Chicago-based Northwestern Medicine just expanded ambient artificial intelligence to clinicians across its health system, but its CIO is already looking at what comes next.
The 11-hospital system said Aug. 15 it was offering the DAX Copilot, an app that listens to medical appointments and uses AI to draft clinical notes, to all of its primary care physicians treating adult patients. Northwestern Medicine CIO Doug King told Becker's that other providers are already pining for the automated documentation tool.
"There are specialties and the ED where the product is not as mature," Mr. King said. "That's one of the asks of our physicians: They want to try it, but it's not ready yet."
Northwestern Medicine hopes to expand the AI assistant to orthopedics, obstetrics and oncology next and eventually the emergency department. The health system's goal is for half of its approximately 3,000 physicians to adopt the tool, which is not mandatory.
Beyond that, Mr. King said he can also foresee a day where the technology queues orders or automates charging for providers, or generates visit notes for patients in terminology they will understand.
With the DAX Copilot, developed by Microsoft subsidiary Nuance, upon getting patient consent a clinician presses record on a smartphone app that ambiently listens to and transcribes the medical visit then quickly generates a draft clinical note for the Epic EHR once the appointment is completed.
"It understands the context of the conversation with the physicians and the patients to where it can omit things — if we're talking about, for example, my child's softball game, it won't include that — but it will add a med to the med list, or problem to the problem list," Mr. King explained. "It's not just another technology that's transcribing the conversation. It holds it within the context of medicine."
During a roughly four-month pilot, clinicians who used the app reduced their physical documentation load by about 30%, Mr. King said.
"Patients are being seen faster by physicians when they need to get in for care, our physicians are able to document faster, and they have less of that [documentation] burden on their shoulder, so they get to do what they want to do, which is see those patients," he said.
Organically, as a result, those physicians have also been able to see more patients, which would be a quantifiable return on investment (Mr. King said the financial terms of the deal are confidential). "But ultimately, the reason we did this was for the wellness of our physicians and combating administrative burden," he said.
Farther out into the future, Mr. King envisions generative AI becoming a daily tool for providers across healthcare, from creating reports for radiologists after reading images to conversing with and giving advice to members of the care team, furthering the goal of precision medicine.
"The doctor will say, 'What are you thinking?' And the AI will respond back, 'I would prescribe this, this or this,'" he explained. "And the doctor might say, 'Well, what about this?' And the AI would respond back with, 'Well, this patient's genome might not metabolize it that well.' And it will just be able to pull that data instantly to help that physician make a better decision for that unique patient."