Relationship of health care practices and social status of mothers with various levels of child mortality: An application of multinomial regression model

Statistical Modelling is used in child mortality research since a long time in order to develop predictive models as well as to know the factors which affect the child mortality. There have been various socio-economic and biological variables identified from previous research which significantly effect child mortality, the role of these factors vary in case of child mortality in different ages of death of the child, so to find out such differences we classify the women according child mortality experience in four categories or levels i.e. Neo-natal mortality, Post neonatal mortality, toddler mortality and women who not experience child mortality in childhood period, further multinomial logistic regression model was applied to find out the effect of health care practises and social situation of mothers on various levels of child mortality. It was found that the colostrums feeding at birth protects from death at neonatal mortality and if a woman wants to have another child then the present child has risk to die in neonatal or post-neonatal period. The poor wealth index was came a risk factor for toddler mortality.