Application of Survival Analysis of Pregnancy women and related post-delivery using Logistic Model Approach.

Abstract

Maternal mortality has been described as one of the foremost and most neglected health problems and human rights abused in the world. It is defined as the probability or percentage of mothers dying as a result of pregnancy related complications or post-delivery complications every year. About 830 women die from pregnancy or child related complications around the world. This research project is aimed at monitoring the delivery process performance and studying the factors that influence its outcome. A total number of two thousand nine hundred and ninety six (2,996) cases out of which eleven thousand and eighty two (11,082) cases were delivered at Caesarean and Spontaneous Vaginal Delivery sections were used in this research. The Cumulative Sum (CUSUM) control chart scheme is used for monitoring while the logistic regression model is fitted to determine the factors contributing to maternal death using Mother Child Hospital, Akure as a case study. Backward variable elimination procedure was used to generate a reduced form of the model that contains only the significant variables at a respectable alpha-value, 0.05. The result from the CUSUM chart shows that all the plotted points are within the control limit and this suggests that the delivery processes have improved over time. The result from the binary logistic regression shows that out of nine factors that were investigated in this research, parity, period of labour (measured in hours), previous caesarean section and length of baby are statistically significant and might influence the delivery process of a pregnant woman. The relationship between the selected risk factors and maternal survival status was then carried out through interpretation of odds ratios from logistic regression.