Introduction: Healthcare expenditures among patients with chronic spinal conditions and those undergoing surgery continue to grow at a notable rate. Identifying patients at risk for increased postoperative healthcare utilization is therefore critical for optimizing care outcomes and resource allocation. The aim of our study is to leverage a range of preoperative patient characteristics – including comprehensive medical, surgical, and social histories – to develop machine learning (ML) models capable of granular prediction of postoperative healthcare utilization among cervical spine fusion patients. Materials and Methods: A cohort of anterior cervical decompression and fusion and posterior cervical decompression and fusion surgeries was queried from a tertiary academic medical center from 2002 to 2022. Patient and operative characteristics were systematically extracted. Several different ML algorithms were employed and optimized to predict high healthcare utilizers using an aggregate 90-day index. SHAP feature importance values were computed for the top-performing model. Results: A total of 4480 cervical fusion surgeries were analyzed, of which 12% were identified as high health utilizers. All models outperformed the American Society of Anesthesia benchmark, with the Balanced Random Forest model demonstrating the best discriminative performance (Area under the curve: 0.772 ± 0.007). The top-three predictive features revealed through SHAP analysis were increased OR duration, 90-day preoperative neuromodulator usage, and 90-day preoperative opioid usage. Conclusion: This study demonstrates the successful creation of a prognostic ML model for the prediction of high healthcare utilization within 90 days of cervical spine surgery. These models, after external validation, have the potential to be instrumental aspects of a spine surgeon’s workflow.
Khazanchi et al. (Thu,) studied this question.