ipopba - stock.adobe.com
Providence Health's ACI test shows burnout improvements
Nuance's DAX ambient clinical intelligence software meaningfully cuts clinical documentation time and streamlines clinician workflows, according to a controlled trial at the health system.
Providence Health's first randomized controlled trial of ambient clinical intelligence shows the technology may help curb burnout, with clinicians reclaiming an average of 2.5 hours each week from after-hours documentation.
"We didn't roll this out as a productivity thing. We rolled it out as a focus on clinician burnout," said Scott Smitherman, associate vice president and chief medical information officer of Providence Clinical Network.
Burnout is no secret in the healthcare industry, with study after study showing the growing burden on providers and the need to address the issue. According to a Mayo Clinic study, nearly half of U.S. physicians are feeling strained.
At Providence, researchers used Epic signal data and identified more than 100 family medicine providers across seven states who were struggling with documentation burden. Of those, 24 were then randomized to use either Nuance's Dragon Ambient eXperience Copilot — an ACI tool — or continue with standard documentation practices. The study measured changes in documentation time, self-reported burnout scores and productivity.
Compared with those using standard documentation, DAX Copilot users spent 29 fewer minutes per day on after-hours charting, culminating in 2.5 hours per week with burnout scores decreasing by 30.3%.
Providers also saw marked gains in patient connection. Before Copilot, nearly 1 in 4 reported struggling to engage with patients; after adoption, that number fell to 4.2%. Physicians reported that the technology allowed them to look patients in the eye, and capture richer narratives in the note while lowering the mental burden of remembering details hours later.
According to Smitherman, the impact has been profound for providers using ACI. "I've had doctors tell me, 'I was thinking about quitting medicine, but actually, I'll work for a couple more years because this has made it more fulfilling,'" he said.
Weaving into existing workflows
Smitherman says a key factor in the success of this technology was tight integration into Providence's Epic EHR via Haiku — an app their physicians already use. After recording in the app, the workflow is familiar for Epic users, using a dot phrase shortcut that inserts the AI-generated text into the clinical note.
Smitherman emphasizes the decision not to fund the technology by requiring higher patient volumes was important to smooth adoption. "Whether you're asking [clinicians] to pay the fee or whether you're asking them to virtually pay for it by seeing more patients… we did neither," he said.
Although Providence did not require higher patient volumes, clinicians in the intervention group saw an average increase of 1.7 patients per day, a change the study authors believe was due to reduced backlog and smoother workflows.
Overcoming challenges
Although there appears to be some clear benefits, ACI doesn't solve every clinical documentation problem at Providence. Some clinicians were dissatisfied with the format and flow of the generated notes, requesting more customization to better reflect their personal style.
Specialty and inpatient settings also pose a challenge. Smitherman says these areas have been slower to ramp up utilization due to difficulties fitting the tool into more complex workflows. Nursing documentation also remains a work in progress.
"We just launched our inpatient pilot, and while there's overlap with outpatient workflows, inpatient and nursing documentation are highly structured and not what the AI is currently designed for. Companies are still figuring out the right fit, but the opportunity to help nurses is clear," he said.
Providence also emphasizes ongoing evaluation of AI safety and accuracy. "We're very careful to look for places where it's safe to use AI… higher risk technologies need a lot of scrutiny," Smitherman said.
What's next for Providence and DAX
Long-term, Providence envisions AI not only drafting notes but curating charts to surface relevant labs, imaging and guidelines during visits. "All of this would have been sci-fi five years ago — and is absolutely happening right now," Smitherman said.
Providence leaders say broad, committed rollout has been essential to gain buy-in. "This is not just a flavor of the week. This is something we're making a commitment to," said Staci Wendt, director of the Providence health research accelerator and a study author. "Hearing that from the organization is really important in terms of people wanting to make that investment."
Smitherman agrees: "Offer it to all of your clinicians. This is transformative… it's absolutely clear that this is the future."
Elizabeth Stricker, BSN, RN, comes from a nursing and healthcare leadership background, and covers health technology and leadership trends for B2B audiences.
