginal douching and perineal cleaning Chlorhexidine vaginal douching Description of included studies

There was one systematic review identified. This review included three RCTs (n = 3012) in the USA, comparing chlorhexidine vaginal douching during labour with sterile water as a placebo control.117 [EL = 1++] Review findings
Women's outcomes

Three trials reported the incidence of chorioamnionitis, including 1514 and 1498 women in the chlorhexidine and placebo groups, respectively. There was no statistically significant difference between the two groups (RR 1.10 [95% CI 0.86 to 1.42]). The same three trials also reported the incidence of postpartum endometritis. Although the data suggested a small reduction in the risk of postpartum endometritis with the use of the chlorhexidine vaginal wash, the difference was not statistically significant (RR 0.83 [95% CI 0.61 to 1.13]). There was no report of other maternal outcomes or side effects of chlorhexidine in these three trials.

Newborn outcomes

Three trials reported on neonatal outcomes, involving 1495 and 1492 neonates in the chlorhexidine and placebo groups, respectively. One trial (n = 910 neonates) indicated that there was no significant difference in neonatal pneumonia (RR 0.33 [95% CI 0.01 to 8.09]). For neonatal meningitis, one trial with 508 and 513 neonates in the intervention and control groups, respectively, did not show significant difference (RR 0.34 [95% CI 0.01 to 8.29]). Two trials, involving 1038 and 1039 neonates in the intervention and control groups, respectively, found neither significant difference in blood culture confirming sepsis (RR 0.75 [95% CI 0.17 to 3.35]) nor in perinatal mortality (RR 1.00 [95% CI 0.17 to 5.79]). No significant difference was found for neonatal sepsis (three trials, n = 2987; RR 0.75 [95% CI 0.17 to 3.35]). There was a trend suggesting that the use of vaginal chlorhexidine during labour might lead to a higher tendency for newborns to receive antibiotics, but this association was not statistically significant (RR 1.65 [95% CI 0.73 to 3.74]). No other neonatal outcomes or side effects of chlorhexidine were reported. Perineal cleaning Description of included studies

There was one UK controlled study (n = 3905) identified which compared cetrimide/chlorhexidine for perineal cleaning during labour with tap water.118 [EL = 2+] The allocation of intervention/control was by alternate months. The study population included women who had a caesarean birth (17.2% for cetrimide/chlorhexidine and 16.3% for tap water). Review findings
Women's outcomes

The findings (cetrimide/chlorhexidine group n = 1813; tap water group n = 2092) showed no evidence of a difference in the number of women who developed fever (temperature > 38 °C) (OR 1.2 [95% CI 0.8 to 1.9]), use of antibiotics (OR 1.02 [95% CI 0.86 to 1.9]), perineal infection (OR 1.4 [95% CI 0.77 to 2.7]), perineal infection (OR 1.4 [95% CI 0.77 to 2.7]), perineal breakdown (OR 5.8 [95% CI 0.3 to 999]) or caesarean wound infection (OR 1.3 [95% CI 0.86 to 1.9]). There was one maternal death in each arm: both were considered to be due to anticardiolipin syndrome.

Newborn outcomes

The results for babies' outcomes showed no difference in eye infection (OR 1.1 [95% CI 0.78 to 1.7]), cord infection (OR 1.3 [95% CI 0.7 to 2.1]), other infections not specified (OR 0.87 [95% CI 0.65 to 1.2]), admission to SCBU (OR 1.1 [95% CI 0.9 to 1.4]), use of antibiotics (OR 0.99 [95% CI 0.82 to 1.2]) or fever (temperature > 38 °C) (OR 1.4 [95% CI 0.66 to 3.0]).

There were 27 perinatal deaths reported in the cetrimide/chlorhexidine group (total n = 1813) and 21 perinatal deaths reported in the water group (total n = 209). The causes of death were reported as one due to uterine rupture and three due to intrapartum asphyxia in the cetrimide/chlorhexidine group, and one due to necrotising enterocolitis and one due to neonatal septicaemia in the water group. Other deaths were considered to be due to either congenital abnormality or birthweight less than 1000 g. Evidence statement

There is evidence that the use of cetrimide/chlorhexidine is no more effective than water for perineal cleaning.

No evidence exists to provide advice on the use of sterile gowns, sterile packs or vulval cleansing prior to vaginal examination or vaginal birth in reducing maternal or neonatal morbidity. Recommendations on vaginal douching and perineal cleaning


Tap water may be used if cleansing is required before vaginal examination. [2007]

4.5.5. Double gloves during episiotomy and other procedures Double gloves during episiotomy Description of included studies

There were two RCTs conducted in Thailand comparing the use of double gloves with single gloves while performing an episiotomy. Outcome measures were perforation rates only. The earlier study included 2058 sets of gloves (double-gloving n = 1316; single-gloving n = 742),119 and the later study included 300 sets of gloves (double-gloving n = 150; single-gloving n = 150).120 Review findings

The earlier study reported perforation rates of double inner gloves as 2.7% (P < 0.05), outer as 5.9%, compared with single gloves as 6.7%. The later study reported perforation rates of double inner gloves as 4.6% (P < 0.05), outer as 22.6%, compared with single gloves as 18.0%. Evidence statement

Wearing two gloves appears to reduce perforation rates in inner gloves compared with single-gloving. However, caution needs to be taken in interpreting these results as there was no concealment. Arm sleeves Description of included studies

One case series conducted in the UK (n = 80) has evaluated the effectiveness of wearing a sterile arm sleeve on top of the gown to prevent contamination during obstetric procedures.121 [EL = 3] Review findings

The contamination of arms and hands was 3.8% and 5%, respectively. Evidence statement

There is insufficient evidence on the use of a sterile arm sleeves in preventing contamination. Recommendations on double-gloving


Routine hygiene measures taken by staff caring for women in labour, including standard hand hygiene and single-use non-sterile gloves, are appropriate to reduce cross-contamination between women, babies and healthcare professionals. [2007]


Selection of protective equipment musto be based on an assessment of the risk of transmission of microorganisms to the woman, and the risk of contamination of the healthcare worker's clothing and skin by women's blood, body fluids, secretions or excretions.p [2007, amended 2014]

4.5.6. Identification of women and babies who may need additional care

The GDG members decided to use the criteria below (the list is not exhaustive) to identify women and babies who may need additional care, and therefore would need referral to specialist care not covered in this guideline:

  • intrapartum haemorrhage
  • placental abruption
  • ruptured uterus
  • ‘suspected’ amniotic fluid embolus
  • ‘suspected’ pulmonary embolus
  • eclampsia and severe pre-eclampsia
  • cord prolapse
  • shoulder dystocia
  • massive obstetric haemorrhage
  • maternal collapse
  • monitoring suggesting fetal compromise
  • undiagnosed breech.



This recommendation is adapted from Infection: prevention and control of healthcare-associated infections in primary and community care (NICE clinical guideline 139).

Copyright © 2014 National Collaborating Centre for Women's and Children's Health.
Bookshelf ID: NBK328272


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (16M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...