Determinants of Antenatal Care Visits in Indonesia with Synthetic Minority Over-Sampling Techniques for Imbalance Data

Abstract

Maternal mortality and infant mortality are two indicators that describe the degree of public health as well as indicators of sustainable development in Indonesia. The acceleration of reduction these two indicators must be supported by antenatal care services since pregnancy for the safety of mothers and babies. Based on the results of the IDHS 2017, antenatal care coverage in Indonesia (77.4%) is still far from the target in 2024 (95%.) This study used logistic regression analysis with Synthetic Minority Oversampling Technique (SMOTE) resampling method because of imbalance data to explore the determinants of complete antenatal care visits in Indonesia and descriptive analysis to find out an overview of complete antenatal care associated with factors that are considered influencing it. Data that was used in this study is the Indonesia Demographic and Health Survey (IDHS) 2017 with unit of analysis for women of childbearing age who are married or live together and gave birth to their last child in the period 2012-2017. The logistic regression results of the SMOTE method show that the variables of mother's education, husband's work status, knowledge of pregnancy danger signs, distance to health facilities, timing first antenatal check, mother's age, economic status, birth order, and number of problems during pregnancy significantly affect the completeness of antenatal care visits. The policy recommendations in this study are expected to be adopted by government to increase antenatal care visits in Indonesia as an effort to reduce maternal and infant mortality.