Estimation of the Cox Proportional Hazards Model Using the Most Important Variables Affecting the Risk of Neonatal Death in Sudan

Abstract

This paper aimed to identify the important factors that affected neonatal mortality and estimate Cox proportional hazards model with these factors. Data were collected from Omdurman Maternity Hospital, from the record of pregnant women from the first follow-up until the delivery and whether the neonate was alive or dead. The data focused on demographic variables (mother's age, number of previous delivery, city, number of neonate and the sex of neonate) and health variables (neonate weight and mode of delivery). Through log rank test the variables (age, previous delivers and weight of neonate) had a significant effect and the estimated single Cox proportional hazards models with these significant variables are significant and also the multiple Cox proportional hazards model is significant.