Abstract

Background: Postoperative infections drive excess healthcare costs. Predicting which patients are at risk of infection enables targeted prevention and earlier intervention strategies plus resource optimization.

Objectives: To quantify healthcare costs, (billed) hospital length of stay ((B)LOS), and the difference between financed length of stay (FLOS) minus BLOS in surgical patients with and without postoperative infections, and to evaluate whether PERISCOPE® AI predictions identify patients at risk for excess costs.

Design: Retrospective observational cohort study.

Setting: A large non-university teaching hospital in Belgium.

Methods: Surgical patients from five specialties (neurosurgery, orthopedic surgery, general surgery, urology, and vascular surgery) were analyzed. PERISCOPE® postoperative infection risk predictions (low, medium, high) were applied to all patients. Healthcare costs, length of stay, antibiotic costs, and FLOS-BLOS were compared between infected and non-infected patients and across risk categories.

Main outcome measures: Total healthcare costs (including antibiotics), length of stay, FLOS-BLOS difference, and cost distribution across PERISCOPE® risk categories.

Results: Among 31,735 surgical procedures across five specialties between 2022 and 2024, the 30-day postoperative infection rate was 7.3%. Infected patients were older, more often male, more likely to undergo emergency surgery, and incurred markedly higher healthcare costs (€36,196 vs €3,768), longer hospital stays (27.9 vs 4.0 days), and net-financial losses for the hospital. PERISCOPE® stratified patients into low- (0.7%), medium- (1.4%), and high-risk (12.4%) infection groups and demonstrated clear cost and resource gradients, effectively identifying infection risk and associated financial burden across specialties.

Conclusions: Postoperative infections impose a significant financial burden on healthcare systems. PERISCOPE® accurately identified high-risk surgical patients, enabling cost-efficiency through early intervention and prevention. Artificial intelligence (AI) guided early intervention in high-risk patients may reduce costs and may support value-based surgical care. Meanwhile, low-risk patients may benefit from less invasive monitoring and earlier discharge.