Socio-Economic Determinants of Skilled Birth Attendant Utilization in South Asia: A Machine Learning Approach

On a global scale, the mortality rates among mothers and children represent a significant concern within the realm of public health. Competent birth attendants have demonstrated a substantial impact on decreasing maternal and newborn mortality rates. This research aimed to clarify the socioeconomic determinants that impact the use of skilled birth attendance in maternal health while also recognizing the diverse factors that influence this utilization throughout South Asian countries. This study included 196,234 women aged 15 to 49. The study employed machine learning algorithms to identify the socioeconomic factors most strongly associated with skilled birth attendance in South Asian countries. Our findings indicate that approximately 83.0% of childbearing women received skilled birth attendance during delivery. The Random Forest model was recognized as the most effective machine learning method, with an accuracy of 94.71% and an AUC of 94% after thorough testing and assessment. The main socioeconomic determinants discovered were site of delivery, wealth status, maternal education, maternal age, age at first childbirth, age at first sexual intercourse, birth order, type of solid fuel, source of drinking water, sanitation facilities, and media exposure. The ML model predicted SBA and its topmost associated factors well. The RF classifier could more accurately distinguish between skilled and unskilled birth attendants. This RF is the most preferred classifier for predicting SBA.