Development and Internal Validation of Machine Learning Algorithms for Predicting Discharge Disposition Following Surgery for Metastatic Spine Tumors. (16th November 2020)

• Record Type: Journal Article
• Authors: Fatima, Nida; Shin, John H
• Is Part Of: Neurosurgery. Volume 67(2010)Supplement 1
• Publication Date: 2020-11-16
• Subjects: Nervous system -- Surgery -- Periodicals; 617.48005
• DOI: 10.1093/neuros/nyaa447_907
• Languages: English
• ISSNs: 0148-396X

Abstract

Abstract: INTRODUCTION: Prognostication of the discharge disposition in patients following surgery for metastatic spine tumor can improve the efficiency of the hospital management and quality of care. These factors can guide to implement targeted treatment to identify the potential barriers and lay out a groundwork to overcome them. METHODS: The patient cohort was identified from the American College of Surgeons, National Surgical Quality Improvement Program (2014-2016). We developed eight machine learning algorithms, and the algorithm with the best performance across discrimination, calibration and overall performance was used for predicting the discharge disposition. RESULTS: Statistical analysis included 3, 578 patients with discharge disposition to home following surgery for metastatic spine tumor in 71.3% (n = 2, 552) of the patients, while nonroutine discharge was observed in 28.7% (n = 1, 026) of the patients, which included skilled care (n = 366, 10.2%), rehabilitation (n = 525, 14.7%), acute care (n = 74, 2.1%), and others (n = 61, 5.9%). The model with 10-predictive factors which included: age, gender, functional status, disseminated cancer, any bleeding disorder, preoperative serum albumin, preoperative hematocrit, preoperative white blood cell count, preoperative serum alkaline phosphatase, and preoperative platelet count-performed well on the discrimination, calibration, Brier score and decision analysis to develop a machine learning algorithm. Logistic regression showed higher AUCs than least absolute shrinkage and selection operator across these different models. The predictive probability derived from the best model was uploaded on an open access web application which can be found at: https://spine.massgeneral.org/drupal/DischargeDisposition-MetastaticSpineTumor. CONCLUSION: Machine learning algorithms provide promising results for prediction of discharge disposition in spine surgery. Hence, these algorithms can provide useful factors for patient-counselling, accurate risk adjustment, and quality metrics.