Generative artificial intelligence for surgical site infection surveillance

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Infection Prevention and Control

Background:

• Surgical site infection (SSI) surveillance is resource-intensive.
• Generative artificial intelligence (GenAI) offers potential for improving SSI detection and classification efficiency.

Purpose of the Study:

• To investigate the efficacy of GenAI in augmenting the detection and classification of SSIs in spinal surgery patients.
• To compare GenAI-assisted SSI review with traditional methods used by infection prevention (IP) nurses.

Main Methods:

• A case-control study involving patients who underwent spine surgery, comparing SSI cases with non-SSI controls.
• GenAI analyzed clinical data to identify potential SSIs, with IP nurses utilizing GenAI output for case review.
• Classification of SSIs included superficial, deep, and organ space infections.

Main Results:

• GenAI achieved 100% sensitivity in identifying all SSIs (16/16) and 93.7% specificity (15/16).
• Discordance in infection classification occurred in 37.5% of cases, with GenAI improving accuracy upon physician review.
• GenAI-assisted SSI reviews had a median completion time of 9 minutes, indicating enhanced efficiency.

Conclusions:

• GenAI demonstrates significant promise as a tool to enhance surgical site infection surveillance in spinal surgery.
• The use of GenAI can lead to improved efficiency in infection prevention processes.