Menu
Menu
Menu
Menu
Background: The number, rates and causes of early neonatal deaths in South Africa were not known. Neither had modifiable factors associated with these deaths been previously documented. An audit of live born infants who died in the first week of life in the public service could help in planning strategies to reduce the early neonatal mortality rate. Methods: The number of live born infants weighing 1000 g or more, the number of these infants who die in the first week of life, the primary and final causes of these deaths, and the modifiable factors associated with them were collected over four years from 102 sites in South Africa as part of the Perinatal Problem Identification Programme. Results: The rate of death in the first week of life for infants weighing 1000 g or more was unacceptably high (8.7/1000), especially in rural areas (10.42/1000). Intrapartum hypoxia and preterm delivery are the main causes of death. Common modifiable factors included inadequate staffing and facilities, poor care in labour, poor neonatal resuscitation and basic care, and difficulties for patients in accessing health care. Conclusion: Practical, affordable and effective steps can be taken to reduce the number of infants who die in the first week of life in South Africa. These could also be implemented in other under resourced countries. © 2005 Pattinson et al; licensee BioMed Central Ltd.
Data were collated from 102 sentinel sites within the public health service in South Africa. All perinatal deaths (stillbirths and neonatal deaths of 500 g or more) at these health care facilities were recorded over four years from 1st October 1999 to 30th September 2003. At each site all perinatal deaths were discussed at regular mortality audit meetings. While most first week deaths in infants weighing less than 2000 g at birth were captured, some deaths that occurred between the time of discharge from the health facility and 7 days of age in heavier infants may have been missed. Therefore, there may be an underestimation of deaths due to infections in infants weighing 2000 g or more. After review by the medical and nursing staff involved in the maternal and neonatal care, the probable primary and final causes of death were identified as well as any modifiable factors. The primary cause was defined as the underlying obstetric factor or condition which started a train of events that resulted in the death (why the death occurred, e.g. placental abruption) while the final cause was defined as the pathological process in the infant that actually caused the death (how the infant died, e.g. hypoxia). Primary causes identify pregnancy complications that can often be prevented (e.g. eclampsia) while final causes highlight areas of inadequate neonatal care (e.g. immaturity related deaths). The listed modifiable (avoidable) factors included missed opportunities for good care and examples of substandard care. These draw attention to areas of maternal and newborn care where improvements are needed. The MRC unit contacted all services using PPIP and requested them to electronically send their data for collation, using PPIPWIN v2 (Simply Software®). This software package utilises a simple, user-friendly computer-based programme. Once basic perinatal data are entered, the programme calculates various perinatal care indices, describes the medical conditions that led to the perinatal deaths and lists modifiable factors associated with the deaths. Each site was categorised as metropolitan (the large amalgamated cities such as Cape Town), cities and towns, or rural areas. This categorisation was chosen as it grouped the hospitals and clinics into naturally comparable units, covered most of the institutional deliveries occurring in those areas and was thought to be more representative of population based data than any other combination. Most metropolitan areas are served by teaching hospitals and represent a fully functioning, tiered health care system, with all patients in the area having relatively easy access to tertiary care if needed. The city and town grouping represents areas where patients usually have easy access to primary and secondary level institutions, but where there is some difficulty in accessing tertiary hospitals. Finally the rural grouping represents primary care facilities, with the patients having to be referred for either secondary or tertiary care. Often health care facilities in cities and towns and rural areas had to provide levels of care beyond their means due to an inability to refer these patients. It was decided not to combine the data by levels of care across the country because of the very different referral patterns. All data were therefore categorised by site of delivery and not area of residence. The number of unrecorded deaths after home births is unknown, but is estimated not to be large. Only probable modifiable factors related to deaths are included in this analysis. Data on live born infants weighing 500 to 999 g were excluded as the reliability of these data were questionable. Some very small infants were coded as stillbirths in error, or their data were not recorded, as they were regarded as non-viable. Late neonatal deaths were also not considered as many infants are not closely followed by the health care services after the first week of life. Data were not available from the private sector where most of the community with health insurance receive care. Infants with low-weight unrelated to maternal hypertension or any other identified obstetric cause were coded as “idiopathic SGA”. Infection as a final cause of death included both prenatal infections (e.g. syphilis) and infections after delivery (e.g. necrotising enterocolitis). Patient, administration and health worker related modifiable factors, which probably lead directly to the death, were considered. Sometimes more than one modifiable factor could be identified for a single death. Ethics approval for the initial studies using PPIP was obtained from the University of Pretoria’s Faculty of Health Sciences. The programme has since been taken over by the national and provincial Departments of Health and is approved by all health institutional Chief Executive Officers where it is used. Patient anonymity is assured at all times. As this paper specifically addresses infant deaths in the first week of life, data on stillbirths have been excluded but are available on the PPIP website [7].
N/A
