Clinical Performance of a Machine Learning-Based Artificial Intelligence System Compared With Anesthesiologist Assessment in Preoperative Patient Evaluation
Clinical Performance of a Machine Learning-Based Artificial Intelligence System Compared With Anesthesiologist Assessment in Preoperative Patient Evaluation
Preoperative evaluation is essential for identifying patient-related risks before elective surgery and for planning safe anesthesia management. Traditionally, this evaluation is performed by anesthesiologists based on clinical history, physical examination, comorbidities, and laboratory findings.
This observational study aims to compare the clinical performance of a machine learning-based artificial intelligence system with anesthesiologist assessment during preoperative patient evaluation. The artificial intelligence system independently analyzes patient data and generates risk assessments, which are then compared with evaluations performed by anesthesiologists.
The primary objective of the study is to assess the level of agreement between the artificial intelligence system and anesthesiologists in preoperative risk assessment. Secondary objectives include evaluating the accuracy and consistency of the artificial intelligence system and exploring its potential role as a decision-support tool in preoperative clinical practice.
The findings of this study may contribute to understanding the potential benefits and limitations of artificial intelligence-assisted decision making in preoperative evaluation
Preoperative evaluation is a critical component of perioperative care, aimed at identifying patient-specific risks, optimizing patient safety, and guiding anesthetic planning prior to elective surgical procedures. This process traditionally relies on the clinical judgment of anesthesiologists, who integrate medical history, physical examination findings, comorbid conditions, and relevant laboratory data to assess perioperative risk.
Recent advances in artificial intelligence and machine learning have enabled the development of clinical decision-support systems capable of analyzing complex clinical data and generating predictive risk assessments. Despite increasing interest in these technologies, their clinical performance and reliability in real-world preoperative settings remain insufficiently evaluated.
This observational study is designed to compare preoperative risk assessments generated by a machine learning-based artificial intelligence system with routine anesthesiologist-led evaluations. Adult patients scheduled for elective surgery will undergo standard preoperative assessment performed by anesthesiologists as part of usual clinical care. Independently, anonymized patient data will be processed by the artificial intelligence system to produce preoperative risk assessments. The artificial intelligence output will not be available to clinicians and will not influence patient management.
The primary outcome of the study is the level of agreement between the artificial intelligence system and anesthesiologists in preoperative risk stratification. Secondary outcomes include the consistency, concordance, and overall performance of artificial intelligence-generated assessments compared with clinician evaluations.
This study involves no interventions and does not alter standard patient care. All anesthetic and perioperative management decisions will remain entirely under the responsibility of the treating anesthesiologist. By systematically comparing artificial intelligence-based assessments with clinician evaluations, this study aims to clarify the potential role, strengths, and limitations of artificial intelligence as a supportive tool in routine preoperative evaluation
Inclusion Criteria:
Exclusion Criteria: