Bayesian small area estimation of anemia prevalence using a generalized linear mixed model

Anemia is a public health problem that can result in unfavorable outcomes for reproductive-age women. In Ethiopia, anemia is a health problem, particularly for women and children. The demographic and health survey (DHS) provides the data that is used to create reliable estimates of anemia status at the regional and national levels. However, because of their small sample sizes and high sampling variability, these surveys are not reliable for direct estimates at the zonal level. In order to obtain reliable and precise estimates of anemia prevalence for Ethiopian administrative zones, this study delineates Bayesian small area estimations (SAE) based on a generalized mixed effect model by combining DHS and census datasets. The empirical evidence indicates that the estimates of anemia prevalence of women generated by the SAE approach are reliable and more precise. The hierarchical Bayes (HB) estimates are anticipated to offer irreplaceable information to administrative decision-makers and policy experts for identifying the zones requiring more attention. The disaggregate level estimates of anemia are directly relevant towards achieving the health priorities of the Sustainable Development Goals (SDGs) indicator 2.2.3. Therefore, these estimates will be useful for fund allocation as well as monitoring and policy formulation at zonal level administrations.