Study Area and period
The 2016 Ethiopian Demographic and Health Survey data was employed for this study. It was a nationwide survey intended to provide up-to-date estimates of key demographic and health indicators. Ethiopia, with a projected total population of 118,218,249 by 2021 according to World population review [19], is the most populous country in East Africa, and the second-most populous country in Africa where the majority (93%) Ninety-three percent of households use some type of solid fuel for cooking mainly would for domestic energy needs [20]. The country is divided into 9 regions (Oromiya, Amhara, SNNPR, Benishangul-Gumuz, Tigray, Afar, Somali, Gambela, and Harari) with two city administrations (Addis Ababa and Dire Dawa) which is further sub-divided into zonal administration then into Woreda for administrative purpose. The Woreda is also divided into kebele which is the smallest administrative unit.
Study design, sampling technique, and data collection
We have conducted a cross-sectional study using the survey of the 2016 Ethiopian demographic and health survey (DHS) data. This survey was carried out by the Ethiopian central statistics agency (CSA) with the Ministry of Health (MoH) under the ICF technical assistance using a census frame of a complete list of 84,915 enumeration areas (EAs) created for 2007. the survey employed a two-stage stratified sampling to select eligible residential households across 645 enumeration areas (EAs). Each region was stratified into urban and rural areas, giving 21 sampling strata. Samples of EAs were selected independently in each stratum in two stages.
Dependent variable
The data collected regarding blood hemoglobin level was used as a base for categorizing the anemia status of non-pregnant women of reproductive age. The blood samples for anemia testing were taken from voluntarily consented non-pregnant by capturing a drop of blood from their finger prick and collected in a microcuvette. The analysis to determine the blood Hemoglobin level was conducted on-site using a battery-operated portable HemoCue analyzer. After the onsite analysis, the concentration of blood hemoglobin from each sample was measured in g/dl. The World Health organizations (WHO) criteria were used to categorize the non-pregnant women's anemia status using their hemoglobin (Hb) level. Hence, the non-pregnant women were categorized as not anemic (Hb ≥ 12.0 g/dl), mildly anemic (10.0–11.9 g/dl), moderately anemic (7.0–9.9g/dl), or severely anemic (Hb < 7.0/dl) [21]. Afterward, the non-pregnant women were further categorized as not anemic and anemic.
Exposure variable
The fuel type for household energy demand was the main exposure variable. The non-pregnant women were asked during the survey “What type of fuel does your household mainly use for cooking?” to assess the main fuel types used in the household and there were eleven fuel types namely, (electricity, liquefied petroleum gas (LPG), natural gas, biogas, kerosene, coal/lignite, charcoal, wood, straw/shrubs/grass, agricultural crop, animal dung) as a response. Therefore, these were further re-categorized into biomass fuels (wood, straw, animal dung, and crop residues) and clean fuels (electricity, liquid petroleum gas, and natural gas) [22]. For this analysis medium polluting fuels were removed from the analysis, and only biomass fuel users and clean fuel users were assessed for anemia status and other predictor variables. Because of the small number, moderate and severe anemia are categorized into one variable, as ‘moderate to severe anemia’.
Other predictor variables
The predictor variables included in this study were, women-related factors (the age of the women, BMI, education level, chat chewing, and parity) and socio-demographic factors (sex of head of the household, residency, and wealth index). Maternal age and BMI of the non-pregnant women in reproductive age were used as a continuous variable for the analysis and the highest education level achieved by the non-pregnant women in reproductive age was categorized as no formal schooling, primary school, secondary, and higher (university/college and above). The tobacco smoking status was labeled as ‘yes’ or ‘No’. The residency was grouped into urban and rural while the wealth index was a five-category variable as poorest, poorer, medium, richer, and the richest.
Inclusion and exclusion criteria
Of the total of 15683 women, 4722 of them were excluded because they didn’t fulfill the inclusion criteria for this study as shown in (Figure 1) and a total of 10,961 women were enrolled for the final analysis in this study.
Data analysis and sample characteristics
We have used SPSS analysis software version 25 for the analysis. Multinomial logistic regression was employed to determine whether there was an association between the use of different fuel types and the anemia status of the non-pregnant women in reproductive age in Ethiopia. Adjusted Odds Ratio was used to measure the association and their 95% Confidence Interval was calculated. A p-value of less than 0.05 was considered for statistical significance. An independent sample t-test was used to assess the mean difference in blood Hemoglobin level (g/dl) between biomass and clean fuel users.