Abstract:
Introduction: Maternal death is one of the basic problems of women on their health during their
reproductive ages (15-49). Worldwide, 99% of deaths of women in their reproductive ages are
due to childbirth and pregnancy complications. In a world Ethiopia was one of the countries
affected by this problem.
Objectives: This study has been aimed to analyze the spatial pattern of maternal death in
Ethiopia and identify the factors that affect the maternal death.
Methods: This study was conducted in Ethiopia and the data was basically secondary which is
obtained from 2016 Ethiopian Demographic and Health survey (EDHS). The Bayesian Geoadditive regression model is used to identify the major risk factors and spatial effects (spatial
pattern) on maternal death in Ethiopia. The DIC model selection criterion is used for model
selection.
Results: A total of 10,009 women aged 15 to 49 were included in the study of which 1.43% died
due to childbirth or pregnancy related complications. Based on the DIC, Bayesian Geo-additive
regression model suits data over the two usual generalized linear regression models fitted in this
study (Bayesian generalized linear model and semi parametric regression model).Based on the
results obtained using Bayesian Geo-additive regression model, place of delivery, number of
antenatal care visit, marital status, wealth index and continuous covariate (age of mother and
number of birth order) significantly determines maternal death. Based on the evidences of
spatial variation in a model, higher risk of maternal death is found in Afar, Somali, Benishangul
gumuz and Gambela regions.
Conclusions: The results of this study suggested that there are complex social, demographic and
geographic processes operating in maternal mortality. This result can be more clearly
understood using the appropriate statistical models. There have been geographical differences in
patterns of maternal death.
