Introduction: In low income and middle income countries, neonatal mortality remains high despite the gradual reduction in under five mortality. Newborn death contributes for about 38% of all under five deaths. This study has identified the magnitude and independent predictors of neonatal mortality in rural Ethiopia. Methods: This population based nested case control study was conducted in rural West Gojam zone, Northern Ethiopia, among a cohort of pregnant women who gave birth between March 2011 and Feb 2012. The cohort was established by Maternal and Newborn Health in Ethiopia Partnership (MaNHEP) project in 2010 by recruiting mothers in their third trimester, as identified by trained community volunteers. Once identified, women stayed in the cohort throughout their pregnancy period receiving Community Maternal and Newborn Health (CMNH) training by health extension workers and community volunteers till the end of the first 48 hours postpartum. Cases were 75 mothers who lost their newborns to neonatal death and controls were 150 randomly selected mothers with neonates who survived the neonatal period. Data to identify cause of death were collected using the WHO standard verbal autopsy questionnaire after the culturally appropriate 40 days of bereavement period. Binomial logistic regression model was used to identify independent contributors to neonatal mortality. Result: The neonatal mortality rate was AOR(95%CI) = 18.6 (14.8, 23.2) per 1000 live births. Neonatal mortality declined with an increase in family size, neonates who were born among a family of more than two had lesser odds of death in the neonatal period than those who were born in a family of two AOR (95% CI) = 0.13 (0.02, 0.71). Mothers who gave birth to 2-4 AOR(95%CI) = 0.15 (0.05, 0.48) and 5+ children AOR(95%CI) = 0.08 (0.02, 0.26) had lesser odds of losing their newborns to neonatal mortality. Previous history of losing a newborn to neonatal death also increased the odds of neonatal mortality during the last birth AOR (95%CI) = 0.25 (0.11, 0.53). Conclusion: The neonatal mortality rate in our study was three times lower than the regional neonatal mortality rate estimate, indicating community based interventions could significantly decrease neonatal mortality. The identified determinants, which are amenable for change, emphasize the need to improve quality of care during pregnancy, labour and delivery to improve pregnancy outcome.This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
We conducted a population-based nested case control study among a cohort of pregnant women in three districts of West Gojam zone (North Achefer, South Achefer and Mecha). The zone is located 500 kms away north of the capital city Addis Ababa. Twenty-four kebles (i.e. the smallest administrative unit) were selected from the three districts (7 from North and South Achefer districts each and 10 from Mecha district). The selected three districts were among the highly populous districts of the zone with a total population count of 292,250 in Mecha, 155,863, in South Achefer and 173,211 in North Achefer districts [9]. Each kebele has one health post, providing disease prevention and health promotion services to 2500–5000 population. Each health post is staffed with two health extension workers (HEWs) and reports to the next level called health centers [10]. The cohort was first established in 2010 by Maternal and Newborn Health in Ethiopia Partnership (MaNHEP) project. The project was led by Emory University in collaboration with the Ethiopian Federal Ministry of Health and Addis Ababa University [11]. The intervention implemented Community Maternal and Newborn Health (CMNH) care package through the existing Health Extension Program (HEP) in six woredas (districts) of Amhara and Oromia regions (three from each region) [11,12]. The project was aimed at improving the capability and performance of HEWs to provide targeted maternal and newborn health (MNH) services; increasing demand for targeted MNH services and improve self-care behavior and quality of MNH services in lead woredas. A lead woreda is the one that has the commitment and capacity to continuously improve MNH service delivery to meet the needs of mothers and children [12]. The project used three basic approaches to achieve its objectives: 1) Behavioral change communication on issues of maternal and newborn health, 2) Maternal and newborn health training for health extension workers, women in reproductive age and potential care givers among family members and 3) Continuous collaborative quality improvement interventions [11]. Pregnant women, in their third trimester, were enrolled into the cohort as identified by community volunteers who had CMNH training by MaNHEP. Once in the cohort, mothers received continuous training on care during pregnancy, labour and delivery by the volunteers. This study focused on data collected from mothers in Amhara Region who gave birth between March 2011 and February 2012. All pregnant women in their third trimester living in the selected kebeles were recruited into the cohort. The cases were mothers who lost their babies for neonatal death at the end of the follow up period and the controls were mothers with a live neonate at the end of the fourth week after birth. The controls were randomly selected from a sampling frame containing list of mothers, in the respective gotes (i.e. a structure smaller than kebele) of the cases, whose pregnancy outcome was confirmed (Fig 1). Age of the mother, maternal and paternal education were among the socio-demographic variables involved in the analysis (Table 1). Three high school graduate females trained for five days on the data collection tool and interview techniques collected the data. Mothers who lost their newborns were interviewed earliest at forty days after death of the newborn to minimize recall bias. Data were cleaned and entered using EpiData version 3.1 statistical software. Further cleaning and analysis was done using SPSS version 20 statistical software. Frequency and proportion were calculated for all variables which were included in the analysis. Neonatal mortality rate was calculated per 1000 live births. Bivariate analysis was conducted to measure the association between the dependent and individual independent variables. To control the effect of confounding variables multiple binary logistic regression models were used. Crude and adjusted OR with 95% CI (Confidence interval) was used to interpret findings of the bivariate and multivariate analysis respectively. Ethical clearance was obtained from School of Public Health, College of Health Sciences, Addis Ababa University, Research and Ethical Committee (REC). After the purpose of the study was explained, participants provided a written consent before the interview. Mothers who lost their newborns were interviewed after forty days of culturally acceptable bereavement period.