Prediction Model of Prolonged and Obstructed Labor in East Nusa Tenggara: A Multivariate Adaptive Regression Splines Analysis

Abstract

Prolonged and obstructed labor are the type of abnormal labor that may lead to maternal and fetal mortality. This study established the model for predicting prolonged and obstructed labor in East Nusa Tenggara. A health facilities-based case-control study was conducted in November 2017 among 570 women who gave birth at public health facilities in East Nusa Tenggara. Data were obtained by reviewing antenatal records, the mother's card, and partographs. In bivariate analysis, all variables with a p-value less than 0.25 determined by chi-square for categorical and independent t-test for numerical variables were included in multivariate analysis. Multivariate Adaptive Regression Splines (MARS) analysis was used to establish the final prediction model. The present study found that women <22, >26, and >34, with Hb levels of <12.5 gr%, and had nulliparity or multiparity (4 times) were reported as a higher risk of prolonged and obstructed labor. Meanwhile women with a fundal height of <34 cm, a height of >156 cm and >149 cm, a history of normal labor, presentation of the fetus behind the head, gestational weight gain of <12.3 kg, and pre-pregnancy BMI of <28.9 kg/m2 were identified as factors decreasing the risk of prolonged and obstructed labor. In conclusion, significant predictors of the outcome were maternal characteristics (age, parity, height, and history of labor method), maternal nutrition status (BMI pre-pregnancy, gestational weight gain, and hemoglobin levels), and fetal status (fundal height and fetal presentation).