Physicians’ use of health care augmented intelligence (AI) is growing rapidly, according to AMA research (PDF), but a recent analysis of survey data has found that a gap exists in AI use between doctors who are in private practice and those working as employed physicians.
Use of health care AI—often called artificial intelligence—stands at 72% for employed physicians and 64% for those in private practice, according to AMA survey findings.
“One issue is always cost,” said Margaret Lozovatsky, MD, AMA’s vice president of digital health innovations.
“With employed physicians, often the organization will take in some of the cost, and that may be why you're seeing higher numbers in that group,” Dr. Lozovatsky explained. “But these tools are now becoming more and more available as the cost is going down—so that difference may become smaller and smaller over time.”
She added that, among employed physicians, the organization often brings the tools to them. Whereas, in private practice, physicians must do their own product and vendor research, and their own contracting.
“A lot of these steps may be lengthy and burdensome in a private practice situation, but are handled by organizations for employed physicians,” Dr. Lozovatsky said.
Private practice physicians, however, having done their own homework and due diligence, may be more committed to the AI tools they choose.
“To a certain extent, there is absolutely more autonomy in a private practice in general for use of any tools and for making decisions on things like scheduling tools and such,” Dr. Lozovatsky said, but she added that cost can also be a factor with employed physicians.
“The tools that we're talking about have been so well received by most clinicians, that even if an organization gives them an option, we're seeing a lot more opt in than opt out,” she said. “And especially for these tools, because they're pricey, most organizations aren't imposing them on people—they don’t want to pay for something that's not being utilized.”
For the survey, private practice physicians included those in solo practice plus single-specialty and multispecialty group practices. Employed physicians included those in faculty practices, hospitals, ambulatory surgery centers, urgent-care facilities, HMOs and managed-care organizations, medical schools and virtual care companies.
Improved efficiency drives adoption
Keshni Ramnanan, MD, chair-elect of the AMA Private Practice Physicians Section (AMA-PPPS), is the chief medical officer of Oak Medical, a Wisconsin-based, private multispecialty group providing post-acute care in nursing homes and skilled-nursing and assisted-living facilities.
“Our group has an EHR that does a lot of this AI stuff but employed physicians are usually going to be using Epic or Cerner,” which—Dr. Ramnanan explained—are loaded with more AI tools and are out of the price range for her group of 30 physicians and 60 nurse practitioners.
“So, of course, there's an element of a cost,” she said. “Cost is not a problem to them because they don't have to worry about it, but—when it comes to us—all those features end up adding up to more money.”
That said, Dr. Ramnanan said the AI tools her group does use have improved efficiency and workflow that has resulted in some operational savings.
Her group uses AI tools for scheduling, billing, processing faxes, facility admissions and other “back-office things.”
“Once somebody gets admitted to a facility, they automatically get scheduled according to particular algorithm that we created,” Dr. Ramnanan explained.
On the clinical side, AI tools provide some clinical-decision support and help to create visit notes from dictation and by pulling relevant information from patient records.
How AI tools are used
AMA survey respondents were asked to declare whether they use one or more of 15 AI functions.
Documentation of billing codes, medical charts or visit notes was the most-used AI application among both groups, as 22% of private practice physicians said they use AI for this purpose as did 21% of employed doctors.
Twenty percent of employed physicians said they used AI for the creation of discharge instructions, care plans or progress notes, as did 19% of private practice doctors.
The biggest gap between the two groups was in the use of patient-facing chatbots for customer-service functions, with 12% of employed physicians using AI for this, compared with only 8% for private practice doctors.
A three-percentage point difference between the two groups was reported for the following AI functions:
- Translation services: 16% for employed; 13% for private practice.
- Summaries of medical research and standards of care: 15% employed; 12% private practice.
- Assistive diagnosis: 14% employed; 11% private practice.
Generation of draft responses to patient portal messages—an AI function seen as a tool for fighting burnout—was only used by 10% of employed physicians and 9% of those in private practice.
That finding is “really, really interesting,” Dr. Lozovatsky said.
Stay up to date on AI
Follow the latest news on AI and its applications and effects for health care—delivered to your inbox.
“Those tools were thought to have a lot of promise because the top thing physicians feel frustrated with usually is the volume of in-basket messages,” she noted. “I'm hearing this from many organizations—that the physicians struggled with the fact that the responses weren't exactly what they wanted—and so the amount of time it took to edit a response that was developed actually often took longer than it did to write it from scratch.”
Dr. Lozovatsky added, however, that the portal-response AI function is being widely adopted by medical assistants and nurses who have “really have loved them in a way that the physicians haven't.”
“So, perhaps, those tools will continue to play a role—but in a different user population,” she said.
The AMA STEPS Forward® “Governance for Augmented Intelligence” toolkit, developed in collaboration with Manatt Health, offers a comprehensive eight-step guide for physicians and health care organizations to establish a governance framework to implement, manage and scale AI solutions.
Why do physicians adopt AI
In a survey section asking what attributes lead physicians to adopt AI, 57% of private practice physicians and 55% of employed doctors agreed with the statement: “It is the standard of care.”
“That is very significant, because I think that number is going to go up exponentially because it is becoming the standard of care now—and, in certain specialties, it has been the standard of care for many years,” Dr. Lozovatsky said.
In radiology, for example, if there is not an algorithm looking for incidental image findings, if something is missed, physicians will be asked “Why did you not have this algorithm?” she said.
Similarly, there are algorithms to detect or predict deterioration of patient conditions.
“If these things are missed, the question will be, ‘Why didn't you have that in place?’” Dr. Lozovatsky said. “There's some very astute physicians that took the survey that recognized that.”
Dr. Ramnanan noted that the survey found 50% of private practice physicians believed AI could be helpful in value-based care arrangements, compared with only 46% of employed physicians.
“It's interesting that value-based care is a little bit higher for private practice, which is something that matters to us more than the employee model,” she said. “If we can show value, we’re better off in our ACOs [accountable care organizations] or when it comes to insurance plans.”
While the overall 72%–64% gap in AI use seems big, most of the survey questions had only small differences between private practice and employed physician responses.
“It’s really interesting that the differences weren’t that significant between the two groups,” Dr. Lozovatsky said. “We worry a lot about tech equity—and it’s not to say we shouldn’t continue to follow this—but these responses tell me that the private practice environments are staying on top of these innovations, which is exciting, and I think we should call that out.”
Learn more with the AMA about the emerging landscape of health care AI. Also, explore how to apply AI to transform health care with the “AMA ChangeMedEd® Artificial Intelligence in Health Care Series.”
