In this episode of Diabetes Care Conversations, we explore how artificial intelligence is transforming diabetes care, from reducing administrative burden to enabling more proactive, data-driven, and personalized treatment. Davida Kruger, MSN, APN-BC, BC-ADM, CDCES is joined by Janice MacLeod, MA, RD, CDCES, FADCES who shares practical insights on how clinicians can responsibly integrate AI, address risks like bias and data limitations, and take an active leadership role in shaping patient-centered care.
Resources and References
Davida Kruger
AI is already changing healthcare from how to document visits to how we identify risk, interpret data, and support clinical decision-making. But what does that actually mean for diabetes care? And how can clinicians use AI in ways that are practical, responsible, and patient-centered? Hello, and welcome to ADCES Podcast, Diabetes Care Conversations. In each episode, we speak with guests from across the diabetes care space to bring you perspectives, issues, and updates that elevate your role, inform your practice, and ignite your passion. I'm your host, Davida Kruger. Our guest today is Janice McLeod, a diabetes cardio metabolic consultant, past chair of the Academy of Nutrition and Dietetics Diabetes Practice Group, and a member of the board of directors for the ADCES. Janice is here to help us unpack what AI means to practice today and what diabetes care and education specialists need to know. Janice, welcome. We're so glad you're here.
Janice MacLeod
Thank you, Davida, for this opportunity. By background, I'm a registered dietitian, a certified diabetes care and education specialist, and much of my work has focused on helping people navigate the complexity of diabetes. And this has included work with numerous diabetes technology companies and several years with a pioneer AI-powered digital health company. You know we know diabetes care generates enormous amounts of data. And I learned in my work in digital health how AI can take all of that data and turn it into tailored real-time feedback. And that empowers self-management for that person with diabetes, but it also keeps them connected with their care team by providing clinical insights and decision support for the team. That enables a pivot to data-informed continuous care that much, much better matches the relentless demands of a 24-7 condition like diabetes.
Now, what also happened about three years ago, I noticed in ADCES discussion board post, and it was about how AI was potentially going to threaten the diabetes care and education specialist job. And I knew it was so important to reframe that discussion. And so I and others that jumped into the discussion were successful at doing that. But it ultimately led to me being invited to speak to a state ADCES diabetes annual meeting. And then I subsequently spoke to several more states and I've also done lots of work with the Academy of Nutrition and Dietetics on this topic, as well as other organizations and other disciplines. In my talks and publications, I talk about how AI might help us address some of the biggest healthcare challenges.
And what I realized is that AI in healthcare isn't a technology story, it's a healthcare transformation story. So my goal is to help my colleagues understand what AI actually is, where it's already impacting diabetes care, where we need to be cautious, and most importantly, why clinicians, and I would say especially diabetes care and education specialists, have such an important leadership role to play in ensuring these tools are used wisely to improve care.
Davida Kruger
Excellent. Thank you so much for that background information. You know, across the spectrum, many clinicians are really hearing about AI now. You've got the background, you've been talking about it, you've been going out to... Why is AI suddenly such an important topic for all of us to grab and understand better?
Janice MacLeod
What I believe happened, we had the release of vastly improved large language models. That's a form of generative AI at the end of 2022. That's when OpenAI released ChatGPT. What that did is it put an open access AI tool into the hands of common ordinary people who were now able to see for themselves AI is here, it's going to change how we live our lives, and it's certainly going to totally transform healthcare. So clinicians reacted on one extreme with fear and on the other extreme blindly believed that AI was going to magically fix healthcare. And of course, neither extreme reaction is helpful. We need a more balanced response and recognizing our responsibility to step up and lead in our places of practice to wisely bring AI into healthcare. AI experts actually predict that there will be more change in healthcare in the next decade than in the previous century.
So, what are some of these challenges that AI can help us with? We have rising chronic diseases, but our healthcare system is acute care focused. And that means we're downstream, reactively managing symptoms when we should be upstream, proactively addressing root causes and preventing those symptoms from ever happening. We also have workforce shortages, clinician burnout. We have administrative burden, waste, inefficiencies, and data overload. Every 72 days, the amount of health information available to us doubles, and clinicians cannot realistically process all of that alone. AI, as we've talked about, is a tool to help transform overwhelming data into actionable insights, and that can support better, more timely, tailored care for entire populations cost-effectively, so at scale. And it's also interesting to me to note that our current evidence-based findings may only be the tip of the iceberg and representing only shallow evidence for the care of a generic patient. So think of our current standards of care. What AI will enable us to do is analyze data from multiple sources, so EHR, lab data, research data, patient-reported outcomes, social drivers, and so forth. And that will enable the next generation of what's being called deep evidence, and that opens up the door to precision medicine. So an exciting time. Yeah.
Davida Kruger
And those are great examples. Can you just share for us an example of, like, I think AI scares the bejeebers out of me because I don't understand it, but certainly the way you do. Could you explain maybe for me and for the audience an example of where I'm already using it without even realizing it?
Janice MacLeod
Well, AI certainly is not science fiction. It is already embedded in tools clinicians and patients are using today. In fact, one of the very first ways that healthcare systems have brought AI into healthcare is to take on many of the repetitive administrative and operational tasks that are just everywhere in healthcare. So think about things like scheduling, triage, supply chain, billing, shortening the revenue cycle, and so forth. But the savvy healthcare leaders are taking this opportunity to completely rethink the patient customer care experience and connect the dots to what needs to happen next. So let me give you an example of this. So an AI tool could trigger a notification to a patient who has lab work due. And then once that lab work is analyzed, a plain language summary is sent explaining to that patient what do their results mean, but also with a link to what needs to happen next. So for example, let's say the patient had a lipid profile, showed dyslipidemia, the system would generate an automated referral to the registered dietician for nutrition counseling to help reduce heart disease risk. And so this is why it's so important for diabetes care and education specialists to be at the table as AI initiatives are planned and being discussed in your clinic, in your health system, wherever it is that you practice, because you're the one who's going to know, hey, that's a great place to build in an auto referral to the diabetes care and education specialist, the dietician, the nurse practitioner, whoever it might be. And AI is also helping reduce clinical burden by assisting with documentation. And this is what Dr. Eric Topol, he's the author of the book, Deep Medicine, calls keyboard liberation. I love that term. AI has the potential to restore time to clinicians so they can focus on what matters most and that's connecting with their patients.
Now, other ways AI is being used in the background, you may not know it, but clinical decision support, summarizing clinical research, translation services, population health risk gratification, which I mentioned a bit about. Also, some are using it increasingly so for assessing and addressing social drivers of health and then linking to local resources. There's a lot of a surge in interest in doing this right now in AI. Then we also, Davida, we're seeing the recent release of consumer-facing health versions of the common large language models. And so what this does is people can use these to integrate their medical record, their EHR record, but then also if they want to connect data from their connected devices and use that to then get some at least preliminary health advice. So very early release, so we'll have to see how these impact our practice.
Davida Kruger
That's phenomenal. And it has so many opportunities like you've outlined and certainly in the world of diabetes to make sure that the diabetes educators are included, the nurse practitioners are included, technology is included. I love how you've described that because it really does bring it home and it makes it real for the diabetes world. You know, we're drowning and if this can take a little bit off of our plate and make sure that the person with diabetes still gets the care they need. I think that's phenomenal. There's excitement about AI out there, but what are the biggest risks and misconceptions clinicians should be aware of? What needs to be in place for AI to truly improve diabetes care for all of us?
Janice MacLeod
There's so much. You know, it's almost like we know the stages of grief, right? It's almost like we're going through these stages of AI. Yes. You know, probably at least some of us have gone through this stage of fear. And then maybe some of us have progressed to this stage of excitement and, you know, maybe almost unbridled enthusiasm and maybe not as intentionally or thoughtfully as we should have jumped into some different AI efforts and so forth, but it's now like we are in this reality stage. So what we've seen is over the past decades, know, waves of technology vendors have rushed into healthcare with different AI products, but many have exited just as quickly because they probably underestimated the complexity of the healthcare landscape. You know, it's one thing to apply AI to other industries, but healthcare, we know it, it's very complex and competing incentives and so forth. And so without a deep grasp of these realities, know, tools, pilots and so forth will fail to take root in scale. In fact, according to a recent Arcadia health survey, they estimate that only 18 % of health systems actually have mature AI programs. And then the other thing in this survey that I thought was so interesting, only 48 % of healthcare IT leaders believe their infrastructure can actually support AI.
Davida Kruger
Whoa.
Janice MacLeod
And so the single biggest barrier to successful AI implementation that everybody's talking about right now is this fragmentation of our health records and our data streams. And what healthcare leaders are learning is that investing time and resources in building a strong data infrastructure is critical to being able to support AI integration putting in place data governance, integrating those siloed data streams, and addressing data security, patient privacy, and so forth. And then also important, and this is the second thing you hear being talked about all the time, healthcare leaders are also recognizing that AI must be integrated into clinical workflows. That's developing the policies, the protocols, the roles and responsibilities, and so forth. Without that, again, these pilots just, they don't scale. We don't really integrate. Now, some other red flags though that just as individual clinicians that are maybe considering using an AI tool in our practice, particularly these large language models, we need to understand some limitations with them. You know, there is a lack of knowledge-based reasoning. There's a risk of incorrect or even falsified responses. What does that mean? Well, that means the clinician using that tool needs to absolutely review all output for accuracy because you will be liable for that. And that means you also need to be knowledgeable about the topic you're asking about so you can recognize what's wrong in that answer. And then there can be some training data set limitations that can perpetuate bias and further the care access gaps that we're always talking about. AI has potential to really help address that, but not if we're not really thoughtful in how we integrate AI. And then also we have to recognize currently the large language models are not regulated and there are legitimate patient privacy and cybersecurity concerns. So, Davida, in short, AI is not gonna fix healthcare. Humans who wisely use AI, recognizing its limitations, will.
Davida Kruger
Maybe I'm wrong, but it makes me think of how many years ago we started with electronic medical records. I think I hid under my desk thinking if I stayed there long enough, it would go away. But I had to learn it. And today I appreciate the value it's brought to clinical practice and what a better clinician I think I can be. And what I'm hearing you say is we need to be at the table as diabetes care and education specialists, nurse practitioners, or whatever it is, we need to be at the table to make sure what we do in life is included in the development of the AI for the system and that that would help us with a better understanding. So I super appreciate that and also the pitfalls that you've shared with us. But what role do you think the diabetes care and education specialist should play in shaping how AI is used in healthcare? Specifically, what is it? Do you really think we should need to be doing to make sure we don't miss that or that there's no gap in what we want to be doing with AI.
Janice MacLeod
Yeah, it's such an important question. And that's really the focus of when I'm speaking to groups and publications and so forth. So the question is, how do each one of us contribute to bringing the promise of AI into diabetes practice? And I believe that diabetes care and education specialists are uniquely positioned because we truly understand the importance of behavior change. We understand those non-clinical factors, the environment that the patient lives in, the reality that they're facing in their day-to-day life. We're also really good team-based care players and leaders, I think. So AI can process the data and help surface the patterns and so forth and help us to understand where we need to prioritize, what patients most need our care when. But we are there to provide that context, the empathy, bring the shared decision-making skills.
So I like to really encourage people to hold your current role and responsibilities loosely because they are going to change completely. You can count on that. A quote you often hear is AI won't replace clinicians, but clinicians who use AI will likely replace those who do not. Or another way of putting that, and this is a little more blunt, but I think it's true, diabetes care and education specialists who can be replaced by AI likely should be.
And so, okay, I know. So what I've observed is that AI is really sorting out those who are true experts in the field. True experts aren't threatened by their role completely changing their confidence and their expertise, and they want to be part of transforming healthcare. And so that's why I knew it was so important that we reframe that discussion board conversation that I mentioned at the outset. And then really developing a value mindset. And what do I mean by that? That's not just about value-based care and how we get paid, but it's how we think about and deliver care. And so in a time of rapid healthcare transformation, which we are in, we need to be able to clearly articulate the value of the services that we're providing. So I like to think of the Institute for Healthcare Improvement's quintuple aim and being able to describe very clearly. How are my services cost-effectively improving the health of the population I'm serving? And what am I doing to improve the experience of care for clinicians and patients alike? And also, what am I doing to expand access to my services? And these questions also become a good way to evaluate the AI tools that I'm bringing and how effective are they as I bring those into healthcare. So earn a seat at the table by beginning to build some AI literacy skills and then be that trusted and go-to clinical expert about the integration of AI in healthcare.
Davida Kruger
That's wonderful. And you know, I always believe if you're not offered a seat at the table, you just bring a chair and sit down. I think that's really important in our roles because sometimes people overlook us. And so you just bring your own chair, sit down and say, this is what I have to offer. And what you shared with us today, truly in our roles, we should be at the table. We should be helping to design it. And what I keep hearing you say, and I'm going to remember this and say it over and over to myself that we should not fear this, we should embrace it and be part of it. And I love hearing that and I love the knowledge you have, so thank you. What would you say to colleagues who feel overwhelmed or unsure about engaging with this kind of technology? You have convinced me I don't need to retire, that I can do this. What would you say to our colleagues so that we can get them on board, so we can get people to want to engage and be at the table?
Janice MacLeod
Yes, you know, people often do ask, how do I get started? You know, this just feels so overwhelming and it really can feel overwhelming and I share that sometimes. But take it one step at a time, right? And just begin to build AI literacy skills. And I give some ideas for how to do that in the handout that is available in the show notes. And remember, when we're working with our patients, we don't have them go home and do 10 things, we ask them to do one thing maybe. And that's what you should do. Just tomorrow, start doing one thing from that list. Just pick one thing. And one of the things that I have on that list is to begin perhaps following AI and health experts. And I have their names, their links listed there that you can follow on LinkedIn. I have found that to be extremely helpful. And then I would hope that diabetes care and education specialists as they begin to build their knowledge in this area would start engaging with these leaders and responding to some of the posts and giving their perspective as well. So becoming part of that conversation. And then also there is a link in the show notes to a way on LinkedIn that you can get a monthly summary of all of the relevant AI and digital health articles that have been published. And it's a very helpful, concise summary of the articles, not just here are all these articles you need to read. And so oftentimes I find it helpful to just read through that summary really quickly. It doesn't take that long. And that way, at least I'm aware of what's happening out there all over the healthcare landscape and even if I don't have time to dig in really big time, which most of us won't, I at least am understanding what is happening in the broader AI and healthcare field. So I encourage people to do that.
Another is thinking through all of the repetitive tasks you do every day. You know, just sit down and think about it at some point. And then what would be some ways you could potentially bring AI to help you to do that. But very importantly, a way we can lead in this and earn that seat at the table that you talked about, pick up our chair, set it down at the table, and become that AI clinical expert of choice. Working with your colleagues, hopefully from multiple disciplines, to think through what priority problems do we need AI to solve in our clinic or our department or our health system? And then what AI tools could we bring to address the problem? And then think about what education and change management would be needed for staff to prepare them for this, because this is a big shift for people. And then how will we evaluate the effectiveness of the tools that we brought in? What's the return on investment? And then, very importantly, what will we need to integrate the AI tools into our team's workflow so we can scale it and get it to be a part of the way we do and deliver care?
And then finally, always keeping an eye on how we can ensure ethical and inclusive use of AI to expand access to our services and serve all of our population, not just a few.
Davida Kruger
Outstanding and what great information. We are actually coming to the close of this amazing podcast and I think the information was wonderful and then we will have some materials in the resource section. But I want to close by some final thoughts and calls to action, Janice. Looking ahead five to 10 years, how do you think AI will change the way diabetes care is delivered?
And if the listeners take away just one message, one message about AI in diabetes, cardiometabolic care, what would you want that to be? What would you suggest?
Janice MacLeod
Well, I'll answer the first part first. What will we see in five to 10 years? I hope and do in fact believe that we will see a shift from reactive to truly proactive diabetes care. So instead of waiting for complications to occur, we're gonna identify risk earlier, we're gonna get out ahead of it. This will prompt earlier and more personalized intervention.
We'll also, I hope and do believe, we'll see greater integration of data. So integrating those disparate data streams that we deal with today. So data from our devices, from medications, from labs, lifestyle and so forth, often live in separate places and AI will help us bring all of that together so we can get a more complete and focused picture of the patient in front of us. And so we can make faster and more informed decisions.
At the same time, I'm also hoping care will become more team-based, more efficient, with AI helping us prioritize which patients need help when, also reducing the administrative burden, as we talked about, so clinicians can focus more on meaningful patient interactions. And most importantly, AI is not going to ever replace the human side of care. If anything, it should create more space for it. So, the future diabetes care is about combining intelligent tools with human expertise so we can deliver more personalized, accessible care.
So what would be perhaps a key takeaway message that I can leave with our listeners is that AI is coming into healthcare, whether we participate or not. The question is not so much whether these tools will shape the future of care, but will you be a part of guiding that transformation. I believe diabetes care and education specialists must be at the table because of our perspective. And if we step forward and engage with these technologies and ask thoughtful questions, help design better tools, advocate for person-centered care, we can be ones that help ensure that AI strengthens healthcare rather than complicates it. So this is our opportunity to lead the next chapter of healthcare transformation.
Davida Kruger
I think that's a great message to our listeners. Thank you so much for your knowledge. It's so impressive. And I want to thank our listeners for listening to the episode of Diabetes Care Conversations and Engaging with ADCES. You can find any resources related to this episode in the show notes. And remember, being an ADCES member gets you access to many resources, education, and networking opportunities.
Learn about the many benefits of ADCES membership at adces.org slash join. The information in this podcast is for informational purpose only. It may not be appropriate or applicable for individual circumstances. This podcast does not provide medical or professional advice and is not a substitute for consultation with a healthcare professional. Please consult your healthcare professional for any medical questions. And again, Janice, I want to thank you for being our guest today and for this amazing information.
Janice MacLeod
Thank you so much, Davida, thanks to our listeners.