Enterprise-Level Emphasis Operational Analytics: Predicting Healthcare Encounter Resolution Success Using Machine Learning
Rajender Radharam
Abstract
Enterprise-Level Emphasis represents a critical operational framework within healthcare administration, designed to ensure accurate claims processing and encounter data management by systematically identifying and resolving discrepancies. This study analyzed ELE’S operational performance using descriptive statistics and machine learning regression techniques to predict settlement success rates. Performance data from 20 observations revealed robust operational metrics: average claim processing time of 12.17 hours, error rates of 4.16%, employee performance scores of 79.8%, system uptime of 97.61%, and settlement success rates of 91.53%. Correlation analysis showed strong negative relationships between processing time, error rates, and operational outcomes, while there were positive relationships between employee performance, system uptime, and settlement success. Two ensemble learning models—AdaBoost regression and XGBoost regression—were developed to predict settlement success rates. AdaBoost achieved a training R² of 0.9993 and a testing R² of 0.9548, demonstrating reasonable generalization despite moderate overfitting. XGBoost showed severe overfitting with perfect training performance (R² = 1.0000) but poor testing results (R² = 0.8907), which reduced the performance of AdaBoost on unobserved data. This analysis confirms the operational performance of ELE, while highlighting the importance of model regularization in predictive analyses. The high generalization of AdaBoost makes ELE more suitable for predicting resolution outcomes, although hyperparameter optimization could improve the practical applicability of the two models in healthcare operational management.