- Email: [email protected]
A comparison of partnership caseload midwifery care with conventional team midwifery care: labour and birth outcomes
A comparison of partnership caseload midwifery care with conventional team midwifery care: labour and birth outcomes Yvonne Benjamin, Denis Walsh and NickTaub Objectives: to compare the e¡ects of partnership caseload midwifery care, with conventional team midwifery care. Comparisons of labour interventions and birth outcomes were made between the two models of care. Design: a prospective, non-randomised clinical trial. Setting: Women’s Hospital at Leicester Royal In¢rmary, Leicester,UK in 1998. Participants: 303 women from the experimental group and 308 from the control group (611 in total) matched for age, ethnicity, marital status, parity, gravida and height who gave birth between April 1997 and August 1998. Intervention: the control group received conventional team midwifery care during pregnancy, labour and birth, and the experimental group received care from midwives working in partnerships that provided continuity of care during pregnancy, labour and birth.
Yvonne Benjamin Midwife, Neonatal Unit, Leicester General Hospital, Gwendolyn Rd, Leicester Denis Walsh Senior Lecturer in Midwifery, De Montfort University, Charles Frears Campus, 266 London Rd, Leicester LE2 1RQ, UK E-mail: [email protected] NickTaub Lecturer in Medical Statistics, Department of Epidemiology, Leicester University, Leicester (Correspondence to DW) Received 17 August 2000 Revised 24 November 2000, 8 December 2000 Accepted 25 January 2001 Published online 29 May 2001
Key ¢ndings: 21% of women in the experimental group had an epidural compared with 32% of the controls (OR 0.56 95%,CI 0.39^0.81, P¼0.002).The normal vaginal birth rate (74% v 66%, OR 1.45, 95% CI 1.02^2.05, P¼0.038), upright birth posture rate (60% v 14%,OR 9.64, 95% CI 5.96^15.61, P¼50.001), intact perineum rate (40% v 30%,OR 1.57, 95% CI 1.05^2.35, P¼0.027), and physiological third stage rate (37% v 1.5%,OR 38.69, 95% CI 11.98^124.89, P¼50.001) were signi¢cantly higher in the experimental group.The induction of labour rate (16% v 23%,OR 0.66, 95% CI 0.44^0.98, P¼0.042) was signi¢cantly lower in the experimental group.Women in the experimental group had more home births (17% v 1.3%,OR 15.38. 95% CI 5.48^43.14, P¼50.001); used the midwife-led birthing suite more often (28% v 12%,OR 2.77, 95% CI 1.82^ 4.22, P¼50.001); were more likely to take an early discharge (two to six hours) from hospital following birth (25% v 3%,OR11.32. 95% CI 5.55^23.06, P¼50.001); and were attended in birth more often by either their named midwife (67% v 5%,OR 39.65, 95% CI 22.38^70.25, P¼50.001) or her partner (known midwife) (84% v 14%,OR 32.74,CI 20.96^51.14, P¼50.001). Implications for practice: partnership caseload midwifery care resulted in less interventionist labour and more normal birth than conventional team midwifery care. Women in the experimental group had more home births, birth in a midwife-led suite and opted for early discharge home postnatally more often than the controls.They also experienced much higher levels of continuity, particularly of a known midwife during labour and birth.The study ¢ndings should encourage other maternity units in the UK to pilot and evaluate the model to see if these bene¢ts are transferable. & 2001Harcourt Publishers Ltd
INTRODUCTION Since the Report of the Expert Committee on Maternity Care (DoH 1993) set as an ‘Indicator of Success’ that 75% of women should know the Midwifery (2001) 17, 234 ^240 & 2001 Harcourt Publishers Ltd doi:10.1054/midw.2001.0257, available online at http://www.idealibrary.com on
midwife who cares for them in labour, many maternity units throughout England have endeavoured to implement continuity of carer schemes. Very few have actually succeeded in reaching the target (Audit Commission 1997,
Comparison of partnership caseload midwifery care with conventional team care
Page et al. 1999, North Staffordshire et al. 2000), causing authors to question the target’s evidence base (Warwick 1997, Green et al. 2000). Although there is very good evidence that continuity of support during labour from a trained professional is advantageous (Hodnett 2000a), to date there are very few evaluations of the merit of a woman being cared for during labour by a professional with whom a relationship has been formed antenatally. In fact, Page et al. (1999) and North Staffordshire et al. (2000) studies are the only published findings to date that achieved the DoH (1993) target. The schemes they (Page et al. 1999, North Staffordshire et al. 2000) evaluated involved midwives working in pairs or triplets, providing all the antenatal, intrapartum and postnatal care of women booked with them. The models were dubbed ‘partnership caseload midwifery practice’. Page et al.’s (1999) study was a nonrandomised comparison with conventional community midwifery care and they found statistically significant results in favour of partnership caseload practice for less epidural use, less episiotomies and perineal lacerations. The North Staffordshire et al. (2000) study had an element of randomisation, albeit geographical areas not research subjects, so was a stronger design. They showed significant reductions in epidural and augmentation rates with caseload care. Women’s views strongly favoured labour continuity from a known midwife in Page et al.’s (1999) study. Other studies of caseload midwifery (Farquhar et al. 1996, Allen et al. 1997, Pankhurst 1997) have been published, but none have achieved the DoH (1993) target for continuity during labour and, therefore, their findings are less relevant to this study. In order to achieve the standard of a 75% known midwife for intrapartum care, midwives’ work patterns and caseload size have to be radically altered from conventional community midwifery practice. The caseload midwife has to be on-call for extended periods of time and caseload numbers of women per midwife need to be around 40 per year (Walsh 1995). Such radical departure from traditional practice has been a major stumbling block for maternity units who either cannot achieve the caseload ratios per midwife or who do not have midwives willing to try the new model. For these reasons, it is imperative that robust evaluations do occur when pilot schemes are launched. The purpose of the study reported here was to evaluate a partnership caseload practice scheme, launched in 1996 at a maternity unit in the east midlands of England. Named ‘Birth Under Midwifery Practice Scheme’ (BUMPS), it consisted of three pairs of midwives who provided total care for a defined caseload of women attached to specific general practitioner (GP)
235
practices so that each midwife attended 40 births/year. The BUMPS project thus enabled women to get to know their two midwives during pregnancy and to be cared for during labour by them. Postnatally, the same two midwives would attend. The project provided midwife-led care for women ‘booked’ as low risk and consultant-led care for women with risk factors. Traditional team midwifery care at the maternity unit provided a community midwife, linked to specific GP practices, who provided the bulk of ante/postnatal care and labour care by one of a large team of 25 midwives which included the community midwife. Midwife-led care and consultant-led care was based on the same risk criteria applied by the BUMPS project. BUMPS midwifery care was compared with team midwifery care.
METHODS Research design A comparative, prospective, non-randomised design was chosen because it was not feasible to mount a randomised controlled trial. As entry to the maternity services was via the women’s general practitioners (GPs), randomisation would have to have occurred at that time. For most women within both care packages, the majority of their care took place in a primary care setting and this is in contrast to earlier randomised controlled trials that evaluated similar aspects of the midwifery care package, e.g. continuity of care, midwife-led care and midwife-led areas (Hodnett 2000a,b). These studies were able to randomise at hospital level and then offer two contrasting packages of care that predominately occurred at the hospital. In this study, if women had been randomised on ‘booking’ with their GP, two parallel systems of care would have had to be in place at this level. This could either have been the same midwife offering different packages of care to a control and experimental group or two separate groups of midwives offering the alternative package to a control or experimental group. Neither of these options were feasible – the first because contamination between the care packages would have been highly likely, and the second because it was not organisationally possible to install parallel care systems at GP level. The primary hypothesis to be tested was that partnership caseload practice would reduce the epidural rate in this group compared with conventional team midwifery. Epidural rate was chosen as a primary outcome measure because it represents a significant marker of medical intervention in labour and birth. Previous randomised controlled trials of midwifery-led care and
236
Midwifery
continuity of carer, components of the BUMPS package, showed a lowering of epidural rates (Hodnett 2000a,b), and two other previously published studies of partnership caseload practice also adopted it as an important outcome measure, Page et al. (1999) as the primary outcome measure and North Staffsorshine et al. (2000) as a secondary measure. Based on the unit epidural rate of 30%, a sample size of 300 women in each group would need to be recruited to demonstrate a reduction of 10% in the experimental group with 80% power at a 5% level of significance. This is one of the accepted benchmarks of clinical significance. Ethical approval was granted by the Local Research Ethics Committee.
Table 1. Characteristics of women at trial entry. Values are given as mean [SD] or n (%) except gravida & parity Characteristics
Age (years) Height (m) Parity Gravida Ethnic group white Afro-Caribbean Indian/Pakistani other/not known Marital status married/widowed single supported (living with partner) single unsupported not known/ divorced/separated
Study group n¼303 29.5 [5.3] 1.55^1.64 52.4%41 60.4%41
Control group n¼308 27.3 [5.2] 1.55^1.64 51.7%41 62.0%41
263 (87) 5 (2) 17 (6) 18 (6)
287 (93) 4 (2) 11 (4) 5 (2)
195 (64) 92 (30)
150 (49) 139 (45)
10 (3) 6 (3)
3 (1) 16 (5)
Outcomes The primary outcome measure was epidural rate. Secondary measures of outcomes were delivery outcome, induction of labour rate, use of syntocinon for augmentation of labour, perineal outcomes, maternal position for birth, thirdstage management, place of birth, neonatal Apgar scores at one and five minutes and admission to the neonatal unit. Data were also collected on continuity of carer.
and 14 weeks’ gestation. At 28 weeks’ gestation women were invited and consented to participate in the study after guarantees of anonymity and confidentiality. Written informed consent was required prior to recruitment. Recruitment continued until 31 July 1998, by which time 303 women had been recruited from the BUMPS project and 308 from the control group.
Data collection and management Eligibility and recruitment Prior to the study, women from the specific GP practices served by the BUMPS midwives were matched prospectively with women from the specific practices serving the control group who were experiencing conventional team midwifery care. Homogeneity was attained for age, marital status, ethnicity, parity and height, but not for marital (64% v 47%) or single-supported status (30% v 44%) (Table 1). However, 94% of the study group and 91% of the control group were living with their partner, which diminishes the likelihood of these differences causing selection bias. Similarity of social class had been previously addressed via the advice given by the Health Authority as to the comparability of the two areas of the city that women were drawn from. All women who booked either with the conventional midwifery team (control group) or the BUMPS project (experimental group) from 1 April 1997 onwards, were eligible for inclusion in the study except those women who subsequently miscarried or gave birth before 28 weeks gestation. Women with any other risk factors were eligible for inclusion. Therefore, women at both high and low obstetric risk were eligible for inclusion in the study. These exclusion/inclusion criterium replicated the McCourt and Page (1996) study. Women were given an information leaflet at their first attendance at the hospital, between 10
It was obviously not possible to keep women unaware as to which arm of the study they were in, but they were recruited from the south and west of the city with a three-mile corridor separating them. Some degree of contamination could have occurred between the groups, although this never came to the attention of the midwives or the researcher. Nor could the researcher (a research midwife appointed from within the unit) be blinded to the arms of the study. Therefore, both maternal behaviour and data collection could have introduced some bias. An external statistician who was blinded to the respective groups rechecked the data analysis. Data were collected via a clinical audit tool previously validated by the Page et al.’s study (1999). Additional parameters, e.g. birth position, were added and these were collected by a very robust hospital IT system, Euroking. In addition to clinical process and outcome data, maternal satisfaction was measured at four weeks and three months postnatally, and these findings will be the subject of a subsequent paper. An ethnographic study of the partnership caseload practice arm of the trial was also carried out and has been reported (Walsh 1999).
Data analysis Logistic regression modelling was used to compare the study and control groups. Odds
Comparison of partnership caseload midwifery care with conventional team care
237
Women cared for by the BUMPS midwives were more likely to adopt an upright birth posture (60% v 14% OR 9.64, 95% CI 5.96– 15.61, P¼50.001), have an intact perineum (40% v 30% OR 1.57, 95% CI 1.05–2.35, P¼0.027), and experience a physiological third stage (37% v 1.5% OR 38.69, 95% CI 11.98– 124.89, P¼50.001), although the postpartum haemorrhage rate for these women was lower in comparison to the control group (0.4% v 3% OR 0.13 95% CI 0.02–1.04, P¼0.054). The episiotomy rates were also lower in the experimental group (6% v 10% OR 0.57 95% CI 0.27– 1.17, P¼0.126) (Table 3). Place of birth findings were contrasting. The experimental group had far more home births (17% v 1.3% OR 15.38. 95% CI 5.48–43.14, P¼50.001) and used the midwife-led birthing suite more often (28% v 12% OR 2.77, 95% CI 1.82–4.22, P¼50.001). They were also more likely to take an early discharge from hospital (two to six hours) from hospital following birth (25% v 3% OR 11.32. 95% CI 5.55–23.06, P¼50.001) (Table 4). There were no differences between the groups in relation to neonatal outcomes. Apgar scores of 56 at one minute and 56 at five minutes respectively were comparable and both groups had 10 admissions to the neonatal unit. Each group had one perinatal death (Table 5).
ratios, confidence intervals, and P-values were calculated using the statistical software SPSS for Windows, Version 7. P-values of less than 0.05 indicated statistical significance. When tested through logistic regression, the effects of potential confounding variables (primigravida, midwife only care, midwife/GP care) reinforced the findings of the unadjusted results. Therefore, only unadjusted results are presented in this paper. Full adjusted results are available from the researcher.
FINDINGS The finding for the primary clinical outcome was that fewer women in the experimental group had an epidural compared with the controls (21% v 32% OR 0.56 95%, CI 0.39–0.81, P¼0.002). These women also had more normal births (74% v 66% OR 1.45, 95% CI 1.02–2.05, P¼0.038) and fewer caesarean sections (15% v 19% OR 0.75, 95% CI 0.49–1.15, P¼0.186), although the latter was not statistically significant. Their induction of labour rate was lower (16% v 23% OR 0.66, 95% CI 0.44–0.98, P¼0.042), but they experienced slightly higher rates of syntocinon augmentation (14% v 13%, OR 1.08, 95% CI 0.67–1.73, P¼0.750) (Table 2).
Table 2 Birth process and outcome data Study group n¼303 n % Epidural Normal vaginal birth Caesarean section Induction Syntocinon augmentation
61 224 44 49 41
Control group n¼308 n %
OR
95% CI
P
98 204 57 71 39
0.56 1.45 0.75 0.66 1.08
0.39^ 0.81 1.02^2.05 0.49^1.15 0.44 ^ 0.98 0.67^1.73
0.002 0.038 0.186 0.042 0.750
21 74 15 16 14
32 66 19 23 13
Table 3 Birth process and outcome data (normal spontaneous deliveries only). PPH¼primary postpartum haemorrhage Study group n¼224 n % Upright birth posture Intact perineum Episiotomy PPH Physiological 3rd stage
135 90 13 1 82
60 40 6 0.4 37
Control group n¼204 n % 28 61 20 7 3
14 30 10 3 1.5
OR
95% CI
P
9.64 1.57 0.57 0.13 38.69
5.96 ^15.61 1.05^2.35 0.27^1.17 0.02^1.04 11.98 ^124.89
50.001 0.027 0.126 0.054 50.001
OR
95% CI
P
15.38 2.77 11.32
5.48 ^ 43.14 1.82^ 4.22 5.55^23.06
50.001 50.001 50.001
Table 4 Place of birth data. (all births) Study group n¼303 n % Home Midwife-led birthing suite 2^ 6 hour discharge
51 85 76
17 28 25
Control group n¼308 n % 4 38 9
1.3 12 3
238 Midwifery
Table 5 Neonatal outcome data. (All births) NNU¼Neonatal Unit Study group n¼303 n % Apgar 56 at1 minute Apgar 56 at 5 minutes Admission to NNU Perinatal death
19 0 10 1
6.7 3.3 0.33
Control group n¼308 n %
OR
95% CI
P
9.6 0.7 3.2 0.32
0.70 0 1.02 1.02
0.38 ^1.28 ^ 0.47^2.48 0.06^16.32
0.245 0.847 0.970 0.991
Control group n¼308 n %
OR
95% CI
P
22.38 ^70.25 20.96 ^51.14 4.16 ^ 8.57
50.001 50.001 50.001
27 2 10 1
Table 6 Continuity of carer data (all births) Study group n¼303 n % Named midwife for labour Known midwife for labour Known midwife postnatally
203 255 242
67 84 80
15 43 123
Significant differences were also apparent in each group’s experience of continuity of midwifery care. Women on the BUMPS Scheme were attended at birth more often by either their named midwife (67% v 5% OR 39.65, 95% CI 22.38–70.25, P¼50.001) or her partner, known midwife (84% v 14% OR 32.74, CI 20.96–51.14, P¼50.001). Thus 86% of the control group were cared for during labour and birth by a midwife they had never met before compared with just 16% of the experimental group. Continuity postnatally was also much better for the BUMPS’ women. Eighty per cent received care from a midwife they knew compared with 40% of the controls (OR 5.97 95% CI 4.16–8.57, P¼50.001) (Table 6).
DISCUSSION Caution must be adopted in any attempt to generalise from these findings, given that this was a non-randomised design. The possibility of selection bias has to be acknowledged, though the homogeneity of both groups is reassuring (Table 1). Women did not have a choice regarding BUMPS project care. To opt out of the scheme, they would have had to book out of area with different GPs. The possibility of selection bias regarding place of birth is key to the robustness of findings because epidural analgesia is not available at home nor in the midwife-led birthing suite. We believe that the BUMPS women did not have a predilection to choose these options. The homebirth and midwife-led birthing suite use in the two years prior to the project was similar in both groups of women. We suggest that the contrasting findings in this study were due to the nature of the scheme, which emphasised continuity of carer through all phases of care. In particular, being
5 14 40
39.65 32.74 5.97
cared for by a known midwife during labour and birth facilitated women making independent choices as to environment for birth. This is clear from the qualitative study previously alluded to (Walsh 1999). The researcher was not blinded to the treatment streams, so this may also have led to assessment bias. However, some data entry and all of the analysis was cross-checked by an independent statistician, reducing possible bias effects. It is also possible that the characteristics of the midwives themselves may explain outcome differences. The BUMPS midwives were a selfselected group and this had to be so because their method of working was so different to a conventional community midwife. It would not have been possible to compel midwives to adopt such a work pattern. Other randomised studies offering alternative care packages comment on this potential confounder (Flint & Poulengeris 1989, Hodnett 2000b). If the midwives at the outset shared a common philosophy of care, it may influence how care is given and affect outcomes. Alternatively, it is also possible that, as an individual midwife practices over time, the nature of the scheme itself, e.g. the relationship she builds with women, influences her approach to care. The findings confirm those of Page et al. (1999) that women cared for by partnership caseload midwifery schemes have less interventionist labours than conventional care. Waldenstrom and Turnbull’s (1998) review of randomised controlled trials that evaluated elements of the BUMPS package of care, i.e. continuity of care, midwife-led care and midwife-led birthing facilities also found this noninterventionist trend. What is different with this study is that a statistically significant higher normal birth rate was found. Though earlier studies have often hypothesised that continuity schemes and midwifery-led schemes would
Comparison of partnership caseload midwifery care with conventional team care
improve normal birth rates, all have failed to confirm this (Flint & Poulengeris 1989, Hundley et al. 1994, Turnbull et al. 1996). Lower induction of labour rates and less use of epidural analgesia might have contributed towards the higher normal birth rate as both have been linked to more operative deliveries (Goer 1995, Howell 2000). The fall in epidural rate is dramatic and may, in part, be explained by women in the experimental group choosing to give birth either at home or in the Home from Home facility, both of which preclude epidural use. McVicar et al.’s (1993) randomised controlled trial of the Home from Home Scheme showed significantly lower epidural use in this facility. The significantly higher intact perineum rate for normal birth with BUMPS care is an interesting finding. Variation in episiotomy rates may explain some of these differences. An alternative explanation may lie with the data on position for birth. Sixty per cent of women in the BUMPS group adopted upright postures for birth compared with just 14% of the control group. Some studies of alternative versus recumbent birth position (Gardosi et al. 1989, Sermer & Raines 1997) have shown a non-significant trend towards intact perineums in the upright posture groups. The episiotomy rates in the study groups are probably among the lowest in the UK (Clinical Standards Advisory Group 1995), comparing favourably with the World Health Organization’s (1996) recommendation of 10% for normal birth. The experimental group had a lower postpartum haemorrhage rate for normal birth despite having a much higher physiological third-stage rate. All the trials on active versus physiological third stage indicate a higher postpartum haemorrhage rate with the latter (McDonald et al. 2000), so this is a surprising finding. Neonatal outcomes showed no differences between the study groups with a neonatal admission rate of 3% in each. There was one perinatal death in each group so no conclusions can be drawn regarding perinatal mortality. In this study the level of continuity of care for women in the BUMPS group exceeded the national target of 75%, with 84% of women receiving care from a known midwife, compared to just 14% of women in the control group. This higher level of continuity of caregiver concurs with Page et al.’s (1999) and the North Staffordshire et al.’s (2000) findings. In attempting to explain the differences in interventions and outcomes between the two groups, Walsh’s (1999) parallel qualitative study of the BUMPS women’s experience of care is enlightening. Women spoke about their birth experiences as empowering and feeling in control. They felt that their midwives acted as their
239
advocates, particularly in representing their desire for non-intervention during labour. Knowing antenatally that they would be attended in labour by a midwife they knew instilled confidence about choosing a home birth and about developing birth plans. The greater level of self-management of their own births is consistent with the less interventionist labours they had and is further illustrated by the common use of upright posture for birth and the frequent choice of a natural third stage. McCourt and Page (1996) and Pankhurst (1997) both found higher levels of maternal satisfaction with the partnership caseload model when compared with conventional care. Maternal satisfaction is an important dimension of evaluations of service developments and should be appraised alongside clinical outcomes when decisions are made regarding the ‘roll-out’ of pilot schemes. The absence of published data in this area for this project weakens its generalisability. In conclusion, although the implications for other UK maternity units are mitigated by the research design, the results should encourage them to set up and evaluate their own partnership caseload schemes. ACKNOWLEDGEMENTS We would like to thank the following people who assisted with this study: BUMPS Midwives, Bradgate Team midwives, Margaret Stockwell (Senior Midwife), Christina Oppenheimer, Roger Neuberg (Consultant Obstetricians), Rosie Garrett, Melanie Thwaites (Midwifery Team Leaders), Fran Gill (IT Manager), all general practitioners from the women’s Health Centres, The Leicester Royal Infirmary NHS Trust Research Fellowship Scheme that funded the Project.
REFERENCES Allen I, Dowling S, Williams S 1997 A leading role for midwives. Policy Studies Institute, London Audit Commission 1997 First class delivery: improving maternity services in England & Wales. Audit Commission, London Clinical Standards Advisory Group (CSAG) 1995 Women in normal labour. HMSO, London Department of Health 1993 Changing Childbirth: the report of the expert committee on maternity care. HMSO, London Farquhar M, Camillerie-Ferrante C, Todd C 1996 An evaluation of midwifery teams in West Essex. Final report. PHRU & HSRG, Institute of Public health, University of Cambridge, Cambridge Flint C, Poulengeris P 1989 The know your midwife report. Available from 49 Peckermans Wood, London Gardosi J, Hutson N, Lynch C 1989 Randomised controlled trial of squatting in the second stage of labour. The Lancet 2: 74–77 Goer H 1995 Obstetric myths versus research realities: a guide to the medical literature. Bergin & Garvey, London Green J, Renfrew M, Curtis P 2000 Continuity of carer: what matters to women. Midwifery 16(3): 186–196
240 Midwifery Hodnett ED 2000a Support from caregivers during childbirth. (Cochrane Review) In: The Cochrane Library, Issue 3, Update Software, Oxford Hodnett ED 2000b Home-like versus conventional birth settings. (Cochrane Review) In: The Cochrane Library, Issue 3, Update Software, Oxford Howell C 2000 Epidural v non-epidural analgesia in labour. (Cochrane Review) In: The Cochrane Library, Issue 3, Update Software, Oxford Hundley V, Cruickshank F, Lang G et al. 1994 Midwife managed delivery unit: a randomised controlled comparison with consultant led care. British Medical Journal 309: 1400–1404 McDonald S, Prendiville W, Elbourne D 2000 Expectant v active management of the third stage of labour. (Cochrane Review ) In: The Cochrane Library, Issue 3, Update Software, Oxford McVicar J, Dobbie G, Owen-Johnston L et al. 1993 Simulated home delivery: a randomised control trial. British Journal of Obstetrics & Gynaecology 100: 3316–3323 McCourt C, Page L 1996 The one-to-one project, final report. Thames Valley University, London North Staffordshire Changing Childbirth Research Team 2000 A randomised study of midwifery caseload care and traditional ‘shared-care’. Midwifery 16(4): 295– 302 Page L, McCourt C, Beake S, Hewison J 1999 Clinical interventions and outcomes of one-to-one midwifery
practice. Journal of Public Health Medicine 21(3): 243–248 Pankhurst F 1997 Caseload midwifery: an evaluation of a pilot scheme, final report. University of Brighton, Brighton Sermer R, Raines D 1997 Positioning during the second stage of labour: moving back to basics. Journal of Obstetric, Gynaecologic and Neonatal Nursing 26(6): 727–734 Turnbull D, Holmes A, Cheyne H et al. 1996 Randomised, controlled trial of efficacy of midwife-managed care. The Lancet 348: 213–218 Waldenstrom U, Turnbull D 1998 A systematic review comparing continuity of midwifery care with standard maternity services. British Journal of Obstetrics & Gynaecology 105: 1160–1170 Walsh D 1995 The Wistow project, final report. Women’s Hospital, LRI NHS Trust, Leicester Walsh D 1999 An ethnographic study of women’s experience of partnership caseload midwifery practice: the professional as friend. Midwifery 15(3): 165– 176 Warwick C 1997 Open dialogue. Can continuity of care be the only answer? British Journal of Midwifery 5(1): 6 World Health Organization 1996 Care in Normal Birth: A Practical Guide. Report of a Technical Working Group. Maternal and Newborn Health/Safe Motherhood Unit, Geneva
Yvonne Benjamin Midwife, Neonatal Unit, Leicester General Hospital, Gwendolyn Rd, Leicester Denis Walsh Senior Lecturer in Midwifery, De Montfort University, Charles Frears Campus, 266 London Rd, Leicester LE2 1RQ, UK E-mail: [email protected] NickTaub Lecturer in Medical Statistics, Department of Epidemiology, Leicester University, Leicester (Correspondence to DW) Received 17 August 2000 Revised 24 November 2000, 8 December 2000 Accepted 25 January 2001 Published online 29 May 2001
Key ¢ndings: 21% of women in the experimental group had an epidural compared with 32% of the controls (OR 0.56 95%,CI 0.39^0.81, P¼0.002).The normal vaginal birth rate (74% v 66%, OR 1.45, 95% CI 1.02^2.05, P¼0.038), upright birth posture rate (60% v 14%,OR 9.64, 95% CI 5.96^15.61, P¼50.001), intact perineum rate (40% v 30%,OR 1.57, 95% CI 1.05^2.35, P¼0.027), and physiological third stage rate (37% v 1.5%,OR 38.69, 95% CI 11.98^124.89, P¼50.001) were signi¢cantly higher in the experimental group.The induction of labour rate (16% v 23%,OR 0.66, 95% CI 0.44^0.98, P¼0.042) was signi¢cantly lower in the experimental group.Women in the experimental group had more home births (17% v 1.3%,OR 15.38. 95% CI 5.48^43.14, P¼50.001); used the midwife-led birthing suite more often (28% v 12%,OR 2.77, 95% CI 1.82^ 4.22, P¼50.001); were more likely to take an early discharge (two to six hours) from hospital following birth (25% v 3%,OR11.32. 95% CI 5.55^23.06, P¼50.001); and were attended in birth more often by either their named midwife (67% v 5%,OR 39.65, 95% CI 22.38^70.25, P¼50.001) or her partner (known midwife) (84% v 14%,OR 32.74,CI 20.96^51.14, P¼50.001). Implications for practice: partnership caseload midwifery care resulted in less interventionist labour and more normal birth than conventional team midwifery care. Women in the experimental group had more home births, birth in a midwife-led suite and opted for early discharge home postnatally more often than the controls.They also experienced much higher levels of continuity, particularly of a known midwife during labour and birth.The study ¢ndings should encourage other maternity units in the UK to pilot and evaluate the model to see if these bene¢ts are transferable. & 2001Harcourt Publishers Ltd
INTRODUCTION Since the Report of the Expert Committee on Maternity Care (DoH 1993) set as an ‘Indicator of Success’ that 75% of women should know the Midwifery (2001) 17, 234 ^240 & 2001 Harcourt Publishers Ltd doi:10.1054/midw.2001.0257, available online at http://www.idealibrary.com on
midwife who cares for them in labour, many maternity units throughout England have endeavoured to implement continuity of carer schemes. Very few have actually succeeded in reaching the target (Audit Commission 1997,
Comparison of partnership caseload midwifery care with conventional team care
Page et al. 1999, North Staffordshire et al. 2000), causing authors to question the target’s evidence base (Warwick 1997, Green et al. 2000). Although there is very good evidence that continuity of support during labour from a trained professional is advantageous (Hodnett 2000a), to date there are very few evaluations of the merit of a woman being cared for during labour by a professional with whom a relationship has been formed antenatally. In fact, Page et al. (1999) and North Staffordshire et al. (2000) studies are the only published findings to date that achieved the DoH (1993) target. The schemes they (Page et al. 1999, North Staffordshire et al. 2000) evaluated involved midwives working in pairs or triplets, providing all the antenatal, intrapartum and postnatal care of women booked with them. The models were dubbed ‘partnership caseload midwifery practice’. Page et al.’s (1999) study was a nonrandomised comparison with conventional community midwifery care and they found statistically significant results in favour of partnership caseload practice for less epidural use, less episiotomies and perineal lacerations. The North Staffordshire et al. (2000) study had an element of randomisation, albeit geographical areas not research subjects, so was a stronger design. They showed significant reductions in epidural and augmentation rates with caseload care. Women’s views strongly favoured labour continuity from a known midwife in Page et al.’s (1999) study. Other studies of caseload midwifery (Farquhar et al. 1996, Allen et al. 1997, Pankhurst 1997) have been published, but none have achieved the DoH (1993) target for continuity during labour and, therefore, their findings are less relevant to this study. In order to achieve the standard of a 75% known midwife for intrapartum care, midwives’ work patterns and caseload size have to be radically altered from conventional community midwifery practice. The caseload midwife has to be on-call for extended periods of time and caseload numbers of women per midwife need to be around 40 per year (Walsh 1995). Such radical departure from traditional practice has been a major stumbling block for maternity units who either cannot achieve the caseload ratios per midwife or who do not have midwives willing to try the new model. For these reasons, it is imperative that robust evaluations do occur when pilot schemes are launched. The purpose of the study reported here was to evaluate a partnership caseload practice scheme, launched in 1996 at a maternity unit in the east midlands of England. Named ‘Birth Under Midwifery Practice Scheme’ (BUMPS), it consisted of three pairs of midwives who provided total care for a defined caseload of women attached to specific general practitioner (GP)
235
practices so that each midwife attended 40 births/year. The BUMPS project thus enabled women to get to know their two midwives during pregnancy and to be cared for during labour by them. Postnatally, the same two midwives would attend. The project provided midwife-led care for women ‘booked’ as low risk and consultant-led care for women with risk factors. Traditional team midwifery care at the maternity unit provided a community midwife, linked to specific GP practices, who provided the bulk of ante/postnatal care and labour care by one of a large team of 25 midwives which included the community midwife. Midwife-led care and consultant-led care was based on the same risk criteria applied by the BUMPS project. BUMPS midwifery care was compared with team midwifery care.
METHODS Research design A comparative, prospective, non-randomised design was chosen because it was not feasible to mount a randomised controlled trial. As entry to the maternity services was via the women’s general practitioners (GPs), randomisation would have to have occurred at that time. For most women within both care packages, the majority of their care took place in a primary care setting and this is in contrast to earlier randomised controlled trials that evaluated similar aspects of the midwifery care package, e.g. continuity of care, midwife-led care and midwife-led areas (Hodnett 2000a,b). These studies were able to randomise at hospital level and then offer two contrasting packages of care that predominately occurred at the hospital. In this study, if women had been randomised on ‘booking’ with their GP, two parallel systems of care would have had to be in place at this level. This could either have been the same midwife offering different packages of care to a control and experimental group or two separate groups of midwives offering the alternative package to a control or experimental group. Neither of these options were feasible – the first because contamination between the care packages would have been highly likely, and the second because it was not organisationally possible to install parallel care systems at GP level. The primary hypothesis to be tested was that partnership caseload practice would reduce the epidural rate in this group compared with conventional team midwifery. Epidural rate was chosen as a primary outcome measure because it represents a significant marker of medical intervention in labour and birth. Previous randomised controlled trials of midwifery-led care and
236
Midwifery
continuity of carer, components of the BUMPS package, showed a lowering of epidural rates (Hodnett 2000a,b), and two other previously published studies of partnership caseload practice also adopted it as an important outcome measure, Page et al. (1999) as the primary outcome measure and North Staffsorshine et al. (2000) as a secondary measure. Based on the unit epidural rate of 30%, a sample size of 300 women in each group would need to be recruited to demonstrate a reduction of 10% in the experimental group with 80% power at a 5% level of significance. This is one of the accepted benchmarks of clinical significance. Ethical approval was granted by the Local Research Ethics Committee.
Table 1. Characteristics of women at trial entry. Values are given as mean [SD] or n (%) except gravida & parity Characteristics
Age (years) Height (m) Parity Gravida Ethnic group white Afro-Caribbean Indian/Pakistani other/not known Marital status married/widowed single supported (living with partner) single unsupported not known/ divorced/separated
Study group n¼303 29.5 [5.3] 1.55^1.64 52.4%41 60.4%41
Control group n¼308 27.3 [5.2] 1.55^1.64 51.7%41 62.0%41
263 (87) 5 (2) 17 (6) 18 (6)
287 (93) 4 (2) 11 (4) 5 (2)
195 (64) 92 (30)
150 (49) 139 (45)
10 (3) 6 (3)
3 (1) 16 (5)
Outcomes The primary outcome measure was epidural rate. Secondary measures of outcomes were delivery outcome, induction of labour rate, use of syntocinon for augmentation of labour, perineal outcomes, maternal position for birth, thirdstage management, place of birth, neonatal Apgar scores at one and five minutes and admission to the neonatal unit. Data were also collected on continuity of carer.
and 14 weeks’ gestation. At 28 weeks’ gestation women were invited and consented to participate in the study after guarantees of anonymity and confidentiality. Written informed consent was required prior to recruitment. Recruitment continued until 31 July 1998, by which time 303 women had been recruited from the BUMPS project and 308 from the control group.
Data collection and management Eligibility and recruitment Prior to the study, women from the specific GP practices served by the BUMPS midwives were matched prospectively with women from the specific practices serving the control group who were experiencing conventional team midwifery care. Homogeneity was attained for age, marital status, ethnicity, parity and height, but not for marital (64% v 47%) or single-supported status (30% v 44%) (Table 1). However, 94% of the study group and 91% of the control group were living with their partner, which diminishes the likelihood of these differences causing selection bias. Similarity of social class had been previously addressed via the advice given by the Health Authority as to the comparability of the two areas of the city that women were drawn from. All women who booked either with the conventional midwifery team (control group) or the BUMPS project (experimental group) from 1 April 1997 onwards, were eligible for inclusion in the study except those women who subsequently miscarried or gave birth before 28 weeks gestation. Women with any other risk factors were eligible for inclusion. Therefore, women at both high and low obstetric risk were eligible for inclusion in the study. These exclusion/inclusion criterium replicated the McCourt and Page (1996) study. Women were given an information leaflet at their first attendance at the hospital, between 10
It was obviously not possible to keep women unaware as to which arm of the study they were in, but they were recruited from the south and west of the city with a three-mile corridor separating them. Some degree of contamination could have occurred between the groups, although this never came to the attention of the midwives or the researcher. Nor could the researcher (a research midwife appointed from within the unit) be blinded to the arms of the study. Therefore, both maternal behaviour and data collection could have introduced some bias. An external statistician who was blinded to the respective groups rechecked the data analysis. Data were collected via a clinical audit tool previously validated by the Page et al.’s study (1999). Additional parameters, e.g. birth position, were added and these were collected by a very robust hospital IT system, Euroking. In addition to clinical process and outcome data, maternal satisfaction was measured at four weeks and three months postnatally, and these findings will be the subject of a subsequent paper. An ethnographic study of the partnership caseload practice arm of the trial was also carried out and has been reported (Walsh 1999).
Data analysis Logistic regression modelling was used to compare the study and control groups. Odds
Comparison of partnership caseload midwifery care with conventional team care
237
Women cared for by the BUMPS midwives were more likely to adopt an upright birth posture (60% v 14% OR 9.64, 95% CI 5.96– 15.61, P¼50.001), have an intact perineum (40% v 30% OR 1.57, 95% CI 1.05–2.35, P¼0.027), and experience a physiological third stage (37% v 1.5% OR 38.69, 95% CI 11.98– 124.89, P¼50.001), although the postpartum haemorrhage rate for these women was lower in comparison to the control group (0.4% v 3% OR 0.13 95% CI 0.02–1.04, P¼0.054). The episiotomy rates were also lower in the experimental group (6% v 10% OR 0.57 95% CI 0.27– 1.17, P¼0.126) (Table 3). Place of birth findings were contrasting. The experimental group had far more home births (17% v 1.3% OR 15.38. 95% CI 5.48–43.14, P¼50.001) and used the midwife-led birthing suite more often (28% v 12% OR 2.77, 95% CI 1.82–4.22, P¼50.001). They were also more likely to take an early discharge from hospital (two to six hours) from hospital following birth (25% v 3% OR 11.32. 95% CI 5.55–23.06, P¼50.001) (Table 4). There were no differences between the groups in relation to neonatal outcomes. Apgar scores of 56 at one minute and 56 at five minutes respectively were comparable and both groups had 10 admissions to the neonatal unit. Each group had one perinatal death (Table 5).
ratios, confidence intervals, and P-values were calculated using the statistical software SPSS for Windows, Version 7. P-values of less than 0.05 indicated statistical significance. When tested through logistic regression, the effects of potential confounding variables (primigravida, midwife only care, midwife/GP care) reinforced the findings of the unadjusted results. Therefore, only unadjusted results are presented in this paper. Full adjusted results are available from the researcher.
FINDINGS The finding for the primary clinical outcome was that fewer women in the experimental group had an epidural compared with the controls (21% v 32% OR 0.56 95%, CI 0.39–0.81, P¼0.002). These women also had more normal births (74% v 66% OR 1.45, 95% CI 1.02–2.05, P¼0.038) and fewer caesarean sections (15% v 19% OR 0.75, 95% CI 0.49–1.15, P¼0.186), although the latter was not statistically significant. Their induction of labour rate was lower (16% v 23% OR 0.66, 95% CI 0.44–0.98, P¼0.042), but they experienced slightly higher rates of syntocinon augmentation (14% v 13%, OR 1.08, 95% CI 0.67–1.73, P¼0.750) (Table 2).
Table 2 Birth process and outcome data Study group n¼303 n % Epidural Normal vaginal birth Caesarean section Induction Syntocinon augmentation
61 224 44 49 41
Control group n¼308 n %
OR
95% CI
P
98 204 57 71 39
0.56 1.45 0.75 0.66 1.08
0.39^ 0.81 1.02^2.05 0.49^1.15 0.44 ^ 0.98 0.67^1.73
0.002 0.038 0.186 0.042 0.750
21 74 15 16 14
32 66 19 23 13
Table 3 Birth process and outcome data (normal spontaneous deliveries only). PPH¼primary postpartum haemorrhage Study group n¼224 n % Upright birth posture Intact perineum Episiotomy PPH Physiological 3rd stage
135 90 13 1 82
60 40 6 0.4 37
Control group n¼204 n % 28 61 20 7 3
14 30 10 3 1.5
OR
95% CI
P
9.64 1.57 0.57 0.13 38.69
5.96 ^15.61 1.05^2.35 0.27^1.17 0.02^1.04 11.98 ^124.89
50.001 0.027 0.126 0.054 50.001
OR
95% CI
P
15.38 2.77 11.32
5.48 ^ 43.14 1.82^ 4.22 5.55^23.06
50.001 50.001 50.001
Table 4 Place of birth data. (all births) Study group n¼303 n % Home Midwife-led birthing suite 2^ 6 hour discharge
51 85 76
17 28 25
Control group n¼308 n % 4 38 9
1.3 12 3
238 Midwifery
Table 5 Neonatal outcome data. (All births) NNU¼Neonatal Unit Study group n¼303 n % Apgar 56 at1 minute Apgar 56 at 5 minutes Admission to NNU Perinatal death
19 0 10 1
6.7 3.3 0.33
Control group n¼308 n %
OR
95% CI
P
9.6 0.7 3.2 0.32
0.70 0 1.02 1.02
0.38 ^1.28 ^ 0.47^2.48 0.06^16.32
0.245 0.847 0.970 0.991
Control group n¼308 n %
OR
95% CI
P
22.38 ^70.25 20.96 ^51.14 4.16 ^ 8.57
50.001 50.001 50.001
27 2 10 1
Table 6 Continuity of carer data (all births) Study group n¼303 n % Named midwife for labour Known midwife for labour Known midwife postnatally
203 255 242
67 84 80
15 43 123
Significant differences were also apparent in each group’s experience of continuity of midwifery care. Women on the BUMPS Scheme were attended at birth more often by either their named midwife (67% v 5% OR 39.65, 95% CI 22.38–70.25, P¼50.001) or her partner, known midwife (84% v 14% OR 32.74, CI 20.96–51.14, P¼50.001). Thus 86% of the control group were cared for during labour and birth by a midwife they had never met before compared with just 16% of the experimental group. Continuity postnatally was also much better for the BUMPS’ women. Eighty per cent received care from a midwife they knew compared with 40% of the controls (OR 5.97 95% CI 4.16–8.57, P¼50.001) (Table 6).
DISCUSSION Caution must be adopted in any attempt to generalise from these findings, given that this was a non-randomised design. The possibility of selection bias has to be acknowledged, though the homogeneity of both groups is reassuring (Table 1). Women did not have a choice regarding BUMPS project care. To opt out of the scheme, they would have had to book out of area with different GPs. The possibility of selection bias regarding place of birth is key to the robustness of findings because epidural analgesia is not available at home nor in the midwife-led birthing suite. We believe that the BUMPS women did not have a predilection to choose these options. The homebirth and midwife-led birthing suite use in the two years prior to the project was similar in both groups of women. We suggest that the contrasting findings in this study were due to the nature of the scheme, which emphasised continuity of carer through all phases of care. In particular, being
5 14 40
39.65 32.74 5.97
cared for by a known midwife during labour and birth facilitated women making independent choices as to environment for birth. This is clear from the qualitative study previously alluded to (Walsh 1999). The researcher was not blinded to the treatment streams, so this may also have led to assessment bias. However, some data entry and all of the analysis was cross-checked by an independent statistician, reducing possible bias effects. It is also possible that the characteristics of the midwives themselves may explain outcome differences. The BUMPS midwives were a selfselected group and this had to be so because their method of working was so different to a conventional community midwife. It would not have been possible to compel midwives to adopt such a work pattern. Other randomised studies offering alternative care packages comment on this potential confounder (Flint & Poulengeris 1989, Hodnett 2000b). If the midwives at the outset shared a common philosophy of care, it may influence how care is given and affect outcomes. Alternatively, it is also possible that, as an individual midwife practices over time, the nature of the scheme itself, e.g. the relationship she builds with women, influences her approach to care. The findings confirm those of Page et al. (1999) that women cared for by partnership caseload midwifery schemes have less interventionist labours than conventional care. Waldenstrom and Turnbull’s (1998) review of randomised controlled trials that evaluated elements of the BUMPS package of care, i.e. continuity of care, midwife-led care and midwife-led birthing facilities also found this noninterventionist trend. What is different with this study is that a statistically significant higher normal birth rate was found. Though earlier studies have often hypothesised that continuity schemes and midwifery-led schemes would
Comparison of partnership caseload midwifery care with conventional team care
improve normal birth rates, all have failed to confirm this (Flint & Poulengeris 1989, Hundley et al. 1994, Turnbull et al. 1996). Lower induction of labour rates and less use of epidural analgesia might have contributed towards the higher normal birth rate as both have been linked to more operative deliveries (Goer 1995, Howell 2000). The fall in epidural rate is dramatic and may, in part, be explained by women in the experimental group choosing to give birth either at home or in the Home from Home facility, both of which preclude epidural use. McVicar et al.’s (1993) randomised controlled trial of the Home from Home Scheme showed significantly lower epidural use in this facility. The significantly higher intact perineum rate for normal birth with BUMPS care is an interesting finding. Variation in episiotomy rates may explain some of these differences. An alternative explanation may lie with the data on position for birth. Sixty per cent of women in the BUMPS group adopted upright postures for birth compared with just 14% of the control group. Some studies of alternative versus recumbent birth position (Gardosi et al. 1989, Sermer & Raines 1997) have shown a non-significant trend towards intact perineums in the upright posture groups. The episiotomy rates in the study groups are probably among the lowest in the UK (Clinical Standards Advisory Group 1995), comparing favourably with the World Health Organization’s (1996) recommendation of 10% for normal birth. The experimental group had a lower postpartum haemorrhage rate for normal birth despite having a much higher physiological third-stage rate. All the trials on active versus physiological third stage indicate a higher postpartum haemorrhage rate with the latter (McDonald et al. 2000), so this is a surprising finding. Neonatal outcomes showed no differences between the study groups with a neonatal admission rate of 3% in each. There was one perinatal death in each group so no conclusions can be drawn regarding perinatal mortality. In this study the level of continuity of care for women in the BUMPS group exceeded the national target of 75%, with 84% of women receiving care from a known midwife, compared to just 14% of women in the control group. This higher level of continuity of caregiver concurs with Page et al.’s (1999) and the North Staffordshire et al.’s (2000) findings. In attempting to explain the differences in interventions and outcomes between the two groups, Walsh’s (1999) parallel qualitative study of the BUMPS women’s experience of care is enlightening. Women spoke about their birth experiences as empowering and feeling in control. They felt that their midwives acted as their
239
advocates, particularly in representing their desire for non-intervention during labour. Knowing antenatally that they would be attended in labour by a midwife they knew instilled confidence about choosing a home birth and about developing birth plans. The greater level of self-management of their own births is consistent with the less interventionist labours they had and is further illustrated by the common use of upright posture for birth and the frequent choice of a natural third stage. McCourt and Page (1996) and Pankhurst (1997) both found higher levels of maternal satisfaction with the partnership caseload model when compared with conventional care. Maternal satisfaction is an important dimension of evaluations of service developments and should be appraised alongside clinical outcomes when decisions are made regarding the ‘roll-out’ of pilot schemes. The absence of published data in this area for this project weakens its generalisability. In conclusion, although the implications for other UK maternity units are mitigated by the research design, the results should encourage them to set up and evaluate their own partnership caseload schemes. ACKNOWLEDGEMENTS We would like to thank the following people who assisted with this study: BUMPS Midwives, Bradgate Team midwives, Margaret Stockwell (Senior Midwife), Christina Oppenheimer, Roger Neuberg (Consultant Obstetricians), Rosie Garrett, Melanie Thwaites (Midwifery Team Leaders), Fran Gill (IT Manager), all general practitioners from the women’s Health Centres, The Leicester Royal Infirmary NHS Trust Research Fellowship Scheme that funded the Project.
REFERENCES Allen I, Dowling S, Williams S 1997 A leading role for midwives. Policy Studies Institute, London Audit Commission 1997 First class delivery: improving maternity services in England & Wales. Audit Commission, London Clinical Standards Advisory Group (CSAG) 1995 Women in normal labour. HMSO, London Department of Health 1993 Changing Childbirth: the report of the expert committee on maternity care. HMSO, London Farquhar M, Camillerie-Ferrante C, Todd C 1996 An evaluation of midwifery teams in West Essex. Final report. PHRU & HSRG, Institute of Public health, University of Cambridge, Cambridge Flint C, Poulengeris P 1989 The know your midwife report. Available from 49 Peckermans Wood, London Gardosi J, Hutson N, Lynch C 1989 Randomised controlled trial of squatting in the second stage of labour. The Lancet 2: 74–77 Goer H 1995 Obstetric myths versus research realities: a guide to the medical literature. Bergin & Garvey, London Green J, Renfrew M, Curtis P 2000 Continuity of carer: what matters to women. Midwifery 16(3): 186–196
240 Midwifery Hodnett ED 2000a Support from caregivers during childbirth. (Cochrane Review) In: The Cochrane Library, Issue 3, Update Software, Oxford Hodnett ED 2000b Home-like versus conventional birth settings. (Cochrane Review) In: The Cochrane Library, Issue 3, Update Software, Oxford Howell C 2000 Epidural v non-epidural analgesia in labour. (Cochrane Review) In: The Cochrane Library, Issue 3, Update Software, Oxford Hundley V, Cruickshank F, Lang G et al. 1994 Midwife managed delivery unit: a randomised controlled comparison with consultant led care. British Medical Journal 309: 1400–1404 McDonald S, Prendiville W, Elbourne D 2000 Expectant v active management of the third stage of labour. (Cochrane Review ) In: The Cochrane Library, Issue 3, Update Software, Oxford McVicar J, Dobbie G, Owen-Johnston L et al. 1993 Simulated home delivery: a randomised control trial. British Journal of Obstetrics & Gynaecology 100: 3316–3323 McCourt C, Page L 1996 The one-to-one project, final report. Thames Valley University, London North Staffordshire Changing Childbirth Research Team 2000 A randomised study of midwifery caseload care and traditional ‘shared-care’. Midwifery 16(4): 295– 302 Page L, McCourt C, Beake S, Hewison J 1999 Clinical interventions and outcomes of one-to-one midwifery
practice. Journal of Public Health Medicine 21(3): 243–248 Pankhurst F 1997 Caseload midwifery: an evaluation of a pilot scheme, final report. University of Brighton, Brighton Sermer R, Raines D 1997 Positioning during the second stage of labour: moving back to basics. Journal of Obstetric, Gynaecologic and Neonatal Nursing 26(6): 727–734 Turnbull D, Holmes A, Cheyne H et al. 1996 Randomised, controlled trial of efficacy of midwife-managed care. The Lancet 348: 213–218 Waldenstrom U, Turnbull D 1998 A systematic review comparing continuity of midwifery care with standard maternity services. British Journal of Obstetrics & Gynaecology 105: 1160–1170 Walsh D 1995 The Wistow project, final report. Women’s Hospital, LRI NHS Trust, Leicester Walsh D 1999 An ethnographic study of women’s experience of partnership caseload midwifery practice: the professional as friend. Midwifery 15(3): 165– 176 Warwick C 1997 Open dialogue. Can continuity of care be the only answer? British Journal of Midwifery 5(1): 6 World Health Organization 1996 Care in Normal Birth: A Practical Guide. Report of a Technical Working Group. Maternal and Newborn Health/Safe Motherhood Unit, Geneva