Cost Drivers of Cervical Disc Arthroplasty: A Multivariate and Machine Learning Analysis (2016-2021)

Introduction: Cervical disc arthroplasty (CDA) has become an increasingly preferred alternative to anterior cervical discectomy and fusion due to its motion-preserving benefits and lower risk of adjacent segment disease. However, understanding how patient demographics, hospital characteristics, and regional factors affect the associated charges remains crucial for cost management.

Methods: A retrospective analysis of the National Inpatient Sample (NIS) from 2016 to 2021 was conducted. Patient demographics, hospital size, ownership type, and regional factors were analyzed for their impact on CDA charges using multivariate linear regression and machine learning models (random forest and gradient boosting trees). The significance level was set at 0.003 after Bonferroni correction.

Results: Data from 4,212 CDA cases were analyzed. Significant predictors of increased charges included large hospitals and private ownership (p <.001). The Western U.S. reported higher charges compared to other regions (p <.001). The gradient boosting trees model achieved the highest predictive accuracy (AUC=86.2%). Length of stay and regional wage index were significant contributors to increased charges (p <.001).

Conclusion : Hospital size, ownership type, geographic location, and patient-specific factors significantly influence the costs associated with CDA. Machine learning models effectively predicted these charge outcomes and could be used to guide future economic decision-making in spine surgery, aiming to reduce costs while maintaining high-quality care.