The Impact of Racial Disparities on Prenatal Care Adequacy: An Algorithmic Fairness Perspective

Prenatal care promotes a healthy pregnancy by reducing the risk of complications and optimizing the chances of a successful and safe delivery. Racial disparities in prenatal care access and quality have long been documented. With the increasing use of machine learning algorithms in healthcare decision-making, concerns have arisen regarding the potential exacerbation of disparities through algorithmic processes. This paper investigates the impact of racial disparities on algorithmic fairness in predicting the Kessner Index - a composite measure of factors relevant to prenatal care adequacy - utilizing classification algorithms. The study leverages data from the Pregnancy Risk Assessment Monitoring System (PRAMS) collected between 2016 and 2020. The proposed methodology involves feature selection, the implementation of multiple classification models, and an analysis of counterfactual fairness. Furthermore, the paper explores the trade-off between accuracy and fairness considerations. The findings of this research shed light on the presence of biases inherent in algorithmic decision-making and their implications for racial disparities in prenatal care, thereby contributing to the ongoing discourse surrounding equitable healthcare practices.