Predictive Modeling of Surgical Outcomes in Minimally Invasive Spine Surgery

Abstract

Background: Minimally invasive spine surgery (MISS) is widely acknowledged for its benefits, including reduced recovery time and minimized complications. However, predicting surgical outcomes accurately remains a critical challenge in clinical practice. Objective: This study aimed to develop and validate a predictive model for assessing postoperative outcomes in minimally invasive spine surgery, focusing on complications, recovery times, and overall success rates. Methods: A total of 128 patients who underwent MISS at the Department of Neurosurgery, Evercare Hospital Chattogram, between January 2023 and June 2024, were included in the study. Data such as patient demographics, preoperative health status, surgical approach, and intraoperative complications were collected. Predictive models, including logistic regression, support vector machines (SVM), and decision trees, were trained to predict outcomes. Statistical analyses were conducted using p-values, standard deviation (SD), and accuracy, sensitivity, and specificity metrics. Results: The model showed an accuracy of 85.3%, with a sensitivity of 82.5% and specificity of 87.9%. The p-value for the overall model accuracy was <0.01, indicating strong statistical significance. The standard deviation for postoperative complications was 6.3%, and the mean recovery time was 15.2 days with an SD of 3.4 days. The p-value for recovery time prediction was 0.03, indicating statistical significance in predicting postoperative recovery. Among the patients, 13% experienced complications, and 87% had successful outcomes. The model predicted complications with 94% accuracy, enhancing clinical decision-making. Conclusion: The developed predictive model offers reliable insights into postoperative outcomes in MISS, demonstrating its potential for improving surgical planning and patient management.