If artificial intelligence in health care offers greener pastures, for many physicians, it’s because there’s a whole lot less paperwork there.
Physician burnout has been one of the central issues for health care institutions in New Jersey and elsewhere, and administrative tasks are — in survey after survey — often what’s blamed for it.
Emerging AI platforms are promising the potential to reduce some of that bureaucratic busy-work for medical professionals. At Garden State health systems, there’s a plethora of examples of those systems already being used.
###
Across its mostly South Jersey footprint, Cooper University Health Care is all putting all the eyes it can on its patients. And it has AI to thank for it.
Snehal Gandhi, who is chief medical information officer and medical director of hospital medicine for Cooper, spoke highly of their growing collaboration with the AI-powered Nuance Dragon Ambient eXperience solution.
The health system has, since early 2021, been expanding to more of their physicians the opportunity to use this platform offered by tech firm Nuance Communications. The system captures physician-patient conversations, transcribes them and converts them into a concise note formatted for health records that the physician can edit and sign off on.
“It’s allowing physicians to talk naturally, sitting one-on-one, in the way patient visits used to go,” Gandhi said. “What’s happened with (health records) is, instead of facing the patient and making eye contact, you wound up interacting with your keyboard and mouse as you typed away during a patient encounter.
“If you were the type of physician to ensure you made eye contact so the patient felt they were paid attention to, you wound up writing notes at the end of the day. That’s when what we call ‘pajama time’ comes in. You’d be sitting in the office or at home staying up several hours in the evening finishing notes after a full panel of patients.”
The use of this Nuance system has allowed Cooper’s medical professionals across more than 20 specialties to cut down on documentation time by upwards of 50%, Gandhi said.
When it comes to what health systems can do to reduce feelings of burnout, Gandhi said cutting down on “pajama time” by half goes a long way. That’s why it’s expanding it for use in the inpatient setting in the health system.
“The feedback (from staff) has been great, especially for those struggling to keep up with documentation and notes,” he said. “They love it. They’re not going home and spending hours completing documentation. They’re spending time with family outside of work, doing the things they like to do to recharge for the next day.”
That’s something physicians really need, he added. By extension, so is more AI automation.
“It’s really the wave of the future,” Gandhi said. “And it has so much potential to ease that (burnout) trend.”
###
Craig Limoli, founder and CEO of Wellsheet, is sure of that, too.
His Newark-based firm offers a predictive clinical workflow platform that boasts a 44% reduction in the amount of time physicians spend dealing with electronic health records, or EHRs, per patient. The implications that could have for physician burnout is highlighted all over their marketing materials.
Limoli described Wellsheet as harnessing the advancements in machine learning abilities to optimize the presentation of EHR data. The platform predicts what would be the most relevant information for physicians, partly based on a medical specialist’s own preferences. It aims to help inform assessment plans with those insights, while also better contextualizing a patient’s different risk factors based on available data.
When you look at innovation in the health care space in the past decade, Limoli said, a large majority of those solutions weren’t directly tied to the day-to-day interaction physicians have with clinical data.
“So, it’s a unique position for us to be embedded in the clinical workflow and really be part of (physicians’) daily processes around interacting with clinical data to reach treatment decisions and conduct their delivery of care,” he said.
One of the company’s clients is RWJBarnabas Health. The platform has been deployed to seven of that system’s hospitals, with thousands of physicians across those hospitals adopting it into workflows on a daily basis, according to Limoli.
“Those physicians have used Wellsheet as a primary tool for accessing patient information in a more physician-friendly way,” he said. “It also improved care coordination when a consult was completed and there’s a hand-off from one physician to another.”
Helping break down barriers in communication across health systems has shown utility in reducing burnout as well. A seamless overnight shift change, in which everyone is properly informed of a patient’s care plan, leads to less frustration, and less staff having to repeat themselves.
For all the reasons that might exist to adopt a solution that eases stress for medical professionals, those professionals still take convincing.
“When creating new technology and trying to have them adopted into workflows … it’s not an easy task to gain the respect and engagement of physicians,” Limoli said.
When Wellsheet was first testing its system out — starting with RWJBarnabas Health’s Newark Beth Israel Medical Center — it had its team on site several days a week to work side-by-side with physicians to see it in use and get feedback for later iterations of it.
Limoli said it takes a “level of humility” that the company was willing to embrace to invest the time and effort required to encourage actual adoption by physicians.
Skepticism is something every tech firm touting a new AI tool will run up against, leaders in the health care space suggest. Physicians, of course, want tools to make their lives easier. But they also have to trust those tools before implementing them into their daily work.
“That can take years to get right,” Limoli said.
###
Some local health systems have assembled tech experts, ethics consultants and clinician teams to implement their own in-house AI-powered solutions.
Kash Patel, executive vice president, chief digital information officer, at Hackensack Meridian Health, said the organization has been promoting ideas within a committee structure that evaluates potential use cases of these technologies.
Out of that has come models that predict at an earlier date when a clinician might need to alert a chronically ill patient’s family to the potential need for palliative care plans. It’s also analyzing the use of AI predictors of advanced kidney disease and asthma.
Between those innovations and a platform it’s looking to implement — similarly to other health systems — that condenses clinician notes and automates lab order referrals to cut down on administrative time, Patel sees a lot of potential to alleviate some of the pressure that family medicine physicians in particular are under due to high patient loads.
But Patel echoed others in saying that there’s still a lot of work left to be done in nudging and incentivizing physicians to use high-tech approaches after years of practicing medicine in the same way.
Also of importance to Hackensack’s team is that further use of these AI-enabled tools is kept in check as “fair, safe and equitable,” Patel said. Among other things, that means not having these predictive systems picking on particular cohorts, or exposing patient data.
Assuming that’s all well accounted for, Patel is thrilled with the potential this technology holds.
“We expect in the next year to see exponential growth in it,” he said. “It’s an exciting time, for sure.”
Examining voices
Traditional stethoscopes can help detect conditions within a few heartbeats or deep breaths. It’s a standby for a reason, said Sandra Elliott, chief innovation and commercialization officer at Hackensack Meridian Health.
Always on the lookout for tomorrow’s medical standby, Elliott’s organization has made a big investment into an insight that has the potential to turn the human voice into a similarly important vital sign — one that can also be an indicator of health-related outcomes.
Last month, Hackensack Meridian Health announced that, in partnership with its Bear’s Den innovation accelerator program, it would be investing in artificial intelligence technology company Canary Speech. That startup has a patented technology that monitors and assesses “digital biomarkers” in the human voice.
“The idea is that, when you’re standing up straight with your shoulders back, your voice is going to sound different in subtle ways than when you’re slumped over,” Elliott said. “When you get to that sub-auditory level, you can analyze waveforms and understand how those subtle changes are tied to clinical situations.”
Those subtle markers in the voice hold potential to be used to make insights into the severity of disorders such as anxiety and depression.
“What would be phenomenal, and I’m oversimplifying here, is if as you could have a child ahead of a pediatric visit speak into a mobile application that could flag that individual for more depression- and anxiety-oriented questions,” Elliott said. “If you can identify whether someone is at high risk for depression, it opens up more potential for an intervention before a crisis occurs.”
Using this voice technology to find clues that might deepen understanding of a patient’s mental health status is an easy-to-scale application that Elliott expects might be seen as soon as this year, pending a FDA approval. Canary Speech is also researching how this technology might be leveraged in the future for clinical evaluations of Alzheimer’s disease, Parkinson’s disease and other conditions.
With this and other AI-driven solutions, Elliott said there have to be physician considerations: They need to be comfortable with the due diligence and clinical research components.
And, she said, that goes for patients, too.
“They need to be comfortable that AI is being used in a transparent fashion, and not as a health care version of Big Brother,” she said.
