Results 1: study population

Site participation

All 26 eligible NHS study sites agreed to participate. Records were received from all 26 sites relating to births that had occurred over the period from 1 January 2015 to 30 June 2022. The number of sites reduced during the study period, as some sites changed their MIS. Participating sites across the UK are shown in Figure 4, and their characteristics are shown in Table 5.

FIGURE 4 Participating study sites. (Figure not reproduced here; see the original article.)

Women opting out of the study

Sites were requested to provide the number of women who had opted out, and 24 sites provided these data. Sixty-five women (range 0–16 women per site) opted out of the study after site opening, with 11 (42%) sites recording no opt-outs during the data collection period.

Pool use and waterbirths throughout the study period

Figure 5 provides information on the proportions of women using a pool for labour, and the proportions of waterbirths, at all study sites from quarter 1, 2015 to quarter 2, 2022. The proportion of women using a pool during labour was highest in quarter 3, 2016, when 11.9% of women used a pool during labour and 6.0% of woman gave birth in water; in quarter 2, 2022, the proportions were 7.7% and 4.0%, respectively.

FIGURE 5 Rates of pool use and waterbirth across the study period. (Figure not reproduced here; see the original article.)

Study populations

The primary study aim was to compare the maternal and infant outcomes for ‘low-risk’ women who gave birth in water (group 1) against ‘low-risk’ women who left the water prior to birth for reasons other than clinical need (group 2). Since the secondary study objectives were to set pool use and waterbirth in the context of NHS care, we also defined:

• group 3 – women who used a pool during labour, with no complexities identified in the antenatal records/or by the midwives at the time of pool entry, and who were transferred to obstetric care during labour
• group 4 – women who used a pool in labour, who had recorded complexities in their antenatal record or (for births after site opening) recorded at pool entry
• group 5 – women with no record of pool use during labour or birth.

The flow chart of birth records by these groupings are shown in Figure 6.

FIGURE 6 Flow chart of the study. AMU, alongside midwifery unit; FMU, freestanding midwifery unit; Q, quarter. (Figure not reproduced here; see the original article.)

A total of 956,307 records were received from 26 sites relating to births from 1 January 2015 to 30 June 2022. A total of 87,450 (9%) birth records had a record of using water immersion as labour analgesia/alternative pain relief, or for birth (groups 1–4: women who used a pool); and 868,857 (91%) birth records had no record of water immersion (group 5: women who did not use a pool). After exclusions for duplicate records, missing data within records, intrauterine deaths prior to labour and unattended births [410 (0.2%) women who used a pool; 85,065 (9.8%) women who did not use a pool], 87,040 (99.8%) birth records of women who used a pool and 783,792 (90.2%) women who did not use a pool remained for analysis. Linkage to the NNRD occurred in 5659 birth records (6.5% of all women who used a pool). Few records were true duplicates in the same data set. For some sites, data relating to births without pool use, in one or more period, were sent to the study team more than once.

From 87,040 eligible birth records for women who used a pool, at least one POOL study field was completed in 24,911 (28.6%) records, indicating that the birth occurred after the site opened to the study.

Pool (groups 1–4) and non-pool (group 5) users

The study planned to describe the proportion and characteristics of women who used a pool for labour or birth compared to women who do not use a pool. The characteristics of women (and their infants) who used water during labour or birth and those that did not are summarised in Table 6. Women using water immersion for analgesia were more likely to come from a White ethnic background than women not using a pool for analgesia (81.0% vs. 70.1%, respectively), with disparities seen in women of an Asian or Asian British ethnic background (4.2% vs. 12.6%, respectively) and in women of a Black, Black British, Caribbean or African ethnic background (1.3% vs. 4.0%, respectively). Women who used a pool were also more likely to understand the English language (85.0% vs. 81.0%, respectively). There was some indication that women who used a pool were more likely to reside in more affluent areas than women who did not use a pool (least deprived quintile = 20.7% vs. 15.2%, respectively). Pool and non-pool users were comparable for maternal age at birth (mean 30 years), gestational age at birth (39–40 weeks), infant’s sex (51% male) and head circumference (mean 34.6 cms).

Group 4: women with complexities recorded during pregnancy or at the time of pool entry

Defining women with complexities

Complexities (or risk factors) were defined using a combination of data sources, using (1) the existing historic fields and (2) the new ‘POOL’ fields completed by the midwife following birth, to ensure a consistent definition of risk in study populations across all outcomes.

1. When the existing historic maternity fields were applied to the cohort of 87,040 birth records, 12,787 (14.7%) were deemed as having a recorded risk factor (see Appendix 9). The number of women identified as high risk from historic records was likely to have included some women unnecessarily, and for this reason, no further analysis was undertaken for group 4 identified by the existing fields in the maternity records.
2. From the 24,199 birth records completed after the site opened to the POOL Study, by the midwife following birth, a total of 23,800 (98.4%) birth records had utilised the new POOL fields to record risk factors present at the time of pool entry [recording either condition(s) were present or that no risk was present (no/none)] (Table 7). A total of 2725 (11.4%) records were flagged as the mother having at least one complexity at pool entry.

When historic and new midwife-completed fields were used in combination, alongside any mention of conditions identified from other fields, 13,811 (15.9%) of the whole cohort were found to have at high-risk conditions (group 4), with 73,229 (84.1%) remaining in groups 1–3 (see Figure 6).

Characteristics of women with complexities

The study planned to establish and describe the characteristics of, and outcomes for, women with known risk factors, who used a pool during labour. For these 2725 births flagged by the midwives as the mother having at least one risk factor at pool entry (‘high-risk’), 1310 (48.1%) had a waterbirth recorded and 1415 (51.9%) did not, indicating that they had left the pool to birth (Table 8). These proportions differed by parity; nulliparous women were less likely to have a waterbirth than multiparous women (29.8% vs. 59.8%, respectively). Eighty-three per cent of women (n = 2270) had a spontaneous birth, although this rate was lower for nulliparous women when compared to multiparous women (68.8% vs. 92.6%, respectively). A higher proportion of nulliparous women had an emergency caesarean section and forceps (11.0% and 13.7%) compared to 3.3% and 2.8%, respectively, of multiparous women. A higher proportion of nulliparous women (with a spontaneous birth) experienced OASI compared to multiparous women (with a spontaneous birth) (25.7% vs. 1.9%, respectively). Similar patterns were observed, with their infants experiencing ‘adverse infant outcomes or treatment’ (6.8% vs. 4.2%, respectively).

Group 3: women using a pool without complexities (‘low risk’), but with additional monitoring or interventions before birth

Of the 73,229 women without antenatal complexities, 12,827 (17.5%) received additional monitoring or interventions before or during birth. These included women who received analgesia incompatible with pool use and left, or did not return to the pool for clinical indications, and were recorded as receiving additional monitoring, obstetric and/or anaesthetic interventions prior to birth (Table 9).

Groups 1 and 2: intervention – women giving birth in water and comparator – leaving the pool to give birth

Women who used a pool in labour, with no risk factors in their antenatal record or recorded at pool entry, without additional monitoring or interventions before birth, who had a waterbirth, were allocated to group 1.

Women who used a pool in labour, with no risk factors in their antenatal record or recorded at pool entry, without additional monitoring or interventions before birth, who gave birth out of water, were allocated to group 2.

Of 60,402 birth records that formed groups 1 and 2, 39,627 (65.6%) were recorded as waterbirths (group 1) and a total of 20,775 (34.4%) were recorded as leaving the pool before giving birth (group 2). As only 0.4% (n = 234) had null entries for waterbirth, and with no other indication of a waterbirth recorded in other fields, these women were assumed to have given birth out of water. Births partially born into water (e.g. head/buttocks) were recorded for 323 (3.0%) of all waterbirths recorded after site opening, including 142 woman who stood up between birth if their baby’s head and body without shoulder dystocia.

Groups 1b/2b: ‘low-risk’ women with midwife concerns prior to birth

To identify women in groups 1b/2b, we used a combination of existing fields (e.g. maternal/fetal intrapartum problems during labour) and a new MIS field (‘If the woman left pool prior to birth – did any of the following develop between leaving the pool and birth’). A total of 2128 of 60,402 (3.5%) records [group 1b: 268/39,603 (0.7%); group 2b: 1860/20,565 (9.0%)].

Groups 1a/2a: ‘low-risk’ women with no midwife concerns prior to birth

Within these groups were women for whom midwives had not recorded clinical concerns that were present prior to birth (groups 1a/2a).

Site variation

Table 10 shows the number and rate of women who did not use a pool (group 5), those using a pool by risk status (groups 1–4) and the rate of waterbirth by sites. After data cleaning, including of duplicate records, a total of 869,744 birth records were received over the study period of which 87,040 (10.0%) related to women who used a pool and 782,704 (90.0%) were related to women who did not use a pool. The number of births that an individual site contributed to the study ranged from 6804 to 65,667. Rates of pool use varied by site (between 1.6% and 22.9%) and in the proportion of women using a pool who gave birth in water (range 32.7–82%). Overall, 15.9% of women using a pool had a risk factor recorded in their antenatal record or at the time of pool entry (group 4), although considering the issues with the MIS, this may be an overestimate as they may have included some women unnecessarily. Among women without risk factors recorded, 14.7% of women who used a pool developed complications in labour with additional monitoring or interventions before birth (range between 7.6% and 21.7% by site).

Characteristics of women using a pool

The study planned to characterise and describe outcomes for women with identified risk factors at labour onset, who used a pool during labour (group 4) and women who received additional monitoring or interventions before birth (group 3) (Table 11). The linkage rate of infants to NNRD data was higher in infants of mothers who received additional monitoring or interventions before birth (group 3) (13.6%) than those in the antenatal period with risk factors (8.9%) (group 4) and mothers with no risk factors or monitoring or interventions (group 1: 3.7%; group 2: 5.8%). Waterbirths were recorded in 48% (n = 6581) of women with antenatal risk factors. Among the 12,827 women who had recorded additional intrapartum monitoring or interventions, including where required transfer to an OU (group 3), 62 women (0.5%) were recorded as subsequently having a waterbirth.

Maternal age, gestational age and ethnicity were comparable across all groups with lower pool useamong most deprived women. Most births without complications (groups 1 and 2) occurred in midwifery-led settings (83.9% and 75.4%, respectively), with a smaller proportion in OU (11.1% and 22.1%, respectively); there was a higher involvement of obstetric care among women in group 4 (48.6%).

Table 12 shows the breakdown of mode of birth for all low-risk mothers (groups 1–3) by parity. Nulliparous women are more likely to have an instrumental birth, with 9.9% having forceps, 6.0% vacuum extraction, with 5.9% resulting in an emergency caesarean section, compared to 0.7% of multiparous women. Rates of vaginal breech births were comparable.