Predictors for extended hospital length of stay and 90-days reoperation in elective spinal stenosis surgery: a retrospective analysis using neural networks.
Abdulmajeed A Aljabr, Bader K Alqahtani, Abdullah M Alghamdi, Omar F Alsalem, Mohammed A Almanna, Nada T Alnefaie, Mohammed M Alshardan, Abdullah M Alshehri, Majed S Abalkail, Fahad H Alhelal, Sami I Aleissa
Abstract
Open AccessStudy design: A retrospective comparative study. Background: The length of hospital stay and the setting to which patients are discharged after surgery can vary, impacting both healthcare resource utilization and the patient's recovery process. Identifying the factors that affect these outcomes is essential for enhancing patient care and efficiently managing resources in elective spinal stenosis surgery. Objectives: This study aims to identify the predictors associated with extended hospital length of stay and 90-day reoperation in elective spinal stenosis surgeries, to optimize patients' postoperative outcomes. Methods: We conducted a retrospective study at King Abdulaziz Medical City, Riyadh, Saudi Arabia, from January 2016 to January 2024. Statistical analyses, including multivariate analyses and neural networks, were performed using SPSS. Results: A total of 389 patients were included, with a median length of stay of 6 days (interquartile range [IQR] = 3), with 113 patients (≥75th percentile) in the extended hospital length of stay group (n = 113; 29.3%). Age ≥60 years (P = 0.021), reoperation within 90 days (P = 0.027), body mass index (BMI) ≥35 (P = 0.001), diabetes (P = 0.004), and American Society of Anesthesiologists (ASA) classification (P = 0.019) were found to be significant predictors of extended hospital stay. In the multivariate analysis, BMI ≥35 (P = 0.002) and diabetes (P = 0.030) remained significant predictors of extended hospital length of stay. ASA classification (P = 0.001) and prolonged hospital stay (P = 0.027) were significant predictors of 90-day reoperation risk in the bivariate analysis; however, these associations were not retained in the multivariate analysis. Conclusion: Our study highlights the need for preoperative optimization of comorbidities, BMI, and overall health to improve postoperative outcomes and reduce costs in elective spinal stenosis surgeries.