Audit of a new model of birth care for women with low risk pregnancies in South Africa: the primary care onsite midwife-led birth unit (OMBU)

In developing countries, perceived poor quality of care and the medicalization of childbirth are both important barriers to facility-based birth [1]. A recent community-based survey of 1898 women living in rural, low income settings in the United Republic of Tanzania reported that 42% had bypassed their local primary care facility without a referral to give birth at a higher level of care, usually a government hospital [2]. Despite logistical challenges and added costs, bypassers were more likely than non-bypassers to report being very satisfied with the overall birth experience, and to rate the quality of care higher. It was concluded that gradually shifting birth care from primary care clinics to health centres and hospitals in this setting might improve health and experiential outcomes, as well as improving health system efficacy [2]. However, such a shift could overload the higher level of care services and result in unnecessarily medicalized and expensive care for healthy, low-risk, pregnant women. Against a backdrop of persistently high maternal mortality ratios in Sub-Saharan Africa, there is an urgent need to re-evaluate the models of birth care available to women in this region.

In South Africa, a middle income country, the national health system is based on the primary care model. The objective of this model is that the majority of people with less serious conditions receive care in a primary care setting. For women with ‘low-risk’ pregnancies, primary care clinics provide antenatal and postnatal care, and community health centres provide both antenatal/postnatal care and a 24-hour birth care service; women with ‘high-risk’ pregnancies receive doctor-led antenatal and birth care at secondary or tertiary care hospitals [3],[4]. This model works well for antenatal care; however, the theory that women can be antenatally triaged to receive low-risk birth care at a primary care community health centre has two fundamental flaws in practice:

One of the consequences of this dilemma is that women who prefer to deliver in what may be perceived to be a ‘higher quality’ facility may use various strategies, such as arriving at the hospital in advanced labour in the hope that it will be too late to be sent away to deliver at the community health centre. Such strategies may result in birth before arrival at the hospital and put both mothers and babies unnecessarily at risk of adverse pregnancy outcomes.

Inevitably, there are some low-risk women who receive birth care in secondary and tertiary hospitals in South Africa. Therefore, in some of the larger hospitals, the obstetric unit (OU) is divided into ‘low-risk’ (midwives’) clients and ‘high-risk’ (doctors’) patients (some of whom have normal births). However, this model undermines the principle of primary care for low-risk pregnancies. In addition, low-risk women who are admitted to the OU are cared for by hospital staff and use expensive hospital resources. A study from New Zealand found that low-risk women receiving midwife care in a secondary care setting compared with a primary care setting had higher rates of caesarean section and admission of their babies to the neonatal intensive care unit [12]. Thus, once in an OU, there is an increased tendency for obstetric interventions to be utilized which are not appropriate for low-risk women, leading to a cascade of unnecessary interventions [13].

To enhance the normality of childbirth, many industrialized countries have introduced midwifery units which provide holistic, high quality, midwife-led care. This type of midwife-led care compared with doctor-led or shared care models has been found in a Cochrane review of 13 randomised controlled trials conducted in the UK, Australia, Canada, Ireland and New Zealand to be associated with an increased likelihood of a spontaneous vaginal birth, fewer obstetric interventions (instrumental delivery, episiotomy and epidural analgesia), and a trend towards greater maternal satisfaction and cost-effectiveness [14]. In the UK, depending on the location, women can choose to access midwife-led care in the form of ‘alongside midwifery units’ (AMUs) which provide midwife-led care based at the same site or in the same hospital as doctor-led obstetric units (OUs) [7]. Midwife-led care may also be provided on sites geographically separated from hospital OUs; these are known as ‘free-standing midwifery units’ (FMUs) [7]. Whilst the primary care birth service in South Africa may be staffed by midwives and provided in ‘free-standing’ centres, it is qualitatively very different from the UK model of midwife-led units in that it lacks the emphasis on ‘normality, continuity, advocating autonomy and building relationships with mothers’ [15] that is key to midwife-led models.

A commentary on the organisation of maternity care globally highlights the importance of sharing good practice models among countries and calls for research on effective ways of ‘revitalizing normal birth’, particularly in middle income countries [15]. In this paper, we propose a model of birth care adapted to the South Africa health system called the primary care onsite midwife-led birth unit (OMBU). Similar to AMUs in some respects (location and approach), it combines the benefits of giving birth on the premises of a hospital, with those of a midwife-led, low-cost, and low-intervention primary care birth. A unique aspect of the OMBU is that it is staffed, administered and funded by the primary care services, not the hospital. This makes it a useful model for countries where primary care is distinct from secondary/tertiary services, as is the case in many low-income countries such as Tanzania [2].

This audit used routinely collected data submitted to the Provincial Health Service on a monthly basis. The data were collected in the same way by dedicated staff of the OMBU and hospital labour ward and analysed retrospectively by the authors. The reliability of data sets before and after the intervention were thus comparable. Total number of births, maternal and perinatal mortality data, as well as mode of delivery data for the audit were collected retrospectively from Frere Hospital for the 12 month period from 1 January 2011 to 31 December 2011 before the OMBU opened, and for the 12 month period 1 July 2012 to 30 June 2013 after the OMBU opened. We allowed three months on either side of the March 2012 OMBU opening to ensure that the audit periods were a good representation of the situation.

The total number of births on the hospital premises increased from 6352 during the earlier 12 month period to 7375 after the OMBU was established, representing a 16% increase. The number of births in the hospital OU decreased from 6352 to 5764 births, representing a 9.3% reduction, with 1611 women giving birth in the OMBU (22% of total births, excluding women referred to the OU due to complications arising during labour). Despite the increase in the total number of onsite births, the number of maternal deaths decreased from 14 deaths during the earlier 12 month period, to six deaths following establishment of the OMBU, equivalent to maternal mortality ratios of 220 and 81/100000 pre-and post-OMBU, respectively. Similarly, there was a reduction in perinatal mortality rates from 43 to 34 per 1000 births pre- and post OMBU (see Table 1 and Figure 1). Caesarean section rates were lower after the OMBU was established, at 35% of deliveries compared with 38%. These results should be interpreted with caution as various factors may have contributed to the observed reductions. These include changes in the risk profile and improvements in services unrelated to the OMBU.

These data are presented for illustrative purposes; we have not performed any statistical comparisons as these would not be meaningful, given the potential for bias in observational data. The high mortality and caesarean section rates reflect the fact that Frere is a high risk referral unit. In the South African public health system, the distinct separation of primary from secondary/tertiary services is usually geographic. There is an understandable reluctance of health authorities to have primary and secondary/tertiary services on the same site because of the risk of an inappropriate shift of low-risk cases to the higher level of care, with increased cost. However, as set out in this paper, care during labour is and must be an exception to this principle. We have shown that having the primary care services within the hospital can, in fact, result in a reduction in the number of births in the secondary/tertiary setting. Although the interpretation of the data from this audit is limited, it is possible that the OMBU contributed to the observed reduction in perinatal and maternal mortality rates. Other factors could also have contributed, such as a change in the patient profile, improved medical staffing, improvements in treatment for women with HIV due to the progressive government policy, and other unknown factors. What has become apparent to OU staff at Frere Hospital is that the quality of care in the OU has improved due to the reduced number of low-risk women giving birth. Furthermore, there appear to be many other advantages, both for staff and for the women using this service. We plan to conduct further research to evaluate these advantages and refine the model to further enhance the normal philosophy of childbirth. To evaluate whether the observed reductions in maternal and perinatal deaths represent real risk reductions, a cluster randomised trial would be necessary.

By incorporating many of the features of normalised midwife-led care with close proximity to doctor-led emergency care if required, the OMBU model has great potential to improve the quality of care provided in primary care settings in South Africa. There are many advantages for pregnant women, hospital staff and primary care staff with this model, which may also prove more cost-effective than the standard primary care model currently used in South Africa. A unique aspect of the OMBU is that it is staffed, administered and funded by the primary care services. We suggest that primary care OMBUs could be considered for any secondary/tertiary care hospitals in South Africa and possibly other countries with similar primary care health system models.