Study setting
The study was conducted in Ethiopia (3° -14° N and 33°-48° E), which is located at the horn of Africa. The country covers 1.1 million sq. km and has a great geographical diversity, which ranges from 4550m above sea level down to the Afar depression and 110m below sea level. There are nine regional states and two city administrations subdivided into 68 zones, 817 districts, and 16,253 Keeble’s (lowest administrative units of the country) in the administrative structure of the country (2016 EDHS).
Data source and sampling
The Demographic and Health Survey (DHS) Programme provides publicly free access to survey data for responsible researchers. Therefore, accessed the datasets using the website (http://www.measuredhs.com). After the reasonable request of the Demographic and Health Survey (DHS).
Accordingly, the Ethiopian Demographic and Health Survey (EDHS) was used for the current study, which has collected data on national representative samples of all populations, including reproductive-age (15–49) women, every five years. To date, four surveys have been conducted, and anemia has been included as a key indicator since the 2005 survey.
Study Design.
A community-based cross-sectional study design was conducted at the national level as one part of the periodic EDHS. The survey was conducted with nationally representative samples from all of the regions of the country.
Data Extraction Methods
EDHS data were downloaded, with permission, from the Measure DHS website in STATA. After a review of the detailed data coding, further data recoding was completed. In the 2005, 2011, and 2016 EDHS datasets, there were 1,133, 1,104, and 1,113 pregnant mothers, respectively. Information on a wide range of sociodemographic, economic, household, and obstetric characteristics, anemia level, and other indicators was extracted.
Measurement and Operational Definition
Anemia status was determined based on haemoglobin concentration in the blood adjusted to the altitude. Anemia was defined as the occurrence of a haemoglobin level of less than 11 g/dL. It was further categorised into mild, moderate, and severe anemia with a haemoglobin range of 10–11 g/dl, 7–10 g/dl, and 7 g/dl, respectively.
Study Population.
Secondary data analysis recorded from the archive of EDHS 2005, 2011, and 2016 data identified the 15–49 reproductive age group from this group as all pregnant mothers who have haemoglobin status records included as study population, and based on the level of haemoglobin, the pregnancies were classified according to the level of haemoglobin stated in the operational definition.
Study variables
Dependent variable
Anemia in pregnant women
Independent variable
pregnancy, age, lactation, region, number of parities, residence, gestational age, education, wealth index, smoking, and family size.
Inclusion criteria
All pregnant women whose anemia level was recorded
Exclusion criteria
Pregnant Women whose anemia level were not recorded
Data quality control and analysis
An initial exploratory data analysis was conducted to check for outliers, missing data, and consistency. All the results of this study were weighted for sampling probabilities using the weighting factor in the EDHS data. The complex sampling procedure was also considered using STATA. The data were analysed using STATA version 14.0 after the three datasets were appended.
Spatial analysis
The spatial autocorrelation (Global Moran’s I) statistic was used to evaluate whether the anemia patterns were dispersed, clustered, or randomly distributed during the 2016 survey periods in Ethiopia. The decision was made based on the calculated Moran’s I values. When the Moran I value is close to 1, it indicates anemia is dispersed, whereas a Moran I value close to +1 indicates anemia is clustered in the study area. However, Moran’s I value of zero shows a random distribution of anemia. Local Moran’s identify hotspot clusters (high-high) and cold spot clusters (low-low). It also measures outliers in which high values were surrounded primarily by low values (high-low) and outliers in which low values were surrounded primarily by high values (low-high) (12). This spatial analysis technique was employed to detect the local-level risk areas of anemia and their outliers on a separate map. Hotspot analysis computes the Z-score and p-value to determine the statistical significance of the clustering of anemia over the study area at different significance levels simultaneously (12). In this analysis, the p-value associated with a 95%, 90%, and 99% confidence level would have been used to determine the existence of significant clustering. Areas at high risk (hotspot) of pregnant anemic mothers (the statistical output with high Gi) and areas at low risk (cold spot) of anemia during pregnancy (the statistical output with low Gi) were detected (12, 13). The spatial interpolation technique was applied to predict the unsampled areas from sampled measurements (14). The ordinary Kriging spatial interpolation method was used to predict raster surfaces from point data. Therefore, a smooth surface for the risk areas of anemia among pregnant mothers was indicated on the anemia risk map. Identifying the most likely clusters was done using the spatial scan statistical method, a method that is widely recommended as it is very important in detecting local clusters and has higher power than other available spatial statistical methods. (14). Women with anemia were taken as cases, and non-anaemic ones were considered controls to fit the Bernoulli model. The default maximum spatial cluster size of 50% of the population was used as an upper limit, which allowed both small and large clusters to be detected and ignored clusters that contained more than the maximum limit. For each potential cluster, a likelihood ratio test statistic was used to determine if the number of observed anemia cases within the potential cluster was significantly higher than expected or not. The primary and secondary clusters are identified and ranked based on their likelihood ratio based on 999 Monte Carlo replications. Therefore, the most likely risk areas for anemia among pregnant women in the 2016 survey were indicated in the spatial map.