Leveraging Big Data Analytics to Identify and Address Health Insurance Enrollment Disparities Among Vulnerable Populations: A Machine Learning Framework

Abstract

Despite ongoing healthcare reforms and Medicaid expansion, significant disparities persist in health insurance enrollment among underserved populations. This article presents a big data–driven framework that identifies and addresses enrollment gaps by integrating multiple datasets—healthcare claims, U.S. Census demographics, and social determinants of health (SDOH)—to produce actionable insights at granular geographic levels. Using unsupervised machine learning techniques, including k-means and hierarchical clustering, the framework uncovers hidden patterns of under-enrollment across ZIP codes in Medicaid expansion states. Factors such as limited digital access, language barriers, and low health literacy demonstrate statistical correlations with reduced insurance uptake. The framework employs predictive modeling to forecast communities with the highest risk of continued under-enrollment based on historical and demographic trends. Data-informed interventions are proposed, including multilingual outreach programs, mobile enrollment units, and culturally competent assistance initiatives, with potential impact evaluated through simulation models and ROI forecasting under policy-adjusted scenarios. Built on scalable, privacy-preserving architecture incorporating de-identification standards and role-based access controls, the framework integrates with cloud-native platforms such as Databricks and AWS for real-time data processing and visualization. This work demonstrates how AI and big data analytics can drive policy innovation, resource optimization, and health equity, offering public health officials, Medicaid administrators, and data strategists’ evidence-based solutions for improving healthcare access across socioeconomically disadvantaged populations.