Q&A: Nurses Are Key Stakeholders in Healthcare’s AI Journey

HEALTHTECH: Which AI use cases do you think will be most adopted and supported by nurses in the next two or three years?

RYAN: There are two specific things. One is administrative tasks in the patient room, but even bigger than that is large language models and what they’re doing for us. There’s so much data in the electronic health record, and no individual physician, nurse or anyone else has time to dig through all of that data. So, bringing forward the data that’s most pertinent to the patient at that given time is going to be one of the things that we can leverage through AI in the next couple years.

The other area is education. We have a nursing shortage and we’re looking at how we can bring more nurses into the workforce. One of the issues is a lack of nurse educators. However, bringing AI into simulation labs to help to train young nurses will help them improve the transition to using AI in the bedside practice. If they go through all of their education without ever really being introduced to AI and then they go to the bedside, they’re not going to understand it. So, let’s use it in their education process so that when they come to the bedside, there’s much better adoption.

HEALTHTECH: Why is it important to involve nurses in decision-making related to IT and AI solutions being used by clinical staff?

RYAN: It’s a pretty simple answer: If you’re involved in it, you’re more likely to adopt it and use it. Nurses are the best at workarounds because we know our workflows. So, if you put a nurse in a room with AI scientists, we can probably come up with a pretty good solution to solve the problems that we need solved. If you’re solving AI in a lab and then taking it to clinicians and saying, “Here’s what’s going to solve your problem,” you may not be solving the right problem to begin with.

Having nurses at the table should begin with identifying the problem to be solved. They can articulate the problems and then, being a part of the process, nurses can help because they are really good problem solvers. Allowing nurses to participate increases adoption, and the faster you can get adoption, the faster your ROI for that AI investment.

RELATED: Set nurses up for success with artificial intelligence.

HEALTHTECH: What’s the risk of nurses not being involved in these conversations?

RYAN: I’m going to use the EHR example. Going back to the late 1990s, when we started putting EHRs in hospitals, it was something that was done to nurses, physicians and clinicians. The situation was that we had to do this. So, we took everything that was on paper and we put it in the computer. That’s what we did. Then, companies started developing electronic health records without really involving clinicians.

There are risks to not involving nurses. There’s money, because an organization is investing a lot of money in AI. If you don’t get the adoption, it could potentially be a lot of wasted money and time. Retention is another one. The more nurses are involved in any aspect of their practice, it leads to better retention and support for whatever the initiative is. There are several different components, depending on the organization and the person.

It’s important for those who participate to educate their peers and say, “Here’s the meeting, here’s what we talked about, here’s why we talked about it and why it’s important.” AI is no different. There needs to be some governance in the organization, and nurses have to be a part of that.

HEALTHTECH: In what ways are nurses key to AI success, both as supporters and as skeptics of AI tools? For example, hundreds of nurses protested AI use in front of Kaiser Permanente’s San Francisco Medical Center in April.

RYAN: I think it sparks passion. For one, it’s created the conversation. There isn’t a nursing influencer on LinkedIn who isn’t talking about those protests. Hopefully, it sparks people to educate themselves about the passion behind it. Getting healthcare to slow down so that we can speed up is important. If you don’t have nurses involved, get them involved. I believe that in any change process, we should have the naysayers at the table. You can’t only bring the people who support you. You have to bring the people who are going to be the slow adopters or the change resistors. That way, you know what you’re up against and you can really formulate a solution that fits the organization and the people in the organization.

HEALTHTECH: How can healthcare organizations gain nurse buy-in for AI initiatives?

RYAN: Education and participation are important. There needs to be education about what AI is and isn’t. There are a lot of misconceptions about what AI is. AI will never be a nurse or a doctor, and that is a pedestal I’ll stand on because clinicians are licensed professionals with critical thinking skills. AI can learn a lot of things, but it isn’t able to decipher the difference between people and personalities. AI is not about replacing the clinician. It should be all about supporting clinicians and prioritizing patient safety.

HEALTHTECH: Is there anything that you think our readers should know about AI from Artisight’s perspective?

RYAN: One of the things I’m most proud about at Artisight is that we don’t develop AI in a lab where scientists are sitting there trying to solve problems. We are developing those solutions with our clients and asking them about the problems they’re trying to solve.

There’s a big focus on nursing quality indicators, things like falls, pressure ulcers, hospital-acquired infections, etc. AI can’t solve those for us, but it can bring the data forward to help nurses make better decisions and reprioritize their work. If you have a patient who’s at risk for all of those things, it’s crucial to give the nurse that information in a way that isn’t just a superficial assessment at the beginning of every shift. Instead, AI can bring data forward that tells nurses what they need to pay attention to for a patient so they can provide better and safer care.

We are working with our customers to figure out which of their challenges can be solved with AI. Then we can do trial and error. Try something; if it doesn’t come to fruition the way you want it to, let’s try something else. I’ll give an example. We did a whole study around pressure ulcers. What’s challenging is defining what is and isn’t a turn. Well, is 10 degrees a turn or is 30 degrees a turn? The solution became a situation where the AI was telling the nurse if the patient did or didn’t turn independently rather than documenting that a turn occurred. You have to peel away the layers of the onion and figure out what is best for a camera to do and what is best for a voice speaker to do, then leave those things that a nurse needs to do to a nurse.

EXPLORE: Mitigate nurse and doctor shortages with clinical automation.

HEALTHTECH: Do you think that healthcare will take meaningful steps toward AI adoption, or do you think that interest in it will wane?

RYAN: I think that it isn’t just hype. However, we have to truly define what it is and what it isn’t. Everybody is so quick to say that something is AI when, in reality, it may be something we’ve been doing for a long time. Maybe it technically is AI, but it really isn’t innovative in certain senses.

Down the line, we would really be doing ourselves a disservice if we didn’t find ways that AI can truly support clinicians and patient safety. This is really all about patient outcomes. We should aim for better patient outcomes despite the direction the workforce is going, which is toward fewer and fewer clinicians and more and more patients. We’re too short-sighted if we don’t at least try to use AI in the future.

Though I also think that five years from now, we’re probably not going to call it AI. It’ll become a normal part of our work. It’ll just be something that we do, and there’ll be another buzzword out there. For now, we need to focus on everything as a step in a phase. We’re not going to go toward rapid deployment of AI. You have to get the adoption, the belief and the support. The industry needs to slow down and make sure there’s governance in place. We have people who understand what they’re doing, and we need to work with our clinicians on how to implement AI into their practices.