Under-resourced hospitals face a big artificial intelligence adoption learning curve, but it’s surmountable if they can keep up with rapidly advancing analytics and AI tools in an industry increasingly turning to automation to lower costs and improve outcomes.
AI can improve health equity and access to care, an organization’s overall financial performance and even recruitment efforts.
But providers serving small, rural and medically underserved communities face distinct challenges in making AI modernization work. The Health AI Partnershipan initiative of the Duke Institute for Health Innovation and Duke University School of Medicine, aims to help.
The partnership, which launched in 2021, has been working with five under-resourced healthcare organizations to increase their expertise in trusting AI models and safely managing long-term usage through a 12-month program it calls the Practice Network.
The organizations in the Health AI Partnership’s first cohort are implementing ambient scribes, a “no-show” algorithm, sepsis warning code and retinal diabetic retinopathy scanning. Through the partnership, they have gained access to best practice guidance, industry mentors and implementation support and meet regularly to focus on troubleshooting their specific AI implementation challenges.
In session at next month’s HIMSS AI in Healthcare Forum, scheduled for July 10-11 in Brooklyn, New York, members of the partnership’s inaugural practice network of community health centers will share how they’re deploying AI into day-to-day routine care.
Leaders from the partnership spoke with Healthcare IT News to discuss common problems associated with evolving AI implementation challenges and AI’s value to low-resourced healthcare organizations.
Bridging knowledge gaps
About 10 months into their engagement, HAIP’s leaders say the mentoring and peer learning program has driven the rapidly growing number of use cases each of these organizations is considering.
There is a wide spectrum of technology involved – large language models, clinical decision support and more being undertaken by the Community-University Health Care Center in Minnesota and four federally qualified health centers.
Those are North Country Healthcare in Arizona; San Ysidro Health in San Diego County, California; Health Center of Southeast Texas; and the WakeMed health system serving North Carolina’s Research Triangle.
Key to the program is an eight-key decision point framework with 31 best practice guides for implementing health AI.
“As we started talking to sites, we soon realized that this needs to be a dynamic community that constantly thinks about best practices as tools get integrated,” said Alifia Hasan, innovation portfolio manager at DIHI.
“They have lots of pressure on them from various spaces on just surviving,” she added, and that is driving organizations like these to approach AI in clinical care and operations.
One example is in recruiting physicians, Hasan noted.
“If they’re not providing these kinds of services – like ambient scribe – they face [recruitment] challenges.”
But participating organizations face difficulty evaluating AI products, even with vendor assistance, due to the organizations’ expertise gaps. They may also lack seasoning in their negotiations with vendors, leading to unfavorable contracting terms.
“Oftentimes, they are seeking advice for how to respond to and navigate contracting and requirements with vendors,” said Mark Sendak, population health and data science lead at DIHI and HAIP co-lead.
“There is a big resource gap for sure, which everyone understands,” added Suresh Balu, DIHI director, associate dean of innovation and partnership for Duke’s School of Medicine and HAIP co-lead. “How do you put things into practice? When we convene these organizations to talk to the experts, they can actually address the knowledge gap.”
Implementation help
At the upcoming forum, Hasan, who manages day-to-day coordination and organizes technical assistance for Practice Network participants, will give an overview of HAIP’s long-term vision.
The partnership envisions scaling the Practice Network program nationally through a hub-and-spoke model that would enable other institutions to provide similar technical assistance to reach more under-resourced healthcare organizations across the U.S.
The digital divide these organizations face is “intimidating,” Sendak said.
“It’s not just the number of organizations. There’s 1,600 community health centers, and we’re working with four of them.”
He noted that what has been especially beneficial is the program’s “office hours,” where program participants meet with HAIP’s AI experts to advise on specific AI implementation challenges they are having.
Representatives from the five participants will join HAIP’s leaders for a panel discussion to share their AI adoption challenges and real-world approaches and how they are measuring value on investment.
The plan for the session is to address “the typical set of challenges, which are nuanced when it comes to adoption of AI, and all the challenges throughout the life cycle of development, deployment and monitoring of AI solutions,” Balu said.
Hasan, Balu and Sendak added that a lot of the credit for having under-resourced and community hospitals be part of the conversation around AI implementation goes to HIMSS, the parent company of Healthcare IT Newsfor leading efforts to address the digital divide.
The HIMSS AI in Healthcare Forum is scheduled to take place July 10-11 in Brooklyn. Learn more and register.
Andrea Fox is senior editor of Healthcare IT News.
Email: afox@himss.org
Healthcare IT News is a HIMSS Media publication.
Leave a Reply