The community-based cross-sectional study was carried to select respondents from September to June 2019 in Finote-Selam district, which is found in Northwest Ethiopia, to assess the maternal knowledge, attitude, and practices (KAP) of exclusive breastfeeding among mothers of children aged 0–6 months in Ethiopia. This method was so important to afford numeric descriptions of knowledge, attitudes, and practices of mothers who have children in the age between 0 and 6 [30, 31].
The target population of the current study included the mothers of children aged 0–6 months but lived in Finote-Selma district of Amhara National Regional State, Ethiopia. A multi-stage random sampling technique was used to identify the actual respondents by firstly selecting mothers of children aged 0–6 months in Ethiopia. Then, within these groups, a random sample of smaller sub-groups such as mothers of children who lived in Finote-Selam and four local administrations-kebeles (2 urban kebeles: 03 & 04 and 2 rural kebeles: Shenbekuma & Bakel) were selected. Afterward, a sample of these smallest sub-groups can be randomly selected to form a target population for the current study. After assessing the potential respondents and the study areas of the present study, the researchers undertook a systematic random sampling technique in the condition that nth aged 0–6 months of mothers at Finote-Selam district was selected. The simple random sampling technique was used to obtain the required sample size calculated by
n = (Zα/2)2 p (1-p), whereby, d2
the prevalence of exclusive breastfeeding in Ethiopia at the national level (P) was considered as 58%, at a 95% confidence level, and 5% degree of desired precision (d). From this assumption, the sample was 375. But, the total sample size for the study area was 413 by considering 10% of the non-response rate.
After selecting the actual respondents of the current study, the researchers used a paper and pencil questionnaire. The current study adopted some questions from the Ethiopia Demographic Health Survey (2016) and modified it to meet the basic objectives and considered the contextual understanding of the study area. The modified questions have checked the consistency of the questionnaire by conducting a pre-test in Burie town (a town that has a similar case of the study area), which was not included in the actual data analysis. After testing the sample questionnaire, corrections were made on unclear questions for the interviewers and interviewees. This reliability of the data was validated by Cronbach’s alpha = 0.81 indicated the quality of data was good.
The researcher-administered questionnaire contained a total of 33 questions and was divided into four parts that addressed the participants’ sociodemographic characteristics (12 questions), breastfeeding knowledge (8 questions), attitudes (5 questions), and practices (8 questions). The knowledge questions have the response options using dichotomous variables: 0 = no, 1 = yes whereas the items represented by five points to reflect the frequencies of maternal attitude on breastfeeding which includes (strongly disagree = 0, disagree = 1, neutral = 3, agree = 4, and strongly disagree = 5), and the questions of maternal practices offered the response options using multiple choices that asked in different manners.
The participants’ scoring system was as follows:
- Knowledge: Good (4–7 points), Fair (2–6 points), Poor (0–2 points);
-
Attitude: Good (14–25 points), Fair (5–13 points), Poor (0–4 points); and
-
Practice: Good (4–6 points), Fair (2–3 points), Poor (0–1 point).
Questions envisioned to assess the breastfeeding knowledge comprised the information the mothers have, the source of information, types of food given for the infant immediately after delivery, duration of EBF, importance of EBF to maternal and child health, and colostrum feeding for the infant. Questions intended to consider the breastfeeding attitude included importance of EBF for the infant’s health promotion, mother’s health improvement, EBF that continued for (0–6) months, cheese swallow, and colostrum feeding that be necessary for the infants. Whereby, questions anticipated to examine breastfeeding practices held up a time that starting breastfeeding, frequency of breastfeeding per day, duration of breastfeeding, colostrum feeding for the infant, the types of food that infant can take in addition to BF under 6 months and the like.
The current study obtained informed consent from respondents after the purpose and procedure of the study were explained. The current study assured participants they would be protected from physical and psychological harm whilst being participants.
Data analysis
Quantitative data were analyzed starting from checking up the accuracy of the transcription of data and the translation of them from Amharic (local language) to English to checking the accuracy of the labels, the values, and level of measurement for all variables. After finishing this process as intended, the researchers conducted univariate analysis using the Statistical Package for Social Sciences (SPSS) version 20 to identify the respondents’ socio-demographic variables such as age, religion, ethnicity, marital status, place of residence, education, employment status, household monthly income, and distance between house and the health center, and to analyze mean, median, standard deviation of the quantitative data.
Binary logistic regression was used to test the association between the dependent variable of exclusive breastfeeding (mothers of children aged 0–6 months carried out breastfeeding was poor Coded as 1 while not is coded as 0) and sociodemographic variables. Again, binary logistic regression was used to analyze breastfeeding KAP of mothers of children in the age between 0 and 6 months (1 was coded as good and 0 as not good). Breastfeeding KAP and the sociodemographic variables were entered into the regression model and then were checked for exp(B) to ascertain 95% of the confidence intervals (CIs) for the odds ratio of predictor’s contribution to the equation. The goodness of fit of the model was checked by Hosmer and Lemeshow test model using backward likelihood ratio method.