Our findings indicate that risk selection and outcomes for midwife-led birth units are satisfactory. A total of 6.9 % of the women who had planned to give birth in a midwife-led birth unit were transferred intrapartum and 3.8 % were transferred postpartum, due to circumstances relating to the mother or the child. Most transfers were uneventful, and the reasons most commonly cited were slow progress in the opening phase, the need for further pain relief, no delivery imminent more than 24 hours after ruptured membranes, discoloured amniotic fluid or changes to the fetal heart sound. Among nulliparous women who had planned to give birth in a midwife-led unit, 19.5 % were transferred intrapartum. This is a considerable proportion, and nulliparous women who wish to give birth in a midwife-led birth unit must be informed of this. However, the transfer rate in our study is clearly lower than the reported rate for midwife-led birth units in England, where observations indicate that 30–43 % of nulliparous women in midwife-led birth units are transferred to a hospital (10, 11).
The most common reasons for postpartum transfers were retained placenta, suspected tears that required repair by a doctor, or respiratory problems in the newborn infant. The use of vacuum/forceps, the incidence of sphincter injuries and the incidence of Apgar score < 7 after 5 minutes demonstrated satisfactory outcomes for deliveries in midwife-led birth units as well as for maternal transfers.
The reasons why 7.7 % of deliveries in midwife-led birth units were intended to take place elsewhere, may be that the women waited too long before setting off to hospital, or that the birth progressed so fast that they would be unable to get to hospital and that the only option available if an unplanned home birth or delivery in transit was to be avoided, was to travel to the midwife-led birth unit. Adverse weather conditions or non-availability of suitable means of transport were rare reasons for unintended deliveries in midwife-led birth units. It is also somewhat surprising that 1.4 % of births in midwife-led birth units are breech deliveries.
The organisation of maternity services and the number of maternity institutions in Norway will depend on factors such as maternal accessibility, quality of service, economy and political decision-making. It has been decided that we should offer a differentiated and decentralised service provision in a three-tier system of maternity institutions (midwife-led birth units, hospital maternity wards, maternity clinics) (4–6). These categories were formerly based on delivery rates and staffing requirements, but they are now based on quality requirements (12).
It is important that all types of maternity institutions are familiar with their own results. Information reported to the Medical Birth Registry is recorded under the actual birthplace. Consequently, the data that are routinely submitted to the Medical Birth Registry do not always provide an adequate basis for assessing the quality of midwife-led birth units and small maternity wards that requisition maternal transfers to larger maternity institutions. Each institution's results must be analysed according to the intention-to-treat principle, i.e. outcomes must be reported for all birthing women admitted, including those who are transferred intrapartum or postpartum. If a midwife-led birth unit can demonstrate good results for the deliveries that take place within the institution, but poor results for deliveries in transit, then the quality of the services provided by that institution is not necessarily good. This study sought to examine this particular issue.
The findings show satisfactory results for mother and child following deliveries in midwife-led birth units. Only 0.6 % of the children had an Apgar score of < 7 after 5 minutes, and we found no data to suggest that serious injuries or deaths in mother or child were related to the midwife-led birth unit delivery per se. The operative vaginal delivery rate was 0.4 % in standard midwife-led birth units. These results match earlier Norwegian and international studies (13)– (20). The study demonstrated a conspicuously high incidence of caesarean deliveries among planned births in modified midwife-led birth units; contrary to the intention, these were not limited to emergency caesarean sections.
The study sheds light on the reasons for intrapartum and postpartum transfers to hospital, at what point in the birthing process transfers took place, and the means of transport. A good ambulance service is clearly important for the midwife-led birth units.
In 2008, at the start of the project period, there were 55 maternity institutions in Norway, of which 14 were midwife-led birth units. In the course of the study period 1.4 % of all childbirths took place in midwife-led birth units. There are currently 47 maternity institutions, of which six are midwife-led birth units. The following eight units have been closed down:
Lykkeliten (2008), Føderiket, Lærdal and Steigen (2011), Valdres and Odda (2013), as well as Hallingdal and Mosjøen (2016). In 2016 a total of 0.7 % of all childbirths took place in midwife-led birth units (9). Norwegian health authorities have now recommended the winding up of modified midwife-led birth units (12), and there is currently only one such modified unit in the country (Lofoten).
Some maternity services are provided by institutions similar to midwife-led birth units (low-risk units) at Oslo University Hospital (the ABC clinic, Ullevål hospital), Stavanger University Hospital (Fødeloftet) and Haukeland University Hospital (Storken). These units were not included in the study, but it will be important for them to be able to document results for birthing women who are transferred from these low-risk units to ordinary maternity wards.
In Denmark, Sweden and Finland, maternity care services are more centralised than in Norway. In these countries there are very few maternity wards with less than 1 000 deliveries per year, and Norway has more than twice as many maternity institutions relative to the birth rate. Geography and settlement patterns may well be different, but this can hardly explain the large differences in the number of maternity institutions.
It is a strength of our study that it is population-based, national in scope and prospective. We were able to supplement the collected material with data routinely submitted to the Medical Birth Registry. The supplementary forms provided details about virtually all deliveries in midwife-led birth units during the study period, including the reasons for transfers to hospital, at what point in the birthing process transfers took place, what means of transport was used, and information about deliveries in transit. Furthermore, the supplementary forms provided detailed information about births that had been planned for a midwife-led birth unit, but that took place elsewhere.
An important weakness of the study is that the number of deliveries included is low, and that all findings associated with rare events such as severe maternal haemorrhage, still births, very low Apgar scores and neonatal death, must be interpreted with care.