Analysis of Childhood Diseases and Malnutrition in Developing Countries of Africa

The objective of this work is to examine the impact of
socioeconomic and public health factors on childhood diseases and
malnutrition in mentioned countries. The causes of child's illness
or child's undernutrition are multiple. This work focuses on some
risk factors which are assumed to cause the child's diseases and
malnutrition as suggested by some previous works (see Kandala,
2001; Adebayo, 2002). Our analysis started with a large number of
covariates including a set of bio-demographic and socioeconomic
variables, such as current working status of mothers, place of
residence, access to toilet facilities, etc (see chapter 2). The
analyses are based on data from the 2003 household survey for Egypt
and Nigeria for the Demographic and Health Surveys (DHS). More
details about the data set are mentioned in the first chapter. The
statistical analysis in this thesis is based on modern Bayesian
approaches which allow a flexible framework for realistically
complex models. These models allow us to analyze usual linear
effects of categorical covariates, nonlinear effects of continuous
covariates and the geographical effects within a unified
semi-parametric Bayesian framework for modelling and inference. A
first step of this work is to analyze the effects of the different
types of covariates on response variables, diarrhea, fever, and
cough which represent the child's diseases in our application. In
this step, a Bayesian geoadditive logit model for binary response
variables is used (see Fahrmeir and Lang, 2001). In a second step,
we employ separate geoadditive probit models (instead of logit
models used in the previous step) to the binary listed variables.
Based on the results of the separate analyses, we applied
geoadditive latent variable probit models (recently suggested by
Raach, 2005; Raach and Fahrmeir, 2006) where the three observable
disease variables are assumed to be indicators for the latent
variable "health status" for the children. In this step, we also
compared the results of the separate geoadditive probit models with
the results of the latent variable models. As a third step, we used
geoadditive Gaussian regression and latent variable models to
analyze the malnutrition status of children in both countries.
Finally, we used latent variable models for diseases and nutrition
indicators together. In the final step, models with one as well as
with two latent variables have been estimated using mixed
indicators (binary indicators "health status", and continuous
indicators "nutrition status") and the results are compared.