|Risk factors of anemia among preschool children in Ethiopia: a Bayesian geo-statistical model|
||Bilal Shikur Endris, Geert-Jan Dinant, Seifu H. Gebreyesus and Mark Spigt
||BMC Nutrition, Volume 8, issue 2; DOI:https://doi.org/10.1186/s40795-021-00495-3
The etiology and risk factors of anemia are multifactorial and varies across context. Due to the geospatial clustering of anemia, identifying risk factors for anemia should account for the geographic variability. Failure to adjust for spatial dependence whilst identifying risk factors of anemia could give spurious association. We aimed to identify risk factors of anemia using a Bayesian geo-statistical model.
We analyzed the Ethiopian Demographic and Health Survey (EDHS) 2016 data. The sample was selected using a stratified, two- stage cluster sampling design. In this survey, 9268 children had undergone anemia testing. Hemoglobin level was measured using a HemoCue photometer and the results were recorded onsite. Based on the World Health Organization’s cut-off points, a child was considered anaemic if their altitude adjusted haemoglobin (Hb) level was less than 11?g/dL. Risk factors for anemia were identified using a Bayesian geo-statistical model, which accounted for spatial dependency structure in the data. Posterior means and 95% credible interval (BCI) were used to report our findings. We used a statistically significant level at 0.05.
The 9267 children in our study were between 6 and 59?months old. Fifty two percent (52%) of children were males. Thirteen percent (13%) of children were from the highest wealth quintile whereas 23% from the lowest wealth quintile. Most of them lived in rural areas (90%). The overall prevalence of anemia among preschool children was 57% (95% CI: 54.4–59.4). We found that child stunting (OR?=?1.26, 95% BCI (1.14–1.39), wasting (OR?=?1.35, 95% BCI (1.15–1.57), maternal anemia (OR?=?1.61, 95% BCI (1.44–1.79), mothers having two under five children (OR?=?1.2, 95% BCI (1.08–1.33) were risk factors associated with anemia among preschool children. Children from wealthy households had lower risk of anemia (AOR?=?0.73, 95% BCI (0.62–0.85).
Using the Bayesian geospatial statistical modeling, we were able to account for spatial dependent structure in the data, which minimize spurious association. Childhood Malnutrition, maternal anemia, increased fertility, and poor wealth status were risk factors of anemia among preschool children in Ethiopia. The existing anaemia control programs such as IFA supplementation during pregnancy should be strengthened to halt intergenerational effect of anaemia. Furthermore, routine childhood anaemia screening and intervention program should be part of the Primary health care in Ethiopia.