Selected Podcast
Respiratory Depression in Hospital Wards and the Critical Care Implications
Ashish Khanna, MD, FCCP, FCCM, a staff intensivist and anesthesiologist, associate professor of anesthesiology, and section head for research with the Department of Anesthesiology Section on Critical Care Medicine at Wake Forest University School of Medicine in Winston-Salem, North Carolina, discusses respiratory depression on hospital wards and the critical care implications.
Featuring:
Ashish Khanna, MD, FCCP, FCCM
Ashish Khanna, MD, FCCP, FCCM is a staff intensivist and anesthesiologist, associate professor of anesthesiology, and section head for research with the Department of Anesthesiology, Section on Critical Care Medicine at Wake Forest University School of Medicine in Winston-Salem, North Carolina. Transcription:
Dr. Michael Smith: Hello and welcome to the Society of Critical Care Medicine’s iCriticalCare Podcast. I’m your host today, Dr. Michael Smith. And today, we’re going to be talking about respiratory depression on hospital wards and the potential critical care implications of that. And I’m joined by Dr. Ashish Khanna. He’s the Associate Professor of Anesthesiology and Section Head for Research with the Department of Anesthesiology Section on Critical Care Medicine at Wake Forest University School of Medicine. Dr. Khanna, welcome to the show.
Ashish Khanna, MD, FCCP, FCCM (Guest): Thank you. Thank you. It’s a pleasure to be here.
Host: So, very interesting topic. And I think at the end of the day, what we want to focus on is that hospital ward patient. How best to manage what might be going on there, specifically in the context of respiratory depression. So, I thought maybe we could start off, since you’re the expert in this; could you define just in general, what are the demographics of respiratory depression including opioid induced depression on the surgical and on the medical wards?
Dr. Khanna: Yeah, that’s a great question. The demographics of respiratory depression are varied. And we used to think that it only affects those who are a certain age group or have obstructive sleep apnea and so on and so forth. But we now know that almost no one is sort of immune to it. So, almost everyone is at risk. Most of the data we have is retrospective registry data and what that data shows us is that folks are hypotensive, and hypoxemic on the ward or the general care floor much more commonly than we ever perceived them to be.
We always thought that the general care floor is a place where our patients are safe and they’re medically stable and they’re essentially transitioning to go home or get out of the hospital. However, most data suggests that anywhere between 50 to 75% of all hospital mortality happens because of events that are triggered on the general care floor. So, events that don’t happen in critical care units or acute care settings but events that are triggered on the general care floor. This happens across the sexes, across age groups and so on and so forth.
So, for example, the data I like to quote all the time is the Get With the Guidelines Registry which is essentially a registry of code blue events that happen across the country. So, in 2012, they estimated nearly 40,000+ acute cardiorespiratory depression events in hospital systems in the United States and about 40% of these patients who had these acute cardiorespiratory depression events did not make it out of the hospital alive irrespective of the immediate result of the cardiopulmonary resuscitation they got. So, it’s a problem. And unfortunately, we’ve sort of known that it’s there. But we’ve been oblivious to what we can do about it.
Host: So, in your opinion, there’s a great need then for a predictions scoring tool, right? Something that’s standardized, validated that allows us to do a better evaluation of ward patients. Is that kind of where we need to go?
Dr. Khanna: Yeah. Absolutely. I mean if you look at it in another sense, there’s so much data out there that shows that our patients stay for example under an oxygen saturation of 90% or stay under a blood pressure mean arterial pressure of 65 for prolonged periods of time. The way we monitor our patients on the floor right now mostly across the world is every four to six hours a provider will go in the patient’s room to check in on the patient. And knowing that patients have been hypotensive, hypoxemic and respiratory depressed for prolonged periods of time; those snapshots of time surveillance is not good enough. It doesn’t pick up the amount of time a patient is at risk. So, the answer to that, well someone might say, just put everyone on continuous monitoring across the hospital. But then, the flip side of that is that that means alarms all over the place. That essentially means converting your general care floor into an intensive care unit. And clearly, that’s not the answer because that means a lot of alarm fatigue and that means that no one is going to respond to those alarms. Which means there will be no difference in outcomes.
So, to navigate that, we need to understand heh, you know what are the patients that are most likely to decompensate and sort of triage them into strata of low, middle and high risk and that’s exactly what we need to do.
Host: And so yes, so those that maybe are in that high risk; they receive the continuous monitoring, right? And maybe those in moderate and low – maybe moderate gets a little more of a check in, low is the normal every four to six hours, something like that. Is that kind of – if we can risk stratify better; that would then tell us which ones need continuous monitoring at the end of the day, right?
Dr. Khanna: Yeah, right. And I will say that it is not just continuous monitoring. I would say continuous monitoring with actionable items and intervention and continuous proactive interventions which would include monitoring base interventions but other interventions as well. For example, checking in on their beta blockers stay or status, checking in on their fluid balance, is someone going to the room and giving them the incentive spirometer to use again and again and so on and so forth. Some very common sense things but obviously on the highest risk category, those should happen all the time. So, that would be the difference between the highest and the lowest risk in terms of interventions.
Host: How does the PRODIGY Trial fit into all of this? Can you help us correlate the results of that trial with the clinical practice of critical care medicine?
Dr. Khanna: Sure. For years as I practiced critical care medicine, I thought about and published something called the Four A. M. Phenomenon. And really, the Four A. M. Phenomenon was nothing more than that middle of the night, unprecedented admission to a surgical or medical intensive care unit for a patient who wasn’t doing well on the floor. And lots of needless investigations and six hours later in the morning, he or she is doing well, send them back to the floor, find out later they probably got overdosed with opioids or was not breathing well and no one checked in and by the time they found him, he was down and blue and they had to call a rapid. So, unprecedented ICU admissions are bad one way or the other and opioids are still the mainstay of pain medication across the hospital. So, putting both of those things together and all the information we had that suggested that continuous monitoring was one possible way to predict and see these events before they actually happen; we thought that PRODIGY, which essentially stands for prediction of opioid induced respiratory depression on inpatient wards using continuous capnography and oximetry would give us a prediction tool that would then help the providers and provider teams to stratify their patients into who needs that continuous monitoring and continuous intervention versus who can sort of be in a hybrid approach and also move their patients up and down in these risk categories depending on how they’re doing.
And really, that was the birth of PRODIGY. The way we did PRODIGY was also interesting. We did it across not just one part of the world, it was a truly international trial. It was done in North America, in Europe and Asia since cultural practices of the way opioids are dosed and the way patients are monitored are slightly different in different parts of the world. So, we used – we wanted to construct a very generalizable tool. The other thing that was very unique for PRODIGY is the way we did the monitoring data. So, patients who were on the general care floor and were scheduled to get intravenous or parenteral narcotic medications for pain control were all a part of the trial. They were all put on continuous oximetry and capnography monitoring that would include a heart rate, respiratory rate, oxygen saturation and anti-loose CO2 along with a variable called the integrated pulmonary index.
All this monitoring was bought blinded and silenced. So, this monitoring was going on in the background, but the nurses and providers were going in every four to six hours in a patient’s room to check in on the patient. So, normal practices were followed. The monitoring data was collected in the background and then it was sent to a group of experts, a clinical event adjudication committee that looked at thousands of hours of continuous waveform data and separated artifact from respiratory depression to opioid induced respiratory depression and then we used that data to build a score based on multi-variant logistic regression analyses, a score that we call the PRODIGY score and then we able to separate out patients for example, the highest risk category on the PRODIGY score was more than 15, the lowest risk category was all the way down to less than eight and there was a separation of an odds ratio of about 6 on the patients in the highest risk versus the lowest risk category.
And while I’m talking about that, just the PRODIGY score itself, I’m delighted to say, that we came up with a very simple and easy to use scoring tool compared to a lot of scores that are very complicated that need you to go online and construct this and do this calculator and so on, PRODIGY was simply as five variable tool; age, male sex, opioid naivety, sleep disordered breathing and the presence of chronic heart failure. And that’s all it was. So, age was our biggest driver. More than 60 with every decade of age was the biggest driver on this but the bottom line is that PRODIGY was a tool that every bedside nurse as soon as she received a patient on the floor, could essential do a PRODIGY risk score then and there and then the hope is that someday when you and I put orders for an opioid into a patient’s chart, that the PRODIGY tool would automatically then pop up the PRODIGY risk score, alert me and then ensure that I do more to my patient than just prescribe the morphine and walk away.
Host: Fantastic step. So this is a simple tool that came out of this international study. I guess the next question then is where are we at with this? Are more and more hospitals using the PRODIGY score at this point? And if not, what’s the plan to get this implemented in more hospitals?
Dr. Khanna: Yeah so, the where are we at? Well, we just published the paper a few months back and like I told you, it’s a very easy to use tool. What we would like to do now is to externally validate it. so, for the purposes of the trial, it was an internally validated tool where there was 1400 odd patients in the trial and using bootstrapping, we got the score and we internally validated it on the same population that we used to calculate the score. We would now want to have a smart investigator in some part of the world pick this up and say I want to do this on a subset of patients with major abdominal surgery on the general care floor and see if it prevents ICU admissions, if it prevents rapid response calls and so on and so forth. So, yes, next step is operations. It’s very easy to implement. Next step is also validation. For external validation though, we will need hospital systems to adopt continuous monitoring. Because the tool is built on continuous monitoring so we will have to do that.
Host: Yeah and looking at these as you said, different subsets of populations, different types of surgeries, et cetera, that’s where you really will see the impact of this – of the PRODIGY score, right so, that’s fascinating stuff. I think a great place for us to end with this Dr. Khanna is simply like what would you – since you were so involved with this, what would you like the listeners of this show to know about respiratory depression and about the PRODIGY score?
Dr. Khanna: Huh, so, what I would like the listeners of this show to know is that respiratory depression is common, it’s profound and rather unpredictable on the general care floor. In fact, if you are a critical care provider, and you work really hard on a patient in the ICU, then you discharge that patient from the ICU to the floor and that patient bounces back to the ICU as a readmission to this later with respiratory depression; I know that there is nothing more frustrating than that. And I wish that we understand that it is a very commonsense thing of just laying all eyes on the patient at all times that would prevent a lot of these readmissions. And really, make our work as critical care providers more useful were we could safely get out patients out of the hospital. So I want my friends and colleagues in the critical care fraternity to understand that even though our work is limited to an intensive care unit, what happens on the floor has a direct bearing on what we do in ICUs in terms of workload, needless readmissions and better utilization of hospital resources and a simple little intervention like this, improving monitoring on the general care floor, more portable multiparameter continuous monitoring, risk tools like PRODIGY to guide us with interventional arms so reaching our nursing staff and colleagues on how to respond to alarms in an effective manner will all help us achieve better outcomes for our critically ill patients in the ICU.
Host: Excellent. Excellent summary. I want to thank you for coming on the show today. This concludes another edition of the iCriticalCare Podcast. For the iCriticalCare Podcast, I’m Dr. Mike. Thanks for listening.
Dr. Michael Smith: Hello and welcome to the Society of Critical Care Medicine’s iCriticalCare Podcast. I’m your host today, Dr. Michael Smith. And today, we’re going to be talking about respiratory depression on hospital wards and the potential critical care implications of that. And I’m joined by Dr. Ashish Khanna. He’s the Associate Professor of Anesthesiology and Section Head for Research with the Department of Anesthesiology Section on Critical Care Medicine at Wake Forest University School of Medicine. Dr. Khanna, welcome to the show.
Ashish Khanna, MD, FCCP, FCCM (Guest): Thank you. Thank you. It’s a pleasure to be here.
Host: So, very interesting topic. And I think at the end of the day, what we want to focus on is that hospital ward patient. How best to manage what might be going on there, specifically in the context of respiratory depression. So, I thought maybe we could start off, since you’re the expert in this; could you define just in general, what are the demographics of respiratory depression including opioid induced depression on the surgical and on the medical wards?
Dr. Khanna: Yeah, that’s a great question. The demographics of respiratory depression are varied. And we used to think that it only affects those who are a certain age group or have obstructive sleep apnea and so on and so forth. But we now know that almost no one is sort of immune to it. So, almost everyone is at risk. Most of the data we have is retrospective registry data and what that data shows us is that folks are hypotensive, and hypoxemic on the ward or the general care floor much more commonly than we ever perceived them to be.
We always thought that the general care floor is a place where our patients are safe and they’re medically stable and they’re essentially transitioning to go home or get out of the hospital. However, most data suggests that anywhere between 50 to 75% of all hospital mortality happens because of events that are triggered on the general care floor. So, events that don’t happen in critical care units or acute care settings but events that are triggered on the general care floor. This happens across the sexes, across age groups and so on and so forth.
So, for example, the data I like to quote all the time is the Get With the Guidelines Registry which is essentially a registry of code blue events that happen across the country. So, in 2012, they estimated nearly 40,000+ acute cardiorespiratory depression events in hospital systems in the United States and about 40% of these patients who had these acute cardiorespiratory depression events did not make it out of the hospital alive irrespective of the immediate result of the cardiopulmonary resuscitation they got. So, it’s a problem. And unfortunately, we’ve sort of known that it’s there. But we’ve been oblivious to what we can do about it.
Host: So, in your opinion, there’s a great need then for a predictions scoring tool, right? Something that’s standardized, validated that allows us to do a better evaluation of ward patients. Is that kind of where we need to go?
Dr. Khanna: Yeah. Absolutely. I mean if you look at it in another sense, there’s so much data out there that shows that our patients stay for example under an oxygen saturation of 90% or stay under a blood pressure mean arterial pressure of 65 for prolonged periods of time. The way we monitor our patients on the floor right now mostly across the world is every four to six hours a provider will go in the patient’s room to check in on the patient. And knowing that patients have been hypotensive, hypoxemic and respiratory depressed for prolonged periods of time; those snapshots of time surveillance is not good enough. It doesn’t pick up the amount of time a patient is at risk. So, the answer to that, well someone might say, just put everyone on continuous monitoring across the hospital. But then, the flip side of that is that that means alarms all over the place. That essentially means converting your general care floor into an intensive care unit. And clearly, that’s not the answer because that means a lot of alarm fatigue and that means that no one is going to respond to those alarms. Which means there will be no difference in outcomes.
So, to navigate that, we need to understand heh, you know what are the patients that are most likely to decompensate and sort of triage them into strata of low, middle and high risk and that’s exactly what we need to do.
Host: And so yes, so those that maybe are in that high risk; they receive the continuous monitoring, right? And maybe those in moderate and low – maybe moderate gets a little more of a check in, low is the normal every four to six hours, something like that. Is that kind of – if we can risk stratify better; that would then tell us which ones need continuous monitoring at the end of the day, right?
Dr. Khanna: Yeah, right. And I will say that it is not just continuous monitoring. I would say continuous monitoring with actionable items and intervention and continuous proactive interventions which would include monitoring base interventions but other interventions as well. For example, checking in on their beta blockers stay or status, checking in on their fluid balance, is someone going to the room and giving them the incentive spirometer to use again and again and so on and so forth. Some very common sense things but obviously on the highest risk category, those should happen all the time. So, that would be the difference between the highest and the lowest risk in terms of interventions.
Host: How does the PRODIGY Trial fit into all of this? Can you help us correlate the results of that trial with the clinical practice of critical care medicine?
Dr. Khanna: Sure. For years as I practiced critical care medicine, I thought about and published something called the Four A. M. Phenomenon. And really, the Four A. M. Phenomenon was nothing more than that middle of the night, unprecedented admission to a surgical or medical intensive care unit for a patient who wasn’t doing well on the floor. And lots of needless investigations and six hours later in the morning, he or she is doing well, send them back to the floor, find out later they probably got overdosed with opioids or was not breathing well and no one checked in and by the time they found him, he was down and blue and they had to call a rapid. So, unprecedented ICU admissions are bad one way or the other and opioids are still the mainstay of pain medication across the hospital. So, putting both of those things together and all the information we had that suggested that continuous monitoring was one possible way to predict and see these events before they actually happen; we thought that PRODIGY, which essentially stands for prediction of opioid induced respiratory depression on inpatient wards using continuous capnography and oximetry would give us a prediction tool that would then help the providers and provider teams to stratify their patients into who needs that continuous monitoring and continuous intervention versus who can sort of be in a hybrid approach and also move their patients up and down in these risk categories depending on how they’re doing.
And really, that was the birth of PRODIGY. The way we did PRODIGY was also interesting. We did it across not just one part of the world, it was a truly international trial. It was done in North America, in Europe and Asia since cultural practices of the way opioids are dosed and the way patients are monitored are slightly different in different parts of the world. So, we used – we wanted to construct a very generalizable tool. The other thing that was very unique for PRODIGY is the way we did the monitoring data. So, patients who were on the general care floor and were scheduled to get intravenous or parenteral narcotic medications for pain control were all a part of the trial. They were all put on continuous oximetry and capnography monitoring that would include a heart rate, respiratory rate, oxygen saturation and anti-loose CO2 along with a variable called the integrated pulmonary index.
All this monitoring was bought blinded and silenced. So, this monitoring was going on in the background, but the nurses and providers were going in every four to six hours in a patient’s room to check in on the patient. So, normal practices were followed. The monitoring data was collected in the background and then it was sent to a group of experts, a clinical event adjudication committee that looked at thousands of hours of continuous waveform data and separated artifact from respiratory depression to opioid induced respiratory depression and then we used that data to build a score based on multi-variant logistic regression analyses, a score that we call the PRODIGY score and then we able to separate out patients for example, the highest risk category on the PRODIGY score was more than 15, the lowest risk category was all the way down to less than eight and there was a separation of an odds ratio of about 6 on the patients in the highest risk versus the lowest risk category.
And while I’m talking about that, just the PRODIGY score itself, I’m delighted to say, that we came up with a very simple and easy to use scoring tool compared to a lot of scores that are very complicated that need you to go online and construct this and do this calculator and so on, PRODIGY was simply as five variable tool; age, male sex, opioid naivety, sleep disordered breathing and the presence of chronic heart failure. And that’s all it was. So, age was our biggest driver. More than 60 with every decade of age was the biggest driver on this but the bottom line is that PRODIGY was a tool that every bedside nurse as soon as she received a patient on the floor, could essential do a PRODIGY risk score then and there and then the hope is that someday when you and I put orders for an opioid into a patient’s chart, that the PRODIGY tool would automatically then pop up the PRODIGY risk score, alert me and then ensure that I do more to my patient than just prescribe the morphine and walk away.
Host: Fantastic step. So this is a simple tool that came out of this international study. I guess the next question then is where are we at with this? Are more and more hospitals using the PRODIGY score at this point? And if not, what’s the plan to get this implemented in more hospitals?
Dr. Khanna: Yeah so, the where are we at? Well, we just published the paper a few months back and like I told you, it’s a very easy to use tool. What we would like to do now is to externally validate it. so, for the purposes of the trial, it was an internally validated tool where there was 1400 odd patients in the trial and using bootstrapping, we got the score and we internally validated it on the same population that we used to calculate the score. We would now want to have a smart investigator in some part of the world pick this up and say I want to do this on a subset of patients with major abdominal surgery on the general care floor and see if it prevents ICU admissions, if it prevents rapid response calls and so on and so forth. So, yes, next step is operations. It’s very easy to implement. Next step is also validation. For external validation though, we will need hospital systems to adopt continuous monitoring. Because the tool is built on continuous monitoring so we will have to do that.
Host: Yeah and looking at these as you said, different subsets of populations, different types of surgeries, et cetera, that’s where you really will see the impact of this – of the PRODIGY score, right so, that’s fascinating stuff. I think a great place for us to end with this Dr. Khanna is simply like what would you – since you were so involved with this, what would you like the listeners of this show to know about respiratory depression and about the PRODIGY score?
Dr. Khanna: Huh, so, what I would like the listeners of this show to know is that respiratory depression is common, it’s profound and rather unpredictable on the general care floor. In fact, if you are a critical care provider, and you work really hard on a patient in the ICU, then you discharge that patient from the ICU to the floor and that patient bounces back to the ICU as a readmission to this later with respiratory depression; I know that there is nothing more frustrating than that. And I wish that we understand that it is a very commonsense thing of just laying all eyes on the patient at all times that would prevent a lot of these readmissions. And really, make our work as critical care providers more useful were we could safely get out patients out of the hospital. So I want my friends and colleagues in the critical care fraternity to understand that even though our work is limited to an intensive care unit, what happens on the floor has a direct bearing on what we do in ICUs in terms of workload, needless readmissions and better utilization of hospital resources and a simple little intervention like this, improving monitoring on the general care floor, more portable multiparameter continuous monitoring, risk tools like PRODIGY to guide us with interventional arms so reaching our nursing staff and colleagues on how to respond to alarms in an effective manner will all help us achieve better outcomes for our critically ill patients in the ICU.
Host: Excellent. Excellent summary. I want to thank you for coming on the show today. This concludes another edition of the iCriticalCare Podcast. For the iCriticalCare Podcast, I’m Dr. Mike. Thanks for listening.