Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes
Doctors click past most of the alerts their electronic health record throws at them, and the industry has long blamed the doctors for it. Thomas Powell is a physician executive and a chief medical information officer. This episode is based on his article “Alert fatigue is a design problem, not a vigilance problem,” published on KevinMD. You will hear why a medical alert is blown right through 49 to 96 percent of the time, why he says that number says more about the system than about physicians, and how the alerts that matter get buried among the ones that do not. You will hear how he uses alert data to separate important from unimportant alerts, why every alert should come with an action item, and why most organizations never retire the alerts that came out of the box. You will also learn why turning off an alert is literally the flip of a switch yet rarely happens, why every alert needs a clinical owner, and why physicians who skip governance meetings end up living with decisions made for them. Press play for how to make the alerts that matter stand out again.
If this episode helped, follow The Podcast by KevinMD on Apple Podcasts or Spotify, and leave a rating.
Partner with me on the KevinMD platform. With more than 100 million impressions a year and nearly 750,000 followers, subscribers and email recipients across my channels, I give you direct access to the doctors and patients who matter most. Whether you need a sponsored article, email campaign, video interview, or a spot right here on the podcast, I offer the trusted space your brand deserves to be heard. Let’s work together to tell your story.
PARTNER WITH KEVINMD → https://kevinmd.com/influencer
SUBSCRIBE TO THE PODCAST → https://www.kevinmd.com/podcast
Transcript
Kevin Pho: Hi, and welcome to the show. Subscribe at KevinMD.com/podcast. Today, we welcome Thomas Powell. He is a physician executive and a career chief medical information officer. Today’s KevinMD article is “Alert fatigue is a design problem, not a vigilance problem.” Thomas, welcome to the show.
Thomas Powell: Well, thank you, Kevin. I look forward to this. I look forward to the discussion.
Kevin Pho: All right, let’s start by briefly sharing your story, then we’ll jump into why you decided to write this article on KevinMD.
Thomas Powell: OK. So as a physician, I went to medical school at University of Texas in San Antonio and did my residency at Duke in anesthesia. And while I was there was the NIH, National Library of Medicine fellowship, really sort of the beginning base of it for clinical informatics, and immediately I knew I had found what I wanted. So when I graduated there, my first role was at University of Miami, in the anesthesia department. I did both clinical and a lot of informatics work. And then really following that, I’ve been a chief medical information officer. I’ve been in academic centers; I’ve been at pediatrics, adult hospitals. I’ve put in Cerner; I’ve put in Epic. So that’s really where a lot of my energy has been, and actually trying to do the job well because informatics sort of got away from us when the HITECH Act was enacted as it was. Now I own my own LLC called CMIO LLC, where I’m consulting. I do interim executive work. I try to help people fix problems that they bring to me, and I do some industry consultation too. And then that brings me, and I just wrote a book that’s coming out on November 3, really putting it all together because really, as I was starting to put down thoughts of my life, I am concerned, and I think the concern is warranted. And I shared it a little bit, and other people that are reading it shared those concerns. So I thought it was very, very good. I’m happy to have done it.
Real physician voices, twice a week
Free, and one click to unsubscribe.
Kevin Pho: All right. So before we talk about alert fatigue in particular, one of the things that you mentioned earlier was that it’s important that CMIOs like yourself do their job well. Just in your terms, what does it mean for a CMIO to do their job well? There are so many aspects to that, so many intersections between health care and informatics. What does it mean to do their job well?
Thomas Powell: Well, and thank you for asking because that’s literally what the book is about. Clinical informatics is the name, and the role, chief medical information officer or chief health information officer, it’s been misunderstood. And a lot of people think of it as a technologic role. And what it is, is it’s a combination of four pretty significant skill sets that unfortunately are fairly rare to find, and people get installed without that ability to execute. So it’s really, you have to understand the technology, obviously. You have to understand the clinical aim. What are you trying to do with this clinical encounter? What are you trying to do with the thing that you’re executing? You have to understand workflow engineering because a lot of this is sort of manipulating both the clinician’s workflow, the way the clinic or inpatient unit flows, with what’s going on inside the EHR so that you can optimize the data. And really one of the big things, Kevin, that’s been a hard part of this is the executive empowerment. It’s very inconsistent how much energy a chief medical information officer has to actually protect the interests of the informatics, of the tool, of the clinical encounter. So I consider myself a physician to the system. That was how I always sort of explained it, is while the physicians are taking care of the patient, I’m trying to take care of the system that helps them take care of that patient. And that is really how I’ve always seen my role.
Kevin Pho: All right. Today we’re going to talk specifically about alert fatigue as a design problem, not a vigilance problem. Tell us why you decided to write this article, and for those that didn’t get a chance to read it, tell us about the article itself.
Thomas Powell: OK. So when you try to write down what was my focus, a big part of it is the physician experience. And there are a lot of things that your medical officer should be doing, and it’s going to be clinical in all different ways. But really, it’s about the physician experience, and alert fatigue is one of the big ones that just hasn’t been cracked yet. It’s basically the usability, and I’m sure you’ve heard the term now, which is physician wellness, where almost every organization is really putting a lot of energy into clinician wellness, but physician wellness specifically. And when you look at the data around physician wellness, physicians tend to report that it’s based upon two things. It’s their EHR causing them trouble, and it’s their HR causing them trouble. So it’s that combination of the work environment that you’re in and the EHR scores, and they’ve been terrible for a while. By the way, one of the most recent assessments was that the average physician rates their EHR experience at a 45.9 on a 100 scale, so an F. And it’s not that far different from where nurses are, which is even lower for nurses. So that’s not good. And really what the book tried to take on also was we’re aiming at our. So we have problems. Health care has had issues for a long time, and in ’99 when the Institute of Medicine came out and said about 98,000 people a year are actually killed in adverse events and things like that, everybody started saying, “OK, well, that’s what HITECH was about. We need an electronic health record.” But for why? Because I would not disagree with them, but they never aimed our technology at anything. It’s just put it in. And so what happens is you have a lot of hospital administrations and people thinking of the electronic health record as like a treatment. Like, it’s the thing that’s going to fix. It’s the solution. And actually, it’s more like an imaging modality. As a physician, I read it, and it tells me the image of what’s going on. I can actually, from the data, if I do this well, can reconstruct the workflows that are broken. I can reconstruct the workflows that are working. I can look at why things are the way they are. And so a big part of my heart was that a lot of people have for years blamed alert fatigue or blowing through alerts on the physicians. And so just to give you how bad it is, because this says more about the system. I think first off, I do want to back up and just say that whenever there’s been a study looking at what is going on from a clinical health, safety, and quality standpoint, the clinicians certainly have a part in it, but it’s a substantially smaller part than you believe. It always comes down to system. Almost every time, the system is the most heavily regarded contributor to adverse events, to all these things. And so when it comes down to alerts, the physician numbers are horrible. When you give a medical alert in an electronic health record system, 49 to 96 percent of the time it’s just blown right through. And that was interpreted by the industry, by the electronic health record industry, by IT, by administration as, “Well, look at those technophobes, the physicians that are not so happy with this, and they’re going to blow through all this.” And what I said, actually an alert, when I start distracting my physician from a clinical encounter, there’s actually an art to it. And so there’s a lot about the alerts themselves that are broken. There’s a lot about the volume of alerts that are broken. So that is the main thing. So I manage alerts. Wherever I go, I go in and I take both the pulse of the clinicians, but you can actually look at every alert. You can tell how many times it fired. You can tell which is being followed. And what you find out is many of these alerts are based on old advice. They’re sometimes now even conflicting. And now in the world of AI, you have these separate tools that are feeding in a whole nother world of alerts, advice, whatever. And what I believe is a central part of the quality safety issue is we have just confounded the clinical encounter. We’ve got a physician now doing data entry. We pay them; they spend 50 percent or more of their time doing data entry in today’s world. And they’re not doing a great job, which is because it was not their job. So they have a record that no one can use, even themselves. I hear that all the time because the notes are getting longer and longer and saying less and less. And the data that’s used, the whole idea was to get data that we could run the system and get better, be a learning health care organization and make our quality better, make our safety better, run our hospitals more efficiently, all those kinds of things. So the data is horrible. The physicians are burned out. Forty percent of physicians say that they are burned out, and it is related to their electronic health record experience.
Kevin Pho: And I agree with you. I have that firsthand knowledge. In my primary care clinic, in our electronic medical records, I would say almost always there’s some type of interaction, alert, boxes that obstruct me from executing my order. And if everything is alert, nothing is alert, right? So I think that that is definitely something that I face every day. Now, in your studies you said up to 96 percent of alerts are ignored, and among those 96 are some legitimate alerts, right? So from the patient’s standpoint, how does that harm patients, the fact that almost all of these alerts are blown through by clinicians?
Thomas Powell: Yeah. That’s actually a really important part of this, and the way that I frame it is it’s a problem that you’re making something and adding such inefficiency to the system, right? So we can talk about if you are just wasting my time, right, which is, that’s what that number’s saying. If I’m showing you a number of alerts that 96 percent of the time you ignore, something’s wrong. Right? But I say the real, the most important issue with it is that there’s no distinguishing feature to the important one, to the one that you’re going to miss. So buried inside of those, it’s not like everything that’s being told to you is unimportant. So that’s the thing, is the hope would be that you get to a world where I only let come through our system the alerts that help, right? I don’t need to teach you how to be a physician. You already know that part. And that’s unfortunately one of the things people did. They’re trying to remind you of how to, don’t forget this and don’t forget that, and administration would want you to do this. And so that’s where a lot of those, they’re overwhelmingly insignificant, even in the libraries of medication alerts. There’s a lot of drug interactions that, yes, theoretically or in some way it’s known that it does, but it’s an insignificant interaction. But if I get to you the right information that helps you make a good clinical decision, that is the goal, and you do that by aimed and targeted. And in a world of somebody blowing through 96 percent of the alerts, those just get blown through, and you find out about it when a clinical event happens. That’s sort of the problem.
Kevin Pho: So from the perspective of the EHR designers who think that physicians who complain about this are just Luddites or technophobes, as you put it, now, what is their perspective when physicians respond to this? Are a lot of them acknowledging this concern? Are they doing something about these blanket alerts, triaging them, stratifying them? What are they doing from their perspective?
Thomas Powell: So great question. So my call in the book is for exactly what you’re talking about. What I can say is there’s not been a lot of attention, because it takes a lot of work. So as I said, right now what I believe is the case in the United States is people are throwing technology undirected at things, hoping that Epic fixes it, or Cerner fixes it, or your AI product fixes it. And it’s never the technology. There’s a lot of operational clinical work that needs to happen. We actually have to buckle our belt and do some hard work, and that’s governance, and it’s decision-making, and it’s making sure there’s an owner. And so that is the hardest part of this, is that it’s so much work that it often gets abandoned. And as you said, if you’re a person who bought a $250 million electronic health record, you think it’s done, right? I mean, I’m putting it in there; it’s fixed, and why is this person complaining about the best or one of the top EHRs in the country? And it was because it wasn’t crafted for any specific purpose, especially for your local environment. That’s, I think, the key.
Kevin Pho: So tell us about the next steps in terms of stratifying and interpreting these alerts. Is it an AI problem? Is it up to someone to reinvent the workflows? Who’s the onus on in terms of solving this alert problem?
Thomas Powell: Yes, and that’s where the diagnostician is, and the diagnostician is my generic term for me for the role, because some people call it CMIO, some people call it a physician informatics leader, some people call it a CHIO. And I say the term is not the issue. What we have is a lot of names and no empowerment from the organization at all. And I say one of the best ways to see how much power a role has, not influence, is the ability to say no to something. To turn something off and to say no. And when was the last time you have seen your physician executive be able to say, “No, that’s not a good idea”? And so they’ve got to have the right staffing. They’ve got to have people with the right training. You have to realize that this is hard and heavy work, now especially in the days of AI. AI, 99 percent of AI is coming into our market completely untested in any significant way before with any particular certification. So you’re taking it in. The vendors themselves are telling you that it’s a proprietary AI model, so they’re not going to tell you if they could. Even if they could tell you what they did, they wouldn’t tell you because it’s proprietary. And so how do you govern and manage and make sure in that world? That’s an ongoing set of a lot of work. And the EHR was the same. And so if you look around and your informatics is not an army, your organization is missing to some extent what’s going on now. And so if you look at some of the most credible health systems in the country, they have now armies of informatics aimed at things, and you’re doing informatics. And that informatics is just what I said. Let’s get out there and evaluate the problems, find the problems, find the workflows, work on wellness, work on efficiency, work on all the things that work, get the data right. It’s a full-time, all-time job, and I’m trying to get that out that we have just sort of undervalued the human. We’ve spent millions on electronic health records and very little on the people who can actually tune them.
Kevin Pho: So you’re a consultant, of course, and you work with many, many health care organizations, I’m sure. Give us a story, case study, or an example of a successful addressment of this problem. How has an organization successfully managed this alert fatigue problem from your standpoint?
Thomas Powell: Yes, so I’ve spent a lot of my career as the person who had to do it, and in some organizations, they did empower. So the way that you address this is, let’s just pick, like you said, let’s narrow it to just the alert problem right now because there are many problems. But the first is data. Don’t want to get too technical, but there’s a world of metadata. And so the data is, did an alert fire? And the metadata is when and who and how many and all those other. It’s all the data that’s tied to it. We can know all that. So I’ve gone into places, and I have made sure that we have steady ongoing reports. And I see a lot of places where once you get that turned on, they just take the reports and read them and put them to the side. “Oh, yeah, we do have a problem.” Because what are you going to do about it? Well, there are algorithms of how you can work down that list. The first is separate them into important versus not important. Get rid of all your alerts that are just telling people something. An alert should be there with an action item to it. So we see that your patient looks like they’re septic, and if you believe so, I’m giving you the tools to do something about it right now. And so those are the kind of alerts. The other is put them in the frequency of being blown through. So your 96 percent that are being nearly always gone through, is it important? Is there a reason for that motion that I can fix, right? Was it just configured wrong? Is it old? Is it not quite right? Is it conflicting with something else? And what you’re going to be able to find is you can reduce the library of alerts, and then you get down to a workable world. Because I’m telling you, most people don’t even retire the alerts that came out of the box. So these vendors will pull up out of the box with a set of alerts that the nation uses in general. Their broad base of users. And whether that works real well on your data in your place, the way you do business, has really never been evaluated. So that’s the thing. It’s just targeted. Use your data. Actually have people that go in and turn things off. And when you put something new in, if ever there was a need, you’re at the. Informatics at the heart of let’s craft this to go to the right people at the right time with the right information so that they can do everything right. And so that’s really what it is, is it’s mostly been neglect, to be honest with you. Most of this has just simply been there’s stuff out of the box, and it just grows. It grows unchecked.
Kevin Pho: How difficult is it to retire an alert? Because what you’re saying sounds like common sense. Why aren’t more organizations doing what you’re doing? For instance, when you say retire an alert, how easy or difficult is it for many health care organizations?
Thomas Powell: It’s literally the flip of a switch. So an alert is a build. The build is the content and the algorithm that would kick it off, and then the connection to the orders or whatever things that you were going to do to respond. And then it becomes, it’s really a container. And that’s an alert container for all the stuff that that alert needed. And you can turn it on or turn it off.
Kevin Pho: But from a bureaucratic standpoint, is it?
Thomas Powell: Oh, that’s the. Yeah, no, then that is the key, is that everyone is afraid, right? Everyone is afraid of making things go away. Because they’re afraid that everything must be there for some really good reason. And when they understand, that is not the case. The vast majority. So that is why you do clinical governance. To me, I wanted to make this point as important as I can. If you’re in an organization like I am describing, these are not IT decisions. These are not administrative decisions. These are clinical decisions. Like every clinical thing that goes on in an organization, you need clinical ownership of an alert. If there’s nobody in your group says, “I agree with that, and I’ll be the sponsor of that alert, and when anyone has a question, I’ll be the one that you can help,” if you don’t have those kinds of things, you’ve got a mess, is what you have. So I will also turn off alerts that are not owned. So you go down, and then you have clinical people making clinical decisions. I try to be, as much as I can, the doctor to the system, and every time I would fool with primary care, for example, I would come to you, Kevin. I would say, “All right, Kevin, they’re going to fiddle with your clinic, your environment. I would like you to be point on this.” And one of the hardest things we’ve had is physician involvement. And so I do say that part, we all have a part of the problem that is going on right now. From the country, the HITECH Act, the hospitals themselves, the physicians, everything is involved in this. And one of the most key things, and I know you’ve probably heard it, if you’ve been in an electronic health record system, nobody wants to pay the physician. You’ve got nurses and other people that are actually being paid to represent their areas and do those things, and so they get relatively good governance and recognition and change management and things like this. And so physicians are still being held to their RVU, and they want them to escape for a meeting, and it’s painful, so they don’t. And then I guarantee you that in that meeting, someone is going to make a painful decision for you, and that’s the key thing. That involvement and getting your physician staff to know that, “By the way, I’m here because I’m trying to make the whole clinical system better, but for you as a physician, if you’re not there, often what’ll be passed to you is something that you don’t like.” And that’s really what’s happened, is there’s a lot of documentation that we don’t want our most expensive resource in the house wasting time on this documentation the way they did. So there’s a lot of that kind of thing that you have to get right. It’s a lot of governance, a lot of meeting and thinking and having smart people in the room with you. And then the rest is just the technical parameter, right? The next is just turning it on, turning it off.
Kevin Pho: We are talking to Thomas Powell. He’s a physician executive and a chief medical information officer. Today’s KevinMD article is “Alert fatigue is a design problem, not a vigilance problem.” Thomas, as always, let’s end with your take-home messages to the KevinMD audience.
Thomas Powell: I think you did such a good job. You led me right to that’s what it is. I mean, I think the main thing to get to the audience is when you sit down and write a book, and I started gathering. My premise was health care is sick, and I did the clinical method. I, as the physician of the organization, did the clinical method and ran down what it did and what the diagnosis is and then how you treat it. And that was the whole thing. And what you realize is America is not just a little sick. Our quality, our safety, our margin, two out of five hospital systems are now functioning in the red. It’s the money, the workforce, the burnout. All of that’s going on. It is real, and it is fixable. I was surprised at how much of these come back to just things that have never been diagnosed and treated. That’s really what it is. You diagnose it and treat it, and there may be some local variation that’s important to you, but it can be fixed. So I highly recommend that you join forces. If you’re the chair or a leader of a specialty or a leader of some other clinical aspect of your hospital, get your people in the rooms and sort of demand to speak for the clinical role, because that’s what I am trying to say, is somehow health care got away from us a little bit. And the physician in the room is an important voice, and I have one in every room, and it’s crafted to be for whose specialty makes that most important. So I highly recommend that you show. I highly recommend you participate. Now, it would be great if you loved it. It would be great if you were interested. But what I’m trying to say is, do it because without it, your organization is going to make that decision anyway. And if you’re not in the room, it will not be a decision that you probably enjoy.
Kevin Pho: Thomas, thank you so much for sharing your perspective and insight. Thanks again for coming on the show.
Thomas Powell: Thanks, Kevin.
Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes


