Using Generative Artificial Intelligence for Healthcare-Associated Infection Surveillance

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Infectious Disease Surveillance

Background:

• Generative artificial intelligence (GenAI) is rapidly advancing but has seen limited integration into medical applications.
• Healthcare-associated infection (HAI) surveillance relies heavily on analyzing extensive clinical documentation.
• GenAI's capability to process large volumes of text makes it a potential tool for improving HAI surveillance.

Purpose of the Study:

• To review the current evidence on applying GenAI for HAI surveillance.
• To identify necessary steps for integrating GenAI into clinical practice for HAI detection.
• To highlight the potential benefits and challenges of using GenAI in this domain.

Main Methods:

• Review of existing literature on GenAI applications in HAI surveillance.
• Analysis of small-scale studies involving simulated patients and retrospective patient records.
• Evaluation of GenAI's accuracy in identifying HAI criteria and making determinations compared to human experts.

Main Results:

• Preliminary studies indicate GenAI can accurately identify HAI criteria and make determinations, comparable to human experts.
• GenAI tools are under development to enhance HAI detection capabilities.
• Evidence suggests GenAI can reduce inter-facility variability and improve the efficiency of HAI reviews.

Conclusions:

• GenAI demonstrates potential to enhance HAI surveillance efficiency and accuracy.
• Widespread adoption requires further validation through prospective clinical studies to ensure safety and efficacy.
• Successful implementation could free up clinical resources for direct infection prevention efforts.