Darrell Bodnar, CIO at North Country Healthcare
Photo: North Country Healthcare
Hospitals and health systems everywhere are implementing artificial intelligence tools in earnest for a variety of use cases. Many are getting results and even seeing ROI.
But a big question remains for many of these healthcare organizations: whether they know with certainty how their generative AI models are working, and whether they can trust all the conclusions those technologies make.
Introducing new risks
One of the most important issues today is responsible AI in healthcare, says Darrell Bodnar, chief information officer at North Country Healthcare, an alliance of three critical access hospitals and a home health and hospice agency serving the northernmost region of New Hampshire.
"AI has the potential to meaningfully improve how we deliver care, support clinicians and manage complex health systems – but it also introduces new risks if it is implemented without strong governance and oversight," said Bodnar.
"I know there is a lot of attention being paid to this and justifiably so. The long-term implications of not deploying AI technologies without the proper governance could lead to many challenges down the road.
"In rural healthcare especially, technology decisions carry a significant weight because we often operate with limited resources and smaller teams," he added. "AI tools – from ambient documentation to predictive analytics – can help reduce administrative burden and support better clinical decision-making, but they must be evaluated carefully for safety, data integrity and clinical impact."
The challenge is not whether AI will be used in healthcare but ensuring it is implemented responsibly and in ways that truly benefit patients and caregivers, he added.
Structured governance model
North Country Healthcare approaches AI and emerging technologies through a structured governance model that focuses on clinical value, patient safety and operational impact, Bodnar explained.
"Before adopting any new system, we evaluate it through several lenses: data security, clinical workflow integration, measurable outcomes and alignment with our broader strategic goals," he said. "We leverage a submission process that uses the nursing model of an SBAR – Situation, Background, Assessment, Recommendation.
"This provides some due diligence upon the requester," he continued. "Then we leverage an internally developed evaluation framework from a diverse multidisciplinary team to determine if we want to move forward with a pilot."
Through this process, North Country has evaluated and continues to evaluate several practical applications of AI that can meaningfully support its workforce. One example that has been completely implemented is ambient voice documentation, which allows providers to focus more fully on the patient conversation while the system assists with generating structured clinical documentation in the background.
"We are also exploring technologies that can summarize clinical documentation while providing contextual references back to the source material in the medical record, allowing clinicians to quickly verify the information and maintain trust in the documentation," Bodnar said.
"Beyond the clinical workflow, we are also assessing opportunities for automation within the revenue cycle," he added. "AI-assisted coding review, documentation analysis, and claims preparation can help smaller rural teams manage increasingly complex reimbursement requirements while improving accuracy and reducing administrative burden."
For North Country, the key is making sure these technologies are implemented thoughtfully and in ways that genuinely support clinicians, staff and patients, he said.
Responsible AI advice
Leaders at hospitals and health systems should focus on three things: governance, collaboration and practicality, Bodnar advised.
"First, organizations need clear governance around emerging technologies, particularly AI," he said. "This includes defining how tools are evaluated, who is responsible for oversight, and how outcomes are measured after implementation.
"Second, these decisions should involve multidisciplinary collaboration," he continued. "Technology leaders, clinicians, compliance teams and operational leaders all bring different perspectives that help ensure the technology supports real clinical needs."
Finally, leaders should remain pragmatic.
"The most valuable technologies are often those that solve everyday problems for staff and patients," Bodnar said. "If we remain focused on improving workflows, supporting clinicians and maintaining patient safety, we can take advantage of innovation while avoiding many of the pitfalls that come with rapidly evolving technologies.
"I would add that a recurring evaluation process should be implemented to ensure the expected outcomes are occurring," he concluded.
Follow Bill's health IT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.
WATCH NOW: Healthcare leaders offer HIMSS26 insights
