Intelligent evaluation system for anticoagulant and antiplatelet therapy in surgical patients

Preoperative evaluation of bleeding and thromboembolic risks in patients receiving anticoagulant or antiplatelet therapy remains a clinical challenge. Decisions regarding drug discontinuation, withdrawal timing, and resumption must carefully balance the risks of hemorrhage and thromboembolism. We developed an innovative, integrated intelligent evaluation system to assist clinicians in minimizing intraoperative bleeding, postoperative ischemic stroke, myocardial infarction, and diabetic ketoacidosis.

The system's functionality was enhanced through the following: (1) Integration with hospital information systems and national cloud-based medical record retrieval for currently used medications; (2) Identification of high-risk thromboembolism patients using ICD-10 coding; (3) Bleeding risk assessment based on the 2018 European Heart Rhythm Association Practical Guidelines; (4) Automated anticoagulant discontinuation recommendations by evaluating bleeding and embolic risks; (5) Calculation of optimal withdrawal timing using patients' creatinine clearance and body weight; (6) Automated prompts for appropriate therapy resumption.

Between January 2018 and September 2021, five patients experienced postoperative ischemic stroke or myocardial infarction within 14 days postoperatively. Following the implementation of the intelligent evaluation system in October 2021, no such adverse events were reported through March 2025. Additionally, the combined time spent on evaluation and instruction was reduced by 3.16 min compared with the previous process.

This cross-disciplinary decision-support platform enhances perioperative safety through comprehensive drug data retrieval and real-time pharmacological calculations for patients receiving anticoagulants, antiplatelets, and SGLT2 inhibitors. By offering structured, automated guidance on drug management, the system fosters consensus among physicians, surgeons, and pharmacists while aligning with clinical workflow needs.