22 hours to 5: AI reorders the inbox at Southern California Permanente Medical Group

Advertisement

Pasadena-based Southern California Permanente Medical Group has cut median response times for high-acuity patient portal messages from 22 hours to five with AI.

The medical group, with 8,000-plus physicians serving more than 4.9 million members, rolled out the Smart Messaging Tool in 2023, automatically flagging urgent patient queries by moving them to the top of providers’ inboxes.

“From a professional satisfaction and joy-in-medicine and burnout perspective, this tool has been wonderful,” said Khang Nguyen, MD, chief medical officer for care navigation at the Permanente Federation and chief healthcare information officer at Southern California Permanente Medical Group, part of Oakland, Calif.-based Kaiser Permanente. “Physicians are like, ‘Hey, I’m actually catching this at the right time,’ versus telling a patient, ‘I’m so sorry we missed your window for this treatment.'”

Industrywide, providers’ inboxes started getting inundated with portal messages during the pandemic’s shift to virtual care, with responses taking up to three days. So Southern California Permanente Medical Group physicians, not finding a solution on the market that fit their value-based care model, built the Smart Messaging Tool.

Besides the 77% improvement in median response times, the natural language processing-based application has classified messages at an accuracy rate of 81%, nearly double the previous legacy system, according to a study of over 3 million messages from 1 million-plus patients published Jan. 15 in JAMA Network Open.

In the past, the medical group would rely on members picking the category for the message, receiving them in chronological order. Urgency often got buried in free text; a message labeled “medication refill” might have included a symptom of chest pain. The solution has also reduced cognitive load for care teams monitoring inboxes.

Even though the medical group’s portal has a disclaimer for when people should seek emergency care, members might not always understand the severity of a given symptom.

“I’ll give you a good example: A patient who says, ‘I’m having weakness and I can’t move my arm.’ They may not know that they have a very short window to get to the ER,” Dr. Nguyen told Becker’s. “So this system pulls that to the top. We get to them quick.”

Every message still gets answered — they’re just now prioritized by acuity. The medical group continuously monitors the tool to detect drift and clinically recalibrate as needed.

“If we’re looking for financial ROI, I would actually say we haven’t found one,” Dr. Nguyen said. “But the long-term ROI is basically patient quality of life, salvaging that clinical situation, saving someone from having a major stroke.”

At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.

Register to Attend Webinar

AI literacy & clinical expertise: Keeping human judgment sharp as AI scales

Wednesday, August 12
11:00 AM - 12:00 PM CDT

Presenters: Lisa Ivanjack, MD, MHCM, FACP, ProvidenceJohn-Paul Mead, MD, Centralus HealthWilliam Gustin, MD, Moab Regional HospitalAmanda Heidemann, MD, FAAFP, FAMIA, Wolters Kluwer HealthYaw Fellin, Wolters Kluwer Health

Advertisement

Next Up in Artificial Intelligence

Advertisement