Study setting and participants
This study was conducted in Raya Kobo district of North Eastern Ethiopia from December 28, 2013 to January 18, 2014. Raya Kobo district is one of the thirteen districts of North Wello Zone, which is located 570 kms North-East of Addis Ababa. There are forty two kebeles (the smallest administrative units next to district in Ethiopia) under Raya Kobo district: five urban kebeles and thirty seven rural kebeles. In this district, there are forty two health posts and seven health centers. Each health post has two health extension workers .
A quantitative community based cross-sectional study was employed to survey mothers of children aged less than 24 months. The sample size was determined using a formula for estimation of single population proportion as follows:
Where n = required sample size, Z = critical value for normal distribution at 95% confidence level (1.96), P = prevalence of prelacteal feeding in Amhara region (47.8%) , d = 0.05 (5% margin of error), D = 1.5 (design effect), and an estimated non-response rate of 10%.
The sampling procedure was started from the stratification (assuming that the rural kebeles are relatively homogenous) of the thirty seven kebeles as rural and the five kebeles as urban (assuming that the urban kebeles are relatively homogenous). Out of the forty two kebeles, one urban kebele and seven rural kebeles were randomly selected using lottery method. Fourteen percent of the under-two population lives in urban areas and the remaining 86% lives in rural . Therefore, based on population proportion, 544 mothers of children aged less than 24 months were taken from rural areas and 89 from urban areas.
Pre-survey was done before the actual day of data collection to know which households have the target mother-child pairs. As a result, there were 1,965 households having the targeted mother-child pairs in the selected eight kebeles (1,663 and 302 from rural and urban settings respectively).The total households divided by the sample size (the sampling fraction) was three. Then the first mother-child pair to be included in the sample was chosen randomly by picking one out of the first three houses, numbered one to three. At the time of survey, from each household unit one eligible mother who had a biological child aged less than 24 months was selected. Non-biological and mothers who are unable to communicate were excluded from the study.
During qualitative study, focus group discussions (FGD) and in-depth interviews were conducted to explore cultural beliefs about prelacteal feeding. The participants were selected purposively based on their role in the community. Two focus group discussions were undertaken in a group of grandmothers (woman who had at least one grandchild), each comprised of eight discussants. The in-depth interviews were undertaken with four traditional birth attendants (two trained and two untrained).
In this study, the outcome variable was prelacteal feeding practices among mothers of children aged less than 24 months. Prelacteal feeding was understood as providing foods and/or drinks other than human milk for the infant before the initiation of breastfeeding .The independent variables were maternal characteristics (age, educational status, religion, ethnicity), household characteristics (area of residence, household head and family size), husband educational status, child’s sex, antenatal care utilization, postnatal care utilization, place of delivery, mode of delivery, breastfeeding initiation and maternal knowledge on the risks associated with prelacteal feeding.
Antenatal care utilization: having at least one visit of health institution for checkup purpose during the pregnancy of the index child .
Family size: everybody living permanently in the same house was counted as family members.
Postnatal care utilization: receiving the care provided to the woman and the index child at least once during the six weeks period following delivery .
Trained traditional birth attendant: traditional birth attendant who receive training on mother-child cares during delivery.
Untrained traditional birth attendant: traditional birth attendant who can provide the delivery services without knowing the basic mother-child cares (did not take any training program).
Data collection instrument, methods and process
Quantitative data were collected using a pre-tested, structured, interviewer-administered questionnaire adapted from Ethiopian Demographic and Health Survey  and the national nutrition survey questionnaire . The adapted questionnaire was modified and contextualized to fit the local situation and the research objective. The EDHS definition of prelacteal feeding was “giving anything to drink other than breast milk in the first three days”. We operationalized prelacteal feeding eating or drinking something other than breast milk before initiation of breastfeeding. We assessed prelacteal feeding by this question “Before initiation of breastfeeding, was (NAME) given anything to drink and/or eat other than breast milk?”. We adapted the questionnaire to our study district culture for the type of prelacteal foods used. The questionnaire was prepared first in English, translated into Amharic, and then back into English by fluent speakers of both languages to check its consistency. The final Amharic version of the questionnaire was used to collect the data. The quantitative data were collected by eight high school graduates. The data collectors and the supervisors (two nurses having Bachelor of Science (BSc) were trained for three days (including practical work) by the principal investigators on the study instrument, consent form, how to interview and data collection procedures. Qualitative data were collected using focus group discussion and in-depth interview. Focus group discussions were facilitated by BSc nurse with the assistance of two nursing students (note takers). The in-depth interviews were also conducted by a female BSc nurse.
The quantitative data were checked for completeness and inconsistencies. It was also cleaned, coded and entered into EpiData version 3.02, then exported to the SPSS 16.0 statistical package for analysis. Binary logistic regression analysis was performed; the crude odds ratio (COR) with 95% confidence interval was estimated to assess the association between each independent variable and the outcome variable, and to select candidate variables for the multivariate logistic regression analysis. Variables with p-value < 0.3 in the binary logistic regression analysis were considered in the multivariable logistic analysis. The Hosmer-Lemeshow goodness-of-fit with enter procedure was used to test for model fitness. Adjusted Odds Ratio (AOR) with 95% confidence interval was estimated to assess the strength of associations, and a p-value < 0.05 was used to declare the statistical significance in the final logistic model.
Qualitative data were transcribed in to an English text by the principal investigator (ML). Different ideas in the text were merged in their thematic areas and a thematic framework analysis was employed manually. The results were presented in narratives in triangulation with quantitative data.
A letter of ethical approval was granted from the Institutional Health Research Ethics Review Committee (IHRERC) of Haramaya University. An official letter was written from School of Graduate Studies to Raya Kobo District Administration Office. Then permission and support letter was written to each selected Kebele. Informed written consent was taken from the participants before the interview. Illiterate mothers were consented by their thumb print after verbal consent. The participants were also assured about the confidentiality of the information they provided.