Latest / The Clinical Realist / Who Gets to Say No in Health IT, with Anthony Guerra

Transcript

Who Gets to Say No in Health IT, with Anthony Guerra · The Clinical Realist

• 0:00
  SPEAKER_00: My guest today has asked the same question in some form about 600 times.
• 0:05
  SPEAKER_00: So Anthony Guerra founded Health System CIO in 2010 and has been covering health information technology since 2005.
• 0:13
  SPEAKER_00: So 16 years of sitting across from the people who sign off on what tech actually enters the hospital.
• 0:19
  SPEAKER_00: So CIOs, CMIOs, chief information security officers, 600 published interviews.
• 0:25
  SPEAKER_00: So this is maybe three or four full technology cycles, depending on how you count them.
• 0:29
  SPEAKER_00: So this makes Anthony one of the people I can ask this particular question of.
• 0:34
  SPEAKER_00: Everyone in health tech is fluent in what a CIO says about AI right now, but almost no one can tell you how the answer has moved.
• 0:42
  SPEAKER_00: And not just the vocabulary.
• 0:43
  SPEAKER_00: The vocabulary changes every 18 months, but the posture, where the hedging went, what they stopped apologizing for, and what they're still not willing to say into a microphone.
• 0:52
  SPEAKER_00: So I spend my time on both the business and the clinical side of that decision, where the thing a CIO approved becomes a thing a physician has to use at two in the afternoon with 14 people waiting.
• 1:03
  SPEAKER_00: So I'm going to take his 16 years seriously and then push on exactly where he reads and mine actually comes apart there.
• 1:10
  SPEAKER_00: So Anthony, welcome to the Clinical Realist.
• 1:13
  SPEAKER_01: It's a pleasure to be here, Dr.
• 1:14
  SPEAKER_01: Matt.
• 1:14
  SPEAKER_01: Thanks for inviting me.
• 1:16
  SPEAKER_00: Absolutely.
• 1:16
  SPEAKER_00: So I want to start with a thing, maybe only you can answer.
• 1:19
  SPEAKER_00: You've asked hundreds of CIOs about emerging technology across maybe three or four big tech cycles.
• 1:26
  SPEAKER_00: So when the topic became AI, what changed in how they answered that?
• 1:30
  SPEAKER_00: And I don't mean just the words they use, but the posture.
• 1:33
  SPEAKER_00: Where did the hedging move?
• 1:36
  SPEAKER_01: Well, uh, I don't know if uh hedging would be the right word, but in terms of how things have have changed, in my opinion, is there's there's two elements here.
• 1:47
  SPEAKER_01: One is uh a degree of sort of cybersecurity risk that I think is kind of through the roof.
• 1:54
  SPEAKER_01: Uh it's really unknown.
• 1:56
  SPEAKER_01: Um and so that's where you could possibly see some hedging starting or some caution.
• 2:01
  SPEAKER_01: Really, the big fix here that I've seen is uh governance, and that solves a lot of downstream problems.
• 2:09
  SPEAKER_01: So what they're what they're trying to do more and more is create a unified governance structure because so much of this depends on deep integration and agreement and shared responsibility with operations.
• 2:25
  SPEAKER_01: So, you know, we're far from the days of selecting a technology and it's IT's role to get it up and running, and we're done.
• 2:34
  SPEAKER_01: It's much more of an integrated kind of operation where everyone needs to be at the table making decisions about who's responsible for what.
• 2:44
  SPEAKER_01: We just put out a story today, which is one of sort of the major issues on the table, which is who owns the responsibility for agentic activity?
• 2:55
  SPEAKER_01: Uh, who owns the model, right?
• 2:58
  SPEAKER_01: So um, and this is a gray area.
• 3:00
  SPEAKER_01: You know, gray areas to me are the most interesting areas.
• 3:04
  SPEAKER_01: Things that are on either end of the spectrum are usually easily resolved.
• 3:08
  SPEAKER_01: And I especially am deeply interested in gray areas of responsibility, where again you have sort of clear definitions on one extreme or the other of who owns what, and it's that gray melding in the middle where sometimes it's I thought you had it, no, I thought you had it, and and nobody had it, especially when nobody wants to own it, right?
• 3:28
  SPEAKER_01: When everyone's like, I'm doing a lot.
• 3:31
  SPEAKER_01: So this is leadership.
• 3:32
  SPEAKER_01: Leadership is not saying because I see a problem, I'm going to own it, but it's identifying it to everyone and saying there's something here, and I don't that's you know, it's not like when you ask a question in school and the teacher says, Well, now do a report on it, and you kind of bought yourself a problem.
• 3:51
  SPEAKER_01: We don't want that.
• 3:52
  SPEAKER_01: We want to be able to identify problems and have shared ownership in the fix so that everyone's willing to identify problems.
• 4:00
  SPEAKER_01: So that's some of the issues going on there around agentic AI.
• 4:04
  SPEAKER_01: Extreme risk, which again, this is all in real time, people grappling with this, extreme risk about what these agents can do and will do, uh, and a better coordination with operations required from AI because it's so clinical, right?
• 4:20
  SPEAKER_01: A lot of it is so clinical.
• 4:22
  SPEAKER_01: So those are some thoughts there.
• 4:24
  SPEAKER_00: I guess I'd push on the gray zone because I feel like AI governance is absolutely gray the whole bit of it.
• 4:32
  SPEAKER_00: And we'll see at some large institutions.
• 4:34
  SPEAKER_00: They've taken their procurement processes and they've really kind of pushed it into more of an AI governance model.
• 4:40
  SPEAKER_00: But even so, if you look at titles today, not many people have a chief AI officer, or not many folks have specific departments that are really aiming towards making AI reasonable and being that point of coordination.
• 4:56
  SPEAKER_00: And so while we all talk about yes, everyone needs to come to the table, in theory, that's wonderful.
• 5:03
  SPEAKER_00: I'm seeing in practice, though, it's becoming very difficult.
• 5:06
  SPEAKER_00: And while we may have performative meetings, I'm still concerned that maybe we are coordinating like we should.
• 5:12
  SPEAKER_00: How are you seeing that?
• 5:14
  SPEAKER_01: I think again, it's in real time.
• 5:16
  SPEAKER_01: They're doing their best, they're figuring this out on the fly.
• 5:18
  SPEAKER_01: You need to try, you need to attempt, you need to have discussions with people and look at your governance and say, continually, is it working?
• 5:30
  SPEAKER_01: Is it not?
• 5:31
  SPEAKER_01: This is another area of deep interest to me.
• 5:34
  SPEAKER_01: Is how governance stays flexible.
• 5:37
  SPEAKER_01: Who looks on a governance structure from the outside from top down?
• 5:42
  SPEAKER_01: If everyone's in it, who's looking top down and deciding this is you know, it's a Goldilocks problem with governance.
• 5:49
  SPEAKER_01: You could have too much governance and things go nowhere, and you could have too little governance.
• 5:54
  SPEAKER_01: Right?
• 5:55
  SPEAKER_01: So it's and it's it's something that you need to design it so it's responsive to feedback and improvement, and that, well, we're missing this role.
• 6:07
  SPEAKER_01: You know, we constantly are hitting a wall here, and in decision making, we're missing this role.
• 6:12
  SPEAKER_01: Or we don't have decision makers here.
• 6:14
  SPEAKER_01: We have people that need to go back and have discussions.
• 6:18
  SPEAKER_01: Can we move the ball forward?
• 6:20
  SPEAKER_01: Or are we always tabling everything to the next meeting, which is the worst, right?
• 6:25
  SPEAKER_01: So it's I just think somehow there has to be a continuous learning environment around governance where we're constantly trying to fine-tune it.
• 6:35
  SPEAKER_01: And also the underlying issues are changing constantly.
• 6:40
  SPEAKER_01: So whereas one role or department didn't need to be included now, now things are changing, and now they do.
• 6:46
  SPEAKER_01: We need another voice here.
• 6:48
  SPEAKER_01: So it's just that ob again, that continuous learning environment, getting better and better at governments.
• 6:53
  SPEAKER_01: I do believe those human-to-human conversations, those relationships combined with governance, can solve a lot of problem.
• 7:02
  SPEAKER_01: I think, put it this way, if you have good governance, you have a chance.
• 7:06
  SPEAKER_01: If you have bad governance, you have no chance.
• 7:10
  SPEAKER_01: You're just you're really in big trouble.
• 7:12
  SPEAKER_01: So that that's that's a big key there, I think.
• 7:16
  SPEAKER_00: I think I'll end with I hope folks can figure out if they have good or bad governance so they'll keep moving that ball forward.
• 7:24
  SPEAKER_00: And that's not always easy, as you said, unless you're looking from the outside in.
• 7:27
  SPEAKER_00: So interesting.
• 7:29
  SPEAKER_01: Well, the ball i the ball another way to look at it is the ball is going to keep moving forward.
• 7:35
  SPEAKER_01: The clinicians are gonna keep moving the ball forward, and they do, and this is called gray or shadow IT.
• 7:42
  SPEAKER_01: This is called I need to text with my colleagues, or my colleagues are texting with me.
• 7:48
  SPEAKER_01: This is an efficient way to communicate.
• 7:50
  SPEAKER_01: This happened, you know, five years ago, ten years ago with texting.
• 7:53
  SPEAKER_01: Everybody was texting because the clinicians want to take care of patients, and they're communicating.
• 7:58
  SPEAKER_01: And IT says, Oh my god, this is not secure, this is outside.
• 8:02
  SPEAKER_01: So we come up, IT comes up with tools and rolls out tools and says, Hey, please, would you mind you could do that thing now that you want to do, but do it over here.
• 8:10
  SPEAKER_01: Uh, and I and the clinicians usually go, Okay, sure, we're happy to do it over here, but we're gonna do it because we need to take care of our patients.
• 8:18
  SPEAKER_01: There's arguments on both sides there, but that's just an honest dynamic, and I've heard it said in many interviews I've done where the best IT executives use this shadow IT as an indicator of what they're missing and respond to it positively by, as I said, supplying that tool.
• 8:38
  SPEAKER_01: They use it as an indicator of what's missing, and they use it constructively.
• 8:42
  SPEAKER_00: I like that.
• 8:43
  SPEAKER_00: So, from a design perspective, I always like the analogy of like college campuses, don't bother paving sidewalks, see where the students go, and then pave those places instead.
• 8:52
  SPEAKER_01: That that's good if you can manage the risk, because when they're when they're not walking on the paths, to use your example, there's risk there.
• 9:02
  SPEAKER_01: So there's a time element to remediate and fix that, and that can be, you know, dangerous.
• 9:09
  SPEAKER_01: Ideally, you're getting ahead of things more, you're having better interaction, relationships, governance, conversations with your users, so that they don't have to tell you that way.
• 9:21
  SPEAKER_01: They could tell you, and it's going both ways.
• 9:26
  SPEAKER_01: You're observing, you're in their meetings, you're understanding their pain points, and you're coming to them and saying, Oh, by the way, we've got this new secure texting thing you can do, right?
• 9:37
  SPEAKER_01: So that's getting ahead of it.
• 9:38
  SPEAKER_01: But this takes deep, deep relationships, coordination, and understanding of operations.
• 9:45
  SPEAKER_01: You must get customer service oriented.
• 9:49
  SPEAKER_01: You must realize your role.
• 9:52
  SPEAKER_01: If you realize your role, you can be effective in it.
• 9:55
  SPEAKER_01: And you are there to serve and empower and protect the organization, but to serve and empower.
• 10:02
  SPEAKER_01: So this, and again, I'm I'm funneling all the things I hear that make sense to me, that resonate, that are repeated by people being extremely effective.
• 10:11
  SPEAKER_01: So I take all that, and that's when I tell you about what I think the best practices are, that's what I'm doing.
• 10:16
  SPEAKER_00: No, I get that.
• 10:17
  SPEAKER_00: Well, you know, I'm very passionate about pilots, and I know you are too.
• 10:22
  SPEAKER_00: And everyone's complaining about pilot fatigue.
• 10:25
  SPEAKER_00: And in some of the published work you've done recently, you argued that the problem's not too many pilots, it's too many different kinds of pilots, and that learnings don't accumulate and the portfolio is very scattered.
• 10:37
  SPEAKER_00: So I'd love to hear your thoughts on that, because I want to push on it.
• 10:40
  SPEAKER_00: You know, how do we combine pilots into groups?
• 10:44
  SPEAKER_00: How how would that work?
• 10:45
  SPEAKER_00: What do you mean by too many pilots in the wrong spaces?
• 10:48
  SPEAKER_01: Well, I mean I'd say so I might take it a different way.
• 10:53
  SPEAKER_01: You know, again, what is what is the role of the IT executive?
• 10:56
  SPEAKER_01: It's to understand the pain points, the friction, the difficulties, both with existing tools and with existing workflows that may benefit from new tools, right?
• 11:06
  SPEAKER_01: So the deeper I understand what you're trying to do and what you're going through, the more I can either satisfy your requests or proactively tell you what's available to deal with your problems.
• 11:17
  SPEAKER_01: I want to make your life better and your work better.
• 11:20
  SPEAKER_01: What is a pilot?
• 11:21
  SPEAKER_01: A pilot is an attempt, is it is a step further on that road, right?
• 11:26
  SPEAKER_01: So I understand there's a request for a new product.
• 11:29
  SPEAKER_01: This would be a new product.
• 11:30
  SPEAKER_01: Either I've suggested it or you've suggested it, but somehow it's gotten to that point where we've together come up with a vendor, but we want to try it out on a small scale, right?
• 11:42
  SPEAKER_01: So what we want to do is if if this is something we're going to do continuously, we want to have structure, repeatable structure around it.
• 11:49
  SPEAKER_01: This goes for anything.
• 11:51
  SPEAKER_01: Repeatable structure.
• 11:52
  SPEAKER_01: We don't want to have it be different every time and never learn from what we've done.
• 11:58
  SPEAKER_01: So you might not have one flavor, you might have three flavors of different types of pilots, and then we know, oh, this one's coming in, this is going to be track A or track B or track C.
• 12:08
  SPEAKER_01: What happens when you do that is you learn and you start to incorporate learnings, and you say, okay, next time we just did track A.
• 12:15
  SPEAKER_01: Let's have a post-mortem, figure out what worked and what didn't.
• 12:18
  SPEAKER_01: Now we make track A better.
• 12:20
  SPEAKER_01: So when I talk about probably different kinds of pilots, that's what I'm talking about.
• 12:25
  SPEAKER_01: But in terms of the issue of, you know, pilot fatigue only happens when you're foisting things upon people, right?
• 12:33
  SPEAKER_01: Nobody had pilot fatigue for a product that they wanted and asked for.
• 12:37
  SPEAKER_01: The problem with pilots is uh you get a false sense of security sometimes.
• 12:43
  SPEAKER_01: So when you have a pilot, you're responding probably to a specific department's request for a specific tool, and you're piloting it in that department with users who are interested in piloting it piloting it.
• 12:56
  SPEAKER_01: All that creates sort of a false sense of security because it's a very unique dynamic, right?
• 13:01
  SPEAKER_01: So now we say, unfortunately, sometimes we say it's time to scale.
• 13:06
  SPEAKER_01: And what does scale mean?
• 13:07
  SPEAKER_01: Scale means we're bringing something to people who didn't ask for it.
• 13:11
  SPEAKER_01: And we think just because it worked with the people who asked for it and the people who volunteered to try it, now we're rolling it out to different people that never asked for it with a tool they might not have picked.
• 13:21
  SPEAKER_01: We're trying to get it to work in other places.
• 13:24
  SPEAKER_01: There's many reasons that it won't work, right?
• 13:27
  SPEAKER_01: On paper, it's different.
• 13:29
  SPEAKER_01: So you have to be very sensitive to thinking about does this translate across to other departments?
• 13:36
  SPEAKER_01: And am I giving it the attention it got in the first pilot?
• 13:40
  SPEAKER_01: So do I giving it every right reason to succeed?
• 13:43
  SPEAKER_01: You think of a plant.
• 13:45
  SPEAKER_01: I planted this one with care in the right soil with everything, and now I'm throwing it in a different in sand, and I didn't water it as much and all that, and it didn't work.
• 13:54
  SPEAKER_01: And I go, why?
• 13:55
  SPEAKER_01: Why didn't it scale?
• 13:56
  SPEAKER_01: Well, I mean, these are some things to think about.
• 13:59
  SPEAKER_00: So I would absolutely agree.
• 14:01
  SPEAKER_00: I find that a lot of times, you know, when you are doing a pilot, you are all of a sudden the vendor's number one customer.
• 14:07
  SPEAKER_00: You have all their support on that side.
• 14:09
  SPEAKER_00: You probably have a project team from IT, you have your champions in place.
• 14:13
  SPEAKER_00: And the success of a pilot, like you mentioned, is not proof that scale will work.
• 14:18
  SPEAKER_00: It is proof that the pilot worked.
• 14:20
  SPEAKER_00: And I think that jump from we'll say positive successful pilot to positive successful scaling to other departments, other buildings, other anything that's different than that pilot is really tricky because now that one champion can't be in 10 places.
• 14:35
  SPEAKER_00: Now the vendor is hands off.
• 14:38
  SPEAKER_00: And so that gap, I find, with the leaders I work with tends to be the hardest piece because you need to do that jump to scale to figure out what that takes.
• 14:49
  SPEAKER_00: But if this is for nursing, then that scaling leap might be different than if it's products for physicians or if it's products for RCM.
• 14:58
  SPEAKER_00: And we've done great at doing pilots, but I don't know that we've done great at taking pilots and truly scaling them successfully.
• 15:07
  SPEAKER_00: And so those kind of muscles and the change management required for it to me feels a little nebulous still.
• 15:14
  SPEAKER_01: Absolutely.
• 15:15
  SPEAKER_01: And and it there's a middle ground between it worked in this department and it needs to work everywhere, right?
• 15:22
  SPEAKER_01: Maybe it's going to work in some places and not others.
• 15:25
  SPEAKER_01: So let's try and figure out what this is suited for.
• 15:29
  SPEAKER_01: And maybe again, it's suited for three or four areas, but not for the other one.
• 15:34
  SPEAKER_01: So we need to be, I guess, a little more tailored to your point, a little more nuanced, a little more tailored.
• 15:38
  SPEAKER_01: Understand that we can't go from to use a clinical term, and and I'm sure I'll botch this, to go from like an ICU environment with all the care and attention and the great ratio to, you know, outpatient treatment.
• 15:51
  SPEAKER_01: There's there's a difference in care levels, so we need some sort of middle ground.
• 15:55
  SPEAKER_00: Well, it's interesting.
• 15:56
  SPEAKER_00: I know you've done over 600 interviews.
• 15:58
  SPEAKER_00: That's a lot of interviews.
• 16:00
  SPEAKER_00: So when you think about what CIOs have been telling you, what are the things that, and again, no names.
• 16:07
  SPEAKER_00: What are the things that you're hearing that they don't want to say themselves on an interview, but is the sentiment that's coming through.
• 16:15
  SPEAKER_01: Well, I mean, this is not sort of secret information, but I think the biggest struggle the biggest and and this is part of their job.
• 16:25
  SPEAKER_01: This is what they're doing.
• 16:26
  SPEAKER_01: It's just a a dynamic, an unfortunate dynamic that's the result of sort of technology at a deeper level.
• 16:32
  SPEAKER_01: I don't think anybody's happy with being trapped in a EHR only mindset.
• 16:39
  SPEAKER_00: Alright, just drop the mic.
• 16:41
  SPEAKER_00: I think we're done here, Anthony.
• 16:42
  SPEAKER_00: Thank you.
• 16:42
  SPEAKER_01: That was the best They shouldn't be.
• 16:46
  SPEAKER_01: They shouldn't be happy, and I don't think they are happy, but you know, if you are and most of them will tell you this, the best kind of CIOs are managing this dynamic with care and diligence and sensitivity, and they they want to be as responsive to their users as possible within reason.
• 17:13
  SPEAKER_01: Now, what's the fear?
• 17:14
  SPEAKER_01: The fear on my end is that sometimes and again it may be so difficult on the on the technology level, this integration of dealing with point solutions, that it's just too painful.
• 17:29
  SPEAKER_01: You know, they talk about all these issues, 80%, 20% in terms of usability or features.
• 17:35
  SPEAKER_01: So if it's got 80%, we stay inside the suite.
• 17:39
  SPEAKER_01: There's all these reasons to stay inside the suite.
• 17:42
  SPEAKER_01: I mean, I I was talking I say this on interviews.
• 17:45
  SPEAKER_01: This is not the scenario anybody wants.
• 17:47
  SPEAKER_01: This is not the world anybody wants.
• 17:49
  SPEAKER_01: Nobody wants a world in which we default to what's in our EHR suite.
• 17:55
  SPEAKER_01: We want a plug-and-play world where you take the you take the lamp on my desk, I can go anywhere in this country and plug it into any outlet.
• 18:04
  SPEAKER_01: You could take any lamp and plug it into any outlet.
• 18:07
  SPEAKER_01: This is the world we want when it comes to applications where the the tool is what defines the quality of the tool, not the quality of the integration.
• 18:18
  SPEAKER_01: If we got to a point in technology where that didn't matter, perhaps AI is going to get us there.
• 18:23
  SPEAKER_01: There's some people thinking that AI is going to help with this and sort of free everybody.
• 18:28
  SPEAKER_01: I think that's the dynamic that they struggle with.
• 18:31
  SPEAKER_01: That's probably the one point of limitation on being completely customer focused on my users.
• 18:39
  SPEAKER_01: My users should be able to want product X, and it shouldn't matter on the underlying end, but it does.
• 18:45
  SPEAKER_01: And there are a lot of compromises that need to be made of staying in suite, as we call it.
• 18:50
  SPEAKER_01: Uh, I think that's something that they struggle with.
• 18:54
  SPEAKER_01: So that's that would be a struggle there.
• 18:57
  SPEAKER_00: You know, I have heard the same struggle, and I think the kind of pieces I keep hearing come up, and I wonder if this is similar for you, is really no one gets fired for staying in suite.
• 19:08
  SPEAKER_00: That's number one.
• 19:09
  SPEAKER_00: And number two is it's hard to go with a new vendor or try a startup, etc., because of the risk, because of the security concerns, just because they aren't the suite, which we have vetted to a much larger extent and we have a history.
• 19:27
  SPEAKER_00: So when you think about an organization, large or small, in suite ends up being the default.
• 19:34
  SPEAKER_00: So what would be those occasions where we could go out of suite and the CIO, the CISO, etc., aren't going to be on the chopping block because what do you think those instances are?
• 19:48
  SPEAKER_01: It's it's probably there probably absolutely clear cases of deep clinical department focused tools that are not even remember, I talk about a gray area that are not a gray area.
• 20:03
  SPEAKER_01: Absolutely you see what I mean?
• 20:05
  SPEAKER_01: So there's there's the you're staying in suite a hundred percent.
• 20:10
  SPEAKER_01: There's the you're not in suite a hundred percent.
• 20:13
  SPEAKER_01: Everybody knows this tool is is not going to be from your EHR vendor.
• 20:19
  SPEAKER_01: You know what I'm saying better than me, like cardiology or surgery, or like some deep, deep stuff.
• 20:28
  SPEAKER_00: Absolutely.
• 20:29
  SPEAKER_01: So that's the black and white, and then there's the gray.
• 20:32
  SPEAKER_01: The gray area is where everything gets interesting and gray, and that's governance.
• 20:35
  SPEAKER_01: That's where we have discussions, that's where those things all get hashed out.
• 20:40
  SPEAKER_01: Governance is fascinating to me because some people talk about it as conversations where we come up with sort of I don't I don't know if there's a formal vote or if it's sort of a hey, you know, and we all come to a consensus.
• 20:53
  SPEAKER_01: Healthcare loves consensus.
• 20:55
  SPEAKER_01: They love the idea of consensus, give consensus, you know, the idea of consensus.
• 21:00
  SPEAKER_01: Right.
• 21:00
  SPEAKER_01: Nobody wants to argue in healthcare and disagree.
• 21:04
  SPEAKER_01: So this this those would be some some ideas there.
• 21:08
  SPEAKER_01: There are absolute point solutions that are not even a question.
• 21:12
  SPEAKER_01: There's the other end, and then there's a space in the middle where you got to figure it out, and how you do that is the art, and it's getting all those voices weighted equal, well not equally, but appropriately.
• 21:24
  SPEAKER_01: The IT voice, the security voice, the total cost of ownership voice, the clinical voice, the operating like every like everybody gets put in the pot, the soup, it gets stirred up, and I always say somehow magically what what you would want AI for is to have that recording, that meeting, where everybody got to talk, have a recording, and somehow AI spits out your decision.
• 21:47
  SPEAKER_01: And says based on this and how you wanted things weighted, boom, you're going in sweet, or you're going with the point solution.
• 21:54
  SPEAKER_01: Like, you know, if if governance could be more data driven, it'd be easier for everyone.
• 22:01
  SPEAKER_00: So I I always love a good metric.
• 22:03
  SPEAKER_00: And I think that when we think about success metrics and even decision-making metrics, like you just mentioned, a lot of times we're measuring things that actually don't show us what the decision is or don't show us what success really is.
• 22:16
  SPEAKER_00: We've used some random heuristic or whatever.
• 22:19
  SPEAKER_00: When we think about the data from 10 years ago versus today, I'm thinking today we just have a ton of data.
• 22:29
  SPEAKER_00: We can measure infinitely more things, whether or not they're important, and it's coming faster and in more real time.
• 22:37
  SPEAKER_00: And so the changes I see from a governance perspective from a decade ago or more to today are speed, ability to get information, whether we want to or not, and just the amount of it.
• 22:50
  SPEAKER_00: So when you think about how CEI CIOs and their counterparts are working today, their job's much different than it was 10 years ago.
• 22:59
  SPEAKER_00: What do you think are those skills that are required for a successful CIO in 2026?
• 23:06
  SPEAKER_01: It's so simple.
• 23:07
  SPEAKER_01: It's got it well, you have to be using AI on the personal end to be more efficient now.
• 23:12
  SPEAKER_01: You can use it to just be personally more efficient.
• 23:15
  SPEAKER_01: But I think the number one CIO skill has has very little to do with technology.
• 23:22
  SPEAKER_01: I mean, they need to know technology, but you can have a really good CTO who so you could let technology outrun you a little bit if you have a great CTO, right?
• 23:31
  SPEAKER_01: So you don't have to stay up on all the granular network cloud stuff.
• 23:36
  SPEAKER_01: You could let some of that stuff go to someone else.
• 23:39
  SPEAKER_01: Your job is to understand operations and to develop deep relationships with your users and your clinicians and your C-suite.
• 23:49
  SPEAKER_01: You need to have conversations, you need to have a great relationship, an open line of communication.
• 23:54
  SPEAKER_01: You need to go to their meetings, you need to sh demonstrate care and attention and interest.
• 24:01
  SPEAKER_01: I think that's the number one key.
• 24:04
  SPEAKER_01: Now, the challenge there in healthcare is there's lots of different clinical departments and areas that are quite different, have different needs.
• 24:14
  SPEAKER_01: So there's part of your learning curve.
• 24:16
  SPEAKER_01: You need to like you should have a whiteboard or folders or something with all these departments, one for each department.
• 24:22
  SPEAKER_01: Who do I know there?
• 24:23
  SPEAKER_01: Who's my point person there?
• 24:25
  SPEAKER_01: When's the last time are there key meetings?
• 24:28
  SPEAKER_01: Right?
• 24:28
  SPEAKER_01: And you go, you you can only be in so many places at once, but do I or one of my representatives in IT have a relationship with whoever the best person is in that department who's interested in informatics and having a relationship with IT?
• 24:46
  SPEAKER_01: So that's I mean, I think if you got that matrix right, where you were able to demonstrate in every which way IT's commitment to serving that depart that department, that specific department, that would go a huge way to being successful.
• 25:04
  SPEAKER_01: They say the more the like look at your calendar.
• 25:08
  SPEAKER_01: How much time are you meeting with clinicians?
• 25:10
  SPEAKER_01: How many and other users, not just clinicians, but other users, how many times are you meeting with your users versus meeting with IT people or vendors or things like that?
• 25:18
  SPEAKER_01: The more time you spend with your users, the better off you're gonna be, you know, embed informaticists.
• 25:27
  SPEAKER_01: I mean, that's what it's all about to me.
• 25:29
  SPEAKER_01: And so that's that's the key to success.
• 25:31
  SPEAKER_01: And it probably always has been, but more and more, more than ever.
• 25:34
  SPEAKER_01: Do that and and you have a great CTO and you'll be fine.
• 25:38
  SPEAKER_00: I love the idea of being an internal service and truly having internal customers and treating each department as your customer.
• 25:47
  SPEAKER_00: That sounds like a great way of kind of taking the role of CIO into effect and really making sure everyone feels served, even though they're very different.
• 25:57
  SPEAKER_01: That's right.
• 25:57
  SPEAKER_01: Known and served, right?
• 25:59
  SPEAKER_01: I have to know you to serve you.
• 26:01
  SPEAKER_01: So that's that's it.
• 26:03
  SPEAKER_01: Absolutely.
• 26:05
  SPEAKER_00: Well, I have two questions I ask everyone at the end of our podcast, and I'm gonna go for it with you again.
• 26:10
  SPEAKER_00: I didn't share these with you in advance.
• 26:14
  SPEAKER_00: So, what's the one operational clinical reality that most people in health tech get completely wrong?
• 26:20
  SPEAKER_01: Uh, so that's a good one.
• 26:22
  SPEAKER_01: Um I think that the one of the the biggest things, you know, you talked about a little bit about this idea of consensus.
• 26:29
  SPEAKER_01: Healthcare's got some quirks.
• 26:31
  SPEAKER_01: One is this it's an inability to say no.
• 26:37
  SPEAKER_01: Um this this is spoken of as a problem.
• 26:40
  SPEAKER_01: So there's this idea of not wanting to say no.
• 26:44
  SPEAKER_01: So you either say yes or you're silent and and things move forward.
• 26:49
  SPEAKER_01: I've heard it repeatedly said there's a problem with application rationalization because everybody can a lot of people can say yes, and who could say no?
• 26:58
  SPEAKER_01: Who could shut things off?
• 26:59
  SPEAKER_00: Who couldn't have to be a few things get fired for shutting something off?
• 27:02
  SPEAKER_00: Who doesn't get fired for shutting things off?
• 27:03
  SPEAKER_01: Who's incentivized to say no?
• 27:05
  SPEAKER_01: Right.
• 27:06
  SPEAKER_01: So you when you have this environment, you get an accumulation, uh a continual addition of stuff.
• 27:13
  SPEAKER_01: It's fat.
• 27:14
  SPEAKER_01: Right?
• 27:15
  SPEAKER_01: It gets fat, and and this is a problem.
• 27:17
  SPEAKER_01: So there's there's some of that, and then I think people thinking that the CIO has has more control than they do, especially in a very large a very large health system that's in multiple states.
• 27:31
  SPEAKER_01: I've heard it said that there's a misunderstanding that autonomy, regional or hospital-specific autonomy to a certain degree.
• 27:42
  SPEAKER_01: So this idea of systemness may be lacking, maybe surface-level branding systemness, but there can be quite a deep degree of autonomy where the IT executive, the CIO, does not quite have the degree of control over those autonomous entities as people think.
• 27:58
  SPEAKER_01: So I think healthcare needs to do better on being able to say no, being able to say shut it off, being able to say cancel the contract, stop the pilot, all this kind of stuff.
• 28:09
  SPEAKER_01: Uh, I think that would go a long way.
• 28:12
  SPEAKER_01: And that that dovetails in a little bit with this uh talk about risk and agentic AI.
• 28:19
  SPEAKER_01: Who can shut things off is is gonna become a huge issue.
• 28:23
  SPEAKER_01: We're gonna we're gonna have big problems where these things do things they're not supposed to.
• 28:28
  SPEAKER_01: And we talked about agent ownership and that gray area that's gotta be figured out.
• 28:34
  SPEAKER_01: We need kill switches and kill buttons, and we would love to have those little undo arrows, but I don't know if they're gonna exist because you know I've got AI, you've got AI, it's it's all over the place, it's in all my stuff, it's got permissions to do all this stuff because yeah, I want to see how this would work, and this would be cool if it could do this for me.
• 28:54
  SPEAKER_01: And now it says permissions to go in my email at God knows where.
• 28:58
  SPEAKER_01: And if it ever decided we weren't friends, you know, it could wipe out my website and wipe out my email, and and I would be sitting there going, Well, who do I talk to?
• 29:07
  SPEAKER_01: What do I do?
• 29:09
  SPEAKER_01: Nothing.
• 29:10
  SPEAKER_01: Who?
• 29:10
  SPEAKER_01: I'm gonna call anthropic?
• 29:12
  SPEAKER_01: Who am I calling?
• 29:13
  SPEAKER_00: If you're seeing stuff in your hospital system and someone gets truly harmed, then you have a much bigger problem.
• 29:18
  SPEAKER_01: You've you have big, big problems.
• 29:19
  SPEAKER_01: I think this is a deep underlying fear.
• 29:21
  SPEAKER_01: I mean, look at the hugging face incident, study that a little bit.
• 29:25
  SPEAKER_01: Uh, you know, you've got the people say quest have questions about their motivations, but you've got the heads of the the frontier models saying we need to slow things down, we need to slow things down.
• 29:35
  SPEAKER_01: We're scared.
• 29:36
  SPEAKER_01: If you take them at their word, they're concerned.
• 29:40
  SPEAKER_01: So, I mean, the CIO should be concerned because again, you don't know sometimes what's going on.
• 29:46
  SPEAKER_01: I don't know who's doing what.
• 29:48
  SPEAKER_01: We need to slow everything down a little bit.
• 29:51
  SPEAKER_00: So I think I don't like to say slow down, I like to say be intentional.
• 29:56
  SPEAKER_00: And I think a lot of times we've accepted a pilot or messed with the workflow or this, that, and the other thing, and we haven't really decided what success looks like, or we haven't actually done our homework.
• 30:07
  SPEAKER_00: The the classic example is a random executive goes to a conference, comes back and says we're using this, and all of a sudden you have to figure out how to make it happen.
• 30:14
  SPEAKER_00: As opposed to, well, what do I actually need for my users?
• 30:17
  SPEAKER_00: Does anyone even want this?
• 30:19
  SPEAKER_00: Right.
• 30:19
  SPEAKER_00: Because picking the tool before you decide what problem it's solving is a humongous problem.
• 30:25
  SPEAKER_01: Absolutely.
• 30:26
  SPEAKER_01: Yep.
• 30:27
  SPEAKER_00: Well, the second question I always ask is when's the last time you changed your mind about something in this industry?
• 30:33
  SPEAKER_00: I probably changed my mind all the time.
• 30:35
  SPEAKER_01: I mean, I think that's uh important.
• 30:37
  SPEAKER_01: I don't know if I have my main my mind made up on anything.
• 30:40
  SPEAKER_01: So, you know, change my mind.
• 30:42
  SPEAKER_01: I just continually just try and absorb.
• 30:45
  SPEAKER_01: I try and absorb, I try and see patterns and and pick up trends that's you know, you've asked me about trends, nuances, what's different, what's changed.
• 30:53
  SPEAKER_01: And sort of the one of the the things I think we're we're seeing is a slight pendulum swing back to being a little more open to going out of suite.
• 31:02
  SPEAKER_01: I think we've seen that pendulum swing just about as far as it can possibly go towards staying in suite.
• 31:09
  SPEAKER_01: I I feel like there's a little bit of pendulum swing, maybe just slightly coming back, but not much.
• 31:14
  SPEAKER_01: And we don't we have no idea what's going to be possible in the next month, let alone year.
• 31:20
  SPEAKER_00: One month, yes, absolutely.
• 31:21
  SPEAKER_00: I can't make even that far ahead.
• 31:23
  SPEAKER_01: We have no idea, which makes it a very exciting time to be following things, but you have to have the mindset for that environment.
• 31:32
  SPEAKER_01: So you have to be super flexible, open to things, open to change, comfortable with chaos, and you're just trying to get your arms around what's going on every day to pretend to prevent catastrophe.
• 31:47
  SPEAKER_01: I think that's kind of a little bit what's going on is let's just prevent catastrophe.
• 31:52
  SPEAKER_01: Let's throw slow down as much as we can or be intentional.
• 31:56
  SPEAKER_01: Let me, from a cyber point of view, let me just get my arms around this.
• 32:01
  SPEAKER_01: And at the very least, if if everything goes bad, you want to be able to demonstrate I did everything a prudent person would do with the information I had.
• 32:12
  SPEAKER_01: Right?
• 32:12
  SPEAKER_01: So it's a difference between negligence and just things going wrong.
• 32:16
  SPEAKER_01: You don't want negligence, so do everything you can and do your best and drink from a fire hose.
• 32:21
  SPEAKER_01: You have to absorb information as fast as you can, and of course, we all use AI to help us absorb the information that'll hopefully help us control AI from killing a soul.
• 32:31
  SPEAKER_00: Yeah.
• 32:32
  SPEAKER_01: It's a good time.
• 32:33
  SPEAKER_01: It's a good time.
• 32:34
  SPEAKER_00: Well, Anthony, thank you.
• 32:35
  SPEAKER_00: This conversation's been great for me.
• 32:37
  SPEAKER_00: I think most people in this industry can tell you what a CIO says.
• 32:41
  SPEAKER_00: I love that you've told us how the answers moved and how it's different and how things are becoming more useful over time.
• 32:47
  SPEAKER_00: So thank you for that.
• 32:48
  SPEAKER_00: Before we go, can you tell people where to find you?
• 32:51
  SPEAKER_01: Uh, sure.
• 32:52
  SPEAKER_01: Uh healthsystemcio.com is the website.
• 32:55
  SPEAKER_01: You can find me on LinkedIn.
• 32:57
  SPEAKER_01: Uh, I got a good presence there.
• 32:59
  SPEAKER_01: And uh that's about it.
• 33:00
  SPEAKER_01: Don't come to my house.
• 33:02
  SPEAKER_01: But uh, you know, online's great.
• 33:04
  SPEAKER_01: So that's where you can find me.
• 33:06
  SPEAKER_00: So if everyone listening, if you want the unfiltered version of what your own IT leadership is weighing right now, Health System CIO is where the conversation is happening.
• 33:14
  SPEAKER_00: Go read it, subscribe to it, and then don't forget about the clinical realist wherever you get your podcasts.
• 33:18
  SPEAKER_00: And if you're working through a live clinical AI governance question inside your own organization or a real one that's not a theoretical one, link is in the comments and definitely book a clarity session with me.
• 33:29
  SPEAKER_00: So I'm Dr.
• 33:30
  SPEAKER_00: Sarah Matt, and this is the clinical realist.
• 33:32
  SPEAKER_00: We'll see you next time.
• 33:33
  SPEAKER_01: Thanks, Dr.
• 33:34
  SPEAKER_01: Matt.