The day in the life of a maternity ward: factors influencing perinatal mortality in rural South Africa and Western Australia.

The perinatal mortality rate (PNMR) is the most sensitive indicator of obstetric care. For developed countries such as Australia, in particular, Western Australia, the rate for babies over 1000g is usually less than six per 1000 births. For developing countries, as is the case for South Africa, the PNMR ranges from 30–200 per 1000, a much higher rate which paints a very different picture.

I have been very fortunate in my medical career to have been given the opportunity to work in both a developing country, South Africa, where I grew up and studied medicine, and a first-world country, Australia, where I currently work as a registrar in O and G. In my practice I can appreciate that a maternity ward is not simply a machine where the outcomes can be predicted from the input. Obstetric care is, rather, a complex adaptive system where outcomes or the ‘output’ can be unpredictable and in different hospitals, especially in rural areas, the same inputs might have completely different outcomes. Hence, changing the health system so the quality of care can be improved is a complex intervention, particularly in obstetrics where the obvious differences in perinatal mortality rates are so unmistakably contrasted in the environments where I have worked.

This article aims to share my unique experiences in comparing the two rural hospitals’ maternity profiles from 2008: Bethal Hospital, Mpumalanga, a district hospital in South Africa; and Bunbury Hospital, a similarly sized regional hospital in Western Australia. The comparable directorates are those components that highlight areas addressed in the everyday happenings a maternity ward, such as equipment and staffing. The second component is to compare and extrapolate already published perinatal mortality rates and literature of South Africa and Western Australia. Stillbirths and neonatal death rates were also analysed as well as causes of weight-specific mortality rates comparing these differences in a first- and third-world setting.

Table 1. Comparing PNMR in South Africa and Western Australia in 2008.

The World Health Organisation (WHO) definition differs significantly from the Australian definition of the perinatal period. Therefore Australian data include babies of at least 400g (or at least 20 weeks if birthweight is unavailable) while the WHO definition commences at 500g (22 weeks if birthweight is unavailable). In addition, the WHO defines perinatal deaths as less than seven days (defined as ‘early neonatal period’ in Australia) while Australia includes deaths up to 28 days. Perinatal mortality rates of Aboriginal and Torres Strait Islander babies of Australia are therefore not comparable to rates for Indigenous populations in the other countries, such as South Africa. Owing to the well-documented discrepancies of PNMR, international comparisons are difficult to analyse, as demonstrated by this article.

The PNMR is nonetheless an important and unique health status indicator, since it addresses the two related issues of late fetal deaths and early infant deaths, many of which are considered preventable. In South Africa, in 2008, there were 1 836 439 births (DHIS data 2010) and 64 883 perinatal deaths (PPIP database), giving South Africa a perinatal mortality rate of 35.3. Among the babies born in 2008 in Western Australia, there were 30 234 births, 224 stillbirths and 57 neonatal deaths, providing a perinatal mortality rate of 9.2 (20 weeks and more) per 1000 total births.

The perinatal mortality rate above is calculated by birth weight of >500g, which gives a figure of 6.0 for Western Australia by WHO, whereas it is 9.2 when calculated by Australian definition of >20 weeks. The perinatal mortality rate in 2008 for babies with Indigenous mothers in Western Australia was 19.0 per 1000 total births compared with the rate of 8.6 per 1000 total births for babies with non-Indigenous mothers.

In Western Australia a neonatal death (the death of a live-born baby during the first 28 days of life) is more likely to occur before the end of the first day of life. In 2008, 43.9 per cent of neonatal deaths occurred in babies aged less than one day. Among the neonatal deaths whose final neonatal cause of death was related to immaturity, the most common diagnoses were extreme immaturity and hyaline membrane disease and among those with hypoxia the common diagnosis was hypoxic ischaemic encephalopathy and meconium aspiration.

Table 2. Neonatal mortality rates per weight category in 2008.

South African neonatal death statistics detail the probable avoidable factors related to early neonatal deaths. In early neonatal deaths due to immaturity, administrative factors were the most common, with lack of facilities, transport and staff featuring prominently. In early neonatal deaths due to hypoxia, the skill of the healthcare provider recorded the highest numbers of probably avoidable factors, almost three times those of healthcare provider-related deaths due to immaturity. Lack of skills seems to be the major problem associated with these deaths.

The prevalence of stillbirth is on average three times greater in the less developed areas of the world than in the more developed areas.³ In Western Australia in 2008, the two major primary causes principal of fetal deaths were extremely low birth weight (<1000g) 33.9 per cent and lethal birth defect, 28 per cent. In Australia, the stillbirth rate in 2008 was 7.3 per 1000 live births and in South Africa, just over three times this at 22.3 per cent. There are three principal obstetric causes of stillbirths identified in South Africa. Most perinatal deaths in South Africa fall into the category unexplained stillbirths (22 per cent). The second most common primary obstetric cause of death is spontaneous preterm birth (53 per cent being less than 1000g). The third most common cause is intrapartum asphyxia (labour related asphyxia, meconium aspiration and cord around neck) and birth trauma.

Extrapolating the data, the highest fetal death/stillbirth rate in Western Australia in 2008 was in the teaching hospital (22.7 per 1000 live births). The referral of mothers in Western Australia with a high-risk pregnancy and/or known stillbirths is illustrated by the fact that nearly two-thirds of fetal deaths (63.4 per cent) were delivered in the metropolitan teaching hospitals. In South Africa in 2008, the PNMRs are low in the Community Health Centres 10.41 per 1000 births (1500 perinatal deaths), which suggests the referral system is working reasonably well. The high rates at the tertiary level in South Africa may also be because the sickest patients end up there, as in Western Australia. However, in South Africa some of these hospitals also function as regional hospitals when there are no regional hospitals in the area.

Complications of hypertension and antepartum haemorrhage are the next two categories that still feature as mortality factors in South Africa. With regards to antepartum haemorrhages, 80 per cent were classified as abruptio, the rest being due to placenta praevia, seven per cent, antepartum haemorrhage of unknown origin, nine per cent and the remained as unspecified antepartum haemorrhage at four per cent. If both pre-eclampsia/eclampsia and abruptio placenta are combined the proportion of perinatal deaths is 22.8 per cent, becoming the most prominent cause of perinatal deaths in South Africa.

Table 3. Stillbirth rate per weight category in 2008.

Recommendations

Implementation of national standard perinatal mortality audit programs aimed at improving the quality of care could substantially reduce perinatal mortality in all settings. Some hospitals in South Africa only complete the first section of PPIP, i.e. the number of births and deaths per birth weight category and do not complete the causes of death or avoidable factors sections. Tools have now been developed and produced for the audit of the quality of care, protocols of management and accreditation of services in Southern Africa. The care of the mother – antenatally, intrapartum, postnatally – and of the newborn infant have been shown to be mutually complementary and should be dealt with together as a continuum in any intervention. Better data on numbers and causes of perinatal mortality are needed, and international consensus on definition and classification related to stillbirth and perinatal mortality rates is a priority. All parents should be offered a thorough investigation, including high-quality autopsy and placental histopathology. The proportion of unexplained stillbirths associated with under investigation continues to impede the efforts in stillbirth prevention.

Culturally appropriate preconception care and quality antenatal care that is accessible to all women has the potential to reduce stillbirth rates for both countries. Obesity and smoking are important modifiable risk factors for stillbirths in Australia, and advanced maternal age is also an increasingly prevalent risk factor in both countries.

Problems of the quality of care during childbirth in South Africa and for the immature or hypoxic neonates are the areas where most preventable deaths occur. In South Africa, problems for which there are interventions that can improve the outcome for perinatal mortality are: the recognition and management of preterm labour, hypertension and intrapartum hypoxia. For newborn babies, the major interventions that can improve the outcomes are: the care of the small and sick baby, and resuscitation of the newborn. Avoidable factors show there is a problem in the care given by medical personnel in all aspects of care from the antenatal clinic to postnatal and newborn care. In some instances it has been identified that there are insufficient staff on duty, and that they are not adequately trained to manage the patients delaying timely interventions.

Table 4. Perinatal mortality causes in 2008.

Placental pathology and infection is recognised for the first time as being the most important contributor to perinatal death in South Africa. Placenta/placenta bed diseases are also associated with preterm birth and are a recognisable link to a substantial proportion of stillbirths in Western Australia. Large disparities
(linked to disadvantages such as poverty) in stillbirth rates need to be addressed by providing more educational opportunities and improving living conditions for disadvantaged women in both South Africa and Australia.

Conclusion

Changing the health system, and more so maternity care, so that the quality of care can be improved is a complex process. One can presume that once avoidable and modifiable factors have been identified and made well known there will be a self-correction with a consequent reduction in perinatal mortality. Recommendations alone do not bring about change, they need to be implemented. An audit or literature review is not necessarily going to change practice, it is clinical care that needs to improve and once we change people we can implement change in practice.

Acknowledgements

I wish to thank Bunbury Hospital in Western Australia and Bethal Hospital in South Africa for their support and contribution to this article.

Placental pathology and infection is recognised for the first time as being the most important contributor to perinatal death in South Africa. Placenta/placenta bed diseases are also associated with preterm birth and are a recognisable link to a substantial proportion of stillbirths in Western Australia. Large disparities (linked to disadvantages such as poverty) in stillbirth rates need to be addressed by providing more educational opportunities and improving living conditions for disadvantaged women in both South Africa and Australia.

Conclusion

Changing the health system, and more so maternity care, so that the quality of care can be improved is a complex process. One can presume that once avoidable and modifiable factors have been identified and made well known there will be a self-correction with a consequent reduction in perinatal mortality. Recommendations alone do not bring about change, they need to be implemented. An audit or literature review is not necessarily going to change practice, it is clinical care that needs to improve and once we change people we can implement change in practice.

Acknowledgements

I wish to thank Bunbury Hospital in Western Australia and Bethal Hospital in South Africa for their support and contribution to this article.

Table 6. Comparing neonatal facilities in a district/regional Hospital in South Africa and Australia in 2011.