Study setting and subjects
The study was conducted in Enebsie Sar Midir Woreda, East Gojjam Zone, North West Amhara. Enebsie Sar Midir is one of the woredas in East Gojjam Zone ANRS. The Woreda is located 365 kilo meters far from Addis Ababa, the capital city of Ethiopia and 182 km far from the capital city of Amhara National Regional State, Bahir Dar. The organizational structure of the Woreda includes four urban and 31 rural kebeles. There are eight health centers /HCs/ and one governmental Hospital in the woreda.
Based on the 2007 national census conducted by the Central Statistical Agency of Ethiopia (CSA), the woreda has a total population of 133,855, of whom 66,139 were men and 67,716 women; 12,259 or 9.16% were urban inhabitants.
Study design and study period
Cross sectional community-based study design was conducted from January to February 2020to assess minimum dietary diversity.
All children aged 6–23 months along with mothers or care givers who located in Enebsie Sar Midir Woreda were the source population.
All children aged 6 months to 23 months along with mothers or care givers drawn from the selected kebeles were the study population.
Sample size determination
A single population proportion estimation formula was used to calculate the sample size. With a precision (d) of 3%, a confidence level of 95% (z = 1.96), and an expected prevalence of adequate dietary diversity of 13% , a sample size of 482 was calculated; however, when a 10% non-response rate was added, the total sample size was 530.
Systematic sampling was used to select the study subjects in the community. As households were sampling units the sampling interval was obtained by dividing the total number of households of 19,122 by the sample size. There are 34 kebeles in the woreda, 30 of which are located in the rural woreda and 4 in the urban woreda. Thirteen kebeles, or around 30% of all kebeles, were included in the study, with 11 from rural areas and two from urban areas.
The first household was randomly selected within the interval of ten using lottery method. Then a bottle was tossed in the middle of each kebele after deciding to follow the head of the bottle. Then the households were counted in the direction of the head of the bottle to identify the first household. After that the next households were selected by adding the interval. In the even bet where there is no eligible person within the selected household, the immediate next household was visited and this continued until the whole sample size was fulfilled.
All mothers and care givers with infants six months to 23 months of age were included in the study.
Seriously ill mothers, non-consenting mothers were excluded from study.
To analyze the data, the study focused on the following dependent and independent variables. The dependent variables of this study were Minimum Dietary Diversity, and Independent Variables like Socio Demographic Characteristics, Economic characteristics of parents, Agricultural factors and Wealth Index.
Data collection procedure
Data collection tool
The data collection tool was structured interview questionnaire which had factors for minimum dietary diversity.
Data collection procedure
The data collection procedure started from identifying the study subjects and selected randomly who was included in the study. The data collection was conducted by using 24-hour recall method in the community from mothers or care givers. All participants who fulfill the criteria were participated. Then oral informed consent was conducted.
The data collectors were grade 10 and 12 complete persons who have previous data collection experience in health care program areas.
Before the data collection, the principal investigators had given basic training about the questionnaire. Thus, all data collectors know about the objectives of the study, the data collection approach, the discipline need to have during data collection and other important issues related to the problem.
The principal investigator, and health extension workers were recruited the supervisor purposively. These persons were familiar about the overall information needed, about the purpose of this research as a supervisor to guide the data collectors, about the time of collection, how to manage the study subjects during data collection procedure, and fix the time from when to when completed questionnaire received. The principal investigator closely followed the data collection procedure. Completed questionnaires were rechecked for completeness and consistency starting from the beginning of the data collection to provide feedback for the next steps of the data collection throughout the of the data collection time.
Data quality control
The structured interview questionnaire was adapted from WHO IYCF indicators and other related scholars to the specific settings considering the objective of the study [22, 23]. The questionnaire was modified according to the study variable included and specific context. The questionnaire was prepared in English first and then it is translated in to Amharic by experts who work on this area before data collection. Pretest was done on 5% of the total sample of the study subject to ensure that the language and contents are culturally suitable, acceptable, and clearly understandable for the participants. Before data collection took place necessary corrections were made on the questionnaire to ensure the quality. During data collection, double checking the questionnaires was employed by the principal investigator.
A one-day training was conducted for all research team members on data collection methods, tool pretesting and reflection. The contents of the training included privacy assurance, confidentiality, interview techniques, and reviews of the study protocol and questionnaire. Research team members were also equipped with quality control skills such as rechecking and reviewing the questionnaires after administration as well as resolving issues that might arise during the fieldwork.
Data management and data analysis
The data were entered in Epi INFO version 7 and transferred in to statistical package for social science (SPSS) Version 20. The data were cleaned and analyzed using SPSS for windows version 20 (Illinmoise Chicago). Frequency distribution table was used to show the distribution of variables. Percentages and table were used describe result of the study. Chi square was used to examine the relationship between minimum dietary diversity and associated factors. P < 0.05 was considered statistical significance. Bivariate and multivariabe logistic regression analysis were employed. Variables that had P value less than 0.25 in the bivariate model were entered into the multivariable regression model. Different factors were regressed with minimum dietary diversity as dependent variable to isolate its independent predictors. Model fitness was checked using Hosmer Lemeshaw test at P > 0.05. Multicolliniary was checked using standard error ≥ 2.0. The results are presented using adjusted Odds Ratios and 95% confidence intervals.
Wealth index was generated by asking durable assets including radio, TV, mobile phone, hand cart, plough plow. Ownership of each asset was given a score of “1” and non ownership was given a score of “0”. Principal component analyses was conducted to generate wealth index after checking all assumptions such as sample adequacy (KMO ≥ 0.5, antiimage ≥ 0.5 and communality ≥ 0.5, Bartlets test of Sphericity (P < 0.5) and absence of variable with complex structure. The wealth index score was then rank ordered into tertile and the highest tertile was taken as rich while the two lowest tertile combined were taken as poor.