Search tips
Search criteria 


Logo of bmjThis ArticleThe BMJ
BMJ. 2007 September 29; 335(7621): 643.
PMCID: PMC1995478
Head to Head

Do we have enough evidence to judge midwife led maternity units safe? No

James Drife, professor of obstetrics and gynaecology,

The UK government claims it is trying to give women more choice by converting local maternity units to midwife led services. Lesley Page believes such units improve the birth experience, but Jim Drife remains worried about the risks of delivering outside hospital

Major changes are being made in the UK's maternity services for political, economic, and clinical reasons. Much has already happened. Antenatal care is provided mainly by midwives outside hospital, with general practitioners playing little or no part and obstetricians seeing only high risk cases. Maternity hospitals are being merged because of pressure to increase consultant presence in labour wards and reduce junior doctors' hours. Large units seem more efficient and can offer more back-up when complications occur. Closing small hospitals is unpopular, however, and a politically attractive alternative is to convert consultant units to free standing midwife led maternity units. The NHS, which has a near monopoly of childbirth, is promoting midwife units as a way of offering choice1 and is advising women that they are safe for low risk pregnancies. This advice is not based on evidence.

Problem of defining risk

Everyone knows that safety can never be absolute. Nevertheless, hospital delivery has become steadily safer. In 2003, for example, the risk of fetal death during labour was as low as 1 in 7642 across the three large hospitals in Dublin.2 3 4 Such up to date figures are often ignored when births inside and outside hospital are being compared.

Maternal complications during childbirth are no less frequent than they were in the past. National audits in Scotland report life threatening emergencies once in 200 births. The most common, severe haemorrhage, occurs once in 300 births and is usually unpredictable. Of 156 such cases in 2004, only a minority were antepartum haemorrhage, but 32 women bled during labour and 116 after delivery.5 Prompt treatment saves several lives every day across the UK, and the national maternal mortality is low only because emergencies are managed effectively.

Nevertheless, pregnancies are now being classed as high or low risk (a false dichotomy as most are in between). Risk classification is based on the history given by the woman at booking. This is no easy task. A woman in her first pregnancy does not have an obstetric history. Family history is often incomplete. Complications such as pre-eclampsia and fetal growth restriction cannot be predicted. The result is that women labelled low risk have a higher corrected singleton perinatal mortality than high risk women.6

Research on performance

Evidence on safety of midwife led units is lacking. A 2005 Cochrane review found no trials of freestanding birth centres.7 There was, however, a trend towards higher perinatal mortality in “home-like settings” with a relative risk of 1.83 (95% confidence interval 0.99 to 3.38). An earlier systematic review comparing continuity of midwifery care with standard maternity services found that midwifery care was associated with an increase in perinatal death “bordering on statistical significance” (odds ratio 1.60; 95% confidence interval 0.99 to 2.59).8 In both reviews the confidence intervals included 1.00 (though only just), so the trends were not significant. Nevertheless, they should worry those who want to change national patterns of care.

Many UK maternity hospitals have a consultant unit and a midwife led unit in the same building, and staff prefer this arrangement. Even in such units, however, the evidence is not entirely reassuring. In the midwife led unit of the Stockholm Birth Centre (one floor below a standard delivery ward) perinatal mortality among primiparous women was significantly higher than among Swedish women receiving standard care (relative risk 1.8; 1.06 to 3.00). For multiparous women, rates were not significantly different.9 When data for first pregnancies were recalculated, the rate of fetal death in labour in the birth centre was 1 in 493, over seven times higher than the rate of 1 in 3779 with standard care in Sweden.10

The main worry about free standing midwife led units is that they are some distance from medical help. Transfers may save lives but are often precautionary and have a negative psychological effect on women.11 Rates of transfer before labour in the Cochrane review were 29-67%.7 In Stockholm the transfer rate during labour was 18%.9 In a Scottish unit rates were 30% before labour and 27% during labour for primiparous women, and 22% and 10% for multiparous women.12 In a US study of free standing midwife led units in the 1980s, 7.9% of women had serious complications in labour and transfer rates among primiparous and multiparous women were 29% and 7% respectively.13

The National Perinatal Epidemiology Unit sums up current knowledge as follows: “a structured review carried out in 2005 concluded that high quality evidence was needed about whether there are important differences in experiences and outcomes for women and babies in these alternative locations and systems.”14 Others have also called for better evidence: “if women at low obstetric risk are offered a choice between free-standing MLMU [midwife led maternity units] and hospital, they should be aware that the safety and effectiveness of delivery in the two settings has not been reliably compared.”15

It is disturbing that in an era of evidence based medicine, midwife led units are being promoted before their safety has been established. The attractions of a relaxed environment and non-intervention are easy to understand, but most women put the highest premium on safety for their baby. Last year the National Perinatal Epidemiology Unit began an evaluation of alternative locations for labour and birth.14 Further change should await reliable evidence on safety and must not be driven by political expediency.


Competing interests: JD is an obstetrician in a tertiary centre and routinely works on the labour ward, including nights and weekends. He is an obstetric assessor for the national Confidential Enquiry into Maternal Deaths.


1. Shribman S. Making it better: for mother and baby London: Department of Health, 2007
2. National Maternity Hospital. Annual report 2003 Dublin: National Maternity Hospital, 2004
3. Rotunda Hospital Dublin. Clinical report 2003 Dublin: Rotunda Hospital, 2005
4. Coombe Women's Hospital. Annual clinical report 2003 Dublin: Coombe Women's Hospital, 2004
5. Penney G, Adamson L, Kernaghan D. Scottish confidential audit of severe maternal morbidity: 2nd annual report 2004
6. McKenna D, Tharmaratnam S, Mahsud S, Bailie C, Harper A, Dornan J. A randomised trial using ultrasound to identify the high-risk fetus in a low-risk population. Obstet Gynecol 2003;101:626-32. [PubMed]
7. Hodnett ED, Downe S, Edwards N, Walsh D. Home-like versus conventional settings for birth. Cochrane Database Syst Rev 2005;(1):CD000012.
8. Waldenstrom U, Turnbull D. A systematic review comparing continuity of midwifery care with standard maternity services. Br J Obstet Gynaecol 1998;105:1160-70. [PubMed]
9. Gottvall K, Grunewald C, Waldenstrom U. Safety of birth centre care: perinatal mortality over a 10-year period. BJOG 2004;111:71-8. [PubMed]
10. Fahy K, Colyvas K. Safety of the Stockholm Birth Centre study: a critical review. Birth 2005;32:145-50. [PubMed]
11. Walker J. Women's experiences of transfer from a midwife-led to a consultant-led maternity unit in the UK during late pregnancy and labor. J Midwifery Women's Health 2000;45:161-8. [PubMed]
12. Mahmood TA. Evaluation of an experimental midwife-led unit in Scotland. J Obstet Gynaecol 2003;23:121-9. [PubMed]
13. Rooks JP, Weatherby NL, Ernst EKM, Stapleton S, Rosen D, Rosenfield A. Outcomes of care in birth centres: the national birth centre study. N Engl J Med 1989;321:1804-11. [PubMed]
14. Rowe R. Evaluation of maternity units in England (EMU) research programme October 2006.
15. Muthu V, Fischbacher C. Free-standing midwife-led maternity units: a safe and effective alternative to hospital delivery for low-risk women? Evid Based Healthcare Public Health 2004;8:325-31.

Articles from The BMJ are provided here courtesy of BMJ Publishing Group