Estimation of Health Indicator at Dis-aggregated Geographic Levels

Abstract

This study leverages data from the Demographic and Health Surveys (DHS) conducted in 2017-18 and 2019 to examine the utilization of antenatal care (ANC) services among women across various districts in Pakistan. The analysis employs a mixed-methods ap proach, integrating both area-level and unit-level auxiliary information to enhance the robustness of the findings. Specifically, the Fay-Herriot (FH) model is utilized for area-level data, capturing district-level healthcare infrastructure and socio-economic indicators, while the Battese-Harter-Fuller (BHF) model is applied to unit-level data, encompassing individual socio-demographic factors such as age, education, and wealth index. By merging these models, the study aims to provide a comprehensive under standing of ANC utilization patterns and their determinants at multiple levels. The combined approach not only enables the identification of nuanced relationships between various factors influencing ANC utilization but also facilitates the generation of more reliable and precise estimates. The results indicate significant variations in ANC coverage, highlighting critical areas where interventions are needed to improve maternal health outcomes. The integration of both FH and BHF models allows for a detailed examination of both district-level and individual-level factors, offering valu able insights for policymakers and healthcare providers aiming to enhance maternal healthcare services in Pakistan. This study underscores the importance of utilizing advanced statistical models to address data limitations and improve the accuracy of small area estimates.