Summary of results
This study found that in this restricted cohort of women giving birth in England, 13.6% were recorded as having a waterbirth. Women of ethnic minority origin, younger women and women of more deprived socioeconomic status are less likely to give birth in water. For the mother, waterbirth was associated with reduced PPH and no association was shown between waterbirth and OASI. For the baby, there was no association between waterbirth and low Apgar score, and neonatal unit admission was less likely in the group born in water. This study therefore shows no association between waterbirth and these adverse outcomes for mother or baby.
Comparison with other studies
This study found that increasing age, with the exception of women over 40 years old, is associated with increasing likelihood of waterbirth. This association is consistent with a previous study finding that women under 25 are less likely to use water for analgesia during labour [8, 31].
In this cohort, socioeconomic deprivation was strongly associated with decreasing likelihood of giving birth in water. Women living in the most deprived postcode areas were less than half as likely to give birth in water as those in the least deprived group. An existing study similarly found that socioeconomically deprived women are less likely to labour in water . In our cohort, women of Black and Asian ethnic origin were substantially less likely to have a waterbirth than white women. This persisted even after adjustment for other factors such as birthweight and obesity which are associated with ethnicity. Recent studies from the USA also report ethnic differences between women who labour or give birth in water and women who do not [6, 9].
Although it is possible that the reason for this discrepancy is that women in these groups laboured in water but left the pool prior to delivery, it is probable that these disparities reflect inequitable access to birthing pools in England. It may be that some groups of women are not aware that waterbirth is an option available to them, or that some groups are less empowered and therefore less able to advocate for their own preferences during labour and birth. This is supported by a secondary analysis of the Birthplace cohort study which found that it is not only women experiencing socioeconomic deprivation who are less likely to labour in water, but also those who do not speak fluent English and those who are unsupported by a partner .
Obese women were found to be less likely to give birth in water. This finding is consistent with existing UK guidelines where women with a BMI of between 30 and 35 are not always offered care in a midwife-led setting . It has elsewhere been shown that, among women who intend to give birth in water, multiparous women are more likely to do so . Our study, in which the group who did not give birth in water includes all low-risk women who had a spontaneous vaginal birth, showed that women of higher parity are less likely to deliver in water. This may reflect individual preferences, lower request for analgesia or reduced time available to allow for waterbirth to occur.
Whilst some previous studies identify birth in water as an independent risk factor for OASI [16, 33] and others suggest an association with an increased risk of more minor genital tract tears, [1, 33, 34] the majority of the published literature reports either no association or a decreased incidence of severe perineal tears and episiotomy associated with birth in water [1, 6, 8, 9, 35,36,37]. In agreement with this, our study found no evidence of increased incidence of OASI following waterbirth.
Waterbirth in this cohort was associated with a reduced risk of a PPH of 1500 ml or more. This is to be expected, since a number of existing studies have also found waterbirth to be equivalent if not superior to birth not in water in terms of association with PPH [1, 35, 38]. However, this finding must be treated with some caution; although the finding was robust to a sensitivity analysis where the cohort was restricted to births that occurred in a midwife led setting and therefore without augmentation, we were unable to control for slow labour progress which is associated with PPH. Furthermore, there may be bias in the recording of PPH of 1500 ml or more, as quantifying blood loss in a birthing pool may be challenging. It is a strength of this study that we were able to adjust for parity and birthweight, both factors associated with OASI and PPH [39,40,41].
Although the possibility of rare but serious adverse neonatal outcomes remains, the conclusion may be drawn from the existing published evidence that, for most neonatal outcomes, there is no evidence of any significant differences between birth in water and birth not in water [9, 11,12,13, 34, 37]. This large study found no association between birth in water and low Apgar scores and also that babies born underwater were less likely to be admitted to a neonatal unit, thereby adding to the body of evidence supporting the safety for the baby of delivery in water.
Strengths and limitations
To our knowledge, this is the largest published study of births in water in the UK. Furthermore, since this study makes use of routinely collected data from an unselected population, the risk of selection bias is reduced. Included trusts were broadly representative in size and location of trusts throughout England. These results are, therefore, readily generalisable to the population in England.
For maternal and neonatal outcomes other than PPH, there is no evidence of a statistically significant association between water birth and adverse outcome. It is possible that this is because this study, although relatively large, is not large enough to detect smaller differences in adverse outcomes in women giving birth in water.
The central limitation is data quality and completeness. Data were only available for a minority of trusts in England; in others, the proportion of women for whom the field ‘waterdelivery’ was empty was too high to draw conclusions. Furthermore, no information was available about the women who labour in water but do not deliver in water. There is also uncertainty around the quality of recording of routine data relating to birth in water. We were not able to access paper clinical records and therefore no validation of this electronic data field has taken place. It is possible that a proportion of women who are recorded as having had a waterbirth may have spent time in a pool during labour, but not actually given birth in water.
This is an observational study and thus no conclusions about causation can be inferred from these results. Furthermore, there are likely to be unmeasured confounding factors that cannot be captured in this routinely collected dataset. This study attempts to limit these by restricting the cohort to women without risk factors requiring birth on an obstetric unit who gave birth by normal vaginal delivery and additionally through a sensitivity analysis including only women who gave birth in a midwife-led setting. However, even within this restricted cohort, it is likely that unmeasurable differences remain between women who delivered in water and those who did not.