Innovation in primary care
Primary care teams share lessons learned from testing AI-powered patient messaging
Background: Primary care clinicians are spending increasing amounts of time responding to patient messages via electronic portals, and this task is contributing to burnout. Some health systems are experimenting with using large-scale language models (LLMs) to generate draft responses to patient messages.
Innovation: At West Virginia University, the authors tested an artificial intelligence tool called Augmented Response Technology (ART) that generates a draft response as soon as a message from a patient arrives. The nurse reviews each draft and decides whether to send, edit, or forward it to the physician. Early versions of ART produced less helpful responses that focused on triage rather than directly answering the patient. The team improved the tool by overhauling prompt input, grouping messages into categories by type (results reviews, medication refills, paperwork, common symptoms), and adding a symptom severity library that’s perfect for tailoring responses to enforce tone, safety, and content delivery.
Implications: The authors conclude that tools like ART have the potential to support patient messaging, but that it requires careful and adaptive prompt design. Different specialties, types of messages, and patient populations require different approaches.
Lessons learned from the front lines of AI-augmented patient messaging
Joseph E. Capito, MD, et al.
West Virginia University Department of Family Medicine (Morgantown, West Virginia)
Pre-embargo link (temporary)
/Open to the public. This material from the original organization/author may be of a contemporary nature and has been edited for clarity, style, and length. Mirage.News does not take any institutional position or position, and all views, positions, and conclusions expressed herein are those of the authors alone. Read the full text here.
