Predictors of hospital readmission within 30 days after surgery for thoracolumbar fractures: A mixed approach

Background Readmission followed by surgery to treat spinal fractures has a substantial impact on patient care costs and reflects a hospital's quality standards. This article analyzes the factors associated with hospital readmission followed by surgery to treat spinal fractures. Methods This was a cross‐sectional study with time‐series analysis. For prediction analysis, we used Cox proportional hazards and machine‐learning models, using data from the Healthcare Cost and Utilization Project, Inpatient Database from Florida (USA). Results The sample comprised 215,999 patients, 8.8% of whom were readmitted within 30 days. The factors associated with a risk of readmission were male sex (1.1 [95% confidence interval 1.06–1.13]) and >60 years of age (1.74 [95% CI: 1.69–1.8]). Surgeons with a higher annual patient volume presented a lower risk of readmission (0.61 [95% CI: 0.59–0.63]) and hospitals with an annual volume >393 presented a lower risk (0.92 [95% CI: 0.89–0.95]). Conclusion Surgical procedures and other selected predictors and machine‐learning models can be used to reduce 30‐day readmissions after spinal surgery. Identification of patients at higher risk for readmission and complications is the first step to reducing unplanned readmissions.