The 5 Rights of AI in Healthcare© framework can help guide nurse leaders with program implementation, says this CNIO.
AI is now one of the biggest disruptors in healthcare as it works its way into all corners of the industry.
Many nurses have ethical concerns about generative AI, along with worries about job displacement and the potential loss of human touch and connection with patients. For CNOs and other nurse leaders, implementation is another big concern.
It's critical that CNOs incorporate AI into workflows in a way that will help nurses and patients, and not become a hinderance. According to Robbie Freeman, vice president of digital experience and the chief nursing informatics officer at Mount Sinai Health System, nurses need to have a seat at the table during decision making to make implementation go smoothly.
"One of the things that I think will be important as we start to see AI make its way into patient care is that our care team members, our nurses, have a seat at the table," Freeman said, "and that their voice is included in how we design these AI products, understand how they work, and then implement and ensure that it's a seamless experience."
At Mount Sinai, Freeman said they take a co-design approach to decision-making surrounding AI products, meaning that everyone sits down together at the table, including data scientists, engineers, and frontline team members, to design the best product that will work for everyone.
"Some of our best ideas come from our frontline team and our nurses," Freeman said. "Anyone in our organization can submit a request for an AI concept or idea, and we have many examples where the frontline team has sent us an idea, we would scope it out together, and where it makes sense, we would partner, develop, and bring it to life."
The goal is to help reduce the burden of documentation and improve the ability to keep patients safe while delivering high quality care, Freeman explained.
"I think when done in the right way, we can do a lot," Freeman said, "but we want to make sure that the key is having the nursing voice at the table along the way."
Framework for implementation
Freeman, along with three other healthcare leaders, Angela Ross, Kathleen McGrow, and Olga Kagan, recently published what they call the 5 Rights of AI in Healthcare© along with an article in the Journal of Nursing Management.
"It was inspired by the five rights for medications," Freeman said, "so we took that, and we wanted to look at it through the lens of AI."
Freeman said the authors came up with five things that are important to consider while implementing AI. Leaders should first identify the right objective, meaning the problem they want to solve with AI.
"[Make] sure that [it's] really clear why you're using the tool and the problem that you're solving," Freeman said.
Second, CNOs need to consider the right approach, while making sure that the AI will be integrated into the workflow and that the approach makes sense. Third, CNOs must have the right competency on the team that will be using the AI, and make sure that the team members understand the "why."
"We don't want to just blindly follow the output of an AI model," Freeman said, "because we really look at this as an augmented intelligence tool and it's not going to ever replace our clinical decision-making."
Fourth, the right data is critical. CNOs need to understand how the data is sourced and interpreted by asking questions about how the AI was trained. Fifth, any implemented AI tool must have the proper safeguards.
"We [need to] make sure that we've tested it prospectively, we understand how it works, we've checked for bias, and that we've ensured that those steps were taken before it gets used for clinical care," Freeman said.
For nurse leaders
The goal of this framework is to empower nurse leaders with a set of high-level questions to ask before they implement an AI product, according to Freeman.
"If this was written for an audience of data science experts, we may get into some more technical discussions on hyperparameter tuning, or the way things are working under the hood," Freeman said, "but we tried to keep it at the business leader [and] clinical leader perspective to make sure we're at least addressing those big boulders."
However, health systems still should utilize those teams to work on the details, Freeman recommended.
"[As organizations] think about their AI governance structure, they need to have teams that can do that deeper dive into the very technical evaluations around the model," Freeman said, "like performance and how you're going to measure that over time, and watch for performance drift, and other considerations."
The Coalition for Healthcare AI (CHAI) has a set of guidelines that can help organizations set up their AI governance, according to Freeman. As for the 5 Rights of AI in Healthcare©, they were written with nurse leaders in mind.
"This was really through the lens of a nurse manager or nurse leader," Freeman said, "and it's not meant to replace organization-wide, overarching, robust governance processes that are also critical."
Taking action
To move forward with AI implementation, Freeman recommended that CNOs ask about AI governance and make sure that nurses have a voice at that table. Nurse leaders should have a voice in any tools that are used at the point of care by the frontline nursing team.
"I think the CNO plays a really important role there advocating for the nursing team, advocating for the safe and ethical use from the patient perspective as a patient advocate, and ensuring that they have a voice within the governance process," Freeman said. "I think that organizations who do that well are going to be really well positioned to realize a lot of value from their use of AI."
CNIOs can also be a great partner for CNOs, Freeman explained, since they can help with communication between the technology teams and the clinical teams.
"I would encourage our CNOS and chief nurse executives to lean on their CNIOs and their nursing informatics team members who can play a really important role here," Freeman said.
G Hatfield is the nursing editor for HealthLeaders.
KEY TAKEAWAYS
CNOs and other nurse leaders need to have a seat at the table during AI decision-making to make implementation go smoothly.
Nurse leaders should make sure that AI is implemented with the right objective, approach, competency, data, and safeguards, as outlined in the 5 Rights of AI in Healthcare©.
CNIOs can also be a great partner for CNOs, since they can help with communication between the technology teams and the clinical teams.