In countries where availability of census and vital registration data are limited, estimating subnational health and demographic indicators is challenging. Existing small area estimation approaches from the survey statistics literature often rely upon the availability on high-quality census information. On the other hand, standard model-based geostatistical approaches used in global health research leverage spatial smoothing and covariates derived from satellite imagery to reduce uncertainty in estimates but often assume the survey design is ignorable, which may be inappropriate given the complex design of household surveys typically used in this context. In this talk, I discuss two projects that bridge these two perspectives and develop area level models for estimation of subnational demographic indicators in low- and middle- income countries using household survey data.

In the first project, we propose a smoothed model-assisted estimator that accounts for survey design and leverages both unit-level covariates and spatial smoothing. Under certain assumptions, the new estimator can be viewed as both design-consistent and model-consistent. In the second project, we adapt our approach for improved estimation of prevalence of rare events such as neonatal mortality using an area level model with spatial smoothing of variance estimates. We demonstrate our estimator's performance using both real and simulated data, comparing it with existing design-based and model-based estimators.