Title:	Advanced Techniques for Modelling Maternal and Child Health in Africa
Authors:	Ngianga-Bakwin Kandala Gebrenegus Ghilagaber KENNETH C. LAND
Keywords:	Child Health
Issue Date:	2014
Publisher:	Springer
Description:	The estimation of levels, trends, and differentials in demographic and health outcomes in developing countries has, over the years, relied heavily on indirect methods that were devised to suit limited or deficient data. In recent decades, some worldwide surveys like the World Fertility Survey and its successor, the Demographic and Health Survey (DHS), have played an important role in filling the gap in the availability of survey data in developing countries. These surveys, conducted at enormous costs, are aimed at enabling investigators to make in-depth analyses that could guide policy intervention strategies. However, their utilization remains suboptimal, because optimal analyses of such data demand advanced statistical techniques. Since the use of DHS data in developing countries, many developments in statistical modelling based on hierarchical models have been published, and our primary aim is to bring together the various methodological advances. Naturally, the choice of these recent developments reflects our own teaching and research interests. We try to motivate and illustrate concepts with examples using real data from the DHS, and the data sets are available on http://www.measuredhs.com. We could not treat all recent developments in the area of health and survival in Africa in this book, and in such cases we point to references at the end of each chapter. The book presents both theoretical contributions and empirical applications of such advanced techniques. We cover a range of new developments from both the classical and Bayesian approaches. In the Bayesian framework, Monte Carlo techniques, in particular MCMC, and their application to spatial and spatio-temporal data are covered. These include techniques such as geoadditive semi-parametric models that link individual health outcomes with area variables to account for spatial correlation; latent modelling that deals with the impact of spatial effects on latent, unobservable variables like “health status” or “frailty”; spatial modelling of multiple diseases that enables quantifying the correlation between relative risks of each disease as well as mapping of disease-specific residuals; and Bayesian structured geostatistical regression modelling that permits a joint estimation of the usual linear effects of categorical covariates, non-linear effects of continuous covariates and small-area district effects on health outcomes within a unified structured additive Bayesian framework.
URI:	http://10.6.20.12:80/handle/123456789/52598
ISBN:	978-94-007-6778-2
Appears in Collections:	Population Studies

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.