Sensitivity and specificity of postoperative antibiotic treatment-based algorithm for detecting potential surgical site infections in adult lumbar spine surgery

INTRODUCTION: Surveillance of surgical site infections (SSIs) is a key component of prevention. Semi-automated approaches have emerged to reduce the workload of full chart reviews. This study evaluated the sensitivity, specificity, positive and negative predictive values (PPV, NPV) of a postoperative antibiotic therapy (POAT-SSI) semi-automated algorithm in lumbar spine surgery. METHOD: All patients who underwent lumbar spine surgery in our institution under the French National SSI surveillance program (FNP-SSI) during 2023-2024 were included. Complete chart review served as the gold standard and was compared with FNP-SSI algorithms and the POAT-SSI approach, identifying suspected cases. Diagnostic performance was assessed using sensitivity, specificity, predictive values, and ROC curve analysis. RESULTS: The POAT-SSI algorithm showed higher sensitivity (100% vs 62%) and similar specificity (96% vs 97%) compared to the FNP-SSI method, with both showing excellent NPV (≥99%). The POAT-SSI approach achieved a superior Youden index (0.96 vs 0.59) and a significantly better area under curves (0.978 vs.0.795, P = 0.046). PPV remained modest due to low SSI prevalence. CONCLUSION: The POAT-SSI-based algorithm showed superior diagnostic performance for SSI surveillance while reducing manual workload. These results support integrating such targeted digital strategies into routine infection control monitoring.