Who Goes Without Care? Demographic, Social, and Economic Drivers of Cost-Related Unmet Medical Needs

Document Type

Poster Presentation

Publication Date

4-17-2026

Keywords

fsc2026

Abstract

Utilizing classification methods with machine learning applications, this study analyzes factors that potentially lead to unmet healthcare needs within the United States. Machine learning algorithms such as K-Nearest Neighbors (KNN), Recursive Partitioning (Rpart), and Conditional Inference Trees (Ctree) were used in the data analysis to identify patterns and make predictions based on the given information, to understand which groups of individuals may be at greater risk of going without needed care. The Well-Being and Basic Needs Survey (WBNS, 2023), conducted by Michael Karpman and Elaine Waxman at the Urban Institute and made publicly available through ICPSR at the University of Michigan, consisted of a sample size of 3,888 respondents after thorough data cleaning. Of the three models, the Recursive Partitioning (Rpart) method outperformed both KNN and Ctree, receiving the highest overall accuracy at 93.56% and a kappa of 0.775, with a sensitivity of 96.1% and specificity of 81.3%, meaning the model was accurate and provided correct classifications. Factors such as having to skip meals due to high costs, being unable to pay rent on time, being unable to pay for utilities, experiencing multiple hardships, food insecurity, and being under the poverty level are found as predictors in being in a position of not receiving necessary care. Overall, limitations and restrictions exist for citizens across the country who are unable to afford necessary medical care, as studies have shown. Further research and social policies must be implemented to continue the shrinking of this affordability gap.

Comments

Poster presented at the 2026 Fisher Showcase, St. John Fisher University, April 17, 2026.

This document is currently not available here.

Additional Files

Share

COinS