Welcome to ACNL in Action, brought to you by the Association of California Nurse Leaders. I'm your host, Charlene Platin. Artificial intelligence, or AI, is moving fast in healthcare. And whether we're ready or not, it's already shaping how care is delivered and how nurses do their work. For nurse leaders, the real question isn't if AI will play a role, but who gets a seat at the table when those decisions are made. Our guests today are Narissa Amber, Director, and Carlene Lugtu, manager in the health informatics transformation team within the department of the Chief Nursing Information Officer and integrated in the data science team at Stanford Medicine. They're here today to talk about why nursing leadership matters in the AI space and how nurses can help shape technology that supports patient care. Thank you so much, Narissa and Carlene, for being here today.

Thanks, Shar, for having us.

Thanks so much, Cha for having us. We're really excited to be here. Look forward to the conversation.

I agree. I think this is a really timely conversation. And you know, the first thing that uh strikes me is the name of the team, Health Informatics Transformation. And Narissa, I'm wondering if you could share more about what is the Health Informatics Transformation team.

Yeah, absolutely. So, like you said, we report up to our Chief Nursing Information Officer, our CNIO, and this was really a team that was created out of a need to respond to this emerging AI technology and have a team that could upskill, lead those AI development and the change management that's involved with that. So a lot of our work has been foundational in developing the frameworks, the governance, the monitoring infrastructure to support the AI implementation. And then also in development and implementation of our AI tools, whether they be vendor or um internally built technology. Carlina, the nurse, any else, anything else to add to that?

Yeah, I think we're really fortunate that our we have this um seat at the table, I guess, as you're mentioning, to really help mold our systems infrastructure and how we're approaching, as well as embedding this framework within the culture of our organizations. Um, you know, nurses are the highest number of labor or you know, staff within an organization. And I think it's really imperative that our voices are here to help do the translation to not only um nurse leaders, but also frontline staff and getting their feedback as we are evaluating and deploying tools.

That's very well said, Carlene. Thank you for that. And uh Narissa, I want to turn it back to you to share more about AI in general. So if you can give us an overview of what AI looks like in healthcare right now, because you know most people are familiar with Chat GPT, for example, but we're not talking about just simply drafting patient emails here. Yeah, absolutely.

Um there's actually a paper that does a really good job of explaining sort of the journey of AI in healthcare. Um, it's in DAMA, and it was written by some of the Google executives when AI and and Chat GPT and all of these technologies were emerging. And then it talks about the three epics of AI and healthcare. The first one being this symbolic AI that really started early in like the 50s. This is just old date, best practice alerts. These things have existed for a very long time and are very common, but it's AI that we don't even recognize has been in our in our culture, in our tools, in embedded in our workflows for a very long time. In the early 2010s, is when the concept of deep learning started. And that was really more about like pattern recognition and finding like early warning signs that was another version of AI that sort of came upon in that time. Um, and things like fall risk algorithms and and sort of task-based, very specific point solutions and risk-based models, is what the essence of that kind of AI is. And then the one of today that we all kind of feel and it has now become a real big part of our world, um, was sort of from the from the um launch of Chat GPT and sort of all of the emerging technologies that came from that. And so today, what we see in AI and healthcare, what that focus has become, and what has become really pervasive is more this generative AI technology that has a lot of looks and feels from ambient documentation to summarization of tasks using tools like ChatGPT or now GPT Health that has come out or Gemini or any of these models to summarize or find information more easily. Um those kinds of tools will be super meaningful as we continue to develop and sort of figure out more of where they best can be used to help release, relieve some of our burden as a workforce.

Yeah, and that's a really interesting point, Arissa, because the truth is that artificial intelligence has been around for a very long time and it's evolved as well as our technology has evolved, and that includes in healthcare. And, you know, speaking about today and nursing and what that looks like today, I'm curious, Carlene, if you could share where AI is already meaningfully impacting nursing workflows as it looks like today.

Well, one of the biggest areas, one of the areas that we're seeing is really directed towards nursing documentation. I think we're really in the early stages of designing what that product looks like to help support by use of ambient um intelligence or ambient, ambient, um, an ambient tool to help support uh the documentation of uh nursing workflows directly into the EHR. And um that's all purely by voice. Uh there's a lot of uh the vendors out there in that space. I think that's very early in and where where it needs to be, but I think there's a lot of buzz and a lot of um excitement around that because of an awareness of all of the documentation burden that nurses do face and and how we can alleviate that. Um, there's tools, you know, tagged on to use of generative AI and LLMs that are okay, how can we generate an end a shift report? How can we help support care plan documentation? Um, and as Nursa had alluded to, we also have a host of predictive tools and risk scores that are uh deployed within the EHR. And so there's a huge host of the different types of tools that are embedded or are going to be further embedded into these workflows. I think from the generative AI or the generative AI landscape of tools, there's still a lot of studying and research that needs to be done as these tools mature and they're well embedded into the workflows and how we're actually going to alleviate that burden for nurses as a whole. And um from the ambulatory lens, uh, there are also tools within supporting the in-basket, of course, you know, nurses and also care team members. There's millions of messages that come through uh providing asynchronous care, supporting all of our patients outside when they're outside of the walls of the whore hospitals. And so that is also probably one of the first use cases of jugnitive AI in healthcare. And it's probably the most widely adopted across multiple health systems. And uh that's there's definitely alleviation of burden there uh by use of those tools. It's it's not an automation per se, but it helps alleviate some of that chart biopsy and chart digging and formulating response for our patients. Um, and I think as this development of all of these tools to support nurses, you know, there's also movement in um computer vision and having video uh cameras mounted within uh smart hospital rooms. And those are all very nascent, I have to say, but I think we're really at the cusp of uh where these tools are going to be more prolific in in all of the healthcare settings.

Yeah, thank you for that overview, Carlene. And it's really interesting to think about the many applications and use cases for AI in both, you know, inpatient setting, outpatient or ambulatory setting. And it's also interesting to see that a lot of the use has been in the outpatient area because a lot of the time I think historically we think about innovation impacting the inpatient area. So really interesting to see that wide adoption of it in the ambulatory space. And I think what you mentioned too about relieving the burden is really interesting and something that is in particular really exciting for nurses, because of course, you know, documentation has been burdensome for, you know, this is a very common thing that we've talked about for many, many years. So it's really exciting to see the possibilities there. And because there's just so many tools, right? Inpatient, outpatient, in all these different areas, there are a lot of different, you know, thoughts and perceptions and misconceptions uh misconceptions about what AI is and how it can look like. And Narissa, I'm interested in learning what are some of these common misconceptions or common perceptions that nurse leaders might have when it comes to AI and healthcare. Yeah, that's a great question.

Um, you know, you you're this landscape that we're talking about. Um, there's this concept of the jagged frontier when it comes to AI, in terms of we are still figuring out the spaces where AI makes sense and works well, and also the places where it doesn't and we should not be using AI. And so figuring out that landscape um is something that's gonna be really important. And um, the misconceptions, there are a lot of misconceptions. I think the biggest one that you hear often is like the AI is gonna take my job away, or AI is gonna take my autonomy, or it's gonna take the clinical tasks and all of that away. And I I just I think that's something that we um is a fear, but really AI is designed for augmentation and not to take like nobody is going to take the nurse away. That is like so um, so far from reality, really. Um AI, especially generative AI, it's like I like to think of it as this real it's a it's an assistant. It can augment your workload, it can help you do things faster and find information in this chart that we have been burdened with documenting for so long. It can help you extract and even input into that in a more seamless way. We're still working through some of those details. Like that isn't all figured out just yet. Um, but really, generative AI is it's good and it's getting better. And it's the worst it will be today, like it will continue to get better over time. But at the end of the day, it's still a really smart, I like to think of it as a really smart but very literal like intern. Like that's what we're talking about here. This is not um something that you can rely on without continuing to be that human in the loop. It misses nuance, which is what you need the human for. And so nursing will still always be around. Like that is something that is going to be so core. But the way that we're measuring this impact and thinking about it is how much time can I save from the nurse and doing these tasks that are actually really laborious and tedious? Let's take that off the plate so that we can give more time to doing the thing that nursing is really good at, which is being with the patient. Um, so it's not it's not gonna take away jobs. It's gonna hopefully take away burden.

Mm-hmm. Yeah, I think the reframing of um AI is really a tool to use rather than a replacement for a job, you know, a role, right, in the healthcare organization. And I think that is really important to consider because of how widespread those misconceptions can be. And I think when there's a lot of change, there this is when a lot of these um different misconceptions can pop up. But there's a lot of use cases for AI to help relieve the burden, to help be, you know, to augment the work that the nurses are already doing. And I think when all these conversations come up, that's probably also why it's so important to have nurses and nurse leaders incorporate into incorporated into these conversations, right, surrounding AI. And uh Carlene, on that note, I'm wondering how can we incorporate more of these nurses and nurse leaders into these decisions surrounding AI tools at different health systems?

So the success of AI in healthcare is really going to be dependent upon an interdisciplinary team. And that means having nurses engaged with the full process of how we are looking at these tools and how we're deploying them into their workflow. Um, if AI is, you know, these tools and features are deployed into the workflow without that nursing input, there are just real risks that are gonna be surfaced immediately. Because if we don't understand the granularity of where the human plus machine are interfacing and making sure it's seamlessly embedded into the workflow, that is a huge risk that organizations will will take on. And nurses by nature are our last line of defense per se of uh absorbing some of that risk naturally in in, especially within the four walls of a hospital. And so if nurses are not educated, if we're not upskilled and don't gain that literacy around AI in general, what are the limitations? And then specifically, specifically to the tool, like what is the intended use of this tool and how it's going to hopefully um what are the goals and what are the actual process and outcome measures in the benefit of the not maybe the patient, maybe in the benefit of the nurse or or the system in general. And I think just an overall awareness is is is a necessity in order to make any deployment really successful. And I think as nurse leaders, uh, you know, the voices of nurse leaders are very strong typically within organizations. I think we're very fortunate because we have a nurse leader that is well uh embedded within the technology domain attached with the with the health system. And so uh, of course, appreciating that not all health systems or hospitals are gonna have this infrastructure that we we have. But it doesn't mean to say that you can't self-organize and have a group within uh any domain or any side of care that can actually voice that they want to be part of that discussion. Uh, it doesn't need a formal structure. And I think uh engaging and leaning in instead of turning away and not not, you know, being part of the dialogue, being part of the design, being part of the deployment, you know, that is something that we need to take power into in making sure that whatever system that you're in, that you're part of those conversations.

Right. And it sounds like it's there's really opportunity both ways when it comes to both the nursing staff and also the nursing leaders to engage in to the advocacy of having more of the incorporation into these discussions because there is a lot of opportunity, right, in the different health systems, whether it's in formal structures or formal structures. And one of the formal structures that we have in our organization is actually shared leadership. And Narissa, I was wondering if I could turn it over to you to share more about your experience with that and how nurses have been incorporated into that.

Yeah, absolutely. I I think actually, I think of our shared leadership. We have a pretty robust shared leadership governance council. Um, and I sit within our research and innovation council. And I think one thing that we have done really well is that we have taken that group of nurses, we are using that forum to not only upskill and um teach that AI literacy, but we're also using that group to vet technologies to make sure they're ready for the clinical practice environment. So they are almost the lions at the gate for our new technology innovations. We do surveys with that group all the time on vendor solutions. And in particular, we have a homegrown solution that is internally built at Stanford. We it's called Chat EHR, but we have given access, early access to this tool to this group so that they can improve their understanding of how to use it. Um it's essentially a chatbot within the EHR, but the group is now empowered to try it, iterate on it, and then be that um super user for their unit when we eventually roll this out more broadly. So they've been able to, they've been able to use this group to help um not only educate, but then improve their awareness of what works well with these tools, um, encouraging them to use them in their personal life too. I think a lot of these language model generative AI solutions, the more you sort of play around with them and use them, you sort of understand the limitations. And so this group has now been um given, like they get early access to a lot of our new technologies so that they can bet them, give their feedback. We bring that feedback like directly to our developers and vendors and say, like, this is what they're saying. We need to fix it. Um, and so that group has been super empowered. They're very um, they're very excited, they're always super engaged. We survey them all the time and get their feedback and then use that to improve it so that by the time we roll it out to a larger uh the, you know, the organization or larger groups, we've had some more vetting of those technologies. So with the Chat EHR tool that we've built internally, like I said, it's a chat bot. Um, the nurses are working on building out prompts for their particular areas so that when we do release this more broadly, they'll have something that they can then share with their colleagues, their unit, their department that will be meaningful to their individual practices. And it gives them the opportunity to figure out what works and what doesn't, sort of trial and error and iteration, um, and be um, you know, be leaders for their unit and really empower them to give their feedback and help us improve and help train the trainer so that they can scale that out more broadly. Let me get there.

Mm-hmm. And I like that example a lot because you're bringing some of these tools and some of these examples directly to where the nurses are, right? You know, using the structures that are already in place to get them involved, right? It's not extra effort on the nurses' part to learn about this and to pilot it and to share their ideas because you you're bringing the developers there, you're bringing the team there. So it's really great to see and um a nice place to get real-time feedback. So uh that's a great example. And, you know, Carlene, I'm interested too, when it comes to involvement in AI, in general for both nurses and nurse leaders, what questions do you think nurses and nurse leaders should be asking before they agree to pilot or launch any AI-enabled solution?

I think at baseline, nurses should first and foremost be, nurse leaders should be aware of what governance structure and what is the organization's approach to what everybody is talking about is in terms of responsible AI. At Stanford Healthcare, you know, we have a North Star under the framework of our responsible AI lifecycle, as well as evaluating any of these tools under another framework called FERM, which is uh came out of uh Dr. Nigib Shaw's lab, who is our chief data scientist here. And I think understanding kind of the layers of how your organization is thinking about AI, I think is the first role as a nurse leader. And then if there is a pilot brought to a particular area, they are a key stakeholder, right? They themselves, as a nurse leader, are overseeing their team members and asking the questions of um what is the purpose, really understanding, you know, making sure that there is a fit to the tool, to the problem. A lot of times when people or anybody thinks about it, yeah, they just want to slap it on top of a problem, but really making sure that they're thinking the nurse leaders are thinking through, okay, how does this impact what is our process? What what what are our team currently doing? Um, and then how can this help and augment? And are we kind of all moving in the right direction towards the problem we're trying to solve? Um and I think further understanding when you're thinking about uh how your organization is approaching responsible deployment of AI, I think, you know, nurses, nurse leaguers can serve as really that translation as well. Uh understanding what has been done to evaluate the tool on our patient population, what has been done to de-risk uh use of this tool in our organization, what really being uh asking those questions, what is the risk involved? Is is this you know a clinical clinician decision-making tool, or or is this more of just an administrative tool and a kind of really asking those questions of how is this going to impact their staff? Um and then really also making sure that they themselves are that conduit for their team to report back how the tool is interplaying into their workflow. Um, nurses are great at work around. If things are not working, they're gonna find a way to work around and they're just gonna not use whatever's been deployed, any tool whatsoever. And I think they are a real uh force within making sure and helping us evaluate the success of these tools. And um I think understanding all of those layers will help nurse leaders with education, educating their teams, supporting any changes to workflow, because again, that's a lot of change management, and then bringing in any you know misconceptions and being that that interface with their team to help with with moving in in the direction of deploying these tools, tools.

Yeah, thank you for sharing that, Carlene. And it really sounds like both nurses and nurse leaders can leverage the skills that they already have as nurses, which include asking really thoughtful, intentional questions, using their critical thinking. It's not that different from when you uh you know deploy other innovations, medications, treatments, right, for your patients. And then this is another tool, right? That a lot of it does support both patients and staff. So using a lot of those transferable skills to think when you think of AI as a tool and a solution. And I do want to ask too about what nurse leaders can do to be more knowledgeable about all of this. And so, you know, Narissa, any thoughts on what competencies or knowledge that today's leaders can use when they think about developing AI and when they think about informatics?

Yeah, absolutely. Um AI literacy is gonna become more and more important. Um, it is something that, like I said earlier, we can start learning some of these things in our personal space. This technology lives and exists in many other facets outside of the job. Um, so we can learn some from that, but AI literacy is gonna become a really fundamental skill, as well as data literacy. Um, you know, nurses spend a lot of time inputting data and looking for data and getting better at knowing what's the signal and what's the noise in what you're seeing is gonna become really, really critical. Um the other thing I will say as an informaticist, I feel like much of my job today is teaching people about prompt engineering. Prompt engineering is like a job that people thought that they were gonna have two years ago. That was something, oh, become a prompt engineer and get, you know, this is gonna be a whole new field and career. Um, that was a moment in time. And now prompt engineering is a skill, um and it's iterative and it takes practice, and it's something that I get the most questions about how to prompt these days as an informaticist. That is not something I was expecting, but something that is pervasive right now with the with the world of generative AI. Um, I think also, and nurses already do a good job of this, but like a healthy skepticism about things, that healthy skepticism is going to become really important to sort of think through the technology and the AI implementations that happen, making sure when you're at the table that you're asking the right questions. And those skills build over time. You learn as you as you do um, as you do, you know, implement more and learn from your mistakes and you know fail forward. Um, you learn as you go. And so I think that is gonna be is something that nurses already do, but something that they're gonna continue to need. Um, and and you know, I I want to just underscore the human in the loop and just share that we need to continue to make sure that we are keeping that in mind that the nurse and the clinician is gonna continue to need to be that human in the loop. The term copilot is super appropriate. We hear the term co-pilot with ambient documentation or whatever other office type products you're using. And the implied other part of being using a copilot is that you're the pilot. And I think that reframing is something that we need to continue to message that yes, you can use a copilot. This is your AI tool. It's gonna help augment your work and make it easier for you. And also, you need to keep your critical thinking um up to date. You need to keep that pilot license still on. Um, and that is gonna be something that I think we continue to make sure we underscore.

Yeah, those are all really great insights to keep in mind, Narissa. And I really like the the framing of us as the pilot, right, as we're using AI tools. And, you know, we could talk about this for so much longer. I do want to close out with just giving you the floor, um, both of you, to share any final words of wisdom to our audience, and many of them who are nurses, nurse leaders, or you know, people in the field of healthcare. Um Carlene, I could turn it over to you first for any final insights.

Yeah, I the space is moving completely mind-bogglingly, like just extremely fast. And I think nurse leaders, I it is just this is another uh line of sight that nurse leaders really need to be aware of um and to lean into uh this this tech sort of technology because it is here and it's it's definitely not going to go anywhere. Um nurses also we we also need to start to shift our our mindset of we are we are just not only we're the end users, but we really need to realize that we can be part of design, um deployment, and every phase of the life cycle, they will be key components in in any AI deployment. And I think appreciating that that they are the keys to success and the keys to realizing if we're actually getting any benefit from these tools and realizing if if we're alleviating or our goals are met or or these are actually tools that we need to keep in our system uh at supporting either our clinicians or our patients. I think leaning into that and making sure that that that they have that voice uh to engage with the process wherever they are in the health system is is I think of importance.

That's great. Thank you, Carleen. And Narissa, any final thoughts as well?

Yeah, those are excellent points. I think this concept of governance for nurse leaders and for responsible AI leadership in general, um, governance is going to be pretty critical to a lot of this work. And having nursing at the table in the center with that clear responsibilities and accountabilities, nurses are the biggest piece of our workforce. And they will have to have a voice in this debate. And keeping governance monitoring for safety, quality, performance, all of these pieces, that's gonna get bigger and more robust over time. Um, and we are gonna also need to figure out how we build our muscle on rolling it back. I will say one of the things we tend not to do well from pilot to implementation is cut it when it's not working anymore. And we are gonna have to get better at decommissioning and rollback plans for when we find in our monitoring and through our governance that things are not working anymore, that we have a pathway and create pathways for that decommissioning of these tools so that we don't continue to keep things in our environment and that we, you know, maintain a healthy AI system.

Great. Thank you so much for those final words. And and I just want to say, Carleen and Orissa, thank you both so much for being here today and for sharing all of your expertise with us on AI.

Thank you so much, Shar, for inviting us. This was awesome.

That was so much fun. Thanks, Sarah.

Thank you both so much, and hopefully we'll have you back for some more conversations in the future. Awesome. Thank you. Our guests today were Narissa Ambers, Director, and Carlene Lugtu, Manager of Health Informatics Transformation at Stanford Healthcare. We'd love to hear your feedback. Email us at socialmedia atacnl.org and connect with us on LinkedIn and Facebook at ACNLNurse. And as always, if you'd like the show, please remember to rate, review, and share. ACNL in Action is presented by the Association of California Nurse Leaders with new episodes every month. To learn more about the show or ACNL in general, visit us at acnl.org. Thank you for listening.