Digital assistant uses AI to ease medical documentation at Sutter
Sutter Health is pilot testing a voice-enabled digital assistant that makes use of artificial intelligence to see if it increases clinician efficiency.
The Sacramento, Calif.-based delivery system is testing the device with a group of doctors in Northern California. The device uses a combination of voice commands from physicians and the context in which they are operating to create a clinically accurate note that is pushed to an electronic health record.
The device is made by Redwood City, Calif.-based Suki, which conducted multiple pilots in 2018 in a variety of practices, including primary care, dermatology and orthopedic. That testing demonstrated a 70 percent reduction in the time physicians spend on medical notes, the company contends. As Suki is used over time, it can distill a physician’s conversation with a patient into an actionable plan of care.
“Maximizing the amount of time clinicians spend with patients while reducing the documentation burden on our clinicians is a strategic and tactical priority,” says Howard Landa, MD, vice president of clinical informatics and EHR at Sutter. “Personalized care paired with digital assistant tools will enhance care delivery and have a positive impact on health outcomes for our consumers, which is what really matters.”
For example, a physician can tell Suki, “I did my typical diabetes counseling” for a patient, and Suki knows how to create relevant content for the note, and the resulting note is tuned not only to the doctor’s medical specialty but also to their own vocabulary and style, which will support streamlined documentation in the patient’s EHR, helping to create the most appropriate care plan.
For now, Suki remains in the early adopter stage, working with three different electronic health records systems that are used in seven medical specialties and handling about 1,000 interactions weekly.
Sutter Health is starting to build a data layer on top of physician notes that will further reduce documentation burdens and introduce new use cases, such as clinical decision support, according to Landa.