You are not choosing what a Dutch or English woman is choosing
The reassuring international results come from systems with uniform midwife training, hospital integration and agreed transfer arrangements. That is the most important qualifier in the whole subject.
If you have read that home birth is as safe as hospital birth, the study you read was probably Dutch, English or Canadian. That does not make it wrong. It makes it about a different system.
The Dutch national data covered 743,070 low-risk women in midwife-led care when labor started, of whom 466,112 planned a home birth. Deaths from the start of labor to 28 days after birth:
| Group | Planned home | Planned hospital |
|---|---|---|
| First baby, death to 28 days | 1.02 per 1,000 | 1.09 per 1,000 |
| Later baby, death to 28 days | 0.59 per 1,000 | 0.58 per 1,000 |
| First baby, newborn intensive care | 3.41 per 1,000 | 3.61 per 1,000 |
| Later baby, newborn intensive care | 1.36 per 1,000 | 1.95 per 1,000 |
Before you set that 1.02 against the English 9.3 per 1,000, look at what each one counts. The Dutch figure counts deaths alone. The English figure is a composite of seven serious problems, of which only 12% were deaths. The two are not on the same scale, and the Netherlands is not nine times safer than England.
The authors' own conclusion carries its own limit, and it is the limit that matters to you: We found no increased risk of adverse perinatal outcomes for planned home births among low-risk women. Our results may only apply to regions where home births are well integrated into the maternity care system.
Note what the Dutch comparison actually is. Both groups were in midwife-led primary care when labor started. One group was at home and one was in hospital, inside a national system where referral to obstetrician-led care is a routine, protocolised event. That is not the same comparison as an American home birth against an American hospital birth.
A Canadian provincial study matched 11,493 planned home births to 11,493 planned midwife-attended hospital births and found no difference in serious problems for the baby. Its underlying rates could not be obtained for this guide, so no ratio from it is printed here. Its conclusion is quotable and the conditions in it are doing the work: planned home birth attended by midwives in a jurisdiction where home birth is well-integrated into the health care system was not associated with a difference in serious adverse neonatal outcomes but was associated with fewer intrapartum interventions.
ACOG says the same thing about generalising, in its own words: Some recent observational studies overcome many of these limitations, describing planned home births within tightly regulated and integrated health care systems, attended by highly trained licensed midwives with ready access to consultation and safe, timely transport to nearby hospitals. However, these data may not be generalizable to many birth settings in the United States where such integrated services are lacking.
And on why the results split the way they do: The relatively low perinatal and newborn mortality rates reported for planned home births from Ontario, British Columbia, and the Netherlands were from highly integrated health care systems with established criteria and provisions for emergency intrapartum transport. Cohort studies conducted in areas without such integrated systems and those where the receiving hospital may be remote, with the potential for delayed or prolonged intrapartum transport, generally report higher rates of intrapartum and neonatal death.
The meta-analyses point the same way, and are honest about how weak the evidence is in the American case. A 2019 systematic review of about 500,000 intended home births split its results by how well integrated the setting was. In well-integrated settings the estimates for death sat close to no difference, for first and later babies alike. In less-integrated settings the estimates were higher, but the uncertainty ranges were so wide that they were compatible both with no difference at all and with a large increase. No absolute rates were published in the record available for this guide, so no ratio from it is printed. Its own published conclusion belongs here in its own words, as every other study in this stage has: The risk of perinatal or neonatal mortality was not different when birth was intended at home or in hospital.
One limitation is worth knowing: its evidence is thinnest for systems like the American one, where home birth is least integrated with hospital care. That review was partly funded by the Association of Ontario Midwives, which you are entitled to know, in the same way you are entitled to know who wrote the American studies.
A separate 2018 review of 28 studies reported that severe tearing and hemorrhage occurred less often in planned home births than in hospital labor units, and that There were no statistically significant differences in infant mortality by planned place of birth, although most studies had limited statistical power to detect differences for rare outcomes.
The second half of that sentence travels with the first.
"Have you transferred a woman there before? What happened?"
"When you tell me home birth is as safe, which country's data is that from?"