Changing childbirth: lessons from an Australian survey of 1336 women

Authors

  • Stephanie Brown,

    Research Fellow , Corresponding author
    1. Centre for the Study of Mothers ‘and Children's Health, La Trobe University, Melbourne, Australia
      Correspondence: Ms S. Brown, Centre for the Study of Mothers' and Children's Health, La Trobe University, 463 Cardigan Street, Carlton 3053, Victoria, Australia.
    Search for more papers by this author
  • Judith Lumley

    Professor
    1. Centre for the Study of Mothers ‘and Children's Health, La Trobe University, Melbourne, Australia
    Search for more papers by this author

Correspondence: Ms S. Brown, Centre for the Study of Mothers' and Children's Health, La Trobe University, 463 Cardigan Street, Carlton 3053, Victoria, Australia.

Abstract

Objective To investigate the views and experiences of care in labour and birth of a representative sample

Design Cross-sectional survey mailed to women 6–7 months after giving birth.

Population All women who gave birth in a two week period in Victoria, Australia in September 1993, except those who had a stillbirth or neonatal death.

Results After adjusting for parity, the risk status of the pregnancy, and social and obstetric factors, specific aspects of care with the greatest negative impact on the overall rating of intrapartum care were: caregivers perceived as unhelpful (midwives: adjusted OR 12.03 [95% CI 7–8–1 8.1, doctors: adjusted OR 6.76 [95% CI 4.–10.31); and having an active say in decisions only sometimes, rarely or not at all (adjusted OR 8.0 [95% CI 4.–16–11). In a separate regression analysis including parity, risk status, obstetric and social factors, but not specific aspects of care, factors associated with dissatisfaction with intrapartum care included participation in a shared antenatal care programme (adjusted OR 1.9 [95% CI 1.–3.1) and being of nonEnglish speaking background (adjusted OR 1.0 [95% CI 1.–2.1). The following factors lowered the odds of dissatisfaction: attending a birth centre (adjusted OR 0.34 [95% CI 0.–1.]) and knowing the midwives before going into labour (adjusted OR 0.8 [95% CI 0.–0.]).

Conclusion The survey demonstrates the potential for ‘new’ models of care to have either positive or negative effects on women's experiences of care. Evaluation of innovations in perinatal care taking into account women's views is a prerequisite for improvements in maternity care. of women who gave birth in Victoria, Australia in 1993.

INTRODUCTION

There is less need nowadays to make a case for measuring so called ‘soft outcomes’1. Recent reviews of maternity services in the United Kingdom and in Australia have accepted the need to consult with ‘consumers’ as well as providers, professional associations and health services2–5. Another welcome development has been the inclusion of measures of ‘patient satisfaction’ in recent randomised trials of innovations in perinatal care6–10. Nonetheless, it remains difficult to measure satisfaction11,12, partly because people tend to give very positive responses to questions asking about their views of medical care. In most surveys ⋚ 80% state they are satisfied when asked for an overall rating of care13,14. This may be amplified in the case of maternity care where the birth of a healthy infant, and survival of one of life's major ordeals are bound to influence how women view the care they receive15,16. A more important problem concerns the capacity to design methods of assessing women's views of maternity care that yield information useful for improving services.

A population based Survey of Recent Mothers was conducted in Victoria, Australia in 1993, three years after the release of the Final Report of the Ministerial Review of Birthing Services in Victoria4. It followed up an earlier survey undertaken in conjunction with the Review by the same research group in 198917,18. The Ministerial Review recommended increasing the availability of models of care with the “potential to improve continuity of care within the public system”, in particular shared care, team midwifery, and birth centre care; and promoting “greater involvement of midwives and general practitioners in the care of women at low risk of complication”4. Given the emphasis on new models of care in reviews of maternity services in Australia and the United. Kingdom2–5, we were interested to find out 1. how women using these options viewed their care; 2. the extent to which greater continuity of care had been achieved; and 3. whether this had enhanced women's experiences of care. The 1993 Survey of Recent Mothers was designed to address these questions with a view to informing policy and service development. This paper describes our findings in relation to care in labour and birth. The results for antenatal care have already been reported19.

METHODS

Sample

Questionnaires were sent to 2224 women six to seven months after they had given birth. All maternity hospitals and home birth practitioners in Victoria, Australia were asked to facilitate the mailing of questionnaires to women who gave birth in a two week period in September 1993, excluding those who had had a stillbirth or whose baby was known to have died. Ethics approval was granted by each of the three tertiary level maternity hospitals in Victoria, and by 22 other hospitals with institutional ethics committees. Of Victoria's 130 hospitals with maternity beds, all but three agreed to participate. The final sample included > 91% of births in the survey period.

Procedure

Three mailings took place at two week intervals. Each one included a covering letter inviting women to participate, a copy of the questionnaire, a brief explanation of the study in six community languages, and a freepost return envelope.

A total of 1366 completed questionnaires were returned. Twenty four of these were excluded because the baby's birth date fell outside the study period. Five duplicate questionnaires and one returned by a woman who had had a stillbirth were also excluded. Another 86 questionnaires were undelivered. Excluding these questionnaires from the denominator, the response rate to the survey was 62.5% (1336/2138).

Using the state's perinatal data collection comparison of respondents with all women who gave birth in the study period showed that the respondents were representative in terms of important obstetric characteristics: the number of previous live births, birthweight of infants in the current pregnancy, and mode of birth. Women born overseas of nonEnglish speaking background, single women, and women under 25 years were under-represented. A detailed description of the study sample and methods is given elsewhere20.

Questionnaire

The questionnaire covered antenatal, intrapartum and postnatal care; socio-demographic characteristics; past reproductive history; and information about the baby. As the questionnaire was anonymous (a requirement for ethics approval) no data collection from medical records was possible. Questions therefore covered both factual information (e.g. who was present at the birth, use of procedures), and women's views of care. Several studies comparing medical records with women's accounts of obstetric events have found high levels of consistency21,24.

Women were grouped according to six major models of care (see Appendix 1). An overall assessment of intrapartum care was obtained from the question “On balance, thinking about what happened to you and what the midwives and/or doctors did, how would you describe your care in labour and birth? Very good, good, mixed, poor or very poor”. Other questions covered access to information, involvement in decisionmaking, reassurance and encouragement offered by caregivers, and the helpfulness of midwives and of doctors when present.

The risk status of the pregnancy was ascertained by asking women whether they had any medical conditions or health problems that required regular medication, special care or extra tests during pregnancy. A measure of obstetric intervention during labour was calculated based on the scoring system devised by Elliot et al.25. The overall score takes account of both the number and types of procedures occurring in each woman's labour and birth, with higher scores assigned for more complex interventions such as epidural or general anaesthesia, and caesarean section. The highest possible score was 40, with a range in this study of 0–28, and a median of 6 (see Appendix 2).

Statistical analysis

Data were analysed using SPSS for Windows26, and involved χ2 comparisons, odds ratios, Mantel Haenszel weighted odds ratios, stratification and univariable and multivariable logistic regression. For statistical analysis, women's overall satisfaction with intrapartum care was grouped into two categories: 1. responses describing care as ‘very good’; and 2. all other categories. An a priori decision was taken to consider all responses other than ‘very good’ as indicating some level of dissatisfaction (i.e. that some aspect of care could have been better)27. Results are stratified by parity because of the number of intrapartum events (e.g. length of labour, frequency of interventions) that differ between primiparae and multiparae. As some sub-groups for stratified analysis were very small, Mantel Haenszel weighted odds ratios were used to give the best overall (populatiodrespondent) summary of the association. We also inspected ordered categorical variables for the presence of trends using the χ2 test for linear trend. Models using logistic regression were developed to assess the contribution of social and obstetric factors, and variables describing women's involvement in decision-making and the helpfulness of caregivers to dissatisfaction with care in labour and birth, and to obtain more precise estimates of the effects of the model of care, and continuity of midwife care.

Power calculations undertaken during the design stage of the study using estimates based on the findings of the 1989 Survey of Recent Mothers showed that a sub-group size of 50 with 300 in the reference category and a two-tailed a of 0.05 would have 80% power to detect a difference of 20% in the proportion of women satisfied with care in labour and birth (from 66% to 86%). Subgroup sizes of 50 or higher were achieved for all models of care with the exception of birth centre care, for important social characteristics such as being aged < 25 years, single, or of nonEnglish speaking background, and for women at higher risk of complications20.

RESULTS

Overall rating of intrapartum care

The majority of women gave a very positive overall rating of their intrapartum care: 7 1.% said their care was very good and a further 21.% described it as good, 6.% said their experience of care was mixed, and 1.% described care as poor or very poor.

Associations with social characteristics

In general, women's overall rating of care was not associated with socio-demographic characteristics. The exceptions were women on lower incomes and women of nonEnglish speaking background who were less likely to have very positive experiences of care. Only 65.6% (137/209) of women with very low family incomes (≤ AUS$20,000) described their care as very good compared with 74.% (397/534) of women in the highest income range of more than AUS$40,000 (OR 0.6 [95% CI 0.–0.1). The proportion of women of nonEnglish speaking background rating their care as very good was 58.% (831 142), compared with 73.% (771/1052) of women born in Australia (OR 0. 1 [95% CI 0.–0.1). Women's overall rating of intrapartum care did not relate to any of the following characteristics: mother's age, marital status, place of residence (rural/metropolitan), school based and tertiary education, or parity.

Pregnancy and reproductive history

No associations were found between rating of care in labour and birth and being at higher risk of complications; a history of preterm birth, infant of low birthweight, pregnancy loss (miscarriage, termination, stillbirth); births after a period of infertility of ≥ 12 months; or prior caesarean section, breech presentation or forceps. There was a strong association between dissatisfaction with antenatal care and dissatisfaction with intrapartum care (OR 6.9 [95% CI 5.2-8.9). Hospitalisation during the current pregnancy was associated with a slightly higher likelihood of rating intraparturn care as other than very good (OR 1. [95% CI 1.–1.1).

Model of care and intrapartum events

Table 1 reports the relationships between women's overall assessment of their intrapartum care and events occurring during labour and birth, including the primary caregiver at the birth, model of care for the pregnancy, and continuity of midwife care. Separate columns show the number and proportion of women experiencing each set of ‘events’ as well as the proportion describing their care as other than very good in each category.

Table 1.  Relationship between intrapartum events and dissatisfaction with care in labour and birth by parity. Values are given as n (%), percentage dissatisfied and OR [95% CI]. ARM = artificial rupture of membranes; CTG = cardiotocograph. VE = ventouse extraction; CS = caesarean section; GP = general practitioner.
 PrimiparaeMultiparae 
 n (%)%dissatisfiedORn (%)%dissatisfiedORWeighted OR
  1. * Excludes women who had an elective caesarean section.

  2. † Excludes women who had either an elective or emergency caesarean section.

Onset of labour*
  Spontaneous353 (73.2)29.71.00556 (77.3)30.61.001.00
  ARM23 (4.8)21.70.66 [0.2–1.9]83 (11.5)19.30.54 [0.3–1.0]0.57 [0.–1.]
  Prostaglandins74 (15.3)20.30.60 [0.3–1.2]44 (6.1)25.00.76 [0.4–1.6]0.66 [0.4–1.1]
  Oxytocin32 (6.6)37.51.42 [0.6–3.2]36 (5.0)38.91.44 [0.7–3.1]1.43 [0.8–2.5]
Augmentation*
  No253 (52.9)25.71.00451 (63.2)29.91.001.00
  Yes225 (47.1)30.71.28 [0.8–2.0]263 (36.8)27.80.90 [0.6–1.3]1.04 [0.8–1.4]
CTG monitoring*
  No144 (31.7)27.81.00366 (53.8)28.11.001.00
  External only225 (49.6)25.80.90 [0.6–1.5]240 (35.3)31.31.16 [0.8–1.7]1.06 [0.8–1.4]
  Internal85 (18.7)37.61.57 [0.9–2.9]74 (1 0.9)27.00.95 [0.5–1.7]1.21 [0.8–1.8]
Length of labour*
  < 6 hours103 (21.5)21.31.00397 (55.9)25.91.001.00
  6 < 12 hours166 (34.7)30.11.59 [0.9–2.9]184 (25.9)29.31.19 [0.9–1.5]1.30 [0.9–1.8]
  ≥ 12 hours210 (43.8)30.91.45 [1.0–2.2]129 (18.2)40.31.55 [1.2–2.0]1.82 [1.3–2.61
Analgesia
  None44 (8.6)18.21.00227 (28.1)25.61.001.00
  Nitrous oxide325 (63.7)31.12.03 [0.–4.9]354 (43.6)33.31.46 [1.0–2.2]2.03 [1.5–2.91
  Pethidine307 (60.2)29.61.90 [0.8–4.6]242 (29.8)38.4142 [1.2–2.8]1.83 [1.3–2.7]
  Epidural164 (32.2)32.32.12 [0.9–5.5]159 (19.6)26.41.12 [0.6–2.2]1.26 [0.8–1.9]
  General anaesthetic36 (7.1)36.12.54 [0.8–8.1]50 (6.2)40.01.94 [1.–3.9]2.10 [1.2–3.7]
Experience of pain*
  As expected or better166 (34.8)16.91.00409 (57.8)21.51.001.00
  Worse than expected252 (52.8)34.12.55 [1.5–4.3]280 (39.5)39.62.40 [1.7–3.4]2.45 [1.8–3.3]
  Uncertain what to expect59 (12.4)37.32.93 [1.4–6.0]19 (2.7)47.43.28 [1.2–9.1]3.04 [1.7–5.4]
Satisfaction with pain relier
  very happy152 (32.1)13.81.00209 (30.6)9.61.001.00
  Happy196 (41.4)26.52.25 [1.3–4.1]328 (48.1)31.44.33 [2.3–7.5]3.27 [2.2–4.9]
  Mixed/unhappy126 (26.6)50.06.24 [3.4–11.6]145 (21.3)54.511.31 [6.2–20–6]8.47 [5.6–13.1]
Mode of birth
  Vaginal288 (56.4)26.71.00631 (77.8)28.51.001.00
  Forceps/VE129 (25.2)30.21.19 [0.7–1.9]36 (4.4)38.91.59 [0.8–3.3]1.29 [0.9–1.9]
  Emergency CS67 (13.1)31.31.25 [0.7–2.3]51 (6.3)33.31.25 [0.7–2.4]1.25 [0.8–1.9]
  Elective CS27 (5.3)33.31.37 [0.5–3.4]93 (11.5)22.60.73 [0.4–1.3]0.86 [0.5–1.4]
Obstetric procedure score
  0–6201 (39.3)24.41.00501 (61.8)28.91.001.00
  7–14147 (28.8)34.01.60 [1.0–2.6]144 (17.8)28.50.98 [0.6–1.5]1.21 [0.9–1.7]
  15–28163 (31.9)28.81.26 [0.8–2.1]166 (20.5)27.70.94 [0.6–1.4]1.06 [0.8–1.4]
Place of birth
  Labour ward301 (59.1)28.21.00503 (62.2)32.21.001.00
  Birthing room101 (19.8)26.70.90 [05–1.6]135 (16.7)19.30.50 [0.3–0.8]0.66 [0.5–0.9]
  Birth centre13 (2.6)7.70.21 [0.0–1.5]25 (3.1)12.00.29 [0.1–4.8]0.26 [0.1–0.8]
  Operating theatre94 (18.5)35.11.37 [0.8–2.3]146 (18.1)27.40.79 [0.5–1.2]0.99 [0.7–1.4]
Perineal trauma†
  Intact perineum67 (16.3)14.91.00242 (36.8)24.81.001.00
  Tear149 (36.2)26.82.09 [0.9–4.8]289 (44.0)29.42.09 [0.9–4.8]1.41 [1.0–2.0]
  Episiotomy196 (47.6)33.22.83 [1.3–6.3]126 (19.2)34.91.63 [1.0–2.7]1.95 [1.3–2.9]
Able to hold baby soon after birth
  Yes408 (80.3)26.71.00719 (88.8)27.41.001.00
  No100 (19.7)36.01.54 [0.9–2.5]91 (11.2)37.41.58 [1.0–2.6]1.56 [1.1–2.2]
Model of care
  Private obstetrician271 (53.1)27.71.00442 (54.6)24.91.001.00
  Private GP35 (6.9)25.70.90 [0.4–2.1]69 (8–5)23.20.91 [0.5–1.7]0.91 [0.6–1.5]
  Public GP68 (13.3)23.50.80 [0.4–1.6]116 (14.3)31.91.41 [0.9–2.3]1.15 [0.8–1.7]
  Public clinic66 (12.9)37.91.59 [0.9–2.9]92 (11.4)38.01.85 [1.1–3.1]1.74 [1.2–2.5]
  Shared care52 (10.2)32.71.27 [0.6–2.5]69 (8–5)47.82.77 [1.6–4.8]2.00 [1.3–3.1]
  Birth centre17 (3.3)17.60.56 [0.1–2.1]22 (2.7)4.50.14 [0.0–0.9]0.32 [0.1–1.0]
Primary caregiver for the birth
  Private obstetrician229 (45.2)22.71.00314 (38.8)22.61.001.00
  GP68 (13.4)22.10.96 [0.5–1.9]113 (14.0)21.20.92 [05–1.6]0.94 [0.6–1.4]
  Obstetrician/GP21 (4.1)33.31.70 [0.6–4.8]41 (5.1)17.10.70 [0.3–1.7]1.00 [0.5–2.0]
  Locum24 (4.7)45.82.88 [1.14.1]22 (2.7)36.41.96 [0.7–5.2]2.39 [1.2–4.6]
  Hospital doctor99 (19.5)41.42.41 [1.4–4.1]125 (15.4)38.42.13 [1.3–3.4]2.25 [1.6–3.21
  Midwife66 (13.0)28.81.38 [0.7–2.7]195 (24.1)36.92.00 [1.3–3.0]1.80 [1.3–2.5]
Midwives known to woman before labour
  No377 (73.9)30.81.00520 (64.1)33.31.001.00
  Yes, not very well89 (17.5)25.8078 [0.5–1.4]196 (24.1)21.90.56 [0.4–0.8]0.63 [0.5–0.9]
  Yes, very well44 (8.6)15.90.43 [0.2–1.0]94 (11.6)14.40.35 [0.2–0.7]0.37 [0.2–0.6]

Among women having a first baby, the only factors associated with a lesser degree of satisfaction were: pain in labour being worse than expected or being uncertain what to expect; dissatisfaction with pain relief; being in labour for 2 12 hours; having either a hospital doctor or a locum as the primary caregiver for the birth; and having an episiotomy. Factors associated with lesser satisfaction among multiparae were: labour continuing for ≥ 12 hours; having a general anaesthetic; using nitrous oxide or pethidine for pain relief; pain being worse than expected, or not knowing what to expect; dissatisfaction with pain relief; attending a public clinic or shared care programme; having either a hospital doctor or midwife as the main caregiver for the birth; and having an episiotomy. Factors associated with significantly greater satisfaction among multiparae were: being enrolled for birth centre care, giving birth in a birthing room, and knowing the midwife who cared for you in labour. Three variables were inspected for evidence of linear trends: length of labour, obstetric procedure score and continuity of midwife care. Trends were present for lesser satisfaction with increasing length of time in labour among multiparae (χ2 for linear trend = 9.2, P= 0.03), and greater satisfaction with increasing levels of continuity of care among primiparae (χ2 for linear trend = 4.8, P= 0.3) and multiparae (χ2 for linear trend = 18.3, P < 0.01).

Pooling the data revealed associations of borderline significance with several variables: women who had either an operative vaginal birth or emergency caesarean section were more likely to be dissatisfied, while women who gave birth in a birthing room, attended midwives in a birth centre for antenatal care, and those who knew midwives caring for them in labour were less likely to give negative ratings of care. Although only a very small number of women had a general anaesthetic, pooling together primiparae and multiparae revealed a clear association with dissatisfaction.

Information and views about caregivers

The relationships between women's overall rating of intrapartum care and their views about information, assistance and reassurance received from caregivers are shown in Table 2. Both primiparae and multiparae were more likely to be dissatisfied if they thought that staff had not been very welcoming on their arrival at hospital in labour, if they were not given sufficient information, if either midwives or doctors had not been very helpful, or if they felt caregivers didn't offer them reassurance or encouragement. While 93.% of women agreed with the statement that midwives and doctors were very reassuring and encouraging, 21.% thought that staff had not been very welcoming on their arrival at hospital, and 24.% would have liked more information from caregivers during labour. Also striking is the proportion of women who rated caregivers as other than ‘very helpful’; 23% describing midwifery care, and 39.5% describing doctors’ care this way.

Table 2.  Relationships between access to information, views about caregivers and dissatisfaction with care in labour and birth by parity. Values are given as n (%), percentages and OR [95% CI].
 PrimiparaeMultiparae 
 n(%)%dissatisfiedORn(%)%dissatisfiedORWeighted OR
  1. * Odds ratio is unstable because of small numbers in cells.

Staff welcoming on arrival
  Very welcoming385 (75.6)20.01.00644 (79.6)20.21.001.00
  Fairly/not at all welcoming124 (245)54.04.79 [3.0–7.6]165 (20.4)61.26.24 [4.3–9.2]5.58 [4.2–7.5]
Midwives/doctors reassuring and encouraging
  Agree468 (92.9)24.71.00745 (93.6)24.71.001.00
  Disagree15 (3.0)73.38.34 [2.4–36.4]32 (4.0)81.213.21 [5.2–39.7]11.25 [5.3–24.5]
  Uncertain*21 (4.2)76.29.71 [3.3–34.5]19 (2.4)94.754.9 [8.5–2290.4]17.33 [6.9–46.4]
Wanted more information
  No350 (69.3)18.91.00636 (79.7)22.81.001.00
  Yes155 (30.7)51.04.47 [2.9–6.9]162 (20.3)49.43.30 [2.3–4.8]3.77 [2.9–5.0]
Helpfulness of midwives
  Very helpful398 (78.2)15.81.00612 (76.2)15.41.001.00
  Fairly helpful82 (16.1)68.311.45 [6.5–20.4]140 (17.4)67.911.63 [7.5–18.1]11.57 [8.2–16.3]
  Some/little/no help29 (5.7)89.746.08 [13.4–241.8]51 (6.5)76.517.90 [8.7–37.7]23.73 [12.7–44.5]
Helpfulness of doctors
  Very helpful276 (54.2)11.61.00430 (53.0)8.81.001.00
  Fairly helpful117 (23.0)47.97.00 [4.1–12.2]144 (17.7)56.313.26 [8.1–21.8]9.86 [7.0–14.5]
  Some/little/no help81 (15.9)60.511.68 [6–3–21.8]118 (14.5)60.215.58 [9.2–26.5]13.75 [9.3–20.5]
  Doctor not present35 (6.9)22.92.26 [0.9–5.8]120 (35–0)35.05.60 [3.3–9.5]4.43 [2.8–6.8]

Women's role in decision-making regarding their care

Another aspect of relationships with caregivers is the extent to which women's views are taken into account in making decisions about their care. Table 3 reports on some of the questions in the survey relevant to this issue. Over 96% of women stated they did want to be given a say in what happened during their labour and birth. The extent to which women perceived themselves as having had a say in decision-making was directly related to their overall rating of intrapartum care (i.e. the greater the involvement women had the more likely they were to rate their care positively). Having people present in the room that women would have preferred not to be there also increased the likelihood of less positive ratings of care. A quarter of primiparae and a smaller proportion of multiparae had prepared a written birth plan and discussed this with caregivers prior to labour: they were neither more nor less likely to be satisfied with their care.

Table 3.  Relationship between women's role in decision-making and dissatisfaction with care in labour and birth by parity. Values are given as n (%), percentages and OR 195% CI].
 PrimiparaeMultiparae 
 n (%)%dissatisfiedORn (%)%dissatisfiedORWeighted OR
Birthplan
  No378 (74.1)29.11.00673 (83.1)28.11.001.00
  Yes132 (25.9)27.30.91 [0.6–1.5]137 (16.9)31.41.17 [0.8–1.8]1.05 [0.8–1.4]
Given an active say in decisions about what happened in labour
  Yes, all cases185 (36.3)9.21.00322 (40–2)7.11.001.00
  Yes, most cases184 (36.0)26.63.59 [1.9–6.8]315 (39.3)32.46.23 [3.7–10.4]5.05 [3.4–7.5]
  Sometimes/rarely/not at all115 (22.5)64.317.84 [9.1–35.2]127 (15.9)69.329.30 [16.1–54.0]23.15 [15.2–37.1]
  Uncertain7 (1.4)28.63.95 [0.3–26.2]9 (1.1)33.36.50 [1.0–32.5]5.19 [1.5–17.5]
  Did not want active say19 (3.7)21.12.64 [0.6–9.6]28 (3.5)39.38.41 [3.2–21.8]5.36 [2.6–11.5]
Several times decisions taken without my wishes being taken into account
  Disagree387 (76.8)22.01.00662 (83.0)22.51.001.00
  Agree66 (13.1)54.54.26 [2.4–7.6]78 (9.8)62.85.82 [3.5–9.8]5.05 [3.5–7.4]
  Uncertain51 (10.1)43.12.70 [1.4–5.1]58 (7.3)51.73.69 [2.1–6.6]3.20 [2.1–4.9]
Unwanted people present at the birth
  No484 (95.7)27.11.00775 (96.4)27.61.001.00
  Yes22 (4.3)59.13.89 [1.5–10.1]29 (3.6)44.82.13 [1.0–4.8]2.75 [1.5–5.0]

Comparisons between models of care

In Table 4 comparisons are made between the six main models of care offered in Victoria in terms of factors significantly associated with overall ratings of intrapartum care. The main areas of significant difference between models relate to: the extent to which women knew the midwives who cared for them in labour prior to labour beginning, whether or not women perceived themselves as having an active role in decision-making, and the helpfulness of doctors. Compared with women cared for by a specialist obstetrician, women attending a general practitioner as either a public (OR 2.67 [95% CI 1.431) or a private patient (OR 2.2 [95% CI 1.–4.]), and women attending a birth centre (OR 9.11 95% CI 4.–19.1) were significantly more likely to know midwives very well prior to their labour beginning. Women attending birth centres were significantly more likely to have had an active say regarding decisions about what happened during their labour and birth than women cared for by obstetricians (OR 3.8 [95% CI 1.–34.]), while women receiving standard public clinic care (OR 0.4 [95% CI 0.–0.1) and women in shared care (OR 0.7 [95% CI 0.–0.1) were less likely to have had an active say about decisions regarding their care. None of the women who attended a birth centre agreed with the statement that ‘several times decisions were made without my wishes being taken into account’ (Fisher's exact test P= 0.3, odds ratio unable be calculated because of cell containing zero). In marked contrast, one in five women in standard public hospital care agreed with this statement (OR for comparison with women receiving care from a private obstetrician = 2.5 [95% CI 1.–3.1). There were no significant differences between models of care in terms of how helpful midwives were perceived to be, the extent to which women felt encouraged and reassured by caregivers, or felt welcomed when they arrived at the hospital. Women receiving public general practitioner care (OR 0.6 [95% CI 0.–0.]), standard public care (OR 0.4 [95% CI 0.–0.1) or shared care (OR 0.5 [95% CI 0.–0.1) were significantly less likely to find doctors very helpful than women receiving care from specialist obstetricians.

Table 4.  Comparisons between models of care. Values are given as n (%) GP = general practitioner.
 Private obstetricianPrivate GPPublic GPPublic clinicShared careBirth centre
  1. * Denotes result statistically significant at P < 0.5. Comparison is with women receiving care from a private obstetrician.

  2. † Excludes women not attended by a doctor in labour or at the birth.

Staff welcoming on arrival
  very569 (79.4)81 (75.0)143 (77.3)119 (75.8)88 (72.1)34 (87.2)
  Fairly134 (18.7)21 (19.4)36 (19.6)36 (22.9)31 (25.4)5 (12.8)
  Not at all14 (2.0)2 (1.9)6 (3.2)2 (1.3)3 (2.5)0 (0.0)
Midwives/doctors reassuring and encouraging
  Agree672 (95.2)99 (94.3)167 (92.3)138 (90.2)106 (87.6)36 (92.3)
  Disagree17 (2.4)3 (2.9)9 (5.0)7 (4.6)10 (8.3)1 (2.6)
  Uncertain17 (2.4)3 (2.9)5 (2.8)8 (4.1)5 (5.2)2 (5.1)
Midwives known before labour
  Very well56 (7.8)16 (15.2)*34 (18.5)*9 (5.7)4 (3.3)17 (43.6)*
  Not very well157 (21.9)26 (24.8)47 (25.5)28 (17.6)14 (11.5)15 (38.5)
  Not at all503 (70.3)63 (60.0)103 (56.0)122 (76.7)104 (85.2)7 (17.9)
Helpfulness of midwives
  very551 (77.7)83 (79.0)143 (78.1)117 (74.5)84 (70.6)33 (84.6)
  Fairly117 (16.5)17 (16.2)30 (16.4)29 (18 5)24 (20.1)6 (15.7)
  Some/little/no help41 (5.8)5 (4.8)10 (5.5)11 (7.0)11 (9.2)0 (0.0)
Helpfulness of doctorst
  very432 (66.2)63 (71.6)91 (56.2)*66 (46.2)*44 (40.4)*10 (83.3)
  Fairly143 (21.9)12 (13 6)40 (24.7)30 (21 0)34 (31.2)1 (8.3)
  Some/little/no help78 (11.9)13 (14.8)31 (19.1)47 (32.9)31 (28.4)1 (8.3)
Satisfaction with pain relief
  very happy194 (31.9)31 (31.3)55 (32.2)42 (30.2)23 (22.5)*17 (44.7)
  Happy275 (45.2)47 (47.5)81 (47.4)59 (42.4)50 (49.0)12 (315)
  Mixed unhappy139 (22.9)21 (21–2)35 (205)38 (27.3)29 (28.4)9 (23.7)
Wanted more information
  Did not want more536 (75.7)90 (85.7)146 (79.8)109 (69.9)84 (70.0)27 (69.2)
  Wanted more172 (24.3)15 (14.2)*37 (25.3)47 (30.1)36 (30.0)11 (28.2)
Able to hold baby soon after birth
  Yes624 (87.0)93 (88.6)158 (84.9)127 (79.9)101 (83.5)35 (89.7)
  No93 (13.0)12 (11.4)28 (15.1)32 (20.1)20 (16.5)4 (10.3)
Given an active say in decisions about what happened in labour
  Always or mostly556 (78.9)89 (85.6)149 (81.0)95 (62.1)*83 (70.9)*36 (94.7)*
  Sometimes/rarely/not at all123 (17.4)11 (10.6)30 (16.3)48 (31.4)32 (27.4)2 (5.3)
  Did not want active say26 (3.9)4 (3.8)5 (2.7)10 (6.5)2 (1.7)0 (0.0)
Several times decisions taken without my wishes being taken into account
  Agree75 (10.5)5 (4.8)20 (11.1)30 (19.9)*15 (12.4)0 (0.0)*
  Disagree591 (82.8)92 (87.6)149 (82.7)105 (69.5)85 (70.2)35 (89.7)
  Uncertain48 (6.7)8 (7.6)11 (6.1)16 (10.6)21 (17.4)4 (10.3)
Unwanted people present at the birth
  Yes24 (3.4)3 (2.9)8 (4.3)5 (3.2)7 (5.9)4 (10.3)*
  No688 (96.6)102 (97.1)176 (95.7)151 (96.8)114 (94.2)35 (89.7)

Contribution of birth events, birth experiences model of care and social factors to dissatisfaction

Fifteen factors reflecting birth events, experiences of caregivers and social characteristics were selected to be fitted in a logistic regression model with dissatisfaction as the outcome variable (Table 5, Model 1). Dissatisfaction was defined as descriptions of intraparturn care as other than ‘very good’. Parity, risk status and obstetric procedure score were included for a priori reasons related to their physiological significance and impact on the mother. The twelve other variables entered into the model were all significantly associated with dissatisfaction at a univariate level. Some variables significantly associated with the outcome variable were excluded because they measured a factor that could logically be viewed as a component of another variable; for example, how welcoming staff were on arrival at hospital in labour is clearly one component of the helpfulness of midwives, and decisions being taken without the mother's wishes being taken into account has a logical bearing on responses to the question about having an active say. Two variables (unwanted people being present, not finding midwives/doctors encouraging and reassuring) were excluded because only a small number of women gave these responses (< 5%). Health insurance status, and primary caregiver for the birth were excluded because they were highly correlated with model of care. Model of care was included in preference to health insurance status or primary caregiver, because it gave the most precise information from a policy or hospital planning perspective. Women who had had an elective caesarean section (n= 125), and therefore had not experienced labour were excluded because of missing or not applicable responses to some questions. A further 61 women were excluded because of missing values for at least one item. This left 1150 in the model for analysis; these women were not more or less likely to be dissatisfied than women excluded from the analysis (χ2= 0.2 df, P= 0.9).

Table 5.  Regression models describing the associations of parity, risk status, birth events and experiences with dissatisfaction with care in labour and birth* (see page opposite for footnote). Values are given as OR [95% CI]. GP = general practitioner; NE = not entered; ref = reference category.
 Model 1Model 2Model 3
 Unadjusted ORPAdjusted ORPAdjusted ORPAdjusted ORP
  1. * Unadjusted and adjusted odds ratios are reported for the maximum model (Model 1) containing all variables originally entered. Model 2 reports the best fit for the same group of variables, showing the adjusted odds ratios for variables retained in the model. Model 3 reports the best fit for the data when the following variables were entered: parity, risk status, obstetric procedure score, length of labour, experience of pain, country of birth, income, model of care and continuity of midwife care, showing adjusted odds ratios for variables retained in the final model.

Risk status
  Low1.00 ref0.41.00 ref0.21.00 ref0.21.00 ref0.5
  High0.88 [0.6–1.2] 0.73 [0.5–1.2] 0.75 [0.5–1–2] 1.88 [0.6–1.3] 
Parity
  01.00 ref0.91.00 ref0.31.00 ref0.31.00 ref0.01
  ≥11.01 [0.8–1.3] 1.26 [0.8–2.0] 1.24 [0.8–1.9] 1 50 [1.1–2.0] 
Obstetric procedure score
  0–61.00 ref0.41.00 ref0.31.00 ref0.31.00 ref0.6
  7–141.22 [0.9–1.7] 1.06 [0.7–1.7] 1.08 [0.7–1.7] 0–99 [0.7–1.4] 
  15–281.17 [0.8–1.7] 0.67 [0.4–1.2] 0.68 [0.4–1.2] 0 81 [0.5–1.2] 
Model of care
  Obstetrician1.00 ref0.0041.00 ref0.21.00 ref0.31.00 ref0.02
  Private GP0.93 [0.6–1.5] 1.00 [0.5–2.0] 0.96 [0.5–1.9] 1.04 [0.6–1.7] 
  Public GP1.08 [0.7–1′6] 0.93 [0.5–1.7] 1.03 [0.6–1.8] 1.22 [0.8–1.8] 
  Public clinic1.59 [1.1–2.4] 0.80 [0.4–13] 1.03 [0.6–1.8] 1.38 [09–2.1] 
  Shared care1.86 [1.2–2.91 1.20 [0.6–2.3] 1.40 [0.8–2.6] 1.89[1.2–3.0] 
  Birth centre0.33 [0.1–1.0] 0.17 [0.0–0.7] 0.25 [0.1–0.9] 0.34 [0.1–1.0] 
Midwives known before labour
  No1.00 ref0.0011.00 ref0.9NE 1.00 ref0.02
  Yes0.62 [0.5–0.8] 0.96 [0.6–1.5]   0.68 [0.5–0.9] 
Length of labour
  < 6 hours1.00 ref0.021.00 ref0.9NE 1.00 ref0.06
  6 to < 12 hours1.22 [0.9–1.7] 0.95 [0.6–1.5]   1.25 [0.9–1.8] 
  ≥ 12 hours1.55 [1.1–2.1] 1.09 [0.7–1.8]   1 56 [1.1–2.3] 
Experience of pain
  As expected or better1.00 ref< 0.0011.00 ref0.0091.00 ref0.0081.00 ref< 0.001
  Worse than expected2.30 [1.8–3.01 1.57 [1.0–2.4] 1.63 [1.1–2.5] 2.27 [1.7–3.0] 
  Uncertain2.59 [1344] 3.30 [1.5–7.41 3.06 [1.4–6.81 3.23 [1.8–5.7] 
Satisfaction with pain relief
  very happy1.00 ref< 0.0011.00 ref< 0.0011.00 ref< 0.001NE 
  Happy3.14 12.24.61 1.98 [1.2–3.31 1.95 [1.2–3.2   
  Mixed/unhappy8.52 [5.–12.6] 3.54 [2.0–6.2] 3.52 [2.–6.1]   
  Missing1.09 [0.3–3.8] 0.51 [0.1–2.5] 0.53 [0.1–2.6]   
Able to hold baby soon after birth
  Yes1.00 ref0091.00 ref0.5NE NE 
  No1.36 [1.0–2.0] 0.84 [0.5–1.5]     
Given an active say
  Always1 .00 ref< 0.0011.00 ref< 0.0011.00 ref< 0.001NE 
  Most cases4.98 [3.4–7.3] 2.55 [1.6–4.1] 2.65 [1.7–4.2]   
  Sometimes/rarely/not at all25.61 [16.4–40.1] 8.90 [4.9–16.1] 9.85 [5.5–17.5]   
  Uncertain1.22 [0.2–9.9] 0.31 [0.0–5.4] 0.36 [0.0–5.0]   
  Did not want an active say4.40 [2.0–9.8] 2.05 [0.7–5.7] 2.18 [0.8–6.0]   
Wanted more information
  No1.00 ref< 0.011.00 ref0.1NE NE 
  Yes3.48 [2.6–4.6] 1.44 [0.9–2.2]     
Helpfulness of doctors
  Very helpful1.00 ref< 0.0011.00 ref< 0.011.00 ref< 0.001NE 
  Fairly/some/not helpful10.60 [7.6–14.7] 6.76 [4.4–103] 6.86 [4.5–10.4]   
  Doctor not present3.95 [2.6–6.1] 5.31 [2.9–9.8] 5.24 [2.9–10.0]   
Helpfulness of midwives
  Very helpful1.00 ref< 0.0011.00 ref< 0.0011 00 ref< 0.001NE 
  Fairly/some/not helpful15.44 [11.0–21.61 12.03 [7.8–18.6] 11.91 [7.9–18.2]   
Total family income
  >AUS$30,0001.00 ref0.11.00 ref0.2NE NE 
  AUS$30,0001.32 [1.0–1.7] 1.50 [1.0–2.3]     
  Missing1.00 [0.6–1.8] 0.99 [0.4–2.7]     
Country of birth
  Australia1.00 ref0.011.00 ref0.3NE 1.00 ref0.1
Overseas
  English speaking1.26 [0.8–2.0] 1.37 [0.7–2.6]   1.24 [0.8–2.0] 
  NonEnglish speaking1.92 [1.3–2.91 1.75 [1.0–3.2]   1.70 [1.1–2.6] 
  Missing1.37 [0.5–3.7] 0.96 [0.2–4.3]   1.17 [0.4–3.3] 

Of the fifteen variables entered into the model, only six remained significantly associated with dissatisfaction after adjusting for all other factors: the helpfulness of midwives, helpfulness of medical staff, degree of involvement in decision making, experience of pain, satisfaction with pain relief, and model of care. Model 2 (Table 5) describes the best fit for the data retaining in the model parity, risk status and the score for obstetric procedures. Midwives being less than ‘very helpful’ was associated with an eleven fold increase in the odds of dissatisfaction after adjusting for other factors. Similarly, doctors being other than ‘very helpful’ raised the odds of dissatisfaction almost sevenfold. The odds of dissatisfaction were increased by a factor of almost 10 by having an active say in decisions in only some cases, rarely or not at all, and almost threefold by having an active say in most cases compared with the group of women who said they had an active say in all cases. There were no significant differences between models of care after adjusting for other factors, with the exception of women who attended birth centres who were significantly less likely to be dissatisfied compared with women attending a private obstetrician.

To give a more precise estimate of the effect of model of care and continuity of midwife care, one further model was fitted. Variables entered in Model 3 (Table 5) were 1. social factors and birth events associated with dissatisfaction (income, country of birth, length of labour, experience of pain); 2. the variables included in the first model for a priori reasons associated with their biological and physiological significance (risk status, parity, obstetric procedure score); and 3. model of care and continuity of midwife care as the variables of main interest. In this model, women who knew midwives prior to labour were significantly less likely to be dissatisfied and women receiving shared care were significantly more likely to be dissatisfied after adjusting for other factors. Also of note is the fact that in this model being born overseas of nonEnglish speaking background makes a significant independent contribution to dissatisfaction.

DISCUSSION

Methodological issues

The strengths of this study are that it is population based, and that it is large enough to enable statistical comparison between important sub-groups, including all major models of maternity care offered in Victoria, and women from differing social backgrounds (e.g. low/high income, metropolitadrural, Australian born/women of nonEnglish speaking background). Although there have been a number of innovations in the organisation of maternity services in Victoria over the last two decades, including the establishment of birth centres at all three tertiary level maternity hospitals; the introduction of shared care arrangements for antenatal care; and rotation of midwifery staff between antenatal clinic, labour and postnatal wards, none of these innovations has been evaluated in randomised trials in the Victorian setting. The main advantage of randomised trials as a technique of evaluation is that potential confounders, such as the risk status of the pregnancy and social background of the mother, are controlled for by the design. The inferences that may be drawn from observational studies are limited by the difficulties involved in assessing the role of known and unknown confounders. This is as true for assessments of satisfaction, as it is for ‘harder’ outcomes. Observational studies of women's views about maternity care, with only a few exception28,29, have tended to gloss over this difficulty. In this study multivariate techniques are used to clarify the relationships between variables. This enables us to pose and go some way towards answering several important questions not addressed in other observational studies. In particular, it has enabled us to sort out the extent to which extremely positive views of birth centre care, and more negative views of public clinic and shared care, may be attributed to social or obstetric characteristics of the women enrolled in these models. As with all observational studies, the generalisability of findings remains limited by the extent to which a satisfactory response rate can be achieved. The response rate of 62.% to this survey was disappointing compared with a similar survey in 1989 where the response rate was 71.%, but remained adequate for most sub-group comparisons20.

Various schools of thought exist on whether research on patient satisfaction is best conducted using standardised questionnaires that have been subjected to rigorous evaluation in terms of their validity and reliability30,31, or whether the disadvantages of standardisation (potential lack of relevance to local populations and services, inability to address policy issues as they arise) outweigh the putative advantages of greater technical reliability13,32. We chose not to use a standardised questionnaire in order to have greater flexibility to address issues arising out of the earlier Victorian survey conducted in conjunction with a Ministerial Review of Birthing Services4. Some questions such as the one which asked if women had an ‘active say’ in decision making replicated questions from the earlier survey. Others took up issues relevant to the current policy environment, including questions about degree to which women experienced continuity of midwife care.

Major influences on satisfaction

Two factors emerge in both the 1989 and 1993 Surveys as having the greatest negative impact on women's overall ratings of intrapartum care. These are the extent to which caregivers are perceived as helpful, and the degree to which women are given an active say in making decisions about their care. When the findings of our earlier survey were published, it was suggested to us that a proportion of women do not want an active say in decisions. In the present study we were able to test this hypothesis. The fact that fewer than 5% of women ticked the option to indicate a preference for decisions being made on their behalf is further reinforcement that having an ‘active say’ matters to most women.

Comparing different models of care

One reason for making comparisons between models of care is that if there are differences, it may be possible to learn from these what factors related to the organisation of care contribute to women having more positive experiences. In this study, and in 1989, women who booked care with a specialist obstetrician were no more or less likely to be happy with their intrapartum care than women receiving standard public hospital or general practitioner care18. This contrasts markedly with women's experiences of these models during pregnancy; in both 1989 and 1993 women attending public hospital clinics were far less likely to be happy with their antenatal care than women attending either specialist obstetricians or general practitioners17,19. The reasons for these differences are complex. One factor may be the lesser degree of continuity of care experienced by women attending private specialists once they go into labour. Whereas women going to a public hospital clinic for check-ups may see different caregivers at each visit, women seeing private obstetricians generally see the same specialist throughout their pregnancy, but once in labour may only see their doctor for one or two short periods before the birth, or in a small number of cases their doctor may not be in attendance at all. For women attending public hospital clinics this fragmentation of care is a feature of both antenatal and intrapartum experiences. Some Victorian public hospitals have sought to offset this fragmentation by rotating midwifery staff between public clinics, labour and postnatal wards, but our results suggest that these steps have so far had little impact on the degree to which women are likely to have had an opportunity to get to know labour ward midwives prior to labour.

In 1989 it was not possible to look in detail at the experiences of women in shared care programmes or birth centres as there were too few women in the sample enrolled in either of these models to permit meaningful comparisons. Even though the current study had low power to detect differences in satisfaction for the small group of women enrolled in birth centre care, the results consistently point towards this group having significantly more positive experiences, even after adjusting for potential confounders such as parity, risk status and extent of obstetric intervention. Women enrolled in birth centres were significantly more likely to report having an active say in decisions, to disagree with the statement that decisions were taken without their wishes being taken into account, and to have known midwives prior to labour-all factors strongly associated with greater satisfaction. They were also more likely to rate caregivers (midwives and doctors) as very helpful, although these differences did not reach statistical significance.

What lessons can be learned from the good experiences of women attending birth centres for care in other settings? It may be that women going to birth centres are more inclined to view their care positively because they have actively selected this form of care. However, if this were the sole explanation we might expect women attending private obstetricians–who have also been able to select their model of care–to report similar levels of satisfaction. Possibly the very positive experiences of women enrolled in birth centres have more to do with a good match between individual women and the environment. One of the distinctive features of birth centres as they operate in Victoria is that they each practise according to an explicit philosophy of care emphasising the involvement of parents in decision-making, and the promotion of pregnancy and birth as healthy life events with minimal need for intervention in the birthing process itself for the majority of women4. All centres also operate within the framework of formal guidelines which set out criteria for transfer and the modes of care practised in each centre (e.g. pain relief options, management of the third stage of labour). Issues covered in the guidelines are discussed with parents during antenatal care, and in antenatal classes which are strongly encouraged for primiparae and women attending a centre for the first time. Women enrolling in birth centre care therefore have many more opportunities to know beforehand what they might expect from caregivers in labour than women in other models that do not promote a particular philosophy or set of guidelines. Equally from a caregiver perspective, an explicit philosophy and set of guidelines for care must make it easier to provide continuity of care even when there is not actual continuity of carer.

The results for shared care are disappointing, and perhaps point to the effects of introducing a new model of care without giving sufficient attention to ways of minimising fragmentation of care when several sets of caregivers are involved. In contrast to Scotland and other parts of the United Kingdom where shared care is the standard form of care, shared care programmes were first developed in Victoria about 15 years ago4. Whereas in Scotland shared care is characterised by a high degree of co-operation between agencies about issues such as selection criteria for women at ‘low risk’ of complications, and guidelines for transfer to more specialised care6, standard protocols are yet to be fully developed for shared care in Victoria. Currently in Australia, shared care is being promoted as a way of easing the pressures on busy public hospital antenatal clinics4. We estimate that in 1993 at least 15% of pregnant women were enrolled in shared care, with most of the growth in this model of care accounted for by a decline in public clinic care. Our analysis of the survey findings relating to antenatal care indicated that women were far from happy with shared antenatal care: 33% of women in shared care rated their antenatal care as very good compared with 73% of women attending a private obstetrician, and 48% attending a public clinic19. The implications for women's experiences in labour when they have had limited contact with the public hospital where their baby will be born coupled with negative experiences of antenatal care have not been adequately explored and warrant urgent attention.

Continuity of midwife care does it make a difference?

In Australia, as in the United Kingdom, reviews of maternity services have all called for increased efforts to bring about greater continuity of midwife care2–5. In discussions of this question it is necessary to distinguish between continuity of carer, and continuity in the philosophy and content of care. Both are difficult to measure in a survey of this kind where questions necessarily entail a subjective judgement. Nonetheless, the data could not be more striking. The vast majority of women in all models of care, with the exception of birth centres, had had no contact with midwives caring for them in labour prior to arriving at the hospital. Even in birth centres, only 44% of women knew midwives very well, and 39% moderately well. Country hospitals serving small local communities probably account for the group of women seeing general practitioners who indicated they knew midwives prior to labour. Two questions are thrown up by this data: Why has it been so hard to implement greater continuity of midwife care? and Does it matter? To answer the latter first, our results suggests that knowing your midwife prior to labour does make a significant contribution to satisfaction independent of other factors including the model of care. Obstacles standing in the way of greater continuity of midwife care in the Australian setting mostly relate to the historic organisation of maternity care into separate teams of people providing antenatal, intrapartum and postnatal care8. Two Victorian hospitals (Monash Medical Centre and Royal Women's Hospital) are currently conducting randomised trials of team midwifery care involving small teams of midwives providing care in pregnancy, labour and during the postpartum hospital stay. In one hospital women at high and low risk of complications are eligible for care by the midwife team, while at the other only women at low risk are eligible. In both hospitals, the teams operate in the standard settings of public antenatal clinics and labour wards, with consultant involvement at a level appropriate to the risk status of the pregnancy. These two projects add to a small but growing body of literature which seeks to demonstrate the feasibility of rethinking organisational arrangements within the framework of existing services7,8,33,34,35

CONCLUSION

The survey results demonstrate the potential for new models of care to have either positive or negative effects on women's experiences of care underlining the need for all innovations in perinatal care to be accompanied by rigorous evaluation of their impact on service users.

Acknowledgements

Analysis of the survey data for this paper was undertaken while Ms S. Brown was a Visiting Research Fellow at the Dugald Baird Centre for Research on Women's Health, Department of Obstetrics and Gynaecology, University of Aberdeen.

The authors would llke to thank the following: the women who completed the Survey; the Victorian Health Promotion Foundation which funded the study; the National Health and Medical Research Council which supported Ms Brown with a travelling fellowship; the hospitals and home birth practitioners who distributed the questionnaires; Ms F. Bruinsma for assistance with preparing the data for analysis; and the Victorian Perinatal Data Collection Unit for providing data on obstetric and social characteristics of women who gave birth in the study period. The authors would also like to acknowledge the assistance of the following members of the project reference group: Ms M. A. Biro, Professor S. Brennecke, Ms F. Gardner, Dr J. Gunn, Ms K. Lamb, Dr C. Longman, Ms V. McCutcheon, Ms A. Paul, Ms D. Smith, Dr C. Tippett.

Appendices

Appendix 1: Models of care in Victoria

Private obstetrician care. Women choosing this model (approximately half of all confinements) attend an obstetrician's private consulting rooms for antenatal check-ups, and are cared for in labour by the same obstetrician.

General practitioner care. Women may attend a general practitioner obstetrician as a private patient (private general practitioner care) receiving care in pregnancy, intrapartum and postnatally from the same practitioner. Public general practitioner care is mainly offered in rural areas; women attend a general practitioner or group practice in pregnancy, and receive standard public hospital care for labour and delivery with local general practitioners providing an on call service for public patients. Approximately 20% of women receive general practitioner care, as either a private or public patient.

Public clinic care. Pregnancy care is provided through a public hospital outpatients clinic. Women attend the same hospital for labour and the postnatal hospital stay. Generally different sets of caregivers are responsible for antenatal, intrapartum and postpartum care. In labour, multiparous women are generally cared for by midwives, while primiparous women are attended by medical staff in addition to midwives. This model accounts for approximately 10% to 15% of births.

Shared care. After an initial visit with a consultant obstetrician at a public hospital clinic, the remainder of pregnancy care is provided by a local general practitioner or midwife/focal physician team in a community health centre, with the exception of two or more visits to the hospital at the 28th and 36th week of pregnancy. In labour, women are cared for by hospital staff (midwives, senior doctor and specialist obstetrician if necessary), returning to their local practitioner for care following discharge. We estimate that this model accounted for < 2% of births in 1989, expanding to approximately 15% by 1993.

Birth centre cure. In this study refers to team midwifery care within a separate section of a hospital where midwives provide antenatal, intrapartum and postpartum care unless complications arise requiring transfer to obstetrician led care. An obstetrician is seen at the first antenatal appointment, with subsequent visits booked at one hospital only if complications arise, and at the other two hospitals offering this model at prespecified times late in pregnancy e.g. 36 weeks, post term. We estimate this model accounted for < 2% of births in 1993.

In general only women at low risk of complications are accepted for shared care or birth centre care, although the exclusion criteria vary across centres and tend to be less restrictive for shared care than for birth centre care. For example, a woman who had previously had a ceasarean would be excluded from birth centre care, but not from all shared care programmes.

The group of women with a midwife as their primary caregiver for the birth is a heterogenous category comprising primarily women giving birth in birth centres receiving team midwifery care, multiparous women receiving public clinic or shared care, and women whose private obstetrician was unable to attend the birth. The smaller group of women who stated that a doctor was not present during labour or the birth is a subset of this category.

Appendix 2: Obstetric procedure score

Details of the scoring system applied to obstetric procedures are as follows:

ProcedureScore
Induction/augmentation2
Pethidine1
Epidural4
CTG monitoring
  Routine at the start of labour1
  External monitor only(≤ 1 hour)2
  External monitor only(> 1 hour)3
  Internal monitor4
Vaginal examination
  11
  > 12
Forceps5
Vacuum extraction3
General anaesthesia6

Ancillary