Study Finds AI Can Help Patients Streamline Their Questions to Doctors

A new study has found that a custom-trained large language model “produced follow-up questions with similar clarity and conciseness, and higher utility, than actual follow-up questions written by members of [a] care team.” Vanderbilt University Medical Center's website has the news.

Compared to the custom model, called CLAIR, the GPT-4 model from OpenAI produced follow-up questions that were “more complete but less clear.”

A tool that could prompt patients to “clarify their portal messages” before hitting send has the promise to “streamline communication and improve care efficiency.” This new study was meant specifically to explore how large language models could “help patients craft more effective messages to their healthcare providers through patient portals.”

The authors of the study noted that “auto-prompting better patient questions could prove more advantageous than strategies resting on auto-generated clinician responses. The team next plans to evaluate the clinical impact of implementing this AI-guided patient messaging.”