Johns Hopkins researchers: Time is running out to leverage AI for patients beyond a certain age

Three holdups to broad AI adoption continue to vex healthcare. If these drags aren’t solved soon, older Americans won’t benefit much by the 1,000+ medical devices that, as of 2025, are both equipped with AI and cleared by the FDA. 

The three problems are misaligned reimbursement strategies, fragmented data infrastructures and widening rural workforce gaps, researchers at Johns Hopkins and other institutions attest in a paper published Aug. 20 in Health Affairs. 

What’s more, over the next decade, “demographic pressures will collide with a tighter federal budget and a shrinking clinical workforce, and the opportunity afforded by AI to address geriatric and rural health care will rapidly diminish.” 

The paper’s first author is Esther Oh, MD, PhD, of Hopkins. Last author is Phillip Phan, PhD, also of Hopkins. Here are three key excerpts.  

1. The most pressing need is a predictable Medicare reimbursement pathway in which AI-based technologies are not an add-on cost but rather a catalyst for productivity. 

So far, CMS has provided only sporadic coverage for AI-based products, Oh and co-authors point out. “Overemphasizing imaging tools in a reimbursement portfolio may discourage investments in other viable and necessary clinical tools, such as decision-support and remote-monitoring platforms,” they add. 

Their proposal: Develop time-limited supplemental payments that convert to bundled value-based models upon proven real-world effectiveness. This approach would “align capital incentives with patient outcomes instead of procedural volume.” 

‘This creates an additional pathway to revenues, hence investment options for those investors looking to diversify their portfolios in this domain.’

2. Attempts to earmark federal infrastructure bills for high-speed internet deployment in rural areas have largely fallen short. 

Meanwhile, “the promise of AI and telecommunications technologies to democratize healthcare for rural populations remains unfulfilled,” the researchers write. “Policy reform, including reimbursement standards changes, can encourage market participants to invest in technologies that, for example, increase access to specialist care in underserved areas.” 

‘Delivering healthcare to dispersed communities is the greatest challenge resulting from the chronically short-staffed workforce serving older adults and their caretakers.’

3. Older adults consistently prioritize face-to-face encounters and community involvement over technical novelty. 

Community advisory boards, bias audits and participatory model co-design can “position AI-based technologies within a social contract rather than a simple regulatory checklist,” Oh and colleagues write. “Prior research in rural telehealth has demonstrated that, when local voices steer deployment, adoption rates rise by 30% and satisfaction scores by up to 40%.” 

‘Thus, as the use of AI-based technologies grows, their perceived level of trustworthiness is dependent on transparency and the ability to co-govern their application.’

If developers and regulators want older adults to adopt AI technologies, “they need to think proactively about the stage-of-life needs, co-morbidities, and functional and mental capacity of the older adult demographic,” Oh and fellow co-authors state. 

In mnemonic format, they note, this means focusing on the “4Ms of an age-friendly health system”—What Matters, Medication, Mentation and Mobility.

Read the rest. 

• In other research news: 
   
  • University of Virginia: AI transforming mental health support for breast cancer patients
     
  • Georgia State University: From Alzheimer’s to AI: How the TReNDS Center is advancing brain research
     
  • University of Freiburg: AI analysis prompts international reassessment of multiple sclerosis progression