Maternal and Infant Health in the United States: An Investigation of Pregnancy Outcomes

Abstract

While infant mortality rates have stabilized in recent years, maternal mortality rates have been steadily increasing since the 1980s. Due to the difficultly of directly studying these mortality rates, this thesis focuses on three clinical outcomes that have been shown to be strongly correlated: cesarean sections, preterm births, and low birthweight births. A step-up approach was used to develop logistic regression models for these clinical outcomes and to determine the most significant patient demographics. K-means clustering was then used to determine which hospitals in the Sutter Health network have the most similar relationships between patient demographics and clinical outcomes. Race/ethnicity, payment type, and hospital location classification were all found to have consistently strong impacts on the risks of the clinical outcomes of interest. The cluster analysis revealed that outcomes worsen at hospitals further away from the center of mass of the Sutter Health hospital network.