Estimation of Early-Age Mortality in Different Districts of Pakistan

Abstract

Estimating early-age mortality at more localized administrative levels offers public health researchers a deeper insight into infant well-being and aids in developing health policies. To achieve this, design-based strategies are proposed for estimating the sur vival function, which is then applied to calculate district-level estimates of infant and neonatal mortality rates. The proposed survival function estimate employs an empir ical distribution function approach. The study evaluates four different strategies by comparing their relative bias and coefficient of variation. Using data from the Pakistan Demographic and Health Surveys (PDHS) of 2017-18 and a special 2019 survey, the re search emphasizes the significance of combining health survey data with administrative records to generate small-area health outcomes. The study introduces four methods for producing small-area estimates by integrating data from consecutive surveys through direct, synthetic, and composite methods. Among these, composite regression under Strategy 2 is identified as the most effective in terms of coefficient of variation and bias. The findings of this study are intended to assist public health policymakers in creating informed policies for areas with limited data and providing a clearer picture of infant mortality rates for governments and NGOs focused on neonatal mortality.