The DHS Program

Spatial Modeling for Subnational Administrative Level 2 Small-Area Estimation (English) 

Document Type

Spatial Analysis Reports

Publication Topic(s)

Geographic Information

Country(s)

Cameroon, Ethiopia, Kenya, Malawi, Nigeria, Zambia, Bangladesh, Nepal

Language

English

Recommended Citation

Wu, Yunhan, Zehang Richard Li, Benjamin K. Mayala, Houjie Wang, Peter Gao, Johnny Paige, Geir-Arne Fuglstad, Caitlin Moe, Jessica Godwin, Rose E. Donohue, Trevor N. Croft, and Jon Wakefield, 2021. Spatial Modeling for Subnational Administrative Level 2 Small-Area Estimation. DHS Spatial Analysis Reports No. 21. Rockville, Maryland, USA: ICF.

Publication Date

September 2021

Publication ID

SAR21

Abstract:

Subnational estimates of the health and demographic indicators recorded in the Demographic and Health Surveys Program are of great importance for prioritizing resources and assessing if target levels for indicators are being attained. In this report, we examine subnational variation in the under-5 mortality rate by using small area estimation models with the goal of estimating at the Admin 2 level. We describe spatio-temporal modeling and consider discrete spatial models in detail, the possibility of including covariates, and accounting for urban/rural stratification, model assessment, and visualization of results. We offer recommendations for subnational modeling, describe an analysis pipeline, and include R code to perform the various steps, using the SUMMER package. We illustrate methods and provide results for Bangladesh, Cameroon, Ethiopia, Kenya, Nepal, Nigeria, Malawi, and Zambia.