A recent study published in JAMA Network Open highlights the effectiveness of the Smart Messaging Tool, an AI-driven natural language processing system that improves the triage of patient portal messages. By assigning clinical topics and urgency labels before messages reach clinicians, the tool significantly enhances the management of electronic communications. In a two-year analysis involving over 3 million messages from more than a million patients at the Southern California Permanente Medical Group, the new system achieved an 81% classification accuracy, a marked improvement from the previous system’s 44%. The median response time for high-acuity messages decreased dramatically from 22 hours to just over 5 hours, facilitating swifter clinician intervention. While the tool efficiently prioritizes urgent concerns, it currently faces limitations due to EHR constraints limiting topic detection to one per message. Future enhancements are being considered, including the potential for multi-topic handling to ensure comprehensive patient care.
Source link
Share
Read more