Complementary feeding practice and associated factors among internally displaced mothers of children aged 6–23 months in Amhara region, Northwest Ethiopia: a cross-sectional study

Gendewuha is a district located in the west Gondar Amhara region, 160 km from Gondar city. It has a total host population of 29,874 and 6000 IDPs. From the host population, 14,105 and 15,769 were men and women respectively. Approximaly, 8724 are children of aged less than 14 years. Of the IDPs, 642 are known to be mothers of children aged 6–23 months who live collectively in the Gendewuha IDP camp Amhara region, northwest Ethiopia. A community-based cross-sectional study design was conducted from June to July 2020.

The source population consisted of all internally displaced mothers of children aged 6–23 months who were living in the Amhara region, northwest Ethiopia. The study population included internally displaced mothers of children 6–23 months from those living in Gendewuha camp, northwest, Ethiopia.

Mothers of children aged 6–23 months who visited internally displaced populations during the data collection period were excluded from the study.

The required sample size was calculated using OpenEpi statistics software version 3.03 by considering the assumptions (80% power of the study, 95% confidence level, 1:1 ratio) and the prevalence of complementary feeding practice (19.4%) from a previous study [6]. By adding a 10% non-response rate, the final sample size was 264 study participants.

A systematic random sampling technique was used to obtain to reach the final study units. Since internally displaced populations are not stable, a survey assessment was performed to count the number of IDP mothers of children found at the time of the data collection, and tent numbers were given and used as a sampling frame. “K” was calculated by dividing the total 642 mothers of children of age 6–23 months present at the site at the time of the data collection period with the sample size, 264, and was found to be “2”. Households with mothers of children aged 6–23 months were recruited in every other household. Households with selected mothers of children aged 6–23 months were identified based on the tent numbers given during the quick survey assessment before data collection. In a situation where there was more than one mother of children aged 6–23 months, the lottery method was used to select one of them. Likewise, whenever a mother had more than one child of age 6–23 months of age, the lottery method was used to select one of the children.

Data were collected using a questionnaire that was partially adapted, structured, interviewer-administered, and closed-ended questionnaires. The tool has 4 sections including dietary diversity inquiry for infants and young children which was adopted from a food and nutrition technical assistant (FANTA). A 24-h dietary recall method, which is the most widely used time period for collecting dietary information, was used to collect dietary information. The data collectors were six diploma nurses and one supervisor who is a BSC nurse. Two days of training were given to the data collectors and the supervisor. The questionnaire was translated to the local language [Amharic] by a language translator and retranslated back in to English to check its consistency. The questionnaire was pre-tested on 5% of the study population before the actual data collection at other sites with a similar population (Chillga IDP site) and corrections were made on some of the ambiguities in the questionnaire. Those mothers who were absent for the first visit were re-visited for two additional visits on different days and were considered non-responders after an absence for three visits. Data completeness and consistency were checked by the supervisor as well as by the principal investigator on a daily basis.

After data collection, the responses in the completed questionnaire were coded and entered into Epi data version 4.1 and were exported to Statistical Packages for Social Sciences (SPSS) version 25.0. It was cleaned and edited (check for missing values and outliers) accordingly. The outcome variable was coded as ‘1’ for appropriate complementary feeding practice and ‘0’ for inappropriate complementary feeding practice. Descriptive statistical analysis was performed to compute the frequency, percentage, and mean for continuous independent variables. Before the analysis, the assumptions of the chi-squared test were checked. When smaller expected frequencies were encountered, re-categorization of variables or merger of the levels was performed. Binary logistic regression was performed to determine the crude relationship of each independent variable with appropriate complementary feeding practices and to select candidate variables for multivariable logistic regression analysis. Variables with a p-value < 0.25 in the binary logistic regression analysis were entered into multivariable logistic regression. The assumptions of the logistic regression were also checked. Hosmer- Lemeshow tests for goodness of fit were carried out and found to be a good fit (p = 0.87(> 0.05)). Finally, the results were presented in texts, tables, graphs and discussed using the odds ratio and 95% confidence level.

All methods aforementioned above were carried out in accordance with relevant guidelines and regulations.

A total of 257 study participants participated in the study, resulting in a response rate of 97.5%. Half (52.1%) of the infants and young children were female. Among the respondent mothers, 109(42.2%) were between the age group of 25–29. Almost all (96.5%) of the participants were Orthodox Christian followers. Of the mothers who participated in the study 158(61.5%) were illiterate, 181(70.4%) were housewives and 238(92.6%) were married. Most husbands’ occupations were farmers (49.8%) and the mean number of family members and under-five children in the study area was 4 (±1.7) and 1 (±0.6) respectively (Table 1).

The mean number of pregnancy and antenatal care (ANC) visits was 3(±1.6) and 3(±1.4) respectively. Regarding mode of delivery, almost all (93%) mothers delivered through spontaneous vaginal delivery, four-fifths (80.9%) gave birth at a health institution, and 45.1% of women had a birth interval of 2 years. About 85.2% of mothers started breastfeeding within 1 h of delivery and only 13.6% of women practiced prelactal feeding. Cow milk was the most commonly used type of complementary feeding (33.5%). Approximaly half (50.6%) of mothers reported that they practiced harmful cultural feeding practices and almost all mothers (93.4%) reported that their current feeding practices were not similar to those before they were displaced (Table 2).

Only 14% of the mothers provided a diversified diet for their children aged 6–23 months. Among the seven food groups; grains, roots and tubers were commonly used, accounting for 89.9%. Only 32.7% of the mothers fed their children’s dairy products (Table 3).

The overall prevalence of appropriate complementary feeding practices was 26.8%. Indicators of complementary feeding were assessed and only 14% of mothers offered four or more food to their children which meeting the minimum dietary diversity criteria on the day preceding the survey. One-third (66.9%) of the mothers initiated complementary feeding at 6 months. About 37.7% of mothers met the minimum meal frequency according to child age category and their breastfeeding status on the day preceding the survey. The minimum acceptable diet for the children was 14% (Fig. 1).

Finally, in the multivariable logistic-regression analysis all significant variables in binary logistic regression were controlled for possible confounding factors. The results showed that the mothers whose child age was between 6 and 11 months were 89% less likely (AOR = 0.11, 95%CI: 0.04, 0.27) to practice appropriate complementary feeding practice as compared to mothers whose child age was between 18 and 23 months. Similarly, mothers whose infant age was between 12 and 17 months were 65% less likely (AOR = 0.35, 95% CI = 0.17, 0.70) to practice appropriate complementary feeding than those mothers whose infants were between 18 and 23 months. The other predictor variable found to be significant factor was the harmful cultural child feeding practice. Those women who did not have harmful culture on complementary feeding were 2 times (AOR = 2.04; 95% CI: 1.06,3.96) more likely to practice appropriate complementary feeding as compared to their counterparts (Table 4).