Classification of Hypertension in Pregnant Women Using Multinomial Logistic Regression

Abstract

Maternal Mortality Rate (MMR) is still a crucial problem in Indonesia and other developing countries; one of the causes is Hypertension in Pregnancy (HDK). This study aims to classify hypertension in pregnant women based on the factors that influence it, with a case study of patients from the Obstetrics and Gynecology Specialist Clinic at the Regional General Hospital (RSUD) Dr. R. Sosodoro Djatikoesoemo Bojonegoro. The variables used were age, gravidity, gestational age, obesity, history of abortion, hypertension, and diabetes mellitus. The research method used in this study is multinomial logistic regression because it uses four categories of dependent variables, namely pregnant women without hypertension, pregnant women with chronic hypertension, pregnant women with gestational hypertension, and pregnant women with preeclampsia. The results obtained in this study were from 3 categories of hypertension in pregnant women; the influencing factors were obesity, gestational age > 36 weeks, having a history of hypertension, and diabetes mellitus, with the resulting model classification accuracy value of 79.6%, which means the classification is classified as good. This research contributes to applying statistical methods in the health sector and as a mitigation effort to help minimize the number (prevalence) of maternal deaths, especially those caused by hypertension.