Pregnancy and child

Safety of home births questioned

New research has placed home births “under scrutiny,” said BBC News. It reported that women who plan home births recover more quickly but there is a greater risk of the child dying.

The news is based on a high quality review of data from over half a million births from several Western countries, exploring how planned birth locations affect a number of birth outcomes for both mothers and babies. The findings are complex, and cannot simply be summed up as showing that hospital births are safer than home births. It is important to highlight that although it appears to show a greater risk of newborn deaths with home births, the absolute risk in either location is still very low (0.2% for planned home births and 0.09% for planned hospital births).

The researchers say that some of the higher mortality rate may be attributable to fewer instrumental or interventional deliveries with home deliveries. This theory cannot be proven or disproved by this research, but would need to be established by further study. Notably, although home births appeared favourable for certain maternal outcomes, the study was unable to shed any light on the risk of maternal death for either location. It is also important to highlight that when the analysis only looked at homebirths that had been attended by a certified midwife, there was no difference in risk of neonatal mortality compared to hospital births.

Where did the story come from?

The study was carried out by researchers from the Maine Medical Center, US, and presented at the 30th Annual Meeting of the Society for Maternal-Fetal Medicine in Chicago. Sources of funding were not reported. The study was published in the peer-reviewed, American Journal Obstetrics and Gynaecology.

The newspapers have correctly reflected the findings of this study. However, their reports that home births are ‘good for mothers’ must be interpreted with care. Though home births were associated with lower rates of certain outcomes such as vaginal tear, instrumental delivery, haemorrhage, infection and so on, it must be remembered that mothers with any identified pregnancy complications would be more likely to be planned to deliver in hospital rather than at home.
The study did not consider the mother’s experience of home birth or hospital birth and crucially, the important outcome of maternal mortality could not be assessed, as the researchers themselves highlight.

What kind of research was this?

This was a systematic review of reportedly all Western publications (predominantly cohort studies) that had reported outcomes for babies and mothers in relation to location of birth, e.g. whether at hospital or at home.

A systematic review is the best way of identifying all relevant studies and cohort studies assessing the relationship between a cause (planned location of birth) and an effect (outcome in the mother or child). However, when combining results of multiple studies, the differences in their methods, the populations included and assessments of outcomes, must be taken into account. A review should also consider whether the individual studies have accounted for all possible confounders that could be affecting the association.

What did the research involve?

The researchers carried out a search of the medical databases MEDLINE, EMBASE and Cochrane for studies published in English, with the aim of identifying “all studies, regardless of methods, comparing intended or planned home births to intended or planned hospital births for maternal and newborn outcomes”. The researchers specifically searched using the keywords ‘home childbirth’, ‘obstetric delivery’, ‘hospitalisation’, ‘hospital’ or ‘inpatient’, plus carried out sub-searches within these headings and searches using combinations of these terms. They looked at those studies concerned with the concepts of comparisons, planned births or birth outcomes.  

They looked at a number of interventions and outcome for both mothers and newborns:


  • Interventions: epidural analgesia, electronic foetal heart rate monitoring, episiotomy (surgical incision to widen the vagina and assist birth), operative vaginal delivery (forceps or vacuum), and caesarean delivery.
  • Outcomes: mortality, lacerations (>3 degree tear to the vagina or perineum), chorioamnionitis (infections of the foetal membranes), endometritis (infections of the uterus lining), wound infection, urinary infection, postpartum haemorrhage, retained placenta, and umbilical cord prolapse.


  • Outcomes: Five-minute Apgar score <7 (measurement of health and responsiveness of a newborn), prematurity ( less than 37 weeks), post-dates (more than 42 weeks), low birthweight (bottom 10% for gestational age or less than 2500g), large baby (top 10% for gestational age or more than 4000g), assisted ventilation requirement, perinatal death (stillbirth of at least 20 weeks or 500g, or death of newborn within 28 days of birth), and neonatal death (death of a newborn within 28 days of delivery)

The researchers carried statistical tests to take into account ‘heterogeneity’ (the differences between the retrieved studies) and combined results to give summary risk figures for maternal and newborn outcomes for both planned home or planned hospital delivery.

They also carried out sensitivity analyses to look the effect of including pre-1990 studies, lower quality studies and studies that had not clearly specified the location of birth.

What were the basic results?

Twelve studies (11 cohorts and one randomly controlled trial) were included, which covered a total of 342,056 planned home births and 207,551 planned hospital deliveries. Studies came from US, Canada, UK, Australia and several European countries.

Planned home births were associated with fewer maternal interventions, including epidural analgesia, electronic foetal heart rate monitoring, operative delivery and episiotomy (an incision to widen the vagina). In terms of maternal outcomes, mothers who had home deliveries had fewer infections, vaginal and perineal tears, haemorrhages, and retained placentas (no difference in the rate of umbilical cord prolapsed).

Of outcomes in the newborn, babies born at home were less likely to be premature, less likely to be of low birthweight, and less likely to require assisted ventilation. However, there was greater likelihood of the baby being born post-dates if delivered at home.

Planned home and hospital births were found to have similar perinatal (the period immediately before and after birth) mortality rates, though planned home births were associated with significantly greater neonatal mortality rates (deaths within 28 days of birth). These were two-to-three times as frequent (32 deaths in 33,302 hospital births [0.09%] and 32 deaths in 16,500 home births [0.20%]).

This observation was consistent across studies. The anticipated population-based attributable risk of neonatal death overall was 0.3% (i.e. 0.3% of neonatal deaths could be accounted for by birth occurring in the home rather than the hospital). The researchers noted an increased proportion of deaths attributed to respiratory distress or failed resuscitation in the home birth groups.

Applying sensitivity analyses that excluded poorer quality studies had little effect on the findings. However, when the researchers excluded studies of home births attended by people other than certified midwives, there was no significant difference between the newborn mortality rates associated with the two locations of birth.

How did the researchers interpret the results?

The researchers conclude that less medical intervention during planned home birth is associated with an almost-tripled neonatal mortality rate.


This is a high quality systematic review that appears to have identified all research assessing the differences in newborn and maternal outcomes associated with planned home deliveries and planned hospital deliveries. However, the associations seen should not be considered as a direct cause-and-effect relationship, i.e. it is an oversimplification to assume that planned birth location is directly or solely responsible for the birth outcomes seen.

Indeed, the principal limitation is that of attributing home or hospital birth as the actual cause of the outcome. For example, it is possible that home birth is associated with fewer instances of prematurity, low birthweight and assisted ventilation, not because home birth reduces the risk of this, but because mothers of babies who are identified as having some problem during antenatal care (e.g. growth restriction), would be more likely to be recommended a hospital delivery.

Likewise, mothers who have an obstetric or medical history putting them at higher risk (e.g. past history of postpartum haemorrhage) are more likely to be recommended a hospital birth. Consistent with this, the researchers noted that women planning home births tended to be at lower risk of complications and were less likely to be overweight or obese, giving birth to their first baby or to have history of previous pregnancy complications.

There are other key points to consider when interpreting this research:

  • Although home birth was associated with greater neonatal death (within 28 days), neonatal death is still very rare, and the absolute size of the risk is low (0.2% among planned home births and 0.09% among planned hospital births). The researchers calculated that only 0.3% of neonatal deaths could be attributable to birth occurring in the home rather than the hospital.
  • It is also important to note that there was no increased risk of neonatal death with home birth compared to hospital birth once the analyses excluded those studies of home births attended by people other than certified midwives. That is to say, when the home birth was assisted by a certified midwife, there was no increase in mortality compared to a hospital birth.
  • As the researchers consider, the higher neonatal mortality rates with home birth may be associated with the lower likelihood of instrumental or interventional delivery with home birth. However, this theory cannot be concluded from the research, and more study is needed to try and unpick the reasons behind these possibly associated occurrences.
  • Maternal mortality rates were an important outcome that was not able to be assessed. This was because the four studies that had considered this outcome (covering 10,977 planned home and 28,501 planned hospital births), experienced no maternal deaths. Therefore, more research is needed on this outcome. Additionally low Apgar scores could not be assessed.
  • The researchers were not able to account for certain potentially important demographic factors, particularly women’s age.

As the researchers rightly say, future study needs to be directed at identifying the factors that contribute to the apparently excessive neonatal mortality among planned home births, and also considering the effect upon maternal mortality.

NHS Attribution