Assessing ChatGPT-4s Accuracy in Detecting Surgical Site Infections After Elective Colorectal Surgery Using Electronic Health Records

What this Trial Is About

This research aims to evaluate how well ChatGPT-4 can detect surgical site infections SSI at three anatomical levels using electronic health records from patients who have had colorectal surgery. Surgical site infections are a common healthcare-associated infection that affects patient health and healthcare systems. Current surveillance methods are manual and time-consuming, and this study explores using artificial intelligence to improve detection accuracy. The study compares two methods of SSI detection for patients undergoing elective colorectal surgery the standard manual surveillance performed by infection control teams and an automated approach using OpenAIs ChatGPT-4 chatbot trained with CDC criteria. The comparison is retrospective, using data from a nationwide infection surveillance program as the gold standard. Participants surgical records and infection status will be analyzed to measure the rate of surgical site infections within 30 days after surgery. The study involves reviewing clinical course texts, microbiology reports, and diagnostic codes to assess AIs accuracy. The evaluation focuses on how well ChatGPT-4 identifies SSIs compared to manual methods, aiming to inform future surveillance improvements.

ChatGPT-4 for Surgical Site Infection Detection From Electronic Health Records After Colorectal Surgery.