How Cios Are Prioritizing Ai Investments For The Next 5 Years

By Author: emrindustry
Total Articles: 266
Comment this article

We speak with health IT leaders, primarily CIOs, to learn the path forward through the priorities they set with their investments in six categories: AI and machine learning; interoperability; telehealth, connected health and remote patient monitoring; cybersecurity; electronic health records and population health; emerging technology; and other systems.

This first feature focuses on AI and machine learning. The top IT executives sharing their plans for the next five years in this first installment in the series include:

Dr. Shaun Grannis, vice president for data and analytics at Regenstrief Institute in Indianapolis.
Dr. J. Michael Kramer, chief medical informatics officer at OhioHealth in Dublin, Ohio.
Mike Mistretta, vice president and CIO at Virginia Hospital Center in Arlington.
B.J. Moore, CIO of Providence, which operates 52 hospitals across seven states – Alaska, Montana, Oregon, Washington, California, New Mexico and Texas.
Michael Restuccia, senior vice president and CIO at Penn Medicine in Philadelphia.
Everyone from board members and C-suite officers to IT managers and ...
... workers in the trenches will learn valuable information and guidance from their peers interviewed in these features.

Ambient, imaging, predictive analytics
Moore and his team at Providence have been making significant use of AI and machine learning. It’s an area that will see increased investment over the next five years.

“We use it pretty broadly, including ambient artificial intelligence in our partnership with Nuance,” said Moore. “That allows caregivers to practice their art of care delivery, while the ambient technology makes the updates, versus they’re having to type into the record. We’re evaluating artificial intelligence for imaging, particularly detection of cancer. There are a number of early detection pieces we’re doing with imaging.

“With predictive analytics, we’re using machine learning quite extensively,” he continued. “Everything from re-admittance, to no-shows, to acute care scenarios. That will help enhance patient delivery, decrease costs [and] allow us to more fully use our space.”

The provider organization has been using machine learning for predictive analytics extensively to meet surges and declines with COVID-19, and it plans to continue to invest in areas like that.

“And then we’re working with other companies like Truveta,” he added. “It’s a data consortium that we’re a founding member of, and that partnership involves using AI and machine learning on things like data normalization to help gain insights.

“There are a number of things we’re doing there that are confidential,” he noted. “We see machine learning as a really big-breakthrough set of technologies for us. We’ve been spending the last two-and-a-half years creating the foundation to allow us to do these things. My three strategic pillars apply here. They are: simplify, modernize and innovate.”

Moore does not foresee any issues selling more machine learning investments to the rest of the C-suite and the board.

“They know my background,” he stated. “I came in two-and-a-half years ago from 27 years at Microsoft. So people trusted they made the right hire. So when I talked about big data and machine learning and AI, they didn’t push back. Over time it went from, ‘Hey, we support B.J.,’ to, ‘Wow, this is a really big game-changer.'”

New to the technologies
Virginia Hospital Center has only just begun to work in the area of AI and machine learning.

“Honestly, I’m not sure what we will be doing in this space,” said Mistretta of Virginia Hospital Center. “We have our toe in the water, so to speak, with sepsis agents and some other predictive analytics currently, so we will see what the organization’s receptivity is. Our issue isn’t really selling investments to leadership. The challenges are below that level with adoption directly.

“There are two large challenges I see with this technology,” he continued. “First, you have to have a tremendous amount of data built up to be able to run the algorithm against, then have a method to validate the results. And second, are the users/clinicians ready to be challenged on their previous mental models of care delivery, and are they open enough to [considering] assistance from technology?”