Study setting
This study was aimed at assessing neonatal deaths and evaluating predictors of death in neonates admitted to the NICU of HUCSH. The study area was Hawassa City, which was located in the Sidama region on the shore of Lake Hawassa in the Great Rift Valley, 273 km south of Addis Ababa. Our research has been conducted at Hawassa University's comprehensive specialized hospital, which was established 16 years ago by the Regional Health Bureau. The hospital was intended to serve the 3.5–5 million total population at the beginning and now serves the whole Sidama region, Southern Nations Nationalities and Peoples Region (SNNPR), and part of the Oromia region. HUCSH has a total of seven departments and fourteen units. Pediatrics and child health is one of the departments, where around 15,000 pediatric patients are treated per year
The HUCSH NICU, inaugurated in 2014, has 35 beds equipped with advanced technologies for neonatal care. It is supported by multidisciplinary care and provides level III neonatal care. There are a total of 28 nurses assigned to the NICU. A consultant neonatologist, trained nurses, and pediatricians, together with pediatric residents and medical interns, provide services 24/7. It has four main classes: preterm, term, Kangaro Mother Care (KMC), and backside. It is located near the hospital’s obstetric ward to receive high-risk newborns from this ward as quickly as possible. Apart from the hospital’s obstetric ward, the unit also receives neonates referred from other health facilities and homes.
The unit has 14 radiant warmers, 4 Continuous Positive Airway Pressure (CPAP), phototherapy, and oxygen concentrator machines. The unit provides outpatient and inpatient services, with estimated admissions of 100–150 newborns per month for inborn neonates as well as referred babies from surrounding provinces. Annual admissions at the NICU during the study period were 1044.
Study design and period
A facility-based retrospective cross-sectional study of chart review was performed for the assessment of neonatal death ,and searching for predictors of death at HURH NICU over the past year prior to the study period(May 2021 to the end of April 2022), and the study period was from May 2022 to August 2022.
Source population
All neonates who were admitted to NICU of HUCSH, Hawassa, Ethiopia during the study period
Study population
All systematically selected neonates that fulfilled the inclusion criteria who were admitted to the NICU of HUCSH, Hawassa, Ethiopia during the study period
Inclusion and Exclusion criteria
Inclusion criteria
All neonates who were admitted in the NICU of HUCSH and registered as died or alive within the first 28 days during the study period
Exclusion criteria
Study variables
Dependent variables
1. Socio-demographic characteristics of the mother and the newborn
-
Maternal
-
Age
-
Residence
-
Ethnicity
-
Occupation and income
-
ANC follow up
-
Neonatal
2. Obstetric factors
- Birth weight
- Mode of delivery
- Gestational age
- Apgar/immediate crying
- Resuscitation at delivery
- Duration of labor
- Duration of rupture of the membrane
- Maternal fever
- Parity
- Prior pregnancy loss
3. Neonatal inpatient conditions
Operational definitions[13–15]
Neonatal mortality is the death of the newborn during the first 28 completed days after live birth (days 0–27) after admission to the NICU and before discharge, as confirmed and recorded on the chart.
Early neonatal death is a death during the first 7 completed days after live birth (days 0–6).
Antenatal care visit: any history of visit or follow-up during the current or index pregnancy at any health institution for a checkup of pregnancy and designated or recorded on a chart.
Intrapartum complications are complications that occur after the onset of labor, including intrapartum bleeding, obstructed labor, prolonged labor, eclampsia, chorioamnionitis, and others.
Congenital malformation is a body deformity or deformities, structural or functional anomalies, that occur during intrauterine life and can be identified prenatally, at birth, or sometimes only later in infancy, and that is believed to have an impact on the health of the baby. It is diagnosed and recorded on charts by professionals on admission.
Hypoglycemia is a measure of low blood glucose (40 mg/dl) that was diagnosed and recorded on charts by professionals on admission.
Hypothermia is a low-body temperature measurement (36.5°C), diagnosed and recorded on or during admission of neonates.
Birth asphyxia is diagnosed whenever a neonate has an APGAR score of 6 in the 5th minute and/or if he or she does not cry immediately after birth, has respiratory distress, floppiness, loss of consciousness, the presence of convulsions, and a loss of neonatal reflexes.
Birth weight is classified using the WHO weight classification: very low birth weight is any child with a birth weight of 1,500 g, while low birth weight is any child with a birth weight of 2,500 g.
Premature rupture of membranes (PROM) is the spontaneous leakage of amniotic fluid from the amniotic sac, occurring after 28 weeks of gestation and before the onset of labor. It was diagnosed and recorded on charts by professionals on admission.
Prolonged rupture of membranes (PROM) is considered when the duration of the leakage is more than 12 hours before delivery.
Sample size determination and procedures
The sample size was calculated by using a single population proportion formula with assumptions of a confidence level of 95% = 1.96, a margin of error (d) = 0.05, and the magnitude of neonatal death (p = 0.23) from a previous study conducted at the NICU of Gondar Referral Hospital [16].Considering a 10% incompleteness rate, the final sample size was 235.
The study participants were selected by systematic random sampling using the registration numbers of the neonates. Among a total of 1044 neonatal registration numbers, every fifth participant was selected. The first study unit was selected using the lottery method. Estimation of neonatal mortality rate (NMR): The median follow-up period of selected risky neonates for the study was determined. The product of the number of days of the median follow-up period and the total number of neonates followed gives neonate-days. The overall incidence of neonatal mortality was calculated as a ratio of neonatal deaths to neonate days per 1000 neonate days.
Data collection and materials
A data extraction checklist that was modified and adjusted from various literary works was used to review the medical records and collect the data [16, 17]. The following variables were included in the data using the data collection tool: the clinical diagnosis at admission, sex, maternal and neonatal age, complications during pregnancy, gestational age, antenatal care visits, birth weight, place of birth (HUSCH, other, or home), the neonate's body temperature, random blood sugar at admission, the mode of delivery, the outcome (survival or death) at discharge, and the length of stay prior to discharge. The International Classification of Diseases in its tenth iteration (International Statistical Classification of Diseases and Related Health Problems, ICD-10-WHO, version 2015) was used to classify clinical diagnoses (congenital deformity, prematurity, birth asphyxia, infection).
Data quality assurance
A data-gathering tool that was properly designed and structured was used. During data collection and data input, data collectors were trained and strictly supervised. The data collectors were medical interns and residents undergoing training to improve their comprehension and interpretation of patient medical charts. The lead investigator cross-checked 10% of the gathered data with medical records and ensured its consistency after data collection.
Data management and analysis
At the end of each day, the data were reviewed for completeness and consistency. It was cleaned, modified, programmed, and entered into EpiData software version 4.6. The data were then exported into SPSS 25 for statistical analysis. During the analysis, the frequency distribution and percentage of various variables were computed to describe and summarize the respondents' basic socio-demographic characteristics. To provide an overview of the variables, the findings were presented in the form of frequency tables, pie charts, and graphs. Neonatal survival (lived or died within the neonatal period of 28 days) was defined as the binary outcome for bivariate and multivariate studies examining risk factors. Binary regression analysis was used to first look into the bivariate relationship between each independent variable and the result. To reduce the impact of confounding variables and pinpoint the primary causes of inpatient neonatal death, the variables with a p-value of 0.20 during the bivariate analysis were used as candidate variables for a multivariate logistic regression analysis model. Neonatal mortality was determined using an adjusted odds ratio (AOR) at a 95% confidence interval (CI) to demonstrate the strength of the link, and statistical significance was deemed to exist at a p-value of 0.05. The Chi square and Fisher's exact tests were used to determine the statistical significance of the differences between the groups. The median follow-up time for the study's risky neonates was established. Neonatal days are calculated as the median follow-up period divided by the total number of neonates followed. The ratio of neonatal fatalities to neonatal days per 1,000 neonatal days, which represents the total incidence of neonatal mortality.