This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.

Microsoft's advanced AI and speech recognition tools streamline processes and improve productivity, ensuring better patient outcomes. Experience secure and efficient healthcare delivery with Microsoft's innovative technology. Visit ThisWeekHealth. com slash Microsoft today and transform healthcare systems with the power of AI.

Today on Keynote

(Intro) just being able to take a specialist from one area and make them accessible to other areas has profound implications for healthcare in our country

My name is Bill Russell. I'm a former CIO for a 16 hospital system and creator of This Week Health, where we are dedicated to transforming healthcare one connection at a time. Our keynote show is designed to share conference level value with you every week.

(Main) All right. It is keynote. And today we're joined by Dirk Stanley, Dr. Dirk Stanley, Applied Clinical Informaticist and Chief Medical Information Officer at UConn Health. Dirk, welcome back to the show. Thank you. Thank you for having me.

the last time you were on, I took one of your blog posts and we just walked through the entire blog post. I don't know if you remember that show. It was actually one of my favorite shows. I really enjoyed I forget what the blog post was, but I learned more about what CMIOs do from that episode, I think, than any other episode, even probably my work directly with CMIOs over the years.

That was the top 15 ways that you know you're a clinical informaticist and like you start noticing things about optimizing the workflow for walking your dog. Having workflow design parties with your friends and, having interventions and stuff. Yeah, that hit a lot of nerves for a lot of people.

So I'm glad you enjoyed it.

Yeah. Are you still doing a lot of writing?

I figure that's a good way to help other people understand things or even if I can help somebody not, learn things the hard way, often there's an easier way to learn things. So I try to make it friendly and educational.

That's what I appreciate about you is your desire to really give back to the industry and share your knowledge.

I'm curious as you're walking around and you're having a conversation and you're doing something, you're sitting there, did you sit there and go, Ooh, that would be a good blog post.

Yeah, that's often, I'm like, yeah, there's, I know there's someone else who could benefit from knowing this. So that's what I do.

Yeah.

All right. For those, we talked about this a little bit, but you started your career at IT. you were a physician before you were in it, right? No,

long story short I was one of those sort of, I, I know you're gonna find this hard to believe, but I was one of those geeky kids, right?

And that led me to, one day, I got a job. I was actually 16 when I got my first actual IT job for a small software company. They had about 100 employees. And they were looking for a high school kid to basically help do, hardware installs and, network patch panels.

This is all in the days before Ethernet and, IP addresses and stuff like that. I supported the technology for a small company. It was chemical process software company. And I was the kid in the, server room. I did the security for, creating accounts for people, managing the accounts.

I did file management. I did backups. I did that patch pamper. I'm

It's their market research was like some of the best in the world, things of paper. And I found out that they actually gave us the SQL database, but the team preferred paper. And so I started doing searches in SQL. And all of a sudden the people kept calling me like, how are you getting that information so fast?

It's in the computer. It's it's in there. Yeah, it's in there. Like we can get

it out of there.

it did not matter that you were 18 or 21 years old. They were just like, you know how to use this thing. You're the best.

And so I learned a lot. I was like on the help desk. So I remember I've been on the receiving end of help desk calls where it, like a salesman would be in the middle of, a pitch in another state and it would be like, Dirk, I need the modem, like I'm trying to do, this is a five million dollar deal, like I need to close it.

Other operators in Europe and I grew up in a another weird thing. I grew up in a German speaking household. So they were like, you speak German. We have operators in Germany. We need you to troubleshoot the network with the operators in Germany. So I was suddenly I'm like in this room. A lot of big screens and maps of the world and packets going around.

And, I did that for a while. And then I studied business in college. I didn't, I wasn't totally sure. I thought maybe I'll do something with computers. I don't know exactly what. And then When I graduated from college, I was honestly a little burnt out. I also did some consulting on my own for a while.

But listen, before I speak to your parents, I have to let you know, your parents want you to go out and do something with you. You can't just live in the basement. You have to go out and get a job. don't know what I want to do, like an angsty 20 something, I'm like, I don't know what I want to do.

She said, look, I work in a hospital. Why don't you come and work here where you could just volunteer here in the hospital. So I spent about a year volunteering at the Westchester Medical Center. And of course, as soon as they looked at my resume, they were like, oh, you're an IT person. So I ended up in the IT department there.

And then. Their data quality person quit one day and they said, Dirk, we need somebody who's good at math. Are you good at math? And I said, yeah, I used to tutor math. And they were like, and how about computers? I'm like, yeah, no, I feel comfortable. Do you understand things like databases and stuff? Yeah, no, I understand that.

I didn't understand the words that people were using. So one day I'm actually, I admit I'm the guy. Who raised his hand, as a young 20 something, raised his hand in front of a large group of doctors and said, Hi I'm Dirk Stanley. I just want to understand when we're talking about central line infections.

Why are we always talking about the central line infections and never the infections on the left and never the infections on the right? And they all looked at me and they're like, kid, you're an IT person, you stay on that side of the screen and we'll take care of the medicine, you're, just don't hurt yourself.

Like what why would we ever need an IT person in medicine? And I didn't have a good answer! I really didn't. And they would like, they would look at me and be like this does not make sense to us like, we do medicine and you do IT, right? So After trying unsuccessfully to get into an American medical school, I went overseas.

I'm a graduate of St. George in Grenada. I went overseas to medical school and suddenly I'm, on an island and I'm reading like that's what a central line is and that's what methicillin is and that's what methicillin resistant staph, the staph aureus, like now I understand what all these doctors were talking about.

You're putting something into database A and you're printing it out and putting it, then re entering it into database B, and then you're printing it out and you're re entering it into database C. You can't do that. You're creating duplicate records. You don't even have a patient identifier. You're creating duplicate records.

Some poor quality person, their graph, instead of looking like this, their graph is going to look like this because all you're doing is creating duplicate records. And some now the IT people were like no. We do IT, and you do medical. Oh, man. Dr. Stanley, we, you don't that's how I ended up in clinical informatics.

It's the one place. Where both sides hate you.

going to give you one more question around this and then we're going to go into futures a little bit. I've never gotten that opportunity with you and I'd love to do it. Medical, IT, but more and more I'm talking to IT leaders and they feel like they're sociologists.

What's it like being a CMIO these days?

You

know,

it's it's a little bit of, I would say, all of the above. I think a lot of CMIOs when they start, myself included, you initially, you're, you look at the screen and you look at the data and you go oh, that either, that's right or that's not right.

And you're constantly looking at the, check boxes that are like on the screen or the numbers that are on the screen or the words, but as you start to get into it, you start to learn oh, wait, something isn't right on the screen. Wait, why is it not right on the screen? Like, how did it get built into the screen?

And so if you're not happy about the way it's working, or if you're happy about the way it's working, it's because there's a human being who like, Made a decision, had a conversation, had a meeting, maybe a group had a meeting. Did the right people have the meeting? Did the wrong people have the meeting?

Did they include everyone that they needed to make a good decision? And then you start to realize Oh it's about like how human beings interact and it's about teamwork and governance and then you really start to explore the policies and the bylaws and you're like, and then the regulations and how does that fit in?

And then there's the very real task of what do you do? When there's, 12 different people knocking at your door at the same time, how do you know who to answer, right? And these are common things I think everyone can relate to, but you really start to see like that managing those things outside the computer are so important.

Coming through Meaningful Use, it felt a little chaotic coming through Meaningful Use. There's a lot of things we didn't know, a lot of things that were still outside the EHR, and some things were inside the EHR. We didn't recognize how many conversations, like you just described, we were going to have.

And then if you practiced across state lines, it's yeah you create it that way for California. But if you're practicing in Texas, it looks this way. And a child at this age is that, all these things come together. I'm not sure we appreciated the complexity that was going to be associated with bringing all this stuff together.

But do you feel like we're pretty far away from that now? Do you feel like we have a handle on the basics of getting information in, aggregating that information, defining that information? Are we still challenged with that a little bit?

Certainly progress has been made, no doubt, right?

If you're studying hospital admissions, some hospitals might define an admission differently than other people, and then it gets into like the regulations and what is an admission if it's an observation status versus a true inpatient status? So there's so many like. Things that to get, to aggregate data at, even inside an organization, but even regionally to aggregate data like a lot of states do, or departments of public health, or even nationally to aggregate data, requires a lot more attention to detail that I think most people have planned for.

And. Only if you configure it the right way. And I think that's another thing that people still commonly confuse, the configuration of an electronic medical record from the actual software itself. And the configuration plays a very big role in everything from, reporting to usability to treating patients and stuff.

Yeah, I think that's the thing that would surprise, like when I talk to people who are outside of health care, I think the thing that surprises them is that a system can be implemented differently at this hospital, and this hospital, and that's within the same system. Potentially could be implemented a little different because for various conversations that go on, you have this physician group over here that's look, I don't know how you learned to practice medicine, but I'm not following that.

That protocol or that order set or that whatever.

There's a lot of regulations, but what's interesting is. We don't really have great regulations for some really tiny little details. And most people would walk right by it and think that's not really, that's not a big issue. But I worry about those like little tiny standards. If I had to give you analogy, it would be like defining what an inch is.

In the real world, we have the, whatever the Bureau of Weights and Measurements are, there's, I think there's different, even international groups that worry about what a kilogram is, or what, worry about what a mile is, or, and we're missing some of those sort of standards, and we have things for milliliters, and, kilograms, and those standards we have, but there's standards of what is an H& P?

That is the bread and butter. Every medical student learns how to do a history and physical. So you have this sort of general sense of I know what a history and physical is. I've done them hundreds of times since my first day of medical school. But when you actually look in the real world, there's histories and physicals that you do when you're admitting somebody.

There's histories and physicals you do when you're planning a surgery. There's histories and physicals you do the before the surgery to know if anything has changed. There's even histories and physicals that aren't done by the primary person doing it, but like a secondary specialist.

So does

the context change

But once you create the terminology and the concepts you wind up the concepts. This is the one we do before we, plan a surgery. This is the one we do the day before the surgery. This is the one we do at the time of admission. All of a sudden, the arguments go away. People are like, Oh yeah, that's the one I'm doing.

And that's the one you're doing. And so it's really important that we, have words to describe these concepts, and healthcare in general really struggles with some of those words, so a lot of my job ironically, is translating because it becomes important then when the clinical people are saying, I, I won HNP in EPIC, then I have to go to the EPIC people and help them understand where in the software we're talking about doing the configuration.

It can create the note. Then I heard somebody say, oh, it can create the note and code it and it can take it all the way through to billing. And I thought, Wow that's an amazing efficiency gain that has just happened, but the terminology matters a

lot. It does. And so we're like many other places, we're currently piloting this we're going through our, due diligence and our testing and we're working this out.

I will say just observationally, what I know so far, And I am not, certainly not one of the country's experts at this. So like there, there are people who are much more experienced at this, but I'll just share it from my own perspective. The AI is very good at drafting things, but it's not good at approving things, right?

Sometimes you have to even change your, your pattern of speech for people's different, either educational needs if you're dealing with kids. So for, it's important to get those, pick up those nuances of the conversation, and I've seen some, other informatics people have shared some interesting examples of where the AI is going.

Bye. made certain assumptions of what it was listening to, but it was taken out of context. And, so you still need a human being to

sign off on. So talk to me about cognitive load here. So what's I'm trying to think of the bias I'm looking for, but the cognitive load.

there.

Reading pages, scrolling through it. You're like whatever you accept.

Isn't that a very real problem? Like I'm a very busy doctor and it created the note and I just go, yeah, all yes. Approve.

Yes. And I'll say even those scenarios, there's so many nuances of workflows that impact that willingness to click accept whether or not you've read it.

And it's a challenge. I'm not saying it's going to be like solving all of healthcare. All right, let me,

let me back up. I'm taking you way into the weeds, but where does AI hold the most promise?

Ooh, that's a good question. Think it's very good at summarizing things. I'm very interested in the ambient AI listings.

So I, and many providers in my organization are very interested in that. And that seems very promising, at least the early stuff I've heard from other people and what we're seeing internally, it looks like it's very promising, assuming everyone. It does what they're, reviews and reads it.

And rather than reading 30 years worth of notes, just being able to ask a question like that's the kind of stuff, even just being able to search for that. And it's funny because prior to ai. A lot of the search mechanisms depended on text, and even if you had optical character recognition and scanning, you're still looking for an actual string of text, which sometimes fails in clinical settings, because, one person calls it a CBC.

So That's the kind of stuff that I think there's a big opportunity for time saving. A lot of doctor's time is just searching through the chart trying to understand where has this patient been? What do I need to worry about as we go into the future?

Yeah, there's a reason that most doctors have glasses.

They are just, they're professional readers, almost. And I hope I didn't offend anybody with that, but remember the first time I was in an EHR and I'm sitting there with the doctor and he goes yeah here's all this stuff. And that's back when they were all like PDFs and stuff.

And I'm like, are you gotta be kidding me? He's no. And to be honest with you, that person who's sitting like who's coming in here next, they expect me to know what's in all of these documents

talk to me a little bit about we have computer vision as a starting actually, starting is probably not even the right word.

curious if you guys are thinking or playing around with any of that.

Actually we Right now, we just started a remote sitting. It's it's not really like an AI sort of thing, but there's especially for people who are big fall risks, or people who are real risk of alcohol withdrawal or drug withdrawal.

It's very hard, especially post Everyone is struggling for staff, right? And, there's other trends that are going on in the country, too. It's not even just COVID. The fact that the baby boomers are aging, and, the people to replace them is getting smaller and smaller.

I knew at the time that they'd be looking for technology to really help [00:24:00] augment the staffing issues. Going back to the cameras we did start a pilot where we're actually implementing those cameras, but they're still more or less, it's not really using an AI solution, it's using a human being who's actually there monitoring And helping to watch the patients in a very safe, secure HIPAA privacy, oriented way that really respects the patient's privacy.

But at the same time, if it looks like they're getting into trouble that they have they have a way of notifying the nurse about it.

Yeah, one of our 229 meetings, we just had the, somebody was building a new tower and they shared a video and I was like. Is there a camera in every room?

They're like, there's absolutely a camera in every room. I'm like, Wow. And I had 50 questions, but wasn't the appropriate time, but he's just saying, look we have specialists all over the city and those specialists are able to go right into that room, have a conversation. And I'm like that's really fascinating to me.

Alright, let's get into another area, which is data. Help me understand, how are we doing with data? How is data changing the way healthcare providers approach patient care at this point?

it's funny, this was a topic at one of the last conferences that I went to, and one of my favorite things is sitting around with other CMIOs and we talk about this question a lot in the ocean of data, like how many, what are we using?

There's so much information inside most electronic medical records, so pulling out the What's relevant and useful and accurate it takes a lot of work, but I think it boils down to a maturity of, from an IT strategy standpoint. And I always, this is another thing I commonly talk with other informatics people.

When I meet other informatics people, I love asking them this question. I say so of all the things, that you're interested in. Let's talk about all the data out things, like getting analytics out, and reporting out, and the patient portal, and clinical decision support, and the patient centered home.

All that stuff that you get out of the computer. Let's talk about all the things that you do to get information into the computer, like CPOE, like usability, like even your governance and your policies and your, what are the workflows that you designed to put stuff in? Because you really can't separate the data out from the data in.

I worry more about the usability. And I think like we, As a industry, we have to recognize that both branches are important, and you can't ignore one for the other. So when I approach a lot of workflows, I start off with, let's talk about how are we going to collect the information for that we are, and how are we going to analyze that information.

And you have to really like, I think a mature way of approaching it is, Recognizing both of those at the same time that helps you collect better data and it helps you retrieve better data and analyze better data. So by doing that and standardizing the workflows, , you're not always chasing like data normalization strategies.

So I always try to bring to my designs. I bring both. I recognize both before I start any projects.

I'm not even sure what the question is. Are there emerging trends in workflow design that you're looking at, or is workflow design, just, it is what it is. Or is it changing?

I think that's something that also since Meaningful Use, people are getting much more like savvy about workflow. And I'll just give a shout out to Charles Webster. I don't know. He writes a lot about workflow. He's a doctor who, goes under The social media handle, Wearflow.

And then what are the problems with it? What don't you like about it? What are the things that, give you headaches and heartache? And then you go to re redesign it, and now you're like, oh, so if the nurse does this first, then the doctor can do this, then the pharmacist can do this, then, And so you'd be surprised how there are a lot of workflows that just a simple, reordering of sequence of events has a massive change in usability, in people's satisfaction, in outcomes.

But you need people to do that sort of work not everyone, as you said before, A lot of people are like I got into the computer because I want to type at the computer, but you almost need to take a human factors engineering approach to it. Because once you have the workflow, then you can say, okay, what data elements do I need for this step?

It's a great job. I really enjoy it.

It's interesting. We'll close with this, the CMI, CMIO role. Used to be an incredibly, sure it's still a very challenging role, but it was one of those roles that people didn't last in because we were doing EHR implementations.

There was a lot of very dynamic conversations that needed to happen. And sometimes the systems didn't work and those kinds of things. I remember and this person is no longer a CIO. So maybe this is a odd thing to say, but I remember him essentially communicating that the CMIO was expendable.

Like he put one in. They would last for two years and then he'd put another one in because the, there's an academic medical center and they just chewed that person up and he just knows he's going to have to put another person in two years. I was like, man, is it really that hard of a job?

I've seen plenty of examples of what you're referring to, I think the hard part, especially, early on, and I agree with you, we're in like CMIO 2.

0 now, early on, there were a lot of doctors who got into it. I think they may not have, there are a lot of doctors who are very enthusiastic about technology and I appreciate that. I love meeting people. I talk to doctors all day long and they have, Ideas of how to make things better, and I love hearing good ideas and there's so many people with good ideas.

But you, but learning how to channel those good ideas and turn them into actual like conversations and meetings and agendas and project plans and blueprints. and that's the hard part that I think a lot of CMIO struggle with. So if you focus mostly on the, I'm going to use this word salesmanship, if you will, oh this is going to be great and you're going to love it, but you don't understand.

And how do we, how do I make a great user experience? How do I make great clinical decision support? How do I make something that people don't complain about clicking endlessly? A lot of workflow redesign reduces clicks. One of the workflows I'm working on right now is going to have a dramatic reduction in clicks.

I'm so excited about it. If you enjoy making good user experiences for not just the doctors, but for the nurses and for the pharmacists, and maybe most importantly, For the patients it's so rewarding and I really enjoy it.

Dirk, I love these conversations. I appreciate it. I look forward to our next one already.

It's gonna, it's gonna be fun. Thank you again for coming on the show and sharing your experience. Really appreciate it. Thank you so

much for inviting me.

Good to see you.