Letters

Perinatal death associated with planned home birth in Australia

Home births are not justified in Australia

Study prompts several questions

Authors' reply

Home births are not justified in Australia

EDITORBastian et al report the risk of perinatal death associated with planned home birth in Australia.¹ I have been criticising the role of home births in Queensland for the past two years, as chairman of the Queensland state committee of the Royal Australian College of Obstetricians and Gynaecologists. In Queensland a registered midwife can (and they do) go into independent midwifery practice after the basic 12 months' training and 20 normal deliveries. Although the College of Midwives has rigid criteria, they are neither enforced nor policed. No one mentions maternal mortality or morbidity.

In two years in Queensland I received reports of one maternal death (a second death was admitted to me by the Queensland state president of the Australian College of Midwives) and one near maternal death requiring hysterectomy and dialysis for three weeks. I was assured that the two maternal deaths were due to amniotic fluid emboli, so would have been unavoidable even in hospital, although I have no records of results of postmortem examinations to prove this. During the same period I documented eight perinatal deaths (to my personal knowledge) out of a total of 400 deliveries by home birth. Queensland also has home birth practitioners who have no insight into their own limitations or what is termed low risk, accepting women pregnant with twins (to deliver in the mountains, and who required helicopter evacuation), with previous caesarean section, and with anti-D antibodies.

I do not believe that home births are justified at present in Queensland, or most parts of Australia, because of inadequate controls, training, supervision, and policing and the immense geographical distances. In a trial of birthing at an Aboriginal homeland (Cherbourg) four perinatal deaths occurred in 80 deliveries,² a figure that I regard as unacceptable. Until the training of domiciliary midwives in Australia reaches the standards of the United Kingdom and the Netherlands, for example, and until these independent midwives are properly policed, and receive adequate back up, home births are not justified.

Peter Sullivan, Consultant obstetrician and gynaecologist. 
American Hospital, PO Box 5566, Dubai, United Arab Emirates
wendysul{at}emirates.net.ae

1. Bastian H, Keirse MJNC, Lancaster PAL. Perinatal death associated with planned home birth in Australia: population based study. BMJ 1998; 317: 384-388[Abstract/Free Full Text]. (8 August.)
2. Queensland Government. Birthing in homelands for aboriginal and Torres Strait islanders. Brisbane: Queensland Government , 1997.

Study prompts several questions

EDITORRetrospective analysis lends itself to two obvious issues of biaswhere the author knows what information to look for and therefore unwittingly finds only information that fits the hypothesis; and where the researcher is classifying a cause of death and the underlying pathology may not be documented or searched for, which again allows the author's argument to be strengthened in a certain direction.

The nature and extent of the datawhat percentage of births and deaths were reported and to whomare unstated in Bastian et al's study.¹ Readers are not told which states supplied perinatal data for 1989-90 whether the non-participants in the Homebirth Australia register complied with these data collections. Eleven deaths were excluded for reasons not stated. Thereafter the authors still refer to 50 deaths but it is not clear whether there were 61 deaths or the minimum data were available on 39.

Birthweight specific data were available for 1985-8 (table 4). Is it possible to calculate a five year mortality without birthweight specific data and without giving the number of known births over five years? The paper does not reference the source of national figures on birthweight specific perinatal mortality or give the years for which the data were available.

Can gestational age and cause of death be ascertained with any certainty through retrospective case analysis without confirmation from a postmortem examination by a perinatal pathologist? What percentage of the intrapartum fetal deaths (table 4) might otherwise have been described as inevitable spontaneous abortions?

Is it justifiable to compare perinatal mortality internationally? The baseline measurement for a fetal death in Australia is 20 weeks, for the United Kingdom 24 weeks, for Norway 16 weeks, and for New Zealand (until recently) 28 weeks. The political, educational, and social determinants for home birth differ widely between Australia and the countries compared, as do the exclusion criteria and the discrepancies in collecting study data.

What is the link between late neo- natal deaths and home birth? The five late neonatal deaths (table 3) include death from postviral cardiomyopathy, chromosomal abnormality, and the sudden infant death syndrome. Definitions of late neonatal death and the sudden infant death syndrome are usually mutually exclusive.

Should researchers attempt to draw definitive conclusions with regard to shortcomings in perinatal care, risk assessment, rates and severity of intrapartum asphyxia, cause and time of death, and failure to transfer women safely in a study such as this?

Sally Katherine Tracy, Independent research midwife. 
Glenbrook 2773, NSW, Australia
stracy{at}zeta.org.au

1. Bastian H, Keirse MJNC, Lancaster PAL. Perinatal death associated with planned home birth in Australia: population based study. BMJ 1998; 317: 384-388. (8 August.)

Authors' reply

EDITOROur study shows that low risk home births in Australia have good outcomes but high risk births give rise to a high rate of avoidable death at home. These conclusions may be unpalatable but are supported by an increasing body of evidence.

We agree with Sullivan that home birth practice in Australia needs to be monitored. We also agree that Aboriginal mortality is distressingly high, but this applies throughout Australia, not just to homelands and home births. The conclusion should be to provide birth options and care that are both adequate and culturally appropriate. Our study indicates that most home birth practitioners achieve outcomes that are similar to those internationally.

Sullivan is wrong to suggest that maternal mortality receives no attention in Australia.¹ Tracy is similarly wrong in her understanding of Australian perinatal data. The nature and extent of these data have been fully described ^{2 3} (as referenced in our paper). Perinatal data and registration data on births and deaths have been available nationally for many years.

There is no mechanism by which 11 excluded deaths (two excluded because they were unattended and nine because of transfer before labour ^{2 3} ) can change the 50 included deaths to either 61 or 39. Birthweight specific data were available for over 70% of home births during 1985-90 and close to 80% during 1985-8. "Inevitable spontaneous abortion" is not a term that can be applied to intrapartum fetal deaths (all but two of which occurred at term or after term in this study). Interesting as an international comparison of the lower limits of registration may be, it is not relevant to planned home births or the comparison in our paper. Gestational age is an important predictor of risk in birth, whether at home or in hospital. Postmortem examinations are important in elucidating the cause of death, but contributing clinical factors must also be considered, particularly in establishing whether there is a pattern of avoidable deaths.

Terms such as neonatal death, infant mortality, and the sudden infant death syndrome are not usually mutually exclusive. In Australia any death occurring within 28 days of birth is a neonatal death, regardless of cause. It is disturbing, though, that in Australia the sudden infant death syndrome is seven times more common among births planned at home than among other births (14.6 v 2.1 per 1000; relative risk 6.9 (95% confidence interval 3.1 to 15.3)).⁴ Whether and to what extent this relates to the levels of peripartum oxygen deprivation identified in our study is unknown.⁴

Hilda Bastian, Consumer advocate. 
PO Box 569, Blackwood SA051, Australia
Hilda.Bastian{at}flinders.edu.au


Marc J N C Keirse, Professor. 
Department of Obstetrics and Gynaecology, Flinders University of South Australia, Flinders Medical Centre, GPO Box 2100, Adelaide SA 5100, Australia

Paul A L Lancaster, Associate professor. 
Australian Institute of Health and Welfare, National Perinatal Statistics Unit, University of New South Wales, NSW 2052, Australia