Small Area Estimation of Child Mortality in Sub-Saharan Africa: Administrative Level 2

Jessica Godwin , University of Washington, Seattle
Jon Wakefield, University of Washington, Seattle

We use a spatiotemporal smoothing model to smooth estimates of child mortality at the Admin-2 level. Previous work provided smooth estimates of child mortality at the Admin-1 level using DHS full birth history data, while correctly accounting for the complex survey design. We extend this work by incorporating MICS full birth history data and MICS and census summary birth history data. We model on a yearly scale which allows us to combine estimates on a yearly scale and estimates on coarser time scales. Additionally, we provide a model based adjustment method for adjusting survey-based estimates in areas where we do not observe any deaths due to small sample sizes. We provide results from Kenya, but will apply this method to many countries in sub-Saharan Africa.

See extended abstract

 Presented in Session 47. Spatial Methods