ChatGPT Finding a Role in Hospital Diagnostics


A recent study in AJIC demonstrates AI’s potential in identifying HAIs, emphasising the need for precise language in AI diagnostics to enhance infection surveillance in healthcare settings cost-effectively.

Recent findings from a proof-of-concept study, published in the American Journal of Infection Control (AJIC), shed light on the potential role of artificial intelligence (AI) in the detection of healthcare-associated infections (HAIs), even in complex clinical scenarios. The study “Assisting the Infection Preventionist: Use of Artificial Intelligence for Healthcare-Associated Infection Surveillance” underscores the significance of precise and uniform language in the deployment of AI technologies for diagnostic purposes and suggests the feasibility of integrating AI into infection surveillance systems, offering a more cost-efficient approach.

A ChatGPT generated image of itself fighting HAIs

The persistent challenge of HAIs in the healthcare sector, particularly in acute care settings, underscores the necessity for advanced diagnostic and surveillance solutions. With approximately 687,000 HAIs reported in U.S. acute care hospitals and 72,000 resulting fatalities in 2015, according to the Centers for Disease Control and Prevention, the development of effective, scalable, and efficient tools for HAI detection is critical. Given that about 3% of hospital patients are affected by an HAI at any time, the implementation of robust infection prevention and surveillance programs is paramount, albeit resource-intensive.

This context highlights the potential value of AI in enhancing HAI surveillance capabilities, particularly in resource-limited environments. The study conducted by researchers from Saint Louis University and the University of Louisville School of Medicine evaluated the efficacy of two AI-powered tools, one based on OpenAI’s ChatGPT Plus and the other on the open-source large language model Mixtral 8x7B, in identifying central line-associated bloodstream infections (CLABSI) and catheter-associated urinary tract infections (CAUTI) through the analysis of hypothetical patient scenarios.

The AI tools demonstrated notable accuracy in identifying HAIs when provided with clear, comprehensive prompts, highlighting the potential for AI to augment current diagnostic practices. However, the research also illuminated the critical need for detailed, unambiguous information input to ensure accurate AI performance, indicating areas for further development in AI tool training and data management strategies.

For diagnostic manufacturers and researchers, these findings underscore the importance of continuing to refine AI technologies for healthcare applications. The integration of AI into diagnostic and surveillance systems could not only enhance the accuracy and efficiency of HAI detection but also reduce the overall resource burden on healthcare facilities. Additionally, the study suggests the necessity for ongoing collaboration between AI developers, healthcare professionals, and diagnostic manufacturers to ensure that AI tools are effectively tailored to the complex needs of HAI surveillance and prevention.

The use of retrieval augmented generation techniques, which leverage extensive knowledge repositories for context, as demonstrated in the study, points to a promising avenue for improving AI diagnostic tools. Such advancements could pave the way for the development of AI-assisted diagnostic products that provide significant value to infection preventionists and healthcare providers.

The AJIC study presents a compelling case for the integration of AI in the fight against HAIs, offering diagnostic manufacturers and researchers a foundation upon which to innovate and develop next-generation diagnostic tools. As the healthcare industry continues to evolve, the role of AI in enhancing diagnostic accuracy and efficiency will undoubtedly be a critical area of focus and development.

“Assisting the Infection Preventionist: Use of Artificial Intelligence for Healthcare-Associated Infection Surveillance,” by Timothy L. Wiemken and Ruth M. Carrico, was published online in AJIC on March 14, 2024. Available at: https://doi.org/10.1016/j.ajic.2024.02.007

Staff Writer

Our in-house science writing team has prepared this content specifically for Lab Horizons

Leave a Reply

Your email address will not be published. Required fields are marked *