Study setting and period
The study was conducted in Aroresa district which is one of 23 districts in Sidama Zone, Southern Nations, Nationalities and Peoples’ Region (SNNPR), Ethiopia. It is located at the distance of 181 kms from Hawassa, the capital of SNNPR and 554 kms from Addis Ababa. The district has 30 rural and 3 urban Kebeles (the smallest administrative unit in Ethiopia) with a total population of 220,332 and of this, females constitute 49.8%. The women of reproductive age group account for 51,337(23.3%) of the total population. According to the district health office report, the total number of estimated deliveries in 2017/18 was 7623 and proportion of the utilization of first ANC, institutional delivery, Postnatal care (PNC) services and contraceptive prevalence were 77, 38, 69 and 49%, respectively. The district has one primary hospital, 8 health centers, 33 health posts, 3 private clinics and 4 private drug stores [Aroresa District Health Office, Aroressa District annual plan, unpublished]. The study was conducted from January 1/2018 to March 30/2018.
Study design and population
A community based prospective cohort study was conducted among a cohort of term pregnant mothers and neonates delivered from January 1/2018 and March 30/2018 in randomly selected kebeles. All term pregnant mothers (≥37 week Gestational Age (GA)) who live in the study kebeles were included in this study and followed up until they gave birth and their neonates were followed-up for a total of 28 days. All term pregnant mothers who had a known psychiatric disorder, unable to speak and residents for less than 6 months were excluded from this study.
Sample size and sampling procedure
The sample size was calculated using single population proportion with the assumption that the proportion of neonatal mortality was 2.9% , margin of error 2%, confidence interval of 95%, and a design effect (DE) of 2. Using the sample size correction formula and adding 15% non-response and lost to follow-up rate, and the final sample size became 600.
Multistage sampling technique was used to identify 600 term pregnant women to be enrolled in the follow up for the study (all term pregnant mothers were recruited consecutively until sample size was reached). First, all the Kebeles: in Aroresa district were determined to be 33. Then, 10 kebeles (9 rural and 1 urban) were selected from the district by simple random sampling method using OpenEpi 3.03. The calculated sample size was proportionally allocated to each study kebele based on expected number of term pregnant women per ‘Kebele’. Then the calculated sample was selected consecutively from each kebele.
The outcome variable is neonatal survival dichotomized as (alive =1 and died = 0) The predictor variables include; socio-demographic and economic factors: ethnicity, religion, place of residence, marital status, education status of mother and father, and occupational status of mother, age of mother, maternal factors: age at child birth, maternal complication (excessive bleeding, puerperal sepsis and fever, prolonged labour, eclampsia and preeclampsia malpresentation and malposition, premature rupture of membrane, and obstructed labour), maternal service utilization factors: place of delivery, ANC visit, postnatal care, initiation of exclusive breast feeding (EBF) and neonatal factors: birth size, birth order and interval and neonatal complication like asphyxia, infection, hypothermia, and jaundice.
Neonatal death: a death of neonate within 28 days of life according to report of mother participated in the study. Neonatal survival is defined as being alive up to the end of follow-up period (28 days). Term pregnancy is a pregnancy between 37 completed weeks up to 42 completed weeks of gestation. Birth size is defined as the size of newborn at birth according to the perception of mother. Stillbirth is defined as any fetus born without a heartbeat, respiratory effort or movement, or any other signs of life. Neonatal complication in this study is defined as a neonate experiencing at least one or more of the following conditions; asphyxia, hypothermia, jaundice, convulsion, unable to breast feed or any conditions which endangers the life the neonate. According to maternal estimate, mother’s perception of baby’s size is defined in this study as small equivalent to less than 2500 g, average 2500 g to 4000 g and large greater than 4000 g. This estimate was obtained because birth weight is unknown for most (86%) newborns in Ethiopia however the mother’s estimate of weight is subjective and interpretation of the finding should be viewed with caution .
Data collection tool and procedure
A structured questionnaire, first prepared in English and translated into Sidamu Afoo (local language), were employed to collect the data. All term (> 37 week GA) pregnant women at selected kebeles were identified by Health Extension Workers (HEW). Trained data collectors were contacted the women to obtain informed consent, to perform interviews and later to conduct postpartum follow-up, home visits, at week 1, and 4. All data collectors were contacted with the supervisor by mobile phone every week and on site supervision. Baseline data collected during recruitment were maternal socio-demographic information, medical history and use of health services like antenatal care. Pregnancy outcomes, the circumstances of delivery, date of birth, date of death of neonate, feeding patterns, and illness episodes of neonate were collected during follow-up period. The data collection processes were supervised strictly by trained supervisors and the principal investigator.
The quality of data was assured using a properly designed questionnaire, proper training of the interviewers and supervisors about the data collection and follow-up procedures, proper categorization and coding of the questionnaire. Questionnaires were pre-tested on 5% of the sample outside the study area. Data were double entered and screened for missing, outlier values and data entry errors using the frequency distribution. Errors were corrected against the raw data and the necessary corrections were made before the analysis.
Data were coded, entered, cleaned and analysed using SPSS version 22. Pregnancy outcome variables were explained by descriptive statistics and neonatal outcome variables were examined against all confounding variables using regression analysis. Kaplan–Meier survival curve was used to show pattern of neonatal death in 28 days. Independent and adjusted relationships of different predictors with neonates’ survival were assessed with Cox regression model. The risk of mortality was explored and presented with hazard ratio and 95% confidence interval. P-value less than 0.05 was considered as significant. Multicollinearity between the independent variable was assessed using variance inflation factors (VIF) and VIF greater than 10 was considered as existence of multicollinearity before interpreting the final output.