Predicting Maternal Mortality Rate Using Data Mining Techniques: The Case Of Jimma University Specialized Hospital Maternity Wards.

Abstract

Maternal mortality is the death of women during pregnancy, childbirth or in the 42 days after delivery remains a major challenge to health systems as a worldwide. Recent reports from WHO and UNAIDS indicate that the number pregnant woman died after delivery is increasing from time to time. This number is dramatically increasing in sub Saharan African countries including Ethiopia. The main objective of this study is to develop a predictive model for the maternal mortality status. The overall activity of this study is guided by a Hybrid-DM processes model and used the data of Jimma University specialized Hospital maternity ward. The study has used 4218 instances, seven predicting and one outcome variables to run the experiments. The mining algorithms; J48 decision tree, Naïve Bayes and PART rule induction are used in all experiments due to their popularity in recent related works. Ten-fold cross validation and 70/30 split criteria test option were used to train and test the classifier models. J48 decision tree algorithms were better performance with 98.74 % accuracy running on 10 fold cross validation test option with default parameter using 14 attribute than any experimentation done in this study. The selected attributes is significant for maternal mortality rate status which is after delivery has been identified. Those are Mothers BP, Address, APGAR score, Diagnosis, Mothers Age, Length of stay, Indication and Condition on Discharge. A promising result is observed in applying data mining techniques to build predictive model for maternal mortality using socio-demographic, clinical and biological features. This study is proved that the prediction of maternal mortality can be applicable with help of data mining application in the maternity ward data and predicting the life status of the mothers after delivery had been identified. This study did not include pregnant women life expectancy, Therefore, for the future work developing a model which could predict the life expectancy of pregnant women after labor and delivery would need further study beside the developed model as well as the improvement of J48 models.