birth of cupidAs many of you know, a recent Dutch study

[1] was conducted looking at all low-risk pregnancies by planned location of birth and found that a planned homebirth was as safe or safer than birthing in a hospital.  Of course, this data only pertains to the Netherlands in which homebirth is more common than a hospital birth and midwives are the primary caregivers for all low-risk births.  The system there is fully integrated with midwives and doctors working together instead of in competition (as you rarely see a doctor unless you have complications or a high-risk pregnancy), and frankly it’s a system I wish every country had.

However, a response to the article has suggested that the findings are misleading because of the maternal mortality factor.  Maternal mortality was not specifically discussed in the study at all (though according to the authors it was included in the overall analysis), but the authors have replied to the concern with information.  But first, let’s see what the hoopla is about…

The author of the reply has suggested that if the homebirth death rate is not zero, one cannot claim homebirth to be as safe as a hospital birth.  I don’t actually follow this line of reasoning to be honest, as birth in and of itself can be dangerous, regardless of location.  However, let’s take a look at the data on what the actual death rates were over this period for this low-risk group.  There were a total of 5 maternal deaths over 146,752 women; however, two of the deaths were completely unrelated to pregnancy (one was a severe allergic reaction, the other was a fall that led to a cerebral haemorrhage).  So we have 3 deaths.  Two were in the planned homebirth group (which was comprised of 62.9% of the women), 1 in the planned hospital group (37.1% of births).

Here are the explanations of the deaths:

One woman gave birth at home at 38 weeks gestation in midwife led care and was referred to hospital in the postpartum period by the general practitioner because of signs of HELLP syndrome. She died the following day due to brain haemorrhage.

One woman suffered from sudden collapse during labour, when she was already in secondary care, and died. Although no definite diagnosis was made at postmortem examination, a cardiac cause appeared to be most likely.

A woman who gave birth spontaneously was discharged after one day. On the fourth day postpartum she was readmitted because of profuse vaginal bleeding and shortness of breath. She had a sudden collapse and died. Postmortem examination showed sinus sagittalis superior thrombosis.

The person who is suggesting homebirth is not safe is specifically focused on the first case.  The author argues that this case reflects on the midwife and was a preventable death in the hospital and as such, the claim that homebirth is as safe as in a hospital is unsubstantiated.  But first let’s look at the other two.  The second death was a planned homebirth but had already transferred to the hospital (in Dutch terminology, secondary care) when she died, but it seems as though the cause was cardiac.  I’m not sure we can truly call this a pregnancy-related death but as she was in labour at the time, we shall. (Notably, I’m sure if this woman had been at home when she collapsed, those against homebirth would try to argue it was preventable in the hospital, but clearly we know that not to be the case as this woman was in the hospital at the time of death.)  The last death was in the planned hospital group and in said hospital.  I don’t know enough about sinus sagittalis superior thrombosis to make any comment on whether or not anyone should have been able to catch this or if it was due to care during pregnancy, etc.  I take it as valid that no one could have prevented this.

Now to the first death, which is being used to say that homebirth is not safe.  This one death.  Was the midwife to blame?  Was being at home to blame?  The authors included this bit of information about the case as well:

At 37 weeks she had been sent for consultation to hospital because of upper abdominal pain and headache. Her blood pressure was slightly elevated, she had proteinuria but her blood results were normal. She had been referred back to primary care. All cases of maternal death in the Netherlands are audited in a confidential enquiry. The conclusion of the audit in this case was that the woman should not have been referred back to primary care by the obstetrician and that she should have given birth in secondary care in hospital.

The hospital failed her, and that is the conclusion of those who have far more information than us.  This woman was not a low-risk pregnancy, but rather a high-risk that was missed by the doctor who did the bloodwork and exam.  Her midwife didn’t fail her, her midwife sent her for a consultation to the hospital which failed her.  Would she be alive if she’d been in the hospital?  Probably but not definitely as we know that women die from HELLP in hospitals too (there is a 1.1% maternal mortality rate for it in the United States[2]).  What is certain is that it immediately makes a pregnancy high-risk and absolutely the hospital is the safest place to be.  But it certainly doesn’t speak to low-risk pregnancies.

What can we say from this study on the safety of homebirth?  Well, if we consider the cardiac event death as a valid pregnancy-related death we have a death rate of 2.17/100,000 (1.08/100,000 if that death is unrelated to the pregnancy) and 1.84/100,000 in the hospital group.  The difference is not statistically significant and it is not a matter of power (though the author of the reply misleadingly says the study is “underpowered” to find this significantly different).  To find this difference statistically significant, one would need a total sample of just under 96 million (48 million per group).  The study’s power is fine, it’s that the effect is so tiny and small, it’s not going to be detected.  And there’s a term when you do find that you have enough power to detect tiny effects: the finding is said to be statistically but not practically significant.  But regardless it’s neither here.

Thus we have no differences in death rates in low-risk pregnancies and the remainder of the data shows that other measures of maternal morbidity favour homebirth.  Sadly one woman died who most likely would have lived if she had been properly diagnosed by the doctor at the hospital and then birthed at the hospital as a high-risk pregnancy.  It’s a travesty, but just as we don’t say that people should thus stay away from the hospital because a doctor erred, we also can’t say that homebirth is unsafe because of a hospital’s mistake.

Just a final aside, the Netherlands has the highest rate of homebirth in the Western world and a system that is midwife-led while obstetricians take on high-risk pregnancies and any complications that arise, they also have an overall maternal mortality rate of 6/100,000.  The US has a maternal mortality rate of 21/100,000.  Which system is safer for women?

[Image Credit: “The Birth of Cupid” by Master of Flora]

[1] de Jonge A, Mesman JAJM, Mannien J, Zwart JJ, van Dillen J, van Roosmalen J.  Severe adverse maternal outcomes among low risk women with planned home versus hospital births in the Netherlands: nationwide cohort study.  BMJ 2013; 346: f3263, doi:10.1136/bmj.f3263.