Home Birth Reference Site

Research Summaries - Page Four - Controversies

Scroll down to read the studies, or use the index to jump to each individually.

Data on babies' safety during hospital births are being ignored

By Professor James Drife
Letter to British Medical Journal, 1999; 319: 1008. (9 October)

Although not original research, Professor Drife's letter is included here as part of a valuable discussion about home birth safety.

Professor Drife writes that the safety of home birth has been exaggerated, and that his own statistical analysis reveals that "for a normal birth, hospital delivery is now three to four times safer than home delivery for the baby.". He reached this conclusion by comparing perinatal mortality for low-risk women in the UK (which has a very low overall mortality rate) with perinatal mortality for home birth cases across all risk groups, including high-risk, in the USA and Australia. He mysteriously chose to ignore several recent studies which found excellent outcomes for home birth in the UK itself.

A number of experts wrote to disagree with Professor Drife's interpretations - Doctors Gavin Young and Edmund Hey stated:

"Since the study groups were dissimilar it is about as helpful as saying that a man and a dog have an average of three legs.....(Drife) did not compare like with like, and he merged groups who should be advised differently. Most women can be told that, as long as they continue to accept professional advice, they are as safe delivering at home as in hospital. For others with a twin, breech, or post-term pregnancy the increased risk of home birth is probably even greater than Drife's figure suggests. "

Professor Geoffrey Chamberlain drew attention to the National Birthday Trust Fund's findings:

"We concluded that there was no evidence that women who had been screened properly in the antenatal period and planned a booked delivery for home had any higher risk than a similar group of women who delivered in hospital."

Read the full text of Professor Drife's letter on the BMJ Website

Responses on the BMJ Website from Professor Chamberlain, Drs Young and Hey, Campbell and McFarlane, and others.

Drife later wrote to correct some of his own calculations.

Back to index

Do obstetric intranatal interventions make birth safer?

British Journal Obstet Gynaecol 1986 Jul;93(7):659-74
By Marjorie Tew

Marjorie Tew argues that statistical analysis shows that the shift to hospital birth, and increased obstetric intervention, has not made birth safer, but more dangerous. She suggests that improvements in perinatal mortality are due to healthier mothers, rather than improved maternity care.

Abstract in full:

Impartial analyses of the evidence from official statistics, national surveys and specific studies consistently find that perinatal mortality is much higher when obstetric intranatal interventions are used, as in consultant hospitals, than when they are little used, as in unattached general practitioner maternity units and at home. The conclusion holds even after allowance has been made for the higher pre-delivery risk status of hospital births as a result of the booking and transfer policies. It holds even more strongly for births at high than at low predicted risk. It follows that the increased use of interventions, implied by increased hospitalization, could not have been the cause of the decline in the national perinatal mortality rate over the last 50 years and analysis of results by different methods confirms that the latter would have declined more in the absence of the former. Data are presented which point to the deleterious effect of interventions on the incidence of low birthweight and short gestation and their associated mortality. Also presented are data supporting the alternative explanation of the decline in perinatal mortality, namely the improvement in the health status of mothers built up over several generations. The organization of the maternity service stands indicted by the evidence. Despite the beliefs of those responsible, it has not promoted, and cannot promote, the objective of reducing perinatal mortality.

Read the abstract on Medline

Are hospital confinements really more dangerous for the fetus?

Early Human Development 1988 May;17(1):29-36
Golding J, Peters TJ
Department of Child Health, University of Bristol, U.K.

The authors argue that the higher perinatal mortality in consultant (obstetrician-led) units compared to home birth or other units (eg midwife-led, or GP units) is due to higher-risk cases transferring to consultant care, rather than obstetric intervention causing problems. Looking at the place the mother originally planned to give birth, mortality under consultant care was 'only' 38% higher! The authors believe that, if other factors such as past obstetric history and socio-economic indicators of risk are taken into account, it is possible that hospital birth might be safer than home birth. These details would be helpful as they all contribute to a mother's risk level; the idea is that women who are higher-risk in the first place are more likely to be receive their pregnancy and birth care from an obstetrician than from a midwife or GP. However, the data was not sufficient for them to test this theory.

This paper may be a response to Tew's article, above. Note that more recent studies, eg the National Birthday Trust study, have tried to compare like with like by matching women choosing home birth with women of a similar risk level who chose hospital birth.

Abstract in full:

A large number of publications has reiterated the observation that perinatal mortality rates in Britain are higher among births in consultant units than among those occurring at home or in other units. In this paper we show that whereas these observations are themselves undeniable, the conclusion that hospital confinements are more dangerous to the fetus is probably erroneous. To illustrate the methodological difficulties, we have used as much information as possible on the delivery intentions for a national survey of 16,668 singleton births taking place in the United Kingdom in one week of April 1970. Using these data, we show that although deliveries in a consultant unit had a three times higher mortality rate than those delivered elsewhere, this was due almost entirely to the excess mortality among transfers of women originally intending to deliver elsewhere. Consideration of the place the mother was originally intending to deliver altered the picture considerably, with mortality only 38% higher among the consultant unit group. It is pointed out that if account was taken of risk factors such as past obstetric history, marital status and social class it is likely that booking for hospital delivery may well carry a lower risk of perinatal death, but that present data collection systems combined with high rates of consultant delivery are unlikely to resolve this question.

Read the abstract on Medline

Home versus hospital birth - Cochrane Database Review

Cochrane Database Syst Rev 2000;(2):CD000352
Olsen O; Jewell MD
The Nordic Cochrane Centre, Rigshospitalet, Blegdamsvej 9, dept. 7112, Copenhagen, Denmark, DK-2100 O.

Only one small controlled trial of planned home and hospital birth met the Cochrane Review's strict criteria, and the author's concluded that "There is no strong evidence to favour either planned hospital birth or planned home birth for low risk pregnant women."

Abstract in full:

ABSTRACT: BACKGROUND: A meta-analysis of observational studies have suggested that planned home birth may be safe and with less interventions than planned hospital birth.
OBJECTIVES: The objective of this review was to assess the effects of planned home birth compared to hospital birth on the rates of interventions, complications and morbidity as determined in randomised trials.

SEARCH STRATEGY: We searched the Cochrane Pregnancy and Childbirth Group trials register and the Cochrane Controlled Trials Register. Date of last search: September 1999.
SELECTION CRITERIA: Controlled trials comparing planned hospital birth to planned home birth in selected women, assisted by an experienced home birth practitioner, and backed up by a modern hospital system in case transfer should be necessary.
DATA COLLECTION AND ANALYSIS: Trial quality was assessed and data were extracted by one reviewer and checked by the other reviewer. Study authors were contacted for additional information. MAIN RESULTS: One study involving 11 women was included. The trial was of reasonable quality, but was too small to be able to draw conclusions.

REVIEWER'S CONCLUSIONS: There is no strong evidence to favour either planned hospital birth or planned home birth for low risk pregnant women.

Outcomes of Planned Home Births in Washington State: 1989-1996

This study is very unusual amongst modern research on home birth, because it found a significantly higher mortality rate for presumed planned home births than for planned hospital births. My comments are below the abstract.

Outcomes of Planned Home Births in Washington State: 1989-1996

Obstetrics and Gynecology August 2002:100:253-259.
Jenny W.Y. Pang, MD, MPH,a,c James D. Heffelfinger, MD, MPH,a,d Greg J. Huang, DMD, MPH,a Thomas J. Benedetti, MD, MHA,b and Noel S. Weiss, MD, DrPHa


Objective: To determine whether there was a difference between planned home births and planned hospital births in Washington State with regard to certain adverse infant outcomes (neonatal death, low Apgar score, need for ventilator support) and maternal outcomes (prolonged labor, postpartum bleeding).

Methods: We examined birth registry information from Washington State during 1989-1996 on uncomplicated singleton pregnancies of at least 34 weeks' gestation that either were delivered at home by a health professional (N = 5854) or were transferred to medical facilities after attempted delivery at home (N = 279). These intended home births were compared with births of singletons planned to be born in hospitals (N = 10,593) during the same years.

Results: Infants of planned home deliveries were at increased risk of neonatal death (adjusted relative risk [RR] 1.99, 95% confidence interval [CI] 1.06, 3.73), and Apgar score no higher than 3 at 5 minutes (RR 2.31, 95% CI 1.29, 4.16). These same relationships remained when the analysis was restricted to pregnancies of at least 37 weeks' gestation. Among nulliparous women only, these deliveries also were associated with an increased risk of prolonged labor (RR 1.73, 95% CI 1.28, 2.34) and postpartum bleeding (RR 2.76, 95% CI 1.74, 4.36).

Conclusion: This study suggests that planned home births in Washington State during 1989-1996 had greater infant and maternal risks than did hospital births.

© 2002 by The American College of Obstetricians and Gynecologists.

Full text of the study available online as at August 2002.


It is always challenging to read a report which contradicts one's deeply-held beliefs, but I have tried to be objective in my comments on this study. Here are some thoughts:

First of all, this study's findings contradict almost all other recent research on planned home birth, which has found that outcomes are at least as good as those for planned hospital birth. For example, the National Birthday Trust Fund study of around 6,000 planned home births in the UK found no difference in death rates, but a great reduction in complications and reductions in injuries to mother and baby when a home birth was planned. Olsen's meta-analysis of the safety of home birth looked at outcomes for a total of over 24,000 women, from six trials with good methodology, comparing planned home birth with planned hospital birth for women of similar risk levels. It found no difference in perinatal mortality rates, but improvements in other outcomes among the planned home births - eg fewer low APGAR scores and caesarean sections. We have to ask why the Washington State study found different results. Is something different about planned home birth in Washington State during the period studied, or was something different about the way this study was conducted?

I find it strange that the full text of the article mentions one Australian study which found worse outcomes for planned home birth (where the planned home births included a disproportionate number of high-risk births, and where the facilities for transfer to hospital were poor), and another study which found poorer outcomes for planned home births in Missouri, 1978-1984 which will be reviewed below. Yet, as noted above, it does not seem to mention any of the European studies which found good outcomes for planned home birth, despite these studies having sound methodology and being published in highly respected journals (eg see the British Medical Journal's special edition on home birth). If the authors were restricting themselves to mentioning studies conducted in areas which were near Washington State, or which experienced similar conditions then this would be understandable - but no such rationale is given. Often when a research paper reaches drastically different conclusions from similar studies, there is an acknowledgement of this and a discussion of why this might have happened is offered - but there is none in this paper. In fact, the only other studies on home birth which seem to be explicitly mentioned in the text (as opposed to the references) are the very small number which produced relatively poor outcomes.

The methodology of this study, although understandable, is not as convincing as those of many other studies reviewed on this site. The authors attempted to guess which home births were planned retrospectively, by looking at birth outcomes for babies born out of hospital or for those born in hospital where there was some indication that a home birth was planned. This is in contrast to prospective studies such as the National Birthday Trust Fund study, where women planning a home birth were recruited into the study at 37 weeks of pregnancy, and the outcomes of their pregnancies were carefully recorded specifically to provide data for the study. The authors of the paper about home births in Washington State explain in the full text of the study that they tried to guess which births were planned home births by excluding results for all women who did not fulfil low-risk criteria. Yet in fact many women who plan home births are not strictly low-risk. Of course, it could be that home births in Washington State would perform worse still if the data were more accurate - but the point is that we do not have all the data which would be needed to make a full assessment.

Personally I feel that women in Washington State, USA, should be worried about this study - was it just the study design, or is home birth there really more dangerous than it is in other parts of the developed world? However, women living elsewhere - and especially in the UK - will probably find that it has little relevance for them. The bottom line is that the overwhelming conclusion of the large body of research on home birth in Europe has shown excellent outcomes. When a large number of well-designed, relevant studies all agree with each other on the safety of planned home birth - how much weight should we really give to this one?

There is a comprehensive critique of this study on the website of the American College of Domiciliary Midwives.

On to Page Five of the Research Summaries

Site Contents