Conversational AI/Big Language Model Can Accurately Diagnose and Triage Health Conditions, Without Introducing Racial and Ethnic Biases – Internal Medicine

Conversational AI/Big Language Model Can Accurately Diagnose and Triage Health Conditions, Without Introducing Racial and Ethnic Biases – Internal Medicine

the findings

Conversational AI (GPT-4) has the ability to diagnose and triage health conditions similar to those provided by board-certified physicians, and its performance does not vary by patient’s race and ethnicity.

background

While GPT-4, a conversational AI, “learns” from online information, the accuracy of this type of AI in diagnosing and triaging has not been investigated, and whether the AI’s recommendations include racial and ethnic biases that may have been drawn from those. the information. Even as the use of technology in healthcare settings has grown in recent years.

road

The researchers compared how GPT-4 and three board-certified physicians diagnosed and triaged health conditions using 45 typical clinical vignettes to determine how each provided the most likely diagnosis and determine which of the levels of triage—emergency, non-emergency, or self-care—was most appropriate.

Study some limitations. The clinical vignettes, although based on real-world cases, provided summary information for diagnosis, which may not reflect clinical practice that typically gives patients more detailed information. In addition, GPT-4 responses may depend on how queries are phrased and GPT-4 may have learned from the clinical vignettes used in this study. Also, the results may not be applicable to other conversational AI systems.

impact

Health systems can use the results to deliver conversational AI to efficiently improve patient diagnosis and triage.

comment

“The results of our study should be reassuring to patients, because they indicate that large language models such as GPT-4 show promise in providing accurate medical diagnosis without introducing racial and ethnic biases,” said lead researcher Dr. Yusuke Tsugawa, assistant professor of medicine. in the Department of General Internal Medicine and Health Services Research at the David Geffen School of Medicine at UCLA. “However, it is also important for us to constantly monitor the performance and potential biases of these models because they may change over time depending on the information provided to them.”

Authors

Additional study authors are Naoki Ito, Sakina Kadomatsu, Mineto Fujisawa, Kiyomitsu Fukaguchi, Ryo Ishizawa, Naoki Kanda, Daisuke Kasugai, Mikio Nakajima, and Tadahiro Goto.

magazine

The study was published in the peer-reviewed journal JMIR Medical Education.

You may also like...

Leave a Reply

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