Planned private homebirth in Victoria 2000–2015: a retrospective cohort study of Victorian perinatal data

Only 1 in 200 women have a homebirth in Victoria, Australia. This is despite international evidence that for healthy normal pregnant women, home birth is not associated with an increased rate of adverse perinatal outcomes [1‐7], or maternal morbidity [1, 3‐8] compared to similar women having a planned hospital birth, particularly if they are multiparous [9]. Whether homebirth is safe for women at high risk of adverse pregnancy outcomes is less clear. Findings from studies in populations which were mixed with regards to risk status report higher rates of adverse perinatal outcomes compared to similar women in hospital [10, 11]. In contrast those that specifically looked at high risk women, identified a reduction in adverse outcomes and obstetric interventions [12] or no difference (apart from a higher vaginal birth rate) in outcomes [13]. Homebirth in England however is more common and integrated into the health system. Whether these findings from England apply to Australia is not clear.

In 2013 a perinatal death occurred in Victoria during a planned homebirth in a woman who was identified as being at high risk of adverse pregnancy outcomes. Following the investigation of this case a coronial recommendation (Coroner Parkinson 2013) was made for the Health Minister to investigate and provide information about the rates of outcomes for planned homebirth in Victoria. This research directly addresses the coroner’s recommendation.

We aimed to quantify the number of women who planned to birth at home in Victoria between 2000 and 2015, to quantify transfer rates to hospital, and to compare the rates of perinatal and maternal outcomes and obstetric interventions for women who planned to birth at home compared to those who planned to birth within a hospital setting by maternal ‘risk’ status.

We undertook a study using data that are legislatively routinely reported on all births in Victoria, Australia to the Victorian Perinatal Data Collection (VPDC). For every birth ≥20 weeks gestation(or ≥ 400 g birth weight if gestation not known), regardless of place of birth the VPDC receives a standardized report detailing over 100 items regarding maternal characteristics, obstetric conditions, procedures and outcomes, perinatal mortality and morbidity and birth defects.

De-identified data for all births from ≥37 weeks gestational age from 2000 to 2015 were extracted. Babies with congenital anomalies or where a caesarean birth was planned were excluded. The majority of planned home births in Victoria occur under the care of a privately practising midwife. Two small public homebirth programs, caring for about 90 women a year in total, were established in 2014. Due to the small numbers and because they have only been operating for two of the 16 years studied, we excluded them from analyses. Women were identified as being of low or high risk informed by the Australian College of Midwives(ACM) guidelines for consultation and referral [14]. These are the criteria referenced in public funded homebirth programs to inform eligibility for planned homebirth.

Specifically, a high risk pregnancy was defined as: a multiple pregnancy, a post-term (> 41 + 6 weeks of gestation) pregnancy, a non-cephalic presentation in labour, obesity(BMI Class 2 or greater: data only available from 2009 onwards), a prior caesarean, previous uterine surgery, grand multiparity(≥5 previous births), any significant maternal medical condition such as pre-existing diabetes, hypertension, renal, cardiac, liver, respiratory, endocrine, immunological, renal, or gastrointestinal disease as determined by individual ICD-10 codes. All other women were classified as having a low risk pregnancy.

Data were largely complete. Where data were missing a ‘not reported’ variable was used. BMI data were only available from 2009 onwards. Validation of the accuracy of the dataset has been reported [15]. For the variables used in this study that have been validated (demographics, mortality outcomes and obstetric interventions), all had accuracy above 90% except for BMI.

From 2000 to 2015, 3945 women planned to give birth at home with a privately practising midwife and 829,286 women planned to give birth in hospital. Of those who planned to give birth at home 3202(81%) were identified as having a low risk pregnancy and 743(19%) were identified as having a high risk pregnancy. Of those planning a hospital birth 701,058(85%) were identified as low risk and 128,228(15%) were identified as high risk. The number of women planning a homebirth, regardless of risk status, increased from 131 per annum in 2000 to 315 per annum in 2015. This reflected a 2.6 fold and 1.8 fold increase for low risk and high risk women, respectively (Fig. 1). The characteristics of the women are presented in Table 1. Regardless of risk status, women who planned a homebirth were older, had a normal BMI, were Australian-born, were parous and were of higher socioeconomic status compared to their equivalent group who planned to give birth in hospital. Of the 3202 low risk and 743 high risk mother-baby pairs who planned a home birth, 2888(90%) and 628(83%) gave birth at home.

Table 2 summarises the frequencies of risk factors for women with high risk pregnancies. Among women who planned to birth at home, 75(10%) had more than one risk factor and among the planned hospital group 16,438(12.5%) did. The most common risk factors for those who planned to birth at home were having had a previous caesarean birth (41.9%) and being post term (32.5%). For high risk women who planned to give birth in hospital having a pre-existing medical condition was the most common risk factor (31%) followed by having had a previous caesarean delivery (28.8%).

Overall we found that home birth, regardless of risk status, was associated with significantly lower rates of obstetric interventions and combined overall maternal and perinatal morbidities. For low risk women planned homebirth was also associated with similar risks of perinatal mortality, however for women with recognized risk factors, planned homebirth was associated with significantly higher rates of perinatal mortality.

Until now, the outcomes from home births in Victoria, particularly among “high risk” women was unknown. We found that for women with risk factors home birth was associated with a significantly increased risk of neonatal death and some perinatal morbidities. Our findings broadly accord with those of the Birthplace in England study that showed that planned homebirth among “high risk” women was associated with a 2-fold increased risk of a composite perinatal outcome (intrapartum stillbirth, early neonatal death, hypoxic ischaemic encephalopathy, meconium aspiration syndrome, brachial plexus injury, fractured clavicle or humerus and admission to NICU for greater that 48 h), albeit that increase was not statistically significant [12]. The Birthplace study also reported a moderate (1.20 95% CI 0.41–3.44) but non-significant increase in risk of adverse perinatal outcomes among a small number of women planning a VBAC at home compared to in hospital [13]. Not all adverse perinatal outcomes were increased in the planned homebirth group however. High risk women planning to birth at home experienced significantly reduced rates of birth trauma and intrauterine hypoxia. We also observed that high risk women who planned to birth at home experienced lower rates of maternal morbidities.

Qualitative studies have been undertaken to understand why women with certain risk factors, particularly those who have had a previous caesarean choose to birth at home. These studies identify a recurrent theme that women want to avoid another caesarean [16]. Our results appear to vindicate their fears. In Victoria, high risk women who planned to birth in hospital had over 3.5 times the rate of unplanned caesarean than those who planned to birth at home. We were not able to address whether this difference was clinically justified or not. However, recently, Dutch investigators aimed to determine if the differences in intervention rates between home and hospital births reflected over or under treatment in these groups [17]. They found both. Our findings here and the observation that there is 10-fold variation in the rate of attempted VBAC rates in maternity units in Victoria [18] would suggest the same is occurring here. There is an urgent need for health services to reassess their policies and support for women planning a VBAC to address these discrepancies in care. Previous research has also identified that women with risk factors also choose to birth at home, do so following a traumatic or bad experience in hospital [19]. Addressing discrepancies in care, and providing support and understanding the experiences of women following adverse pregnancy and labour experiences after giving birth in hospital is needed. Interestingly, despite much debate around absolute risk [20], uterine rupture is commonly cited as a reason that women who have had a previous caesarean should not plan a vaginal birth [21]. The only uterine rupture that occurred in the high risk women who planned a home birth was in a woman who had had more than one previous caesarean section.

Reporting the outcomes for low risk women planning a home birth was necessary so that if we found higher rates of adverse outcomes for high risk women, as we did, we could consider whether this was a feature of home birth per se in Victoria. This was not the case. Perinatal outcomes from planned homebirth in low risk women were similar to those of planned hospital birth. The low risk women planning a home birth also had fewer obstetric interventions and lower rates of maternal morbidities. Our findings were consistent with other studies of low risk women giving birth at home [1].

Seventeen percent of high risk and 10% of low risk women required transfer to hospital. Women and/or babies requiring a transfer to hospital in labour are considered at most risk of adverse outcomes [22]. We did not find that. Instead, we found that the majority of adverse perinatal outcomes were among the women who actually gave birth at home. Understanding why this is the case is needed. A recent survey of intrapartum transfers with 13 Australian midwives identified barriers to timely transfer of women to hospital [23]. Midwives reported health services either refusing to accept transfers or health services/hospital staff being hostile to the midwives when they arrived [23]. This was in contrast to experiences some of these midwives had when they worked as a homebirth midwife in the UK [23]. The rate of homebirth in Australia is much lower than the rest of the world. In Australia only 0.3% of women plan to give birth at home. This is in contrast to 3.4% of women in New Zealand, 2% in Canada and the UK, and 20% in the Netherlands. The organisation of homebirth in Australia also differs. In Victoria the majority – almost three quarters –of women having a homebirth do so under the care of a private midwife. This differs from the UK, Netherlands and New Zealand where homebirth is integrated and offered within the public health care system.

We report the rates of selected perinatal, maternal and obstetric outcomes in Victorian women with and without risk factors, by their planned place of birth. Our findings support current guidelines [14, 24‐26] and should be useful to women who are making informed choices about where to have their baby and to health services and policy advisors who are planning the provision of maternity care.