SCCM Resource Library-Quiet, Please! Minimizing Alarm Fatigue and Maximizing Safe Patient Care

Quiet, Please! Minimizing Alarm Fatigue and Maximizing Safe Patient Care

Video Transcription

Good morning and welcome to this session, which is, of course, the discussion about alarm and alert fatigue in the ICU. We're very happy, ladies and gentlemen,  to have you here today. My name is Barbara McLean. I'm the moderator of this session, plus I am your fourth speaker. I have a few housekeeping remarks to make.  First of all, assuming that our audience was so huge that I would ask you all to move to the center because we need to make other seats available, but we have plenty  of seating here, and note the location of the exit doors in case there's a fire or any other reason to be concerned. Over here and over here, I feel like a flight attendant.  There and there are the exit doors. Please silence your cell phone. And if you need to leave the session, please exit quietly. And once the session has concluded, we request  that you leave quickly because there will be another session, although I don't believe there's another session in this room until this afternoon, so that may not be true.  And if you would like to discuss the topic, just if you take it to a different area than up at the front, that would be very beneficial.  Full Congress attendees gain access to Congress on demand for continued learning, and if you missed the session or would like to listen to a session again,  you'll receive an email with access instructions about two weeks after Congress has ended. CE credit is available on the Congress app or the information.  All of our speakers will disclose their conflicts and do not, please do not publish digital photographs of slides, especially those that say no social media,  which I don't think any of us have said. And be mindful that while social media provides an open air forum for presenters, for the conference and for attendees,  as well as the public, good manners and respect should prevail. And the Twitter note is if you are tweeting, please use hashtag CCC49.  Normally, we would have each speaker have a five to ten minute question and answer period, but we are not choosing to do that today  because we think there will be more value in the panel. So, we will move quickly through our speakers, and our speakers will share their abundant data with you.  And at the end, then we will have a moderate, a moderated panel discussion about alarm and alert fatigue. So, thank you very, very much.  And I'd like to have our first speaker come to the podium, and she'll introduce herself.  Hello. My name is Sandra Cain-Gill. I'm a professor at the University of Pittsburgh in the School of Pharmacy. I'm delighted for the invitation to speak  about alarm and alert fatigue. And I have no conflict of interest related to today's discussion. So, I get to set the tone for today's discussion in the sense  that I'm trying to make sure that we're all on the same page when we're talking about alarms and alerts. I've had the pleasure of working with several of our panelists  to develop a report for the Society on Alarm and Alert Fatigue that we have published one on alarm fatigue and one on alert fatigue in critical care medicine.  And so, while there are variations and definitions for alarms and alerts, these are the ones that we have opted for in that report.  So, an alarm is an automated auditory visual notification. That really suggests some instability in the patient's status.  Those things can be deviations in heart rate, blood pressure, alarms for capnography, alarms for mechanical ventilation, and for smart infusion pumps.  So, those are things that we're thinking about when we're thinking about alarms. When we go into thinking about alerts, it's more along the lines  of the clinical decision support systems. Those pop-up warnings that you receive while you're entering in order at the computer that suggest what you're doing may not be in line  with what the rules are behind that alert. And so, it notifies you that it's something you might want to look at or reconsider. If we're talking about alarm and alert fatigue,  either way we look at that, it's about desensitization. So, if you receive too many alarms or if you receive too many alerts,  you become desensitized, meaning that you receive so many that you could overlook or ignore a potentially significant alert or alarm that could compromise a patient's safety.  So, alarm and alert fatigue. I think it's analogous to actually start here for me is to we're going to talk a lot about the negative sides to alarm and alert fatigue.  And I don't think that's probably any surprise to anybody in the room. But I do want to just take a moment and highlight, you know, there's a reason why we put alarms  and alerts in place to begin with. There are benefits associated with these alarms and alerts before we get into the problems associated with the fatigue.  So, just to go through a study on alarms. This is just one study. There are several others. This is about smart infusion pumps. And this is an evaluation over a two-year period  in a neonatal intensive care unit. And in that time, we're talking about an evaluation of 370,000 alarms. And during that time, the alarms were useful in preventing  and detecting 160 events through a hard stop for that smart infusion pump that would have resulted in the patient receiving a dose of a medication  that was more substantial than it should have been. So, again, it's helpful for patient safety reasons because those, that excessive dosing would have resulted  in some type of an adverse drug reaction for that patient. And those are the hard stop alerts associated with the smart infusion pumps. We also have soft alerts.  And the soft alerts means that the nurse can bypass that alert or override that alert and proceed anyways. And even the soft alerts, or the soft stops actually prevented  around 2,000-ish events from occurring. So, there is advantage to these alarms. Now, let's think about the alerts. So, this is an evaluation that was done  over a six-month period of time. And it was pre and post analysis. It was three months before a drug-drug interaction, clinical decision support alert system was put  in place, and three months after. If you look at the time and the number of drug interactions before and after, there was an 18% reduction in the number of drug-drug interactions  that reached the patient during this timeframe. And so, that actually equals about a 43% relative risk reduction in harm associated with a drug-drug interaction.  So, again, there are benefits to these alerts and alarms that I want to make sure we appreciate before we continue our discussion about fatigue.  So, when I think of fatigue, I think of similar to the boy crying wolf, if you remember that story, right? The boy who cried wolf, he was asked to look over the sheep.  And if there was a problem and a wolf came along, then he was to call the townsman so that the townsman could come and assist that boy in making sure  that the wolf didn't attack those sheep. And so, he says, you know, I'm not sure if this is going to work, you know, let me try it out, test it out.  And I'll see if the townspeople really come. And of course, he cries wolf, the wolf isn't really there, but, you know, the townspeople do come. And he tests this a few times  until finally the wolf really does come. And at this point in time, the townsmen are desensitized to that call. And so, they don't arrive, right?  And that's so similar to what we're talking about when we're talking about these alarms, right? Is the point that, is that you hear it so many times,  we think that it's an important, something important that we need to respond to, that we go to respond to. And it's not really that significant, so we become desensitized.  There are many other negative effects associated with alarms that are beyond just desensitization. As you can imagine, that continuous beeping and feeling  a need of urgency to respond to that can be very taxing, mentally taxing to a clinician. That can lead to job dissatisfaction, right?  You know, just that always on a heightened response can lead to job dissatisfaction. It can lead to distraction in the sense  that if an alarm was occurring while a nurse may be caring for one patient but can hear an alarm in another room, the nurse runs to go tend to that alarm because it's supposed  to be very important, and then they kind of forget what they were doing with the patient they were caring for in the first place. And so, maybe there's a patient safety issue  that can arise there as well. So, it can be distracting, which is also compromising patient safety. So, that's why alert fatigue can be so concerning.  What we really want is to erase that boy-crying-wolf or erase those alarms that may be unnecessary so we can really focus on patient-centered care.  And so, while my goal of my discussion today is not to go through and discuss the ways in which we can reduce alert and alarm fatigue, that's up to the rest  of the panel to discuss today. But that's what our goal is here, right, is to make sure that we can focus on the patient and making sure we're delivering the best care possible  without those unnecessary distractions.  So, the question that was posed to me for today's discussion is how many alarms do we face in a day in the ICU? And so, what I want to do is pose that question to you.  So, just with a show of hands, who thinks that you have or a nurse potentially receives less than five alarms per patient per hour? Who thinks A is the right answer,  less than five alarms per patient per hour? Okay. Who thinks it's somewhere between five to ten alarms per patient per hour? Okay. We have one.  How about C, 11 to 25 alarms per patient per hour? Okay. We have many more people who think that's the answer. And D, 26 to 50 alarms per patient per hour, right?  So, as we increase on our spectrum of a lot of alarms, people seem to tend to go in that direction. And so, now let me provide you some data to,  some data that's provided, that's reported in literature about how many alarms actually occur. So, number of alarms per hour per patient.  This is a, first one is an evaluation that was done at five different intensive care units where the experienced nurses actually documented  when an alarm occurred and then they discussed why that alarm occurred and what their response was to that alarm. And in this particular study, it suggests  that one alarm per 37 minutes per patient, so that would be about two alarms per patient per hour. So, those people, no one picked A.  The A would have been a correct answer here, right? The key is why I think this is on the lower end of the spectrum is because nurses had to report it, right?  Anytime we get into voluntary reporting, regardless if it's a study, and they had to write it down and whatever else they're doing, you know, did they have time to write it down?  And so, I think there may be some reporting biases. Maybe they only reported the significant ones. And so, maybe this particular study is on the low end  of the spectrum for those reasons. Here's another study, though, that actually supports this study. And this was an evaluation that was done in a medical intensive care unit  where there were six alarms per hour in a 12-bed unit. So, that would be about less than one alarm per patient per hour. And so, again, on the lower end of the spectrum here.  Then, this is an evaluation, or this is a report that was provided by the Association for the Advancement in Medical Instrumentation. And in that particular report,  they suggest that it's one auditory alarm signal every 66 seconds per unit. So, it didn't give it to me by patient. So, that would be about 55 alarms per hour in the unit,  in this neonatal intensive care unit. And then, so anybody who said it was on the lower end of the spectrum A, which I didn't see many people there,  there's some data to support that. But now, let's go to the other end of the spectrum. So, this is an evaluation that was done in an ICU in Korea.  And they reported, it was through direct observation where somebody noted the number of alarms that were occurring. They noted that there was about 45 alarms per hour per patient.  So, way on the other end of the spectrum, right? So, those people who said your answer was D, this would be data to support that. There's a substantial amount of variation here.  The reason why I'm telling you how things were reported, because I think that's one of the things that plays into this. I think it depends on your size of your ICU.  I think it depends on the acuity of the patients. I think it depends on a whole host of things that you need to consider. So, we have information telling us how often they occur.  But for you to evaluate in your own institution, you need to think about what it is you're evaluating and considering when you're trying to look for these numbers.  The interesting part about this particular last study was that they actually did an alarm fatigue score for the nurses. And it was on a seven item questionnaire.  So, there were seven questions. And it was a five point Likert scale for each one of those questions. So, the maximum score you could have gotten on that was 35.  And the average score was 24. I actually was kind of surprised that the fatigue score was only 24 when these nurses were receiving 46 alarms per hour per patient.  So, there's ways in which you can go about assessing fatigue. And to my surprise, maybe they're used to it. That could have been a reason for it.  But there are ways to assess using a fatigue score. So, the next question I have is how many alarms are valid or actionable? And so, with this, I think when you're thinking  about an actionable or a significant alarm, the first thing you need to think about is the source of that alarm. So, if you're evaluating at this institution, you might want  to think about source. Because, you know, we have alarms that are actually here about 24% of alarms that are coming from staff manipulation, right?  So, if a nurse is, you know, turning or twisting a patient because they need to for patient care reasons, the alarm is sounding.  And those might not be the alarms that you want to focus  on reducing because they're not necessarily significant or actionable or concerning. So, when you think about, you know, what you want to do with your alarms at your institution,  source may be a consideration for you. Clearly, I think we'd be more pertinent to think about the alarms that are associated with technical problems.  If there's, you know, the battery is running out on a machine, you'd want to hear that. You'd want to know that. You'd want to fix that problem. You'd be more concerned about alarms  that are occurring associated with the patient. The other thing you can think about is how valid the alarms are. In this particular study that looked at the validity  of alarms, valid meant that it did what I kind of suggested in the beginning. It looked for a physiologic deterioration in the patient.  And if the alarm was associated with physiologic deterioration, that was considered to be a valid alarm. However, 25% of alarms aren't valid.  They're not going off for that particular reason. So, there may be a place you want to start thinking about how we could reduce those particular alarms.  75% were actually considered to be valid. So, interesting, 25% of alarms were valid. But you know what? How many of these are actionable? They may be valid.  They may be detecting what you want it to detect. But how many of them are actionable? And around 13% of those alarms were actionable. So, they're valid.  They're saying what they're supposed to say. They're detecting what they're supposed to detect. But not necessarily something that you would do anything about.  So, if you look at a systematic review of a series of studies, it was actually reported that actionable alarms are about 5 to 13%. And so, when you think about actionable alarms,  that may be a focus for us to think about at your institution. How many alarms result in negative consequences? We're going on the other end of the spectrum here, right?  I started this discussion and I said, okay, alarms are put in place for a good reason. They're for patient safety reasons.  But maybe they have a negative consequence associated with this. So, I couldn't get a number to quantify how many. But there are reports about the fact  that alarms can result in a negative effect. And the first one is from 2010. And it's about a nurse who had been so, I guess, annoyed with the number of alarms that were going off  that they turned it off. And as a result, they missed the clinical deterioration of a patient and actually resulted in death. And so, that was really bringing national attention  to this concern. A survey of nurses also indicates that there's more adverse events associated with alarms that have been occurring as well.  So, again, there's negative consequences associated with alarm fatigue. How many alerts do we face in a day in the ICU? So, switching over to alerts,  the clinical decision support alerts. Our denominator, our evaluation for rates of this would be slightly different at alerts per patient encounter or alerts per order.  And this is a summary of what's available in the literature. And it's about 16 to 60% of medication orders generate an alert.  And so, that's quite significant if you're on the 60% range. And, of course, to no surprise, I think, to anybody here, that there's more in the ICU, right?  Because we have more drugs, more drug interactions. And so, there's more alerts that occur when entering medication order for a critically ill patient.  During a patient encounter, so in terms of their stay, about 35 to 90% of patients have some type of an alert during their stay. And if we want an estimation of this, that would be  about 64 alerts per day in three ICUs with 58 beds, which was what was reported, that would be about one alert per bed per day that we need to contend with.  Just to give you an idea, on top of all those alarms we're dealing with, here's the number of alerts that we need to deal with as well. How many alerts are overridden?  That's a question for the group. I'll do it quickly just out of convenience for time. But, you know, think in your head here, A, B, or C. Make your choice, maybe.  Here's some data to support that, which is one study indicates that 97% of alerts are actually overridden. So, really, really on the high end.  But if you look at a systematic review, the range is actually somewhere between 50 to 96%. So, very substantial in the number of overridden alerts.  Sometimes for metrics, we need to think a little bit beyond just if the alert was overridden. Because guess what? Maybe you overrode the alert,  but you didn't do it appropriately, right? Maybe you shouldn't have overridden that alert. So, when you think about the metrics for alerts,  you need to think about was it appropriate to override the alert or inappropriate to override the alert? Not just enough to look at overridden alerts.  We need to break that down a little bit further and provide that emphasis as well. So, how many overridden alerts result in a negative consequence to the patient?  That's important to consider. And we actually have that that's been quantified. For an appropriately overridden alert, there's about 2.74 ADEs per 100 overridden alerts.  And for inappropriately overridden alerts, that's about 16.5 ADEs per 100 overridden alerts. So, if an alert is inappropriately overridden,  much greater chance than an adverse drug event would occur. And for this, those data were actually taken to extrapolate to find out the financial burden associated with that.  That would result nationwide at about 5.5 million medication related alerts being inappropriately overridden, which would be about 200,000 adverse drug events and cost  of the healthcare system somewhere between 800 million and 1.8 billion dollars. So, overall, how many alerts do we have in the ICU? That's what I was asked.  I think with no resistance from the audience, we would answer the question too many, right? Too many is the right answer. But you know what? It's a balance.  And it's a balance between putting those alerts in place and alarms in place to promote patient safety and then tipping the balance too much that we're actually end  up compromising patient safety. So, thank you.  I'm just going to set my timer because we are on a strict time. All right, here we go. Good morning, everybody. My name is Jackie Stoyer. I am a nurse practitioner and a manager  of the critical care service APN team at North Shore University Health System, which is a five hospital system just outside of Chicago. And I'm going to be talking to you  about the alarm reduction project we did at my organization in 2013, which was incredibly successful. I have no conflict of interest or financial disclosures.  These are the two staff RNs and quality manager that worked on this project that need to be acknowledged. My objectives, you can read those on your own.  And we're going to sort of get into the meat and potatoes. So, what happens in people's everyday life, right? My best friend lives near Chicago O'Hare Airport.  And her backyard abuts one of the highways. And she has four dogs, not just the two pictured here. And through the highway goes the L, it's actually would be right  in here, goes the L station. And when I am on the phone with her, there is so much noise in the background. It's my best friend.  And I say, how the heck can you have a conversation with me? She doesn't hear it. This is my backyard. Sorry for the tail end of my hound. Back in here is where a freight line goes  through every morning at 4 o'clock in the morning. Which in the summertime, the windows are open, I hear it. But I don't even realize it anymore.  And in the winter, my entire house shakes when the ground is frozen. I don't notice it. When I have company, when my family's in from out of town  or friends are over, they say, why did your house shake at 4 o'clock in the morning? Okay. Who's this guy? This guy's my dad. He was very involved  in the resistance during the Nazi occupation of World War II. And he smuggled cattle. Yes. Very appropriate that he is standing with a cow.  He smuggled cattle across enemy lines to provide food for the Allied forces and to get money for his family. I was walking through a very small town  in the Netherlands as an adult with him. And he said, you see that tower over there? The Nazis were in the tower. The English came through, bombed out the tower.  I said, what did you do? He said, I went in the ditch with the two head cow I had. I said, weren't you afraid? He goes, no, I got used to it. What? So I'm used to it.  My father's used to it. My friend's used to it. It is no wonder that we have an issue with alarm fatigue and not recognizing alarms in our own institutions. I sat at the desk.  The next desk over was one of my collaborating physicians. A vent alarm was going off. Not the hey, come check me alarm. It's the I'm disconnected from the vent alarm.  Nurses are sitting over there. Nurses are sitting over there. No one's getting up. Him and I got up. Patient had extubated themselves. Listen, it happens.  So we know the Joint Commission talks about, you know, there have been a lot of significant patient-related injuries and death related to alarm fatigue.  And in 2014, this became a national patient safety goal, as it should. Alarms are there to protect our patients for a reason. What's the issue? 5,076. A bigger number is 157,380.  That's the number of alarms we had in our 26-bed unit the month of July 2013, which is over 5,000 alarms a day. 200 alarms an hour, 24 hours a day,  which is three alarms every minute, every hour of the day. That's a problem. How did it begin? The staff came to our quality improvement program and said we can't do this anymore.  It's making us crazy. We need some help. And so what we did was we formed a subcommittee. And this is who was on the subcommittee. And what I said was this is imperative  that the nurses run this project because they are the ones that are there. A physician doesn't need to run this project. I'm a nurse practitioner.  I was the chairperson of quality at that time. I don't need to run this project. The nurses need to run the project. And they did. But we included key corporate players.  And I suggest if you're going to put together an alarm reduction project, you do the same. We decided we needed some baseline data.  We wanted staff survey, patient and family survey, perception of alarm. We needed an environmental assessment. We needed baseline or existing alarm parameters.  We needed to look what those were. And then we needed baseline alarm data. So we talked to our staff. And we asked them these questions.  What is your perception of a high level of false alarms? What training, do you feel like you have enough training to configure the settings? Do you disregard yellow alarms, right?  There's two alarms, the red I am dead alarm and the yellow, hey, you should probably come check me out alarm. One in four disregarded.  Sixty-one percent always went in the room to assess. Eighty-eight percent, the alarms are clearly visible to me. And again, only 82% adjusted the alarms.  They review the alarm settings and the alarm settings are included in handoff. We did a patient survey. Never, rarely, sometimes, often, always. Do the frequency alarms bother you?  Did it disturb your rest? Did you look at the monitor? Do you see this as a safety issue and did it cause you stress? You can see those results. We did the same with the families.  What I found most interesting is asking the family, did you look at the monitor? We all know it's critical care. They help us watch those monitors.  So we got some family survey data. Again, this was clearly a problem. Then we did an environmental assessment. We got Biomed up there. We had decibel readers.  And what we wanted to make sure was that we could hear every alarm from everywhere we needed to be without it being too loud to affect patients and families.  And we did that as well. Then we sat down with Biomed. At that time, we had two alarm systems. We now have one. But at that time, we had two alarm systems.  So we had representatives from both alarm systems to review what alarms were active in our monitors. This was like a meeting of the UN.  We met either weekly or every other week for about eight months to just go through this list to figure out what all the alarms were. And this is the list.  It is no wonder that we had that number of alarms. That was everything that could alarm on any patient every day. Are you kidding me?  And so then we got our baseline alarm data, which literally, I had to, you know, give away my two dogs and my car in order to get this data. But we finally got it.  Our new monitoring system now allows us to get this data. But Evanston Hospital had 157,000 alarms. And Highland Park had 61,000 alarms. Again, that's our baseline alarm data.  And that was monthly alarms. I don't memorize data, so pardon me for the papers. This brain is old and tired. I can memorize a lot, numbers I cannot.  So this was our baseline alarm data. And what we decided was we needed to determine what was going to be a critical life alarm and what was a nuisance alarm.  We defined what that was, and then we made action. We made recommendations that was reviewed by key players. And again, these were staff nurses  that ran this program and a quality manager. I was there just to give them guidance. And our recommendations were reviewed by these people and supported unanimously.  In our institution, since 2013, every alarm is now actionable. We, so for example, we got rid of PVCs and bigemini because who cares?  We actually took off the QTC alarm because the nurse measures at every shift anyway, and if we're concerned, we get an EKG. Believe it or not, we turned off the ST segment alarm  because it was constantly going off, and many times it was related to lead placement. So we got rid of it. We put it back on if we need it. We just got rid of it.  And we drilled down the list so that at this point, no alarm can be turned off. A nurse can make an alarm more strict on their own.  They just can't make it less strict without an order. And this is what our alarm parameters look like now. This is it.  We have tachycardia greater than 130, bradycardia less than 40, asystole if it's greater than 4 seconds, VTEC, VFIB, apnea greater than 30 seconds, O2 sat less than 85.  The leads are off or changing a battery. We also went into ICP, CPP, respiratory rate, our noninvasive, our pulmonary artery, and tidal and IMCO2.  The one important thing we said was pressure to disconnect cannot be turned off with invasive. So for example, your Artline or your swan gans catheter  if you have it, or a CDP, that cannot be turned off. Additionally, what just sort of came out of this by accident was we decided we needed a policy to sort  of set the stage because, again, at that time, we were a four-hospital system, and we were piloting this in two hospitals, one hospital had one monitor system,  the other had the other. So again, this is what I had mentioned earlier. The nurses can make it more strict on their own, but they can't make it less strict.  So for example, if I have someone in AFib RVR with a rate of 180, the nurse will say to me, can I up the rate to 200 while we're treating this? Yes, and I will write the order.  May increase heart rate to 200 for 15 minutes. And then they do it. And they're very good about that. We also created in our policy that there is full disclosure  on all alarms 30 minutes before and 30 minutes after every code. And obviously, the 30 minutes after are for the people that survived the code. But cardiology has taken this on,  and they review every patient that codes in the ICU to make sure an alarm has not been missed. And since 2013, we have not had any issues.  We also realized that we needed a transport policy. We needed something to guide us where when we're sending one of our patients from one institution  to another institution, we knew what we needed to do. So we rolled it out, and boy, did we have a problem. We rolled it out in two units at the same time,  and the noise was worse than it had ever been. Even though we had drilled down that list, there were two things that we didn't realize.  One, there was a setting in the monitors called latching. And what latching does is for every alarm, if the nurse doesn't physically address it, it just continues to build.  And then every alarm continues, continues, continues, and it ends up being this huge list that has to be addressed. Took us 24 hours to fix that.  Other than that, we really didn't have any other problems. So my suggestion, the caveat to this is, don't do this on a weekend. Do this when Biomed, I mean, Biomed was there,  and they were on it. And when I tell you, you cannot imagine the noise in that unit. Biomed was there. Clinical managers were there. I were there. All the staff nurses were there.  It was all hands on deck. But it got fixed quickly. So what we did was we wanted post-intervention data. Did we really make a difference? And we really found that we did.  This was our pre-intervention data. We had 157,000 alarms. Here's our intervention. And we brought it down to 62,000 alarms, which is a 60% alarm reduction. That's significant.  And this is in our big 26-bed ICU. In our smaller 12-bed, mostly medical ICU, pre-intervention, 61,000. We did our intervention down to 29,000, 53% alarm reduction.  That's significant. Evanston Hospital, the administrator said to us, yeah, but did the alarms go down because we had less patient days?  So what we did was we took that data and compared it to patient days. This green line, which is pretty hard to see, was our goal to reduce by 50%.  And you can see, based on patient days, we were well under. And we did the same thing at Highland Park, which is our smaller hospital. So we rolled out our next two hospitals.  We did education. We did the same surveys. We got our baseline alarm data. We rolled out. The good thing with the next rollout is we knew about our latching problem.  So that didn't happen. And so then what we did is when all the hospitals were rolled out and we had our data, we surveyed the staff again.  Do you feel there's a high level of false alarms? Down over 10%. Do you feel like you have adequate training? Even higher. Do you disregard yellow alarms? We're down from 25 to 14.  Do you always go in the room when there's an alarm? It's a yellow alarm. It's up a smidge. Are the alarm settings easily visible? Yes. Do you always adjust your patient?  Well, this is down to 64%. Why? Because they don't need to adjust the alarms because we have a system now that every alarm is actionable.  We're not getting all these nuisance alarms. Do they review the alarms at shift and then in handoff communication? Patient survey. This is when the patient answers rarely or never.  Did they notice the frequency alarms? Less. Did it disturb them? Less. Did it cause them stress and anxiety? Less. We did the same thing with the family. Did it bother you?  Less. Did it disturb your sleep or rest? Less. Did it cause stress and anxiety? Less. So did we achieve our goal? Well, in three hospitals, we did.  That is significant alarm reduction in three hospitals.  And again, we took our third hospital that we rolled out and based it on patient days. And here we had a little glitch, but you know what?  At this point, one of three hospitals, not bad. So our summary of my strategy for this team, we had a committee with key players. That's paramount for you.  You have to have administration buy-in, and you have to have administration support.  You have to review your alarm parameters. And trust me when I tell you you're going to get a list like that and just know that it's going  to take an act of Congress to get everybody in that room, but it's necessary. And you've got to review them line by line by line. Again, six to nine months.  Get your baseline alarm data.  You have nothing to compare. You don't know if you've made a difference if you don't have baseline alarm data.  Consolidate your alarms. Write a policy. Roll it out. Review your data. Resurvey. And plan to continue to review your alarm policy and any ERAs or root cause  if there are any issues, challenges. Again, to get everybody in a room was painful. There were also people in the room that have different agendas.  All I wanted to do was reduce alarm fatigue.  We had pages and pages of alarms, as I talked about. And the two monitoring systems, that was a challenge. Because sometimes what one could do, the other could not. Unchallenges.  Nursing led this project. They bought in. They were very engaged. Staff really accepted the change, and they were looking forward to it.  Our families and patients were so eager to participate. We were thrilled with that.  We had support from administration, biomed, and engineering.  What are our wins?  Significant reduction in alarms. Our unit is so much quieter. It's fantastic. Everybody's more satisfied.  And we've maintained patient safety. Because I'm going to jump down to this bottom one.  Since 2013, no ERA incidents since alarm reduction. And with the cardiology review of any code in ICU, no missed alarms.  That's significant. It's 2020.  Our last review was in September of 2019. No issues.  Again, our alarm parameters now are standardized. We presented this data at NTI, I think, in 2015. We're presenting it to you today.  And our organization has received several awards because of this work.  Can alarm reduction go too far? Of course it can. Of course it can. You have to be thoughtful about your alarm reduction.  And again, as Sandy said, as I pointed out, we're looking at a lot of alarms.  So you really, really have to be thoughtful. And it's important to remember that alarms are not just the alarm. It is the alarm.  But it's also what the human does.  Do they, how do they perceive it? And how do they react?  What are your organizational workflows and your technical understanding? And all I can say is you have to monitor your data. You have to have something in place.  Once you drill down your alarms, you have to have something in place  to maintain your patient safety. Again, this is the recognition of the awards that our organization has gotten.  And again, the presentations that we've been asked to do on this project.  Going forward, again, our performance review reviews our alarm data. We also review the ERAs. We still have nuisance alarms. I'm not going to lie to you.  We've not worked on pumps, vents, pulse ox, other than the alarm parameters in the monitor. Because frankly, after this project, I had significant alarm fatigue.  This project took years. It's worth it, but it took years.  Again, thank you for your time. I hope that you found this helpful. And I am going to turn over to Brad, who is next.  Good morning. I'm Brad Winters. I'm from Johns Hopkins Medicine, where I'm the Division Director of the Critical Care Medicine Division at ACCM and with the Armstrong Institute.  And what I'm going to talk about today is going to expound a little bit on what Jackie and Sandy presented.  Specifically, I'm going to try to focus on evidence in the literature for what works in managing alarm fatigue, because it's definitely going to expound a lot  on what Jackie just showed you, showing that their experience at North Shore Hospital is not unique. That other folks have been able to do this as well, and that you can as well.  So, a lot of the stuff I'm going to be presenting came originally from the Society for Critical Care Medicine Alarm and Alert Fatigue Task Force.  And as Sandy pointed out, we published two papers in critical care medicine about a year and a half, two years ago now.  I'm going to focus on alarm fatigue as opposed to alert fatigue. So, I would recommend that if you're interested in CDSS, alert problems, et cetera,  contact Sandy, because she's the expert in that area, or review our article that we had published a couple years ago. And this is the team that helped do this work.  So, as you heard, we are drowning in alarms, okay? Every day, they are everywhere around us.  The ranges depend on what kind of unit you're in, depends on which papers you look at. But clearly, we all know we are drowning in them.  And very few of us are sleeping blissfully like this guy, surrounded by the alarm clocks, because of it.  Alarm fatigue is an ever-increasing problem. But fortunately, we actually have a good definition for it.  It is the desensitization resulting in the dismissal and or delay of clinical response to alarm secondary to exposure to high rates of alarms.  And it is thought that invalid or so-called nuisance or false or non-action alarms are the ones that are really the problem, not just the total alarms,  although there are some folks who believe that the total alarm volume, even if accuracy of alarms is high, is also a problem.  Instances, as you've seen, have been documented where patients have died because of alarm fatigue and alert fatigue.  Clinicians have turned off alarms and missed critical events that have led to patients' death.  As such, alarm fatigue is not only a threat to us as providers, through psychological exhaustion, burnout, and other issues, but also a threat to our patients.  A little bit of word about metrics, okay?  As you might not want to use a micrometer to measure a fish like these guys have caught, while we have a definition of alarm fatigue,  we don't necessarily have a universal metric for measuring it.  People look at total alarms. People look at non-actionable alarms.  People look at the duration that alarm goes on before it gets turned off or addressed. People look at the incidence of disabling and silencing alarms.  Unfortunately, no single one is likely to be the best metric.  And, of course, these metrics don't account for the psychiatric effect, the psychological effect that we have on ourselves.  So far, there have been no studies that measure the effect of an intervention on alarm fatigue using a universally validated tool.  In particular, one study by Deb and Claudio back in 2015 did make an attempt to assess alarm fatigue by measuring what they called performance deterioration.  Unfortunately, while this was a very interesting study, this is not a widely validated tool.  The first and only validated tool that I've seen in the literature and that we were able to come across with our alarm and alert fatigue task force was, in fact,  published about three years ago by Tor Bazzetta in Journal of Clinical Monitoring and Computing. Now, it was a survey-based one, which is not inherently bad,  but it did not include some hard measures like number of times an alarm was disabled.  But it does have some other problems. It was primarily geared towards nursing. So while nurses or the frontline providers usually at the bedside and take the brunt  of alarm fatigue, it didn't measure how alarm fatigue might be measured in physicians, respiratory therapists, PAs, NPs, and other practitioners.  And so far to date, based on the alarm and alert fatigue task force review and my update that I did in preparation for this talk today, I have not been able to find  that this validated tool has actually been used to measure an intervention. Most of the surveys that you see and tools that people have been using are kind of homegrown  and kind of one-offs for those individual studies. And so like Potter Stewart, the Supreme Court Justice who back in 1964 said something  about lewd and lascivious publications, that I know it when I see it.  Well, we know it when we see it, but unfortunately for evidence-based medicine and knowing what interventions we are going to be able to apply successfully, that is insufficient.  So what are some of the components that we might measure?  Well, we've talked a little bit about some of them, alarm pausing, disabling rates. Performance deterioration is an area that really needs to be investigated  because it really gets at the heart of all the parameters, the psychiatric, the physical, etc. It may be, however, influenced by workload and patient-to-ratio numbers  and therefore is going to have some confounders.  Staff turnover is very multifactorial and has resiliency variation but is also potentially part of the mix. And again, coming back to Deb and Claudio in 2015,  noted that alarm fatigue is a lot more complex than just the total number of alarms or the accuracy of the alarms.  And so a metric that is going to be valid for doing this kind of work is going to need to be very all-encompassing.  So as it stands now and for the rest of my talk, we are going to stick with the total number of alarms going down or the non-actual alarms going down are going to be assumed  to be a leader to reduction in alarm fatigue.  While it has tremendous face validity, keep in mind that it is not certain that it attacks the entire problem.  Just one quick word about human factors engineering. We have very much focused on how we interact with the devices.  So for example, the famous control alt delete for rebooting your computer. But if you think about guardrails unlike infusion pumps, they're primarily designed  to prevent the human from reprogramming or interacting with the machine incorrectly. When it comes back the other direction, we have actually done a very poor job in figuring  out how to get the machines to then interact back with us so that we don't experience alarm fatigue and that our patients are best served. So let's focus on what works.  Okay. First half of this, I'm going to talk about a lot of bundled bedside tools that you can use and have evidence to support them.  I'm also going to talk a little bit about some technological approaches which actually you will see have not been applied directly.  First thing to consider to reduce alarm fatigue is whether you should even be monitoring that patient at all.  I can tell you I get asked all the time to admit patients to monitored beds or my ICUs who clearly have no business being there just because somebody is nervous or wants to,  you know, as they say, CYA.  Okay. We should reconsider about who should be monitored for how long and for what.  Timely discharge from monitoring through processes such as alarm huddles or alarm rounds or alarm reviews have been shown to be very, very effective.  Curtailing this use of telemetry to only high risk patients has also been shown to be effective.  Stoltzfus back in 2018 was able to show that they were able to reduce telemetry utilization by almost 18% and improve adherence to AHA guidelines through an alarm review process,  basically asking clinicians to reconsider or putting hard stops on whether somebody should be admitted.  However, they did find that people still tended to try to leave the,  where they were supposed to document why they wanted the patient placed on monitoring blank showing that there was a lot of work that could still be applied to improve the process.  Optimize the technology. Many false alarms actually come from faulty sensors. This is particularly true for things like pulse oximetry and EKG.  There have been several papers that have shown out that timely changing of electrodes and attention to the quality of the interface between the human and the machine  that is monitoring them is really, really effective at reducing false alarm rates.  Choosing the best technology based on the literature is also crucial. That is something you have to do locally, but there's a lot of good evidence out there  about which technologies work and which ones don't.  Utilize that literature before you make a purchase and implement a particular technology.  And again, as I mentioned before about human factors engineering,  optimize the transfer of the data to the humans. Again, an area where we do need to do much better.  Whoops. Customize.  Okay. You can have a sweet ride like you see here on this picture. I hope I don't insult anybody who owns one of these, but I certainly don't have one.  You heard a little bit about customizing the default preset alarms from Jackie. The settings that are based on machines as they come from the manufacturer,  the defaults, have no basis in reality.  They are primarily geared to protect the companies from lawsuits. Okay. Alarms can be customized. They can be customized based on the kind of unit,  the kind of patient, and on an individual basis. In fact, there's quite a bit of literature showing  that you can statistically drop the alarm rates and false alarms by customizing alarms for the patients. This has been done by Fujita and Choi, by Graham and Sivach,  and Ketco et al., Dandoy as well.  Alarm rounds interventions that then take the approach to customize the alarms as well as reviewing whether patients should even be on the monitor can be extremely effective  at reducing nuisance and false alarms.  Contextuality of alarms also matters and can be part of that bedside alarm review or that alarm rounds. For example, one study showed that they took patients who were in the ICU  for primarily other monitoring, such as respiratory, and said, okay, these people are more likely to be hemodynamically stable.  And then they compared them to patients who were in the units primarily for hemodynamic monitoring and customized the alarms depending on why the patient was in the ICU.  This was very effective and led to virtually no difficulties after disabling the alarms. You can also adjust alarm notification delays.  These can be built into your middleware software or into the primary system depending on what you're doing, and studies have shown  that many abnormalities are ephemeral or short-lived, often due to movement artifact, and they tend to go away after a few seconds.  So for an alarm to be truly real, it has to persist or it has to be occurring in the same fashion over and over again.  And by putting in parameters that suppress that, you can reduce false alarms.  In particular, a recent study by Ruppel et al. showed the alarm customization software, something called IntelliView, that prompted the nurses  to reanalyze the customization parameters for that patient's alarms. That is, it didn't actually customize it itself. It worked through a human interface.  We're able to reduce false alarms and total alarms and improve nursing satisfaction.  So many of these have actually been bundled up together, and if you'd like to take a picture of a slide to take home with you to help guide you for how you might try to do this,  along with a lot of the work that Jackie presented, this is the slide to take a picture of. There are many clinical studies that mix and match individual interventions.  Alteration of default monitor presets, which I just talked about, daily electrode change, better battery management, timely discharge from the monitor or not monitoring at all,  alarm customization, education programs, change in policy, notification delays,  and improved displays going from the machine to the nurse or providers. All strategies that were evaluated by the task force found that either singly  or bundled resulted in significant alarm reductions, anywhere from 31 to 89%.  And most importantly, as you'll see at the bottom of this slide, just as Jackie presented for North Shores Hospital, none of these studies reported an increase in adverse events.  These clinical strategies that you can implement now, granted it may take you nine months, as Jackie pointed out, to actually get them implemented,  but these strategies are proven and are safe.  You can reduce alarm fatigue and alarms and nuisance alarms by these individual strategies.  In fact, the AACN has actually bundled an acronym to help guide us in this regard. It's called CEASE, Communication, Electrodes, Appropriateness,  Setup Custom Alarm Parameters, and of course, Education.  So let's turn a little bit to something called Algorithm Machine Learning Based Alarm Suppression.  These are technological features that exist within the alarms, within the waveforms  that we monitor every day that can be used to try to control the process.  Most studied are EKG alarms using arrhythmia databases such as the MIMIC2 database. Many of these involve approaches such as using arterial blood pressure signal quality  to guide alarm suppression, relevant spectrum machine learning, various feature discretization, feature selection and classifiers, wavelet analysis.  Again, there are many, many different computer engineering tacks that can be used here, and I won't go through them in detail.  You can review them yourself, but all these different approaches to suppressing EKG false alarms, which are the most prevalent ones in ICUs  and telemetry units, was successful anywhere between almost 20% up to about 100% depending on which parameter you were looking at.  Incredibly successful technological approaches to suppressing alarms. Unfortunately, as you can see, they missed true alarms as high as almost 20%.  And this is probably the core problem with a lot of these technological approaches and why you have not seen them actually implemented  in commercially available monitoring equipment.  While the clinical bundles seem to have almost no adverse event rate, if you were to apply these to a clinical patient load, you would miss true alarms.  And for that reason, actually, if you look through the literature, you will actually not find a single study that actually did it prospectively on real live patients.  All these studies are based on databases that are either from patients that were collected retrospectively or prospectively, but are not used live.  And this is part of the problem with them, is as you suppress more and more false alarms,  you unfortunately, with some of these strategies, increase your rate of missing true alarms.  We'll skip that slide.  Alarm delay algorithms are also effective and can be built in along with some of these technologies. Again, all these have been done on databases, not on real people.  SpO2, unfortunately, even though it's a common cause of false alarms, has actually not been studied very much.  Although, there was recent one study using the Adaboost algorithm that was fairly effective.  And even intracranial pressure alarms have actually been approached in the same or similar fashion, but like EKG alarms, have had the same problem.  The more you suppress false alarms, the more likely you are to miss true actionable ones. Time series analysis is another approach, which I won't go into details.  But again, very, very effective in suppressing alarms. But like all these other machine learning algorithm approaches,  still misses an unacceptable number of true alarms and could harm a patient.  Smart alarms, where they integrate the alarm technology, so trying to make things a little more the way we think.  Like when you look up at the monitor, and the monitor's telling you, oh, the patient's got some EKG abnormality.  But you look at the arterial waveform and the pulse ox waveform, which look perfectly normal, you know, well, that's not unstable VTAC, and that's not VFIB.  Trying to get the machines to do it the way our brains think is an active area of interest. But again, none of these technologies have actually been implemented for real,  although many of them are very effective in the laboratory, in the computer lab.  Reasoning algorithms, again, as I mentioned, grouping alarms and how alarms need to be familiar and similar, rather, before a notification is sent, so that you have congruency  between different pieces of data, instead of single stream pieces of data,  which inherently, by themselves, are inaccurate.  And motion suppression algorithm development has also been researched recently. This is under the premise that a lot of false alarms, both in pulse oximetry, ICP monitoring,  and EKG, are actually due to movement artifacts, trying to use videography and other technologies to know when the patient is moving around,  and thereby suppress the false alarms, is another area of active interest. But again, none of these are commercially available in equipment.  This becomes really important, an area that I'm very interested in is continuous vital sign monitoring for general ward patients.  A lot of these issues are going to become more problematic, as we try to expand the level of monitoring to hospitals, as patients in hospitals become sicker and sicker,  and we try to expand ICU care beyond the actual ICU walls,  where alarm fatigue is going to be even less tolerated than it is in the ICU.  Super alarm patterns may be very effective here, because again, you have to have multiple parameters of instability to predict that a patient is going  to deteriorate, as opposed to single parameters.  And again, integration, in order to control alarms in this area, is an active area of interest.  Human factors approaches, coming back to how the machine interacts with us, different ways to approach us with the alarm data,  may also help improve alarm fatigue, and is an area of active interest.  So in summary, rigorous, pre-reviewed literature is very limited.  No one has examined the problem of alarm fatigue directly, and metrics are fairly limited, and not universally agreed upon.  All studies to date equate reductions of false alarm and total alarms with reduced alarm fatigue, which again, has faced validity.  Most of the lab studies, while effective, do not suppress alarms without or leading to a risk of missing true events, and therefore, are not, as we say, ready for prime time yet.  But we do know that clinical bedside bundles, and approaches where we do things like alarm huddles, and customization of alarms, and reconsider who  and who should not be alarmed under what parameters, are effective and safe. Thank you very much. Well, I'm very grateful to all of my colleagues.  Excuse me while I set my alarm, and I generally don't have alarm fatigue. I just ignore my own personal alarms. My name is Barbara McLean.  I'm a clinical specialist and nurse practitioner at the Grady Health System in Atlanta, Georgia. And my role there is to bring new initiatives to the bedside,  and to evolve our nursing care practice, and our physician care practice in the utilization of new technologies. So, that's part of what I do.  So, I'm very grateful that I was invited here, and very grateful to my fellow panelists, and to the Society of Critical Care Medicine for inviting me here today.  I am going to disclose, and I'm sorry I don't have a slide with my disclosure, that I am on the Speaker's Bureau for Baxter, and for Edwards Life Sciences.  And because what I'm going to talk about is hope for our future in terms of leveraging technology, I'm actually going to speak to the experience that we have at Grady Health System.  Grady, for those of you who don't know, you often, some of you might confuse us with Emory. We are not Emory. We are Grady.  Grady is a public hospital. It's funded by the Fulton DeKalb County Authority.  We welcome residents, and attending physicians from the Emory School of Medicine, and also from the Morehouse School of Medicine. So, we have two medical schools.  We have now initiated a program that actually employs APPs and physicians, employs them at Grady. But prior to that, we had these multiple medical schools,  very different ideas about how to manage our patients.  We are slotted for 1,000 beds. Normally, we operate, we always have some beds in closure, around 780 beds. We recently had a flood. We lost 200 of our, 250 of our beds.  And we're quite impoverished. So, I think what I'm going to share with you is going to be really surprising.  It's always surprising to me when I review the technology that we have at the bedside,  because we have really elevated our technology in the attempt to provide best care for patients.  So, we've heard really eloquently from my fellow speakers why we care so much about alarms, why they matter to us so much, and what it is that we're considering when we talk  about alarm and alert fatigue, and reminding ourselves that since, actually, 2010, this became a primary initiative and agenda, and now it's, of course,  been incorporated by CMS and by JCA. I always confuse those initials, just to remind us that these are initiatives worldwide  that we are all taking on in order to improve care and to improve patient safety.  But I would venture to say that more education, more rounding, more huddles is an everlasting process. It's not one and done.  It's continuous, and it requires a tremendous amount of time and energy. And what I think is most important for us to remember is that acknowledging  that we have a problem doesn't fix it.  And from my point of view, and I'm quite sure that from my colleagues as well,  is the more we can reduce the possibility for human factor mistake, the better off we're going to be. So, we all know we have this problem, but what are we doing about it,  and how can we move forward in a platform that brings us towards the future? I keep going backwards. I'm so sorry.  So, I'm going to mimic a little bit what Sandy asked you, except now I'm going to talk about per year.  So, 2,000 or less alarms per year in your ICU. Just your ICU, not your hospital.  2,000 or less, 20,000 or less, 200,000 or less, nobody's raising their hand.  Two million or less, two million or more. Okay. Well, those who said two million or more, you're the ones who are right.  More than two million, and actually not per year, but even per month.  That's an incredible number, and the reason I chose two million is because I was really fortunate yesterday to visit the NASA Space Station,  and our astronaut who was our guide said, two million parts must function  within this redundant system of the space shuttle in order for it to operate effectively.  And I loved that two million analogy, and I thought, this is a very good thing for us to bring forward.  Now, a lot of what I want to talk about is what we're doing classically ourselves, but of course there are so many contributors to the evolution of our understanding of alarm  and alert fatigue and methodologies for management. One of the most important ones being Marie Savard, and I always mispronounce, but I listened to how you did it. I still minimise.  So, just really understanding that we've had a focus on this for a really, really, really, really long time.  And as we move forward, I just want to remind you, and so I'm going to just show you, this is what we're managing at the bedside in our standard critically ill patients.  So, you can see all the points of data that are on the bedside monitor. Bedside monitoring alarms contribute about 37 to 40% of the alarms that we have  to actually consider at the bedside.  Ventilator alarms, 47 to 50%.  And then for us, CRT alarms, which is another 10 to 20%.  And for us, minimal invasive alarms on a separate monitor. So, in my room, in a relatively basic patient, we're utilising these four separate monitors  that are all alarming independently. And of course, they could all be relayed through middleware, and that middleware could relay to Epic.  But we're not using much middleware technology in my institution at this moment.  So, I think what you've heard very eloquently, and that really is the point,  is we know how to collect the data. We've done a pretty good job of collecting the data.  We have the data, and we have the technology to capture the data. And all of these different perspectives of making sure that we're utilising all  of our technologies to their best benefit. And we have staff who are performing work.  But I think one of the most important things, and I really appreciate that Brad really did just,  he lent a lot of credence to that by saying, we know how to do it, but we have to find ways to deploy it to the staff so that they can see it, so they can understand it.  So, they can implement their changes in daily practice.  So, first, just to remind you about physiologic alarm management systems.  Every monitoring company right now, every monitoring company that you can utilise has a methodology for this alarm technology management that takes real-time data,  stores historical data, allows you to come back and look at that data, configure the views, configure your alarms to make your alarms appropriate for each patient.  Let me guarantee you that most nurses taking care of highly critical patients,  that's the last thing they think about is, how do I actually make my alarms fit my patient's need? And my patient's need is changing every hour.  So, we're not changing those alarms at the bedside every hour.  What we really would like to have is middleware that has some learning,  machine learning-based efficiency to help moderate the alarms that are appropriate for the patient. So, I'm showing you here a combination of the Philips monitoring devices,  not because I'm advocating for Philips, but because it was a beautiful picture that I could get quickly and display to you.  And I want to also remind you that every, so, every big monitoring company has capabilities that you may not even know you have, and that's really important.  It's to investigate what those capabilities are to help us actually manage our alarms.  But no matter what, this still requires human interface. So, one of the things I remind us is that as we move forward to AI,  AI requires NI, nursing intelligence, and PI, provider intelligence, all together to actually control our alarms and utilize our technology effectively.  So, I'm sorry because it looks like I was out of order. So, I'm going to go back to this. I'm going to tell you something that we're using at Grady.  We have it integrated into our surgical ICU and our neurosurgical ICU. We have 120 ICU beds, direct ICU beds, and then another 150 step-down and telemetered beds.  Now, we're using this tool. It's called Bedmaster. It's from Excel Medical, an alarm navigator. So, what this is is a deep clinical alarm data tool.  It allows us to analyze what alarms are going off, when they're going off, what kinds of issues were related to that, and it helps us to actually go back  and evaluate what alarms occurred, and to be able to utilize our clinical management staff, our educational staff, and other stakeholders to analyze retrospectively what occurred  and make small changes like daily lead, daily electrode placement, making small changes that affect big reductions in alarms.  It allows us to review all real-time waveforms and pull historic waveforms of any alarm condition that occurred. We can look at all those waveforms. And why does that matter?  Because it takes this complex data and brings it to a place where we can evaluate it and then deploy that information to the staff, to work with the staff to help  to reduce those alarms that are unnecessary and non-actionable. And, of course, you know and you appreciate, so I found myself using this a lot  in my discussion, but it really came from ECMO, which is alarms 95% of the time are non-actionable, and then that 5% of the time you have sheer terror,  just like with ECMO, 95% of the time it's boring, 5% of the time sheer terror,  just like your flight, 95% of the time smooth as silk, maybe bumpy, 5% sheer terror. So really important for us to appreciate that this helps us to really focus  on those sheer terror alarms, asystole, ventricular tachycardia, not short bursts but longer and sustained,  reduction in arterial pressure, changes in the PA pressure, changes in CVP and changes in pulse ox, all of which are alarming at the bedside.  This allows us to go back and adjust our clinical workflows. I'm just going to share a few visuals from that physiological arm system. I keep going backwards.  I'm not a very good right and left side. I'm a Mac. Okay. So if we're taking a look at this perspective, at this perspective, what we're seeing here,  I'm just going to call your attention to that column here that tells you about the recent alarms in four hours. So look at all these patients. These are not real names, obviously.  This is all changed to protect the innocent. And I want you to see how many alarms these patients are having over the prior four hours. Four hours, 9,488 alarms.  And the great percent of them were alarms that could have been controlled through human  factor manipulation simply, most simply related to electro change and appropriate leave placement.  So really important when we look forward at that to also show that we can, and again, this is not the real name, that we can go in and actually capture our real alarms.  And of course, this is what we talk about. When we talk about alarm fatigue, I'm not really saying that the nurse is tired of listening.  I'm saying that they desensitize themselves so much that these real alarms that actually resulted in the loss of the pulse, these are the alarms that cannot be ignored.  And when we have all of these false alarms, we do ignore real alarms. And I loved what Jackie said because that's absolutely reproducible in my system.  Five people sitting at the central station, a red alarm is going off and nobody's answering it.  So it is vitally important to remind ourselves of what categorically we're trying to do, which is call attention to real alarms.  So again, I want to show you that this was utilized in another system, not mine, utilizing real alarms and false alarms.  The alarm committee who had said, we have about 220 alarms per month, they had two and a half million alarms.  And through review, through the alarm committee, through stakeholders working together, they were able to reduce their false alarms by reviewing data every day, by sending it to  the staff, talking with the staff about it, engaging the staff to initiate projects rather than those of us who are at a more expert level or designated as experts within our  institution that we really wanted staff ownership. And deploying all this data to them has really helped.  Now when we look at these types of systems, these initiate alarm watching for specific units, for specific types of alarms, and it helps us to drill down to when those occur,  how often they occur, what time that they occur, and to help us view those individual alarms.  So I'm going to move through the rest of this and then bring you to here, which is delivering data to the person who is responsible for the patient.  So I don't think anyone would argue that smartphone technology is not the wave of the future. It is the wave of the future, smartphone technology.  Now a smartphone, I'm a Mac user, so for me the only smartphone is an Apple phone. But that obviously is not true.  And my system has actually, is developing strategies to send all information to rover phones so that instead of a big central station alarm, the alarms will be deployed to the  nurse who's caring for that patient that day. And if those alarms are life-threatening alarms, they will be sent, if the nurse hasn't acknowledged  the alarm within 30 seconds, it will then be sent to all the staff so that someone comes to the bedside when a patient is exhibiting a true red alarm to identify what's occurring.  So this is, I think, the wave of the future. It's a big investigative, or it's a big investment of your system, but can really help us to  deploy that data to take the source of the data, run it through middleware, actually filter the lesser alarms, put them at a lower level, and support the higher alarms being  deployed first to the bedside nurse and then to all the staff. And it should always mimic the multi-parameter view at the bedside.  So it's not just EKG, but all of our pressures and our pulse ox, and eventually, I hope, our ventilator data as well as others.  So keys to an alarm management system is that you have to have an alarm analysis tool that actually gives you a data-driven approach, helps us to access, understand, and manage  those alarms, provides immediate access to that current and retrospective data, allows us to query and make changes in individual units and specific patient alarm limits, and  to guide the team in better decision-making. So along with that, you do actually need yet another committee, another meeting to attend.  You do need an alarm committee to help reduce these non-actionable alarms and trying to achieve 100% specificity and 100% sensitivity for true alarms that occur at the bedside.  So a really beautiful, beautiful article published in 2018, which was about utilization of smartphone technology, particularly on a telemetry unit.  Beautiful article that talks about developing this strategic software for telemetry system monitoring and deploying that information through middleware technology, using fuzzy  logic and other methodologies, and actually deploying those alarms strictly and straight to the patient.  So there are many sites that are utilizing methodologies to deploy information to smartphone.  The last thing I want to talk about, and I'm going to do it quickly, is just we actually at our hospital have just converted to the PRISMAx, not PRISMAflex, PRISMAx.  But I want to just show you what we can pull from our machines at the end of every month or when they're not in use at a patient. And we can identify patients.  We can go back and review their records. And what we talk about a lot is filter life and access and return alarms. Now those alarms are going off at the bedside.  And what we're able to do is actually come back and say, we have a problem with our VASCAS. So we changed our strategy.  You cannot place a VASCAS that isn't at least 14 French in diameter. And we've done a big educational program with our providers about the size of the catheter  that has to be placed in order to try to maintain better filter life. And you can see significant reduction.  And then when we looked at our average filter life and our concerns were that we did not have a defaulted anticoagulation for filter policy. We don't use currently citrate.  We use heparin. So there's always some issues associated with that. But again, looking at what was occurring.  So we're having lots of alarms about filters and access and return and delta pressure and TMP.  But by actually reviewing with our providers the reason we need to anticoagulate filters in order to optimize CRRT. So very, very important.  And then the last thing I'm going to talk about in my last two minutes is utilizing electronic glucose management system.  Now there's more than one provider, probably even in the exhibit hall. We are currently using Glucomander, which allows us to do very significant evaluation of our patient.  It's an external-based website. And it deploys to a central station. So here's the problem, of course. To the central station, these are not real patients.  So please don't think I'm violating HIPAA. But what we see at our central is we see each patient. We see the nurse who's taking care of the patient.  We see what their insulin rate is, what their target range is, and when their next blood glucose is due. And we have a countdown clock. Now that alarms.  And it's a high-pitched alarm. And it alarms at the central station. And it alarms right where our physicians are rounding with their wow.  So in my hospital and all the SUs, I'm the Glucomander clinical lead, I have replaced speakers to our computers 14 times in one unit. And in other units, 7 to 12 times.  Because people say, oh, that alarm is shrieking. It's bothering me. Turn it off. Or they said, oh, those speakers look nice.  Let me detach them and take them home. Either way. So what we've really discovered is this is an incredible technology that we have to take  off of central station, that we have to deploy to our staff. So again, our answer is within a year to a year and a half, we believe our rover phones  will actually be individualized and we'll be able to tailor all of our alarms. But regardless of that, I just want to tell you about the goal of the alarm.  So even when the speakers have been lost, if you look at the total patients we've had on this is in three years, and this is almost my last slide. So in three years, almost 2,500.  Almost 100,000 blood glucoses. And hypoglycemia of less than 40, 14 blood glucoses. We've reduced our sentinel hypoglycemia events to 0.01%.  So really important to understand the leveraging of technology is going to bring us to a much higher level, to a whole new view.  And as Bradford talked about, alarm bundling, alarm filtering. What I'm reminding you is that I think the wave of the future is customizable software tech, smartphone technology.  Multiple filters that can be learned by the machine and that we can then customize. And then most important, and as Jackie has mentioned, and she did so eloquently, you  have to have champions to collect and review that data. And then you have to have real-time relationships with stakeholders.  And not necessarily, if you don't do this, you will get that. But really engaging and really trying to have control.  So fear of the present, which is where we are today, is actually the hope for the future. We also want to look at methodologies.  And we use SPOKE to alert an alarm for STEMI, for CODI, for stroke. And as you can see here, this very highly complex slide is that SPOKE has the capability  to help us to deploy to the individual nurses what's occurring to their patients at the bedside.  So behind this, and unfortunately the animation is gone, this is the Atlantis shuttle. I just want to remind you that these are operational pieces, small components, which can cause  complete failure.  So we do this with caution. We try to create reliable, redundant systems that actually can deploy us to the future and take us much farther forward than we have been so far.  Thank you very much. And now we'll begin the panel. Thank you so much for your attention. So we'd love to have questions from the audience.  We didn't take questions or answers after our individual talks so that we could have panel time together. So any questions? So I have a question, and my question will be for Jackie.  So Jackie, in your system, you talked a lot about how you had this generation of the committee that really looked at that data.  How much engagement and ownership did bedside nurses actually provide? So the project itself, when we started the project, I had four bedside nurse champions,  one from each facility. And then they went back and worked with every one of their colleagues in their individual ICUs. So they really owned it and stuck with it.  Now, I will tell you, because the project took so long, two of the four nurses dropped out of the project.  But the two that stuck with, Dan and Amy, they then took over those other units. And they would go meet with that staff regularly to say, does this work? What about this?  What do you think about this? So there was a lot of involvement. And really, they're the people that are there, right?  People are in and out, in and out, in and out all day, but they're the people that are there. They were very invested. They were supported by their manager.  I had to go to administration for support for those initially four nurses and then two nurses to be paid for all their additional time.  And then I also secured monies for them to present at NTI. Because I said, listen, we're going to submit an abstract. I don't know if it's going to be accepted.  But I do not want these staff nurses paying for this out of their pocket. Administration said, if it gets accepted, we will support all their expenses.  So I mean, that was sort of the enticement. They were excited, A, to be able to present at a big national conference, and B, to be able to travel without expense.  But more so, they really just wanted the unit quiet. They wanted the unit quiet. We all did. They brought the concern to us.  They brought it to critical care performance improvement and said, we can't do this. And in fact, the two nurses that stuck with it were also both the performance improvement  representatives for their respective units. So it really, the project would not have been successful, in my opinion, without nurses driving that. And they really drove it.  In fact, they were the ones at the time, the monitor system couldn't give us that data.  And we had to purchase software. The software we had to purchase was literally $100. And we were turned down by administration repeatedly.  And the two staff nurses said, in a big, huge corporate meeting, we're each going to chip in $50. Can we have this software? And then administration came through.  So they really were committed to this project and really did a great job. They bought you on $100? Yes, $100. Administration bought me on $100. Listen, it's what it is.  Yes, question from the audience. Yes, I'm a manufacturer. What can manufacturers of monitoring equipment do to help you do this? Good. Thank you.  So I'm going to repeat the question. What can manufacturers of technology and the monitoring equipment, right, monitors, what  can they do to help us develop new methodologies and strategies to help us have more reasonable alarms and not miss one actionable alarm? That's really the issue.  Because the issue is, right now, all of our filtering actually can prevent us from receiving our alarms.  So I think this is a good question for Bradford and for Sandy, and then the rest of us can chime in. Sure.  I would say, given what we know from the studies done with the MIMIC database and some of these  other areas, that until we know exactly which parameters we can suppress without missing real events, I'm not sure that manufacturers are in a position to try to integrate some  of these software technologies in, except perhaps, like, for example, asystole is one particular EKG parameter that we are able to suppress 100% without missing true events.  CTAC is the really difficult one to suppress effectively. But in some areas, some can be suppressed.  I know manufacturers would probably like to have a whole package that addresses all EKG waveforms, but we may not be able to get there, and so we may have to pick and choose which  ones we can put into software packages and which ones we can't. But I'd say, right now, probably the best thing we could do is, given what we know about  bedside huddles, alarm huddles, especially alarm customization, is make it easier for the bedside nurses to implement alarm customization.  I can tell you if, I mean, the folks who have done bedside nursing, which I haven't done, will tell you, trying to go into a monitor and customize every parameter for a patient,  no nurse has time to do that. If it was something that could be done very straightforward and more easily, I think alarm  customization would be a great bedside implementation that could happen in the next couple years.  We know it's safe, especially when coupled with reviews, huddles, where we say, yeah, this patient is safe to expand their parameters. This patient, no, we need to tighten them up.  Basically every patient could have their own set of parameters. Until that is easy to program in, nobody's going to do it. It just takes too long.  So that would be one thing I would say right up front. Yeah, Sandy. So my perspective is a little bit more from alerts, so more around the clinical decision  support, but it's actually similar to Brad's in the words about customization. So the first thing is that when systems who put EHRs in place and use clinical decision  support systems within those EHRs, they're highly inclusive, right? I think it was you eloquently put CYA, was that what it was? Yeah, I didn't want to say the whole name.  So they're very inclusive in, you know, go into great detail. And put alerts in place that are very comprehensive and not necessarily the most significant ones.  And as a result, we get drug-drug interaction alerts that are not useful. We have one alert that's in place that's like, OK, this shouldn't be given with alcohol.  OK, well, we have supplying alcohol in the hospital. We have a whole other problem, right? So they're not necessarily tailored for that particular environment.  The other thing is we go into environments and customization. The alerts that are for a CCU are not the same as for a neurointensive care unit. It's not the same for a PICU.  And yet, we have one set of alerts that are for the entire hospital. And they're not specific to that patient population.  And so I think it's similar to the alarms, which is customization and thoughtfulness for those patient populations.  Yeah, and I realize it's very desirable to have a one-size-fits-all. It's very easy from an engineering point of view to do it that way.  But it doesn't really serve the customer, i.e. the providers or the patient really well. Everybody is different. And particularly as medicine moves towards personalized medicine,  as we like to call it, knowing people's genetics and how that interacts. I mean, ultimately, every patient who comes to the hospital  should really have an entire customized template of how they get monitored, how they get treated, how they get evaluated. Because we are all different.  And it could also help in this arena quite a bit. So I just would like to say one thing. And again, I'm speaking only on behalf of what I work with,  that we actually have embedded defaulted profiles. We have a defaulted profile for an adult. We have a defaulted profile for a pediatric patient.  So within the overarching default, we then also have default profiles. And I would tend to say every manufacturer does that.  So I'm surprised to hear that you're not having that operation. I would encourage you to- Not for drug-to-drug interactions. Oh, for drug-to-drug. That's never going to happen.  Okay, I got you. They put one system in place, it's supposed to fit all, but it just doesn't. But the other thing is that I think if I was talking to any company,  I'm going to completely echo what Brad said. And I want it to be voice. But I'm going to tell you about a cautionary tale. So with our CRT, our PRISMAX technology,  they have to try to make the alarms meaningful and constant for the bedside nurse, they use a red, yellow, green approach.  And the red, yellow, green still requires that you've set default, but we can't set default there. The machine actually learns the patient and shifts the red,  and yellow, and green based on the patient. Now you might think that sounds good, but the reality is then the only time  you get an alarm is when you've already clotted your filter because those red, green, yellow alarms are always changing.  So I think the other aspect is to try to find something between the human factor requirement to set default and the machine learning the individual patient  and setting the default that we need something in the middle. And in the meantime, that should be voice activated.  And I should be able to do that by voice because I can do much faster with voice than making changes on my monitor.  Yeah, machine learning so that a machine can learn each individual patient. Of course, that requires the patient to necessarily be there long enough  that machine learning can actually be applied. No one knows how much data on any one individual patient a machine,  a computer system would need to know before it could safely vary parameters. We don't even know that yet, but theoretically, if I put you on a monitor  and on a telemetry unit and over the next six hours, you don't do anything. Nothing real happens to you. Can the machine safely then alter your parameters without a human input?  Well, the IntelliView system that at least that one article doesn't do that, but at least it kind of evaluates the patient and it says to the nurse, you know,  do you want to consider changing these parameters? Again, I haven't used the IntelliView system myself, but I got the impression  that basically it presents data suggesting that a particular change like you could drop the SPO2 alarm to 85 instead of 88 safely based  on this patient's parameters over the last X number of hours. That has potential for, to be very powerful, but again,  like a lot of the other machine learning, it's got to not miss real events. A single missed real event, we have to have zero tolerance for that. Think of it this way.  If you're in an airplane and your pilot's getting, you know, assaulted with a lot of false alarms, and you can get rid of those false alarms,  but they miss the one alarm that tells them the engine's on fire, you know, game's over. So in many respects, we have to have a zero tolerance, which does hobble us to some degree.  I just want to add one other thing. I don't know if you remember the slide where it was, or the seven or eight slides where it was parameters, parameters, parameters, right?  So this is now on a heart monitor. Every one of those parameters can be adjusted by a nurse. Those pages and pages and pages and pages.  So this reiterates what my colleagues have said is that it's not intuitive and it's not easy, right?  So we, in my now five hospital system, have a default set of parameters for every patient, and we've gotten rid of all that stuff.  So that's one thing is that, you know, you come, these companies come with lists and lists and lists of parameters, back to what Brad said, the C continuation.  However, the other important thing is whomever's representing that company has to know what every one of those things do, what triggers it, how to turn it on,  how to turn it off, because I will tell you, I sat with a lot of biomed engineers with those pages and pages and pages, and I would say, okay, what does this mean?  Hold on. And they're Googling, and they're calling, and they're on their, really? That's why it took six to nine months to go through that list,  because the people in the room didn't even know how it worked. So they came up with this astronomical list, which, by the way,  the bedside nurse is supposed to be able to work, but the people who created that list can't even tell me how it works. So, again, we got rid of, you saw what my list is.  What do I have, 20 alarm parameters? Out of those eight or nine pages, we turned everything else off, and we've had not an issue.  So I think not having these massive lists and certainly knowing how your product works  and what it means is going to be really significant as we move forward into all this other technology. Yeah, I don't know what the right number is.  I mean, I don't know what the upper limit of what would be reasonable to ask a nurse to customize on any one patient each day, or at least approve a customization parameter,  but it certainly is not six pages of individual parameters. My guess, what do you think, maybe 20, 25? It depends on the patient. It depends on the patient.  Based on the data that I showed, the nurses are incredibly comfortable with that list, and they're the ones that created it. So what was that, 20 parameters?  Not even, and not every patient has an ICP monitor, right? Not every patient has AB or, you know, an end tidal CO2, so maybe it's only five or six things. They created that list.  We've had no complaints. So I think when we're looking at all these other devices, it has to be a reasonable number, and it has to be, everybody has to understand how it works,  both of the company providing this equipment and, you know, the people at the bedside, and it has to be intuitive for the nurse to do what they need to do,  because I don't want them spending the time with all this equipment. I don't want them going through seven pages of alarm parameters trying to figure  out what to put on this patient. I want them at the bedside. So I do think, I really appreciate what's been said,  but I think first before you initiate anything, you have to actually look at what's happening in your system, and I think for most of us,  most of us have electronic medical records that you may or may not be able to utilize that effectively, because I think without actually capturing real-time waveforms,  it's very hard for me to go back and say, well, this wasn't a real alarm, because I might be speaking with the nurse caring for the patient. They don't really remember.  I need to be able to go back and look at that data, and I think that remarkable capability in whatever system you use will provide ability for us  to evaluate what is actually occurring versus what is actually actionable and life-threatening versus something simple, and I think one of the redundancies  that have been built into our monitoring systems is a time requirement for that generation, because again, sometimes you have short little bursts of a dysrhythm.  Now it's important for us to know about it, but we don't, but we may or may not be responding to it, and by the time we come to the bedside, it's already over with,  and we didn't actually see what was happening. I think the other part of the future, and I'd really be interested to know what the panel says, and please, if you have a question,  is actually looking at predictive software, and is our belief that not today, because predictive software is in its very early stages, but as we move forward,  will predictive software give us a better handle on our alarms? Will that just add more alarms? Will it give us a better methodology for management?  So Bradford, I can tell you're ready to talk about it. Well, because as I said, I have a strong interest in surveillance monitoring for general ward patients.  We've actually been testing some of those systems on our general wards, and I can say without predictive analytics, alarm fatigue will be a burden any time we try  to move out into an area where we have a much higher patient to ratio, you know, six to one, seven to one nurse to, rather, six to one patient to nurse ratio.  Unless you can do predictive analytics, single parameter alarm systems, even if you integrate them into smart alarms, are not likely to reduce the alarm burden.  I think you have to have predictive analytics that say, this pattern of these parameters, the heart rate up, this down,  that going in this direction, predicts 80%, or you'll pick the number that you'd like, that this patient is deteriorating. And therefore, will alarm and cause an intervention.  Whereas, you know, all those other, you know, heart rates of 115 with nothing else going wrong, you know, get ignored  because they are not necessarily predictive that the patient is deteriorating. I think that's really crucial.  I think it would also be of tremendous value in the ICU, but especially when you are moving away from one-to-one nurse to patient ratios and one-to-two nurse to patient ratios,  you've got to have predictive analytics because otherwise you'll never get any good data. But I also think it depends on the parameter.  For example, we're testing, like a lot of folks are, a homegrown sepsis scoring system called True's. It probably works great in ED. Maybe works okay on the floor.  It's absolutely worthless in the ICU. Worthless, absolutely worthless. So, it really depends on what you're measuring as well. Do you have anything to say?  I think I was covering pretty much what you were going to say. I mean, it's more than one parameter. It's what a clinician does, right?  You don't go look at one parameter and say, oh, I think I know what's going on in the patient. You have to look at the whole patient. And so, that's what this is doing.  It's advancing us. And so, it makes sense that it's going to be better in its specificity for the whatever deterioration that's occurring.  Yeah, like for example, just the other day in the burn unit where I was working, you know, I got called up and they're like, oh, this patient's having all these runs of VTAC.  And I said, all right, send me a, because I wasn't physically in the burn unit. I said, send me a picture of the EKG.  And they showed me an EKG that I swear, it looked just like VTAC, except right below it was another lead that looked perfectly normal.  I said, that's movement artifact on one of the leads, guys. One of your leads isn't connected with the patient's shivering or something. And they were all freaking out.  And it was something as simple as looking at the other lead and saying, no, these are incongruent. Right, because you don't do one thing in isolation. Exactly.  So I want to talk a little bit about something that Brad mentioned. So in our EPIC system, we have built a modified, modified early warning scoring system.  And actually, one of our concerns and one of our staff's concerns was, this is going to take power away from the bedside nurse.  Well, the bedside nurse wasn't recognizing patients who were deteriorating. Now, this is not really prediction, but it's a scoring tool based on five parameters.  When a patient meets a score, and at my hospital, we set the score relatively high, or they have a delta change of at least two points,  that from EPIC actually communicates to the rapid response team. Don't really require any overwhelming interface. Now, we have a huge rapid response team.  We have 30 staff for our rapid response team. And there's six on during the day and four on during night. So we have a very big team.  But we get a lot of modified early warning score system alerts. In the alert that comes to their phone, it actually shows all the parameters.  And it shows what the prior score was and the delta change. And so the rapid response nurse can look at that and say, well,  this patient has been looking this way for a period of time. I better get there right now. And then they can actually communicate back through their phone and to the system  that we're going to reduce this relationship on this patient because this is the way they've been for three days. And they've already visited. They round on that patient per shift.  And we are modifying, modifying because we're now adding a 12-hour backwards look at labs to also correlate the labs to the muse.  And I think a lot of people are doing work similar to that. But again, the big concern for us really is in the ICU, we have a lot more vigilance on the ward.  It's not for a lack of desire. It's just the nurse-patient ratio is so wide. And we found also we had our sepsis scoring tool was embedded two years before we started Muse.  And what one would have thought is that Muse will capture everybody who's presenting with a sepsis alert. It didn't. So we now, with those BPA alerts, don't alarm on the floor.  They're alerted to us through either through spoke or through another methodology. And we spoke pretty aggressively to communicate to stakeholders like the pharmacist, the lab,  the nurse, and the primary provider that they need to come to the bedside within 30 minutes to see a patient. And we've actually had a pretty good run so far.  You know what's happening with all this is, first of all, there's a consistent theme amongst everybody, which is, first of all, stakeholder involvement, right?  I mean, that wasn't happening before. You know, I have a problem in the hospital. Look, I'm going to put an alert in place. I fixed it. Done, right?  I didn't ask anybody if it was useful. I didn't go back and check to see if it actually did what it was supposed to do. So it's stakeholder involvement and it's quality improvement.  We can't set it and forget it. We have to go keep checking on it, keep making the modification until we get whatever goal we're looking for. Right. Excellent. Thank you.  Sustainability is key for a lot of these implementations. And implementation science can teach us a lot about how to do this and not do this.  Unfortunately, a lot of us, and I could speak for myself, don't necessarily do it as well as we possibly could have. Could always do better. Yes. So any questions from the audience?  So we're at around 12.15. We're a little bit early. This will allow people to have a minute to eat and do what they need to do unless there are questions from the audience.  And of course, if you want to speak to any of us, we're here. And I don't think there's anyone coming. So in the next round, because I don't think there is a next round.  So until a little later. So we're available to have discussions with you if you'd like to come up and personally ask questions. Thank you very much for attending.  And I hope that this will help you in your practice. Thank you.

Video Summary

Alarm fatigue in the ICU is a significant issue that can compromise patient safety and delay clinical response to alarms. To address this problem, strategies such as assessing the need for monitoring, optimizing technology, customizing alarm settings, conducting alarm rounds, and considering contextuality of alarms can be implemented. Bundling these strategies together can lead to significant reductions in alarm fatigue. A panel discussion also highlighted the importance of collaboration between equipment manufacturers and healthcare providers in developing solutions for alarm fatigue. The role of customization and predictive analytics in improving alarm management was emphasized, along with the need for ongoing monitoring and quality improvement efforts.

Asset Subtitle

Quality and Patient Safety, Procedures, 2020

Asset Caption

The Society of Critical Care Medicine's Critical Care Congress features internationally renowned faculty and content sessions highlighting the most up-to-date, evidence-based developments in critical care medicine. This is a presentation from the 2020 Critical Care Congress held February 16-19, 2020, in Orlando, Florida, USA.

Learning Objectives

Identify the different technologic distractions that clinicians are confronted with in the ICU environment
Compare successful methods that have safely decreased ICU alarms and technologic distractions
Evaluate the tipping point of swinging too far in the direction of too few critical alarms
Explain how to effectively implement and apply these techniques in your ICU
Describe new analytic programs to filter alarms

More Than the Monitor: How Many Alarms Do We Face in a Day in the ICU?

Sandra L. Kane-Gill

Successfully Decreasing Nuisance Alarms: Fantasy or Reality?

Jacqueline Y. Steuer

How Can I Implement Programs to Decrease Alert Fatigue in My ICU?

Bradford D. Winters

Leveraging New Technology to Filter Important Alarms

Barbara McLean

Meta Tag

Content Type Presentation

Knowledge Area Quality and Patient Safety

Knowledge Area Procedures

Knowledge Level Foundational

Membership Level Select

Membership Level Professional

Membership Level Associate

Tag Communication

Tag Monitoring

Year 2020

Keywords

alarm fatigue

ICU

patient safety

clinical response

monitoring

technology optimization

alarm settings

alarm rounds

contextuality of alarms

reductions

panel discussion

collaboration

equipment manufacturers

healthcare providers

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.