Spatial Smoothing for Estimating Health Outcomes at Lower Subnational Level Using Nationally Representative Population-Based Surveys in the Sub-Saharan Africa

Seminário de Estatística e Gestão do Risco 

Orador: Samuel Manda, Biostatistics Research Unit, South Africa Medical Research Council Pretoria, South Africa. 

Data: 25 Setembro de 2018

Hora: 14H00

Local: Sala de Seminários - Edifício VII 

Abstract: Several researchers in the Sub-Saharan Africa have used data collected from nationally representative population health surveys to perform spatial smoothing modelling for small area estimation of health outcome.  Most of these surveys employ complex designs and to estimate reliably the spatial trend, recent research has encouraged accounting for design weights in the estimation process.  Despite these desired modeling efforts, potential bias introduced because of non-response, missing data and self-reporting of health conditions remain in estimating small area spatial health variations.  Several health conditions could be considered for spatial small area smoothing using a survey, and these may share the same set of (both individually and spatially distributed) risk factors or may be linked etiologically. An even greater statistical challenge in using these data for spatial smoothing of health outcomes is the representativeness of data at lower subnational levels. Most of these surveys were designed to collect representative data at national and regional levels, and lower administrative levels may not have been systemically covered sufficiently. In this paper, we explore these design and statistical problems using theory and motivating examples from household and population health surveys in the region.  

Financiado através do projeto UID/MAT/00297/2013.