Determinants of maternal mortality in districts Jhang and Khanewal

Abstract

According to the WHO calculations, approximately 295000 women died globally in 2017 from maternal health causes, with which 94% of all maternal deaths occur in low and lower-middle-income countries. India, Nigeria, and Pakistan are at the top of the maternal and neonatal deaths list. The study determines the risk factors associated with maternal and neonatal health during pregnancy, which helps provide a better understanding to generate preventive measures to reduce the high rates of maternal mortality. For this purpose,  data from the recently established maternal, newborn, and child health program in two Districts in Pakistan was used for the analysis. We used multinomial logistic regression to model the risk factors associated with maternal and neonatal health during pregnancy, such as swelling of feet, Anemia, fetal movements, vaginal bleeding, and white vaginal discharge in the last three months of pregnancy. Categorical Principal Component Analysis (CATPCA) was used to determine the optimal combinations that best explain the non-normalcy of the observed data. Results from CATPCA showed that 17 parameters in the model explained 96.6% variation of the data. Compared to the multinomial logit model, CATPCA provides a more transparent and precise picture of the risk factor associated with the health of pregnant women. 

Keywords: Maternal health, Newborn and Child Health Program, Maternal mortality, Multinomial Logistic Regression, Categorical Principal Component Analysis.