Study area and period
The study was conducted between September 2018 and February 2019 at Debre Tabor General Hospital, South Gondar Zone, Northcentral Ethiopia. The hospital is found 666 km far from Addis Ababa and 105 km away from Bahir Dar city. It is the largest hospital in South Gondar zone serving 2.7 million populations and linked to 7 district hospitals. Neonatal Intensive Care Unit (NICU) of the hospital had a total of 28 neonatal beds and hosted approximately 1159 admissions per year as per the averaged data from three subsequent quarters of 2018. Other than prematurity and perinatal asphyxia, many of the admissions were attributed to the complications of harmful traditional practices like traditional uvulectomy and suboptimal breastfeeding [30].
Study design and participant characteristics
It was a hospital-based cross-sectional study. The study involved any sick neonate admitted to NICU (i.e both home delivered and those born at health facility were considered). NICU set up was purposively selected to show burden of prelacteal feeding among sick neonates of any birth place (home versus health facility) and any residence (urban versus rural). For this study, mothers were the source of data for most of the considered factors (both neonatal and maternal). Thus, maternal mental competency was the first and foremost criterion of inclusion. Besides, mothers’ ability to listen and speak was considered because interviews were conducted. Therefore, 3 mothers having post partal major depressive disorders, one mother with homicidal ideation, 7 mothers with eclamptic presentation and 5 mothers having natural impairment of talking/listening were excluded. Diagnosis of the aforementioned critical psychiatric illnesses was made by psychiatric professionals in the hospital.
Sample size determination and sampling procedure
Sample size was determined using the following single population proportion formula.
$$ \mathbf{n}={\left(\mathrm{z}\ \left(\alpha /2\right)\right)}^2\mathrm{p}\ \left(1-\mathrm{p}\right)/{\mathrm{d}}^2. $$
Where, n = the minimum sample size required for the study,
Z (α/2) = the desired level of confidence interval which is 95% (1.96).
P = a reasonable estimate for the prevalence of prelacteal feeding from a study conducted in Southern Ethiopia (25.5%) [24].
d = tolerable margin of error, 5% (0.05).
Then, the calculated sample size became 292 and after adding 10% none response rate, the total sample size was 292 + 29 = 321.
Therefore, the total sample size considered for this study was 321 postnatal mothers. Using systematic sampling, every fourth (K = 1159/321 ≈ 4) eligible woman-neonate pair at NICU was selected systematically over 6 months to reach the most representative sample by systematically selecting as many diverse study subjects as possible.
Data collection procedures
Data were collected by four well trained BSc neonatal nurses through face to face interview using a pretested questionnaire. An interview was conducted for eligible mothers at the time of admission to Neonatal Intensive Care Unit. The questionnaire contained factors related to maternal socio demography, neonatal health, Maternal and Child Health (MCH) service utilization and maternal information on prelacteal feeding. Once interviewed for the study, a mother was ensured not to be reconsidered by the time she comes at another time during the study period. This was done by putting a ‘√’ mark on the cover of the neonatal chart so that a mother-neonatal yard won’t be reselected if encountered again.
Variables and measurement
Outcome variable
Prelacteal feeding: a neonate was regarded as prelacteal fed if given any anything before breastfeeding during the first 3 days of birth [17,18,19,20,21,22,23,24,25].
Explanatory variables
Utilized antenatal care: having at least one visit of health institution for checkup purpose during the pregnancy of the index neonate [1, 4].
Utilized postnatal care: having at least one visit of health institution for checkup of both maternal and neonatal health of the index neonate [1, 4].
Early initiation of breastfeeding: when the neonate was initiated to breast feed within 1 h of birth [1,2,3,4,5].
Late initiation of breastfeeding: when the newborn was initiated to breastfeed after 1 h of birth [1,2,3,4,5].
Premature: When a neonate was born below 37 completed weeks of intrauterine life [31].
Mature: when a neonate was born at [37–416/7] weeks of intrauterine life [31].
Post date: When a neonate was born at ≥42 completed weeks of intrauterine life [31].
Primiparous: when a mother had only one live birth [32].
Multiparous: when a mother had more than one live birth [32].
Short: When the inter-birth interval between the immediate prior child and the index neonate is less than 3 years [33].
Optimal: When the inter-birth interval between the immediate prior child and the index neonate is ≥3 years [33].
Low birth weight refers to weight of the newborn less than 2500 g at birth [31].
Data quality control methods
Data collection tool
The questionnaire was first prepared in English (Supporting information 1) and then translated to local language, Amharic (Supporting information 2), suitable for data collection. Two professional translators with medical experience were hired during data cleaning in order to explain the questions and correctly translate maternal answers for minimizing the risk of information loss during translation. The tool was adapted from Ethiopian and other African studies after which it was validated for reliability of measurement [1, 3, 15, 27].
One day training and clear orientation was first provided for data collectors and supervisors on the process of data collection. Before the actual data collection, the questionnaire was validated by pretesting on 16 eligible postnatal mothers (5% of sample size) at Nefas Mewucha District Hospital, which is nearby to the study hospital. After pretesting of the questionnaire, clarity of questions, wordings, sequence of questions, skip pattern and respondents’ reaction to the questions were reconsidered and modified.
During data collection, data collectors were closely monitored and guided by two MSc neonatal nurse supervisors for complete and appropriate collection of the data. Reporting of the collected data to the principal investigator was made on a daily basis. Furthermore, the collected data were double entered into Epidata version 4.2 by two data clerks for validation purpose.
Statistical analysis
The collected data were coded, cleaned, edited and double entered into epidata version 4.2 after which it was exported to stata version 14 software for further transformation and analysis. Frequencies, proportion, summary statistics and cross tabulation were used to describe the study population in relation to relevant variables and findings were presented using text and tables. The assumptions for binary logistic regression model were first checked and then bivariable analysis was carried out to identify candidate variables for multivariable analysis at P < 0.25 [16, 18, 19, 21]. Then, multivariable logistic regressions were performed using the candidate variables to investigate factors which have either positive or negative odds of association [(p < 0.05 at 95% confidence interval (CI)] with prelacteal feeding after auto adjustment of confounding effect in the final model. From the final model, modes of delivery and initiation time of breastfeeding were found to have confounding effects. Multi-collinearity between the study variables was first diagnosed using standard error and correlation matrix. Besides, Hoshmer-Lemeshow statistic and Omnibus tests were performed, and Hosmer-Lemeshow’s test was found to be insignificant (p-value = 0.301) while Omnibus tests was significant (P-value = 0.000) indicating the model was fitted.
Ethical consideration
Ethical clearance was obtained from ethical review committee of Debre Tabor University. Following the approval, official letter of co-operation was given to the hospital manager. After explanation of the study, an informed verbal voluntary consent was obtained from each postnatal mother. Moreover, the mothers were told that the information they gave was treated with complete confidentiality and do not cause any physical harm. Mothers were counseled of avoiding prelacteal feeding to ensure healthy growth of their neonates. All the interviewed mothers were counseled about the health, emotional and social advantages of exclusive breastfeeding of their neonates. Similarly, they were taught of the adverse consequences of prelacteal feeding like infection, malnutrition and decreased bonding of mothers to their neonates. The counseling was assisted by video show using local language (Amharic) to ease their understanding. Otherwise, there were no any other gifts provided to the mothers.