AI and patient care are “top of mind” for healthcare executives in 2024. The pairing seems opportune, since the surveyed leaders see the burgeoning technology as a key tool for improving the perennial mission.

This is one takeaway to be inferred from the latest annual “Top of Mind for Top Health Systems” survey conducted for the Center for Connected Medicine at UPMC in Pittsburgh.

Produced in partnership with KLAS Research, the survey report shows 55 respondents participated. Some 93% of these were C-level or suite-adjacent executives. The remaining 7% were directors or managers. 

The report offers three observations with explanations. Here’s a summary.

1. AI continues to excite healthcare leaders.

For the fourth year in a row, healthcare executives identified AI as the most exciting emerging technology for their sector, the authors note. AI was cited by 85% of respondents, suggesting widespread belief that “AI has significant potential to improve administration, operations, clinical care and other areas of healthcare.” More:  

‘While healthcare leaders have concerns about AI security and governance, they recognize the potential for natural language processing, large language models, generative AI and other applications to significantly aid their organizations in realizing positive outcomes for patients, providers and operations. As AI becomes more accessible, reliable, and cost-effective, leaders are trying to identify AI use cases that could transform the industry.’

2. Health system leaders recognize recent improvements in AI models.

Respondents named AI the technology showing the most improvement for the second year in a row, UPMC reports. Executives noted improvement in large language models and generative AI used to augment care delivery and efficiency. More:

‘Though ambient speech and other documentation solutions have been available for several years, the propagation of generative AI in healthcare has led to more organizations leveraging the technology to try to decrease clinician burnout, decrease cost, automate administrative tasks, and increase patient satisfaction and engagement.’

3. Leaders believe that some common challenges—not least patient care, patient access and provider burnout—are best solved with technology.

Ambient speech technology is one solution being used to address these kinds of problems, the authors point out. While many reported challenges have remained the same over the last year, they add, many organizations have “shifted or split their focus to operational challenges, including efficiency, profit margins and staffing.” More:  

‘Another challenge, data aggregation and analytics, involves addressing the problem of siloed data that is not easily available in digestible formats for clinical care. This frustration is likely amplified due to increased desire to implement AI solutions, which first require a strong data aggregation/analytics base.’

The authors comment that the acceleration of AI into the public consciousness following ChatGPT’s splashy introduction in late 2022 “is a very real and potentially game-changing technology for many industries”—including healthcare.

The report at hand is titled “Leaders Hoping to Expand AI Use Cases in 2024-25.” Download it and related resources from UPMC’s Center for Connected Medicine here.

Carle Health chooses Nabla's ambient AI for it's focus on workflow - For a deeper look into Nabla's deployment at Carle Health, watch this insightful interview with Dr. David Lovinger and Ed Lee, MD, MPH, hosted by John Lynn.

Buzzworthy developments of the past few days.

• To each health system its own ambient AI notetaker. Ascension Saint Thomas in middle Tennessee likes Suki for clinical documentation duties. Ochsner Health is partial to DeepScribe. Kaiser Permanente favors Abridge. And at Ohio State University’s Wexner Medical Center, a physician tells why he and colleagues are making a go of it with Microsoft DAX. “We’re making more eye contact, and our body posture shifts away from the computer and faces the patient, says Harrison Jackson, MD. “We’re able to pick up on more nonverbal cues.” The anecdotal rundown of AI scribal preferences and experiences is from MedCity News. Read the rest.
   
• And here’s a doctor for whom ambient AI has been nothing short of a joy-bringer. “For me, applying AI has been magical,” writes Alicia Jacobs, MD, a longtime practitioner of family medicine and academic physician leader at UVM Medical Center in Vermont. Using ambient AI to auto-generate clinical notes, she asserts, “we can help save the primary-care workforce.” Save it from what? Burnout, attrition, career boltings and other bad things. A local outlet published the piece Aug. 25.
   
• She’ll get no argument from this physician/attorney. The possibilities opened by ambient AI are “expansive,” writes Sai Balasubramanian, MD, JD, in Forbes. “From virtual assistants to data-driven insight generation capabilities, ambient AI is a growing field in the world of artificial intelligence more broadly, and has significant potential to transform healthcare for patients and providers alike.”
   
• A free online symptom checker has made an eye-popping difference in population health. Its purveyor says so, anyway. Tokyo-based Ubie estimates its platform has contributed patient quality-adjusted life years by around 27,000 years while creating approximately $1 billion in economic value. And that’s just over a three-year period ending in 2024. The findings are from Japan only, and the company acknowledges that the approximate impact “should be viewed as a reference point” in light of “the evolving nature of impact measurement.”
   
• What are Lyapunov functions and why are they here? They’re mathematical tools that help ensure stability in dynamic systems, and they could be game changers for healthcare AI. So explain research physician L. Joseph Parker, MD, and anesthesiologist Neil Anand, MD, in an intellectually challenging piece at KevinMD. Intelligence—whether natural or artificial—“isn’t just crunching numbers,” the duo write. “[I]t’s making sense of incomplete data, inferring from the unknown and understanding context.” NASA missions rely on precise calculations, they remind, “and that same need for stability and precision is crucial in healthcare AI.” Which is where Lyapunov functions can make a difference. How? Read and learn.
   
• The Department of Energy is concerned about AI’s mammoth appetite for, well, energy. The agency’s advisory board is worried enough that it wants DOE to build a “data-center-scale AI testbed.” With this the agency would study ways to drain less juice from the grid. The board makes the recommendation in a report recently sent to Energy Secretary Jennifer Granholm. Axios concisely conveys the gist.
   
• A Canadian startup creatively named Noze has invented a medical breathalyzer powered by AI. The device, DiagNoze, is said to identify diseases in much the same way the one used by police nails people for drunk driving. Slovenia-based City magazine says the technology “could radically improve access to health services, especially in underprivileged regions where rapid diagnostics are crucial.”
   
• Professional technology tinkerers have used AI to realistically depict the face of Jesus based on the image in the Shroud of Turin. I find the portrait oddly affecting even though it kind of calls to mind the always brilliant if frequently hung-over Eagles up to their Hotel California period. See what you think.
   
• Recent research in the news:
   
  • Mayo Clinic: 3D body volume scanner uses AI to help predict metabolic syndrome risk
     
  • Stanford: ChatGPT shows promise in answering patients’ questions to urologists
     
  • University of Waterloo: Building new bones with help from 3D printing
     
• AI funding news of note:
   
  • Buzzy mental health startup Slingshot AI just grabbed $30M from Andreessen Horowitz (aka a16z)  
     
• From AIin.Healthcare’s news partners:
   
  • Cardiovascular Business: Better than BMI? 3D body scanner uses AI to measure metabolic health
     
  • Health Imaging: Nearly half of FDA cleared AI medical devices have not been validated on patient data
     
  • Radiology Business: AI draws conclusions from interventional radiology adverse events data, helping docs design interventions