Estimating the local-level child full polio vaccination rates in Ethiopia using a hierarchical Bayes small area estimation approach

Abstract

Vaccination is one of the most effective, affordable, and life-saving medical interventions ever created. Child vaccination is fundamental to building a healthy and welfare society, which is crucial in 2063 African and 2030 global agendas. This study combines the 2019 Mini Ethiopian Demographic and Health Survey (EDHS) with the 2007 population and housing census datasets to employ the hierarchical Bayes (HB) small area estimation (SAE) approach for estimating local-level child vaccination rates. In the HB SAE framework, the deviance information criterion (DIC) was used to select the best candidate model among the three different models fitted. The logistic normal mixed model with known sampling variance was chosen over the other two models (Fay-Herriot model and log-normal mixed model). The mean coefficient of variation (CV) for direct survey-based estimates is 44.41, which is higher than that for the model-based HB estimates at 36.40. Similarly, the root mean square errors (RMSE) of direct survey estimates are greater than those of the corresponding model-based estimates. Therefore, the results suggest that the HB estimates show improvement over the survey-based estimates. This finding also contributes to the sustainable development goal for health (SDG3), which aims to ensure healthy lives and promote well-being for people around the globe.