 | ● |  |
|
|
|
| | | News and views you ought to know about:- When CMS came out with its Aligned Networks initiative in June, the Trump Administration signaled its prioritization of patient ownership of patient data. That sounds only right, of course. But what might it come to mean for AI-equipped healthcare orgs that need troves of patient data to improve the health status of entire populations? The question arose this week at Newsweek’s Digital Health Forum. Sharing HIPAA-covered data among and between organizations for HIPAA-covered purposes is “tricky,” panelists in one discussion agreed. “Currently there is a restriction on what payers and providers can share with one another,” reporter Alexis Kayser writes. “If we do, indeed, move to view health records as a community resource, both parties need to agree on a governance model that both parties trust.”
- If that doesn’t happen, “asymmetrical warfare” between each entity’s AI models could result in a net negative for the industry, warned Aneesh Chopra, who served as the first U.S. chief technology officer under President Barack Obama. “I actually worry that if we don’t resolve this, we’re going to see AI as an inflationary good, not a deflationary good in the overall health care ecosystem.”
- China, Russia and Iran don’t have any guardrails around AI. They just don’t care about such things. Asked to comment on this observation in front of a subcommittee of the House Oversight and Government Reform Committee, an expert witness accepted the handoff and ran with it. “Yeah. Well, you said it. China doesn’t have that,” said Kinsey Fabrizio, president of the Consumer Technology Association. “It’s impossible for our member companies. We have 80% small businesses, and they cannot compete when there are a thousand different potential laws that they have to comply with,” added Fabrizio, clearly a skilled lobbyist. “It just stifles innovation completely. For us to win the AI race, we need to remove that barrier.”
- The questioner prompting the response was South Carolina Republican Nancy Mace, chair of the Subcommittee on Cybersecurity, Information Technology and Government Innovation.
- Asked by Mace to comment on the knowledge that China has “stolen a bunch of our data,” Samuel Hammond, chief economist at the Foundation for American Innovation, replied: “Of course, and they also steal [intellectual property]. And so, really it comes down to hardware and energy. China added over 400 gigawatts to their grid last year. They’re about to do the same thing this year.” How much has the U.S. added? “Approximately zero.”
- More takeaways here, hearing video here.
- Your local hospital should have a process in place for employees to confidentially report risks to patients—or to anyone—involving AI. If it doesn’t, it may run afoul of the Joint Commission. The heads-up is from new guidance the organization released Sept. 17 as the first fruits of its new partnership with the Coalition for Health AI, or “CHAI.” The jointly authored document offers recommendations for seven areas of AI-related activity, from governance structures to quality monitoring to user training. The pointers are written to apply to healthcare organizations at every stage of AI adoption. The two orgs state that one motivator for coming out with the guidance was making the nascent JC-CHAI approach transparent. They’ll consider any serious feedback on the document from its audience and may incorporate it into updates or other communications vehicles.
- In an announcement, Joint Commission president and CEO Jonathan Perlin, MD, PhD, remarks that AI is changing healthcare both quickly and “at a scale I’ve never seen in my time as a leader.” Brian Anderson, MD, CHAI’s chief executive officer, adds that the present guidance and all to follow “are about keeping pace with the evolving field, not just by defining responsible AI but also by making it usable in hospitals and health systems across the country—no matter their resource level.”
- Other outputs are on the way from the JC-CHAI partnership. The orgs say we should next watch for governance playbooks, which they’ll produce after conducting workshops with representatives from hospitals and health systems of all sizes and from all U.S. regions. At some point, the Joint Commissions says, it will develop a voluntary AI certification based on the final set of playbooks and open it to its 22,000-plus accredited and certified healthcare organizations nationwide. The JC-CHAI partnership launched in June.
- Picture yourself interacting with an AI model that can tell you when you might develop any of 1,000 disease conditions over the coming 10+ years. OK, now try to see yourself finding out such a model is already here. Well, in Europe, anyway. Close enough. “While not ready for direct clinical use, the AI model offers new ways to study disease and inform healthcare strategies,” according to researchers who trained the algorithmic wonder on anonymized data from 400,000 patients in the U.K. Biobank. To test it, they used data from a completely different population—a swath of the nearly 2 million individuals in the Danish National Patient Registry. The proof-of-concept model shows it’s possible for AI to “learn many of our long-term health patterns and use this information to generate meaningful predictions,” says Ewan Birney of the European Molecular Biology Laboratory (EMBL), which is headquartered in Germany and has satellite operations there and in Spain, France, Italy and the U.K.
- “By modeling how illnesses develop over time, we can start to explore when certain risks emerge and how best to plan early interventions,” Birney adds. “It’s a big step toward more personalized and preventive approaches to healthcare.”
- Learn more from EMBL here and still more from media outlets all over the place.
- Also worthwhile:
- From AIin.Healthcare’s sibling outlets:
|
| | | |
| |
| | | What if AI medical transcription were twice as accurate? When clinical terms are transcribed correctly the first time, handoffs tighten, coding aligns, and audits speed up. Speechmatics’ new medical model cuts overall word errors by 17% vs. the next best system and halves keyword errors on clinical terms. With 96% medical keyword recall and a 4% keyword error rate, you get first-pass notes you can trust. Read more.
The ambient AI playbook: Lessons from two leading health systems At the recent CompassionIT Summit, leaders from Akron Children’s Hospital and Denver Health shared powerful lessons from rolling out ambient documentation to over 1,500 clinicians. Their biggest takeaway? Stories, not stats, drive adoption. Whether it was a heartfelt testimonial that swayed an entire department or a 60-second Nabla demo that eliminated training anxiety, the common thread was simplicity, authenticity, and clinician-centered design. Read more about the way these health systems are navigating ambient AI implementation: https://dhinsights.org/news/the-ambient-ai-playbook-lessons-from-two-leading-health-systems
|
| | | |
|
| | | Emergency medical services are there when you need them. Yet they barely come up in strategy discussions among or between governmental policymakers, healthcare economists and healthcare leaders. The AI revolution could help change that. An expert in the field tells why it should be welcome to do so in a paper published Sept. 17 in the American Journal of Healthcare Strategy. “AI now provides practical levers for dispatch triage, dynamic deployment, documentation automation and indeed clinical decision support,” explains the EMS professional, Edward Bauter, MBA, of Hackensack Meridian Health in New Jersey. All these activities, he states, can “serve to reduce costs per unit, reduce response times and improve downstream outcomes that will show hospital executives the true value of EMS.” The paramedic and EMS educator, who also runs an EMS training company called Overrun Productions, offers a number of insights worth consideration by healthcare decision-makers. Here are five. 1. Integrating AI into EMS can improve efficiency, reduce costs and enhance patient outcomes. Use cases include AI-assisted sepsis triage, speech-pattern recognition for early stroke detection, predictive trauma triage, STEMI recognition from ECGs and dynamic ambulance relocation, Bauter notes. More: ‘Risks [are present] in the areas of data bias, privacy and overreliance on algorithms, underscoring the need for human oversight and governance through steering committees.’
2. AI adoption in EMS is no longer theoretical; it is both practical and effective. “A deliberate implementation roadmap—including pilot projects, clear KPIs, infrastructure development and clinician training—is critical for scalability and trust,” Bauter writes. ‘EMS agencies that fail to adopt AI risk losing competitive advantage and strategic relevance in a rapidly evolving healthcare landscape.’
3. Areas with strong AI potential in EMS are responses, patient care and patient triage. Currently AI can accurately predict critical care needs in the prehospital environment, which in turn helps allocate the appropriate resources to a scene. ‘During transport, AI can enhance the travel route to reduce transport times or direct to the most suitable facility.’
4. Enterprise value in EMS is historically difficult to define. EMS exists as a service that is generally expected to break even or generate a profit, Bauter points out. “In the current healthcare environment, this can prove difficult given current reimbursement rates and institutional obstacles to financing,” he remarks. “For this, AI can serve EMS to reduce costs, improve patient satisfaction, enhance the clinician experience and start to bring EMS in line with the quintuple aim for healthcare.” More: ‘AI can be used to optimize EMS cost levers, such as reducing unnecessary transports via telemedicine, or optimizing deployment or responses for fewer unit hours, less overtime and lower fuel costs.’
5. Such AI initiatives can give EMS executives and leaders solid footing when approaching budget committees, hospitals and government administrators. “Agencies can now reliably use AI to establish better metrics (response times, over- or under-triaging, door-to-balloon times) and track the growth and efficiency of a system,” Bauter writes. ‘AI agents can also be integrated into systems to make data lakes within their organizations and establish local KPIs.’
Read the rest. - Other newsworthy research:
|
| | | |
|
| |
|
|
