AI Assisted Preoperative Assessment for ASA Classification and ICU Admission

N/AActive Not RecruitingNCT07783334

Bakirkoy Dr. Sadi Konuk Research and Training Hospital2,500 enrolled

Overview

This prospective observational study evaluates the performance of artificial intelligence (AI) models in preoperative anesthesia assessment. Preoperative clinical data from adult patients undergoing elective surgery are independently evaluated by clinicians and AI models (ChatGPT and Gemini). The study compares their assessments of American Society of Anesthesiologists (ASA) physical status classification and the predicted need for intensive care unit (ICU) admission within the first 24 hours after surgery. Actual postoperative ICU admission is used as the clinical outcome for evaluating predictive performance. No treatment or clinical decision is determined by the AI models, and patient management is performed according to routine clinical practice.

Study Type

OBSERVATIONAL

Enrollment

2,500

Conditions

Preoperative Anesthesia Assesment

Interventions

AI-assisted preoperative assesmentOTHER

Preoperative clinical data are independently evaluated using ChatGPT and Gemini for ASA physical status classification and prediction of ICU admission within 24 hours after surgery. AI-generated assessments are used for research purposes only and do not influence clinical decision-making or patient care.

Eligibility

Sex: ALLMin age: 18 YearsHealthy volunteers:

Medical Language ↔ Plain English

Inclusion Criteria: * Age 18 years or older * Scheduled for elective surgery * Evaluated in the preoperative anesthesia clinic * Provision of informed consent Exclusion Criteria: * Pregnancy * Presence of an upper respiratory tract infection * Active herpes infection or active wound and/or lesion in the anesthesia-related area * Age under 18 years * Emergency surgery * Planned cardiovascular surgery * Incomplete clinical data

Outcomes

Primary Outcomes

Agreement in ASA Physical Status Classification

Agreement between clinician-assigned and AI-generated ASA Physical Status classifications will be evaluated for ChatGPT and Gemini using linear weighted kappa statistics.

Time frame: During preoperative assessment

Secondary Outcomes

Prediction of Postoperative ICU Admission

The accuracy of preoperative predictions of postoperative ICU admission made by the clinician, ChatGPT, and Gemini will be evaluated against actual ICU admission occurring within the first 24 hours after surgery. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy will be calculated for each evaluator.

Time frame: Within the first 24 hours after surgery