Study design and setting
Community based cross-sectional study was conducted in Sinan Woreda, East Gojjam Zone from February 16 to March 10, 2016. Sinan Woreda is located at 327 km from Addis Ababa in Northwest and 303 km from Bahir Dar in Southeast. The estimated population of the Woreda in 2015/16 is about 114,475. The estimated number of the under-five year children is 15,499. The total number of children aged 6–23 months account 4315 (Sinan worea health office: Health sector woreda base strategic plan report of sinan woreda, unpublished). Quantitative data were collected from mothers.
We obtained ethical clearance from the Ethical review committee of College of Health Sciences of Debre Markos University. Verbal informed consent, which was prepared in written form and dictated to the respondents during data collection, was obtained from the study participants after explaining the purpose of the study and the benefits. Respondents were interviewed voluntarily, anonymously and confidentiality also was assured. All participants were allowed to ask questions throughout data collection and could refuse to answer questions or stop the interview at any moment.
Sample size determination and sampling procedure
Sample size was determined using single population proportion formula. \( N=\frac{{\left(Z\frac{a}{2}\right)}^2p\left(1-p\right)}{d^2} \) [27], where, p represents the proportion of children who took the optimum dietary diversity, which was 12.6% taken from the study done in Ethiopia [28]. To get the optimum sample size, 3% margin of error (d) was considered with 95% confidence interval.
$$ N=\frac{(1.96)^20.126\left(1-0.126\right)}{0.03^2}470 $$
We added 5% for non-response rate and multiplied by 1.5 because of design effect. Then the final sample size was 740.
Of the 17 administrative Kebeles in the Woreda, seven were selected randomly for the study. Proportional to population size allocation was done to select the desired samples from each selected Kebele. Sampling frame, based on community-based health information system of family folder in health posts, was constructed. Lists of all mothers having children age 6–23 months with Community Health Information System (CHIS) number in selected kebeles were used to select the respondents through computer generated methods of random sampling. Children aged 6–23 months who did not start complementary food were excluded from this study.
Data collection procedure and measurements
Questionnaires were composed of dietary diversity score adapted from the World health organization IYCF guideline which contains seven food groups for young child (6 to 23 months old) dietary diversity (Grains, roots and tubers; Legumes and nuts; Dairy products (milk, yogurt, and cheese); Flesh foods (meat, fish, poultry and liver/organ meats); Eggs; vitamin-A rich fruits and vegetables; Other fruits and vegetables) [7] and maternal and child demographic characteristics adapted from EDHS 2011 [15]. Questionnaires were first prepared in English and translated into Amharic Version, which later on, were translated into English. Amharic version questionnaires were used to collect data. Pre-test was done on 5% of the sample, two weeks before the actual data collection at Machakel Woreda. Data on dietary diversity were collected through face to face interviewing of mothers having children aged 6–23 months by allowing them freely to recall the type of food items they feed to their child/children within the last 24 h. Twelve data collectors, recruited based on their previous experience, and two supervisors had participated in data collection process. One day training on questionnaires and methods of data collection procedures was provided to data collectors and supervisors. Supervisors had checked completeness and consistency of the collected data by reviewing each completed questionnaire daily.
The dependent variable is Dietary diversity (dichotomized as optimal /suboptimal). Independent variables include socioeconomic and demographic characteristics (like age, educational status, occupation etc.), health utilization, and child health characteristics.
Operational definition
Optimal dietary diversity: Dietary diversity was defined as optimal if children (aged 6–23 months) received foods from at least four of seven food groups [(1) Grains, roots, and tubers, (2) Legumes and nuts, (3) Dairy products, (4) Flesh foods, (5) Eggs, (6) Vitamin-A rich fruits and vegetables, (7) Other fruits and vegetables, within the preceding 24 h of interview.
Sub-optimal dietary diversity: was defined as receiving three foods or fewer within 24 h [7, 29].
The terms ‘Women and mothers’ are used interchangeably in this study.
Statistical analysis
Data were entered into EpiData version 3.1 with double entry verification. Analysis was performed using SPSS version 20.0. Frequency and cross-tabulation were used to present descriptive data. Both the bivariate and multivariable logistic regression analyses were performed to assess the association between dependent and independent variables. Independent variables that showed P < 0.2 at 95% CI in the bivariate logistic regression analysis were included in multivariable logistic regression model. P < 0.05, with 95% CI, was considered to declare the variables significantly associated with the dependent variable.