A retrospective analysis was conducted among women who delivered at the University Teaching Hospital (UTH) in Lusaka, Zambia, between January 1, 2008, and December 31, 2012. The present study was approved by the ethical review committees at the University of Zambia (Lusaka, Zambia) and the University of North Carolina (Chapel Hill, NC, USA). Individual informed consent was not sought because the data were collected as part of routine medical care.
Located in Zambia’s capital city, UTH serves a primary catchment population of approximately 2 million. This facility is the province’s only tertiary care center and, in this capacity, UTH receives a high volume of transfers; however, it also provides primary-level care for individuals who live in neighboring communities. The 17-bed labor ward offers 24-hour coverage by midwives, doctors in training, and qualified obstetrician–gynecologists. With three operating theatres, including one dedicated to obstetrics, UTH was the main facility providing cesarean delivery in the Lusaka public sector during the present study period. The UTH neonatal intensive care unit (NICU) is located 150 m from the delivery ward and is staffed by specialist pediatricians and a dedicated nursing team. Patients who are referred to UTH from primary care facilities are provided health services free of charge. Individuals who seek care at UTH without a formal referral (i.e. self-referrals) incur a one-time fee equivalent to US$15.
The present study used observational data from the Zambia Electronic Perinatal Record System (ZEPRS), which collects detailed medical information about prenatal, intrapartum, and newborn care across the Lusaka public health sector [11 ]. This system has been implemented across 25 health centers, including 13 with delivery facilities [11 ]. The ZEPRS application employs real-time data entry at the point of care. A unique identification number is automatically generated for all neonates on delivery and is linked with the mother’s medical record. Data are uploaded on a central server and their quality regularly assessed.
The present study included women with pregnancy information recorded in ZEPRS who had delivered at UTH. This cohort was compared with a group of women who had delivered at primary care facilities in Lusaka. Demographic characteristics assessed included medical history, obstetric history, and pregnancy outcomes. At the time of enrollment, fetuses with an estimated gestational age of less than 20 weeks were dated according to the mother’s last menstrual period. For those with an estimated gestational age of at least 20 weeks, a simple algorithm was implemented that considered the last menstrual period and fundal height at clinical examination [11 ]. If the results obtained by these two methods were more than 3 weeks apart, gestational age was determined from fundal height alone.
Maternal outcome measures included maternal mortality, cesarean delivery, and prenatal or intrapartum hemorrhage. Adverse neonatal outcomes included stillbirth, a low 5-minute Apgar score (<7), and NICU admission. Stillbirth was defined as the delivery of a non-viable fetus at a gestational age of 28 weeks or older. This event was further subclassified as “fresh” (suggestive of intrapartum demise) or “macerated” (suggestive of prenatal demise) on the basis of degenerative skin changes on physical examination at birth [12 ,13 (link)].
The data were analyzed using SAS version 9.3 (SAS Institute, Cary, NC, USA). The data were divided into 3-month windows from the first quarter to the fourth quarter of each year. Point estimates with 95% confidence intervals were calculated for each outcome of interest. Graphical representations were generated of the observed percentages over time. The time trend was then modeled using two separate approaches: a LOESS curve (a non-parametric method that fits simple models to localized data subsets without imposing a predefined structure) and a linear regression line (to determine the relationship between quarter and each outcome of interest). P<0.05 was considered statistically significant.
Located in Zambia’s capital city, UTH serves a primary catchment population of approximately 2 million. This facility is the province’s only tertiary care center and, in this capacity, UTH receives a high volume of transfers; however, it also provides primary-level care for individuals who live in neighboring communities. The 17-bed labor ward offers 24-hour coverage by midwives, doctors in training, and qualified obstetrician–gynecologists. With three operating theatres, including one dedicated to obstetrics, UTH was the main facility providing cesarean delivery in the Lusaka public sector during the present study period. The UTH neonatal intensive care unit (NICU) is located 150 m from the delivery ward and is staffed by specialist pediatricians and a dedicated nursing team. Patients who are referred to UTH from primary care facilities are provided health services free of charge. Individuals who seek care at UTH without a formal referral (i.e. self-referrals) incur a one-time fee equivalent to US$15.
The present study used observational data from the Zambia Electronic Perinatal Record System (ZEPRS), which collects detailed medical information about prenatal, intrapartum, and newborn care across the Lusaka public health sector [11 ]. This system has been implemented across 25 health centers, including 13 with delivery facilities [11 ]. The ZEPRS application employs real-time data entry at the point of care. A unique identification number is automatically generated for all neonates on delivery and is linked with the mother’s medical record. Data are uploaded on a central server and their quality regularly assessed.
The present study included women with pregnancy information recorded in ZEPRS who had delivered at UTH. This cohort was compared with a group of women who had delivered at primary care facilities in Lusaka. Demographic characteristics assessed included medical history, obstetric history, and pregnancy outcomes. At the time of enrollment, fetuses with an estimated gestational age of less than 20 weeks were dated according to the mother’s last menstrual period. For those with an estimated gestational age of at least 20 weeks, a simple algorithm was implemented that considered the last menstrual period and fundal height at clinical examination [11 ]. If the results obtained by these two methods were more than 3 weeks apart, gestational age was determined from fundal height alone.
Maternal outcome measures included maternal mortality, cesarean delivery, and prenatal or intrapartum hemorrhage. Adverse neonatal outcomes included stillbirth, a low 5-minute Apgar score (<7), and NICU admission. Stillbirth was defined as the delivery of a non-viable fetus at a gestational age of 28 weeks or older. This event was further subclassified as “fresh” (suggestive of intrapartum demise) or “macerated” (suggestive of prenatal demise) on the basis of degenerative skin changes on physical examination at birth [12 ,13 (link)].
The data were analyzed using SAS version 9.3 (SAS Institute, Cary, NC, USA). The data were divided into 3-month windows from the first quarter to the fourth quarter of each year. Point estimates with 95% confidence intervals were calculated for each outcome of interest. Graphical representations were generated of the observed percentages over time. The time trend was then modeled using two separate approaches: a LOESS curve (a non-parametric method that fits simple models to localized data subsets without imposing a predefined structure) and a linear regression line (to determine the relationship between quarter and each outcome of interest). P<0.05 was considered statistically significant.
