Study setting and period
A community-based cross-sectional study was conducted from February 10, 2019, to March 21, 2019, among currently married women of reproductive age who have children aged 6–59 months.
The study was conducted in three rural kebeles (Weramit, Wereb Qolatsion and Zenzelma) of Bahir Dar City Administration. Bahir Dar is the capital city of the Amhara Regional State, located 565 km northwest of Addis Ababa, the capital of Ethiopia. For administrative purposes, Bahir Dar is divided into nine sub cities, four satellite cities and nine rural kebeles. According to the City Administration Health Department report, the total population of the town in 2018 was estimated to be 339,683, of which 65,224 were living in rural kebeles. In the same year, the estimated number of under-five-year-old children was 46,007, with 8,834 of them were living in rural kebeles [22].
Sample size and sampling procedure
The sample size was calculated using the single population proportion formula with the following assumptions: a 95% CI (z value of 1.96), 5% margin of error, and a 42.3% prevalence of stunting among children aged 6–59 months from a previous study in Gondar, northwest Ethiopia [23]. With an assumption of 10% non–response rate and taking design effect of 1.5 (for multistage systematic random sampling), the final sample size was 619.
To select study participants, multistage sampling technique was implemented. First, three kebeles were randomly selected from nine rural kebeles in Bahir Dar using a lottery method. Then, proportional allocation was used to determine the desired samples from each selected kebele. In the second stage, systematic random sampling was used to select eligible households, i.e., households with mothers having children aged 6–59 months. The sampling frame of all households with eligible under-five children was obtained from Health Extension Workers as they record all mothers and their children in their kebele under ‘the family folder’. Finally, eligible study participants were recruited, and a house-to-house data collection was conducted. If the mother had more than two eligible children, only the index (youngest) child was sampled to avoid the clustering effect and all the information was collected for the index child.
Measurement and variables
The dependent variable was stunting, which was defined based on anthropometric measurments of height and age. The age of the child was based on the mothers response. To reduce recall bias, mothers response about their child’s age was crosschecked with immunization cards. In addition, respondents were asked to relate the childbirth date with major religious and cultural events. Height measurements were carried out using a Shorr measuring board (ShorrBoards®) and recorded to the nearest 0.1 cm. Children younger than 24 months were measured for length in recumbent position, and older children were measured while standing. Thus, stunting (height for age) was classified into a binary outcome variable (z-score below − 2 standard deviations as ‘stunted and ‘normal’ otherwise) based on the WHO growth standard and nutritional status classification criteria [2].
The main independent variables were the overall levels of women’s empowerment and its various dimensions. For this study, we identified five dimensions of women’s empowerment. The empowerment dimensions were identified based on a suggestion from a previous literature review to select women’s empowerment dimensions in context-specific ways [7] and through a review of literature from Ethiopia [20, 21] and globally [11, 14]. The five dimensions were household decision-making, educational status, cash earnings, house/land ownership, and membership in community groups. The five dimensions of empowerment and their categories are presented below (Table 1). The overall empowerment level was grouped into three categories as low (women who received a total score of two or less), moderate (a score of three or four), and high (a score of five or more), as previously suggested [14].
In addition, variables that needed to be controlled in order to estimate the unbiased independent association between the exposure and the outcome were identified from previous literature [9, 11,12,13,14, 20, 24,25,26,27]. Accordingly, variables related to mother’s characteristics (age, age at first marriage, household family size, access to media and employment status) and child factors (age, sex, birth order, breastfeeding status and morbidity status) were identified.
Data collection tools and procedures
Socio-demographic data were collected using a structured questionnaire adapted from the 2016 Ethiopian Demographic and Health Survey [5]. The tool for measuring women’s empowerment was adapted from previous literature [7, 11, 14, 21, 28]. The anthropometric data were collected using the procedure stipulated by the WHO for taking anthropometric measurements [29]. The questionnaire was pre-tested on 30 participants in a similar setting (Addis Alem kebele) after being translated into the local Amharic language. However, there was no any difficulty in understanding the questions and no modifications were made. Trained data collectors with previous experience of anthropometric measures collected the data.
Data processing and analysis
The collected data was coded and entered into Epi-data version 3.1. Anthropometric measurements were calculated using WHO Anthro software version 3.2.2. The data was analyzed using SPSS version 20.0. Binary logistic regression analyses were used to assess the association between the dependent variables with the main independent variable, as well as each control variables. Variables with a p-value of ≤ 0.2 in the binary logistic model were retained in the final multivariable logistic models to control potential confounding effects. Interactions between variables were assessed using variance inflation factor. Model fit was tested using Akaike Information Criteria and additional standard logistic regression diagnostic tests. Odds ratios, together with the 95%CI, were used to report the associations. Statistical significance was declared at a p-value of < 0.05.
Add Comment