In this weeks GeriPal/JAGS Podcast we talked witk Kei Ouchi, an emergency medicine physician, internist, and researcher at the Brigham and Women’s Hospital and Harvard Medical School in Boston. We recorded this podcast in the hallways of the annual meeting. We talked about outcomes following intubation in the emergency department.
Kei published a paper in JAGS that is notable for several things, but perhaps most of all for the innovative use of color imagery to convey a message. The image in the @AGSJournal tweet above is from Dr. Ouchi’s article – this tweet went viral by the way – and notice what it does: (1) convey the main message that outcomes are worse with advancing age, and are not good in general overall; (2) grab your attention and make you want to learn more.
Kei is very thoughtful about how these data should be used – not on the spot in the ED, when a patient is gasping for air, and you pull up the color figure on your iphone Twitter app – no, not then. Better to use this information in advance, when things are calm, outside the ED, for people at risk of going to the ED in extremis.
This is the first in a series of GeriPal podcasts on the GeriPal – ED interface. ED stands for Emergency Department by the way.
Eric: Welcome to the GeriPal Podcast. This is Eric Widera.
Alex: This is Alex Smith.
Eric: We are live at the American Academy of Hospice and Palliative Medicine, AAHPM, the Hospice and Palliative Nursing Association, HPNA, and SWHPN meeting in Boston.
Alex: And we have a guest with us today.
Eric: Who is our special guest today, Alex?
Alex: Today our guest is Dr. Kei Ouchi. Who is an Emergency Medicine Physician and Internist and Research Fellow at the Brigham and Women’s Hospital, right here in Boston. He was up working until 3 AM last night, but he was kind enough to come join us today on the GeriPal Podcast. Welcome to the GeriPal Podcast Kei.
Kei: Thank you so much for having me here. It’s an honor.
Eric: So, I just want to say, I was up till 2:30 last night, but I was not working. We always start off with a song request. Do you have a song request for Alex?
Kei: How about, “Just Breathe”?
Alex: Perfect. Good choice, given the material. [Singing]
Eric: So, that was a very apropos for the discussion we’re going to be talking about today, which is, intubation in older adults in the emergency room. So you just published a paper in the Journal of American Geriatric Society, or JAGS, titled, “Prognosis After Emergency Department Intubation to Inform Shared Decision-Making”. And it’s like an all-star cast of authors that you have on this article, including James Tulsky, Rebecca Sudore, and Mara Schonberg. Really amazing, but before we go to the article, what got you interested in the subject?
Kei: I’ve been a practicing emergency physician and internist for a few years and through my training and my current clinical practice, I always wondered, what is the information that clinicians have to communicate to the patients or surrogates about this critical junction of their life and their illness? And I just didn’t know how to do it. So I started to look through, what’s the available information that’s out there about this. Then I realized that, we don’t really know actually what really happens to all comers in emergency department. And the data was lacking. And I wanted to learn more about that. That’s how this got started.
Eric: So, it looks like what you did is, you took adults, aged 65 years and older, who were intubated in the emergency room from 2008 to 2015, amongst all these different hospitals, 262 hospitals. What did you find?
Kei: So, I think the most important finding that we found is that, we confirmed that in-hospital mortality, after emergency department intubation for older adults, regardless of their comorbidities or the admitting diagnosis, was very high. Which, I think most clinicians knew already, but we put it in numbers. And we tried to come up with ways to better communicate that to the patients or surrogates, by the clinicians.
Eric: So, we’ll have a picture of this amazing figure 2 that you included in this JAGS article of older adults and their outcomes. Both survival in returning home, surviving and discharging to a nursing home versus dying in the hospital. Which is broken down to 65 to 74-year-olds, all the way to greater than 90-year-olds. And it’s a really impressive graph. What was it, 30…33% of all comers, older adults died … Wait, what was the statistic? Do you remember?
Kei: Yes, that’s correct. It’s 33% when you take all comers coming into the emergency department, who are all intubated in the emergency department.
Alex: 33% died before discharge. So, did not survive the hospitalization.
Kei: Yes, that’s correct.
Alex: And among people who are older than 90, that number is much higher, it’s 50%.
Eric: Yeah and in that group, it’s even more impressive, looking at this graph. It’s 50% died of the greater than 90-year-olds, 36% survived but were discharged to a nursing home, and only 14% survived and returned home after that index hospitalization. Now, we have no data what happens right after the hospital, right? Is there any way to get that?
Kei: I have really thought a lot about this because the most interesting information is actually what happens to the patients after they leave the hospital, right? So, what is that quality of life like? And things like that. Which is really, it’s certainly not in this dataset. I think it has to be merged with Medicare dataset and some other dataset to come up with that, which is going to be hard.
Eric: Yeah, because when I think about like having these discussions with people in the emergency room. You know, going to a skilled nursing facility is generally not the big decision point, it’s not ever being able to return back home. So it would be really valuable to know, of these individuals who went to the skilled nursing facility, how many were actually able to return home? I’m guessing it’s incredibly small.
Kei: I really want to learn more about that. But unfortunately, I haven’t had an opportunity to work with the dataset that has all that information yet. I think as clinicians we all know, anecdotally, what happens to some of these patients that we’re taking care of. But as a whole big picture, it’s still slightly unclear.
Alex: Now, other than age, were there other factors that helped you identify patients who are at greater risk of dying before leaving the hospital?
Kei: Yes. So, there are. But the answer is, for example, metastatic cancer, having the comorbid diagnosis of metastatic cancer or admitting diagnosis of stroke or intracranial hemorrhage. And remember, these are people who are all intubated in emergency department. So, intracranial hemorrhage with intubation in the emergency department. Those are, yes, they would increase your chance of not making it in the hospital for sure. What was striking to me was that, none of these things all combined, they still don’t predict the in-hospital mortality with this good discrimination. It wasn’t all that surprising to me for me to find that. Because we know, all the patients are all very different, and it’s very difficult to risk stratify them. That’s why you need a clinician to kind of, take in all their information together to decide that. And I think, at least in this claims dataset, there’s really not enough information to account for that kind of accuracy. And we really need more information about patients’ functional status and things like that.
Eric: So, you also cite a recent JAMA IM article talking about rates of intubation for older adults and how dementia, I think was outpacing all other diagnoses by a factor of four, as far as the increasing rates of intubation. Do you have any data around cognitive status in these individuals? And was that a factor?
Kei: Yes. I agree and I think that probably a very important factor. Unfortunately, in this administrative dataset, there is no variable that accounts for cognitive status. Even the diagnosis of dementia is not well documented. I mean, they are, but it’s hard to figure out how accurate that is, compared to things like diabetes or heart failure. So, yes, that would have been lovely information to learn more about, but we just don’t have that information in this dataset.
Alex: So, here’s a question. Do you have information or if not, what do you think would happen to similarly sick patients, who are not intubated, who are in the emergency department, but not intubated?
Eric: Are there similarly sick patients? It’s intubation. These people are … that’s a marker of sickness.
Alex: Well, I guess my point is, if they weren’t intubated, then mortality probably would’ve been 100%, right? No. Okay, Kei.
Kei: So, we know from some prior studies on this type of issue, they usually take one disease group or one phenomenon, like sepsis or something like that. Then they follow the outcomes. And we know that patients who go to ICU, if they get intubated, their rate of in-hospital mortality jumps dramatically, as you can imagine. So we know that, intubation itself, is certainly associated with in-hospital mortality and people who are not intubated, are less likely to die in a hospital. But that’s just the nature of their illness itself.
And one more comment about people who are not intubated, but they should have been intubated, but they were not, for one reason or another. It could be that they had terminal illness or had some kind of conversation that led to that. There’s really no good information about that, other than a case series studies. And it’s actually not true, not everyone dies in these case series studies. Some people survive. You know, these people who come in with DNR, DNI orders, and they are not intubated, but they go to ICU or hospital, they survive. And we don’t really know exactly how much, because these are all very small numbers.
Alex: So that’s interesting because it helps put this in context. If a different decision were made, what might happen? I guess I’m getting to the implications of this and how you see it as being used in clinical practice. Like for example, do you see emergency providers taking these figures to surrogates of seriously ill patients in the emergency department and saying, “You know, intubation is one option here. I’d like you to have accurate information about what might happen, based on national study of people, like your loved one, who are sick and were intubated in the emergency department”.
That tells one piece of the story. The other piece of the story is, what happens if we don’t intubate them? Maybe a little bit from you about how you see this being used clinically or is it ready for use clinically?
Kei: I think all my mentors would say that this is not ready to be used immediately, clinically, because this is the first iteration of this type of decision aide. And that’s correct. This has to be tested on different clinicians and different patients to see how they understand this for sure. But, one comment that I’d like to make is that, we hope that this information is helpful when a clinician is synthesizing all the information that you see in front of you. Their vital signs, their lab values, their clinical status, and having this conversation with the patient or surrogate. We hope that this information is somewhere back in their mind to share with them about the baseline risk of mortality for older adults, but it certainly does not speak anything about patients who decide not to get intubated, what happens to them. Because we were unable to look at that information.
Alex: So, do you have any examples from your clinical practice that you could share with our audience, in a way that’s anonymized enough to protect the confidentiality of the people you’ve cared for? Where this sort of issues come up and you wish there had been more information available to make a decision about this.
Kei: Can I just clarify one thing before I talk more about that? What I want to clarify is that this type of decision aide, to communicate the probability and likelihood of some event happening, is only useful for goals of care conversations that happen in a subacute setting, I think. And what I mean by that is, people who are really about to get intubated and they’re hypoxic and hypotensive, and when their family arrives and they are crying and yelling at you like, “Why can’t you put my dad on breathing machine right now?” That’s certainly not the time to bring this decision aide and say, “Hey, look at this, don’t you understand this probability?” That’s not what this is meant for. Our hope was that this was meant for people who are more subacute and have time and emotional status that’s controled to think about cognitive issues like this, like the probabilities.
I think there are two types of conversations in the emergency department like this. One is hyper acute, like what I just described. People who have their emotions super high, and they are unable to process this type of numbers, unless their emotions sort of handled in a way that they can. The majority of my patients are people who are very ill and they’re going to get admitted to the hospital. But they’re not at that point yet and they can still, either converse with you or their surrogates can make this decision, near term decision, going forward. Our hope is that this is used for those patients where, I’m worried that your dad is getting sicker and he might have a chance of having to have to go on a breathing machine, during your hospitalization somehow, today. Have you ever thought about this? And if they’re in the right state, in terms of their emotional status, then perhaps these numbers could be helpful for these patients to understand.
Eric: But the numbers we are looking at here, people who get intubated in the emergency room, do we have data on what happens to those people once they get admitted from the emergency room. Are the numbers worse or better, as far as survival or discharged to a skilled nursing facilities for older adults?
Kei: That’s correct. There are prior studies on patients, when they look at patients who are intubated in the hospital, rather than the emergency department. And the numbers are fairly similar. But they don’t, they don’t usually, they didn’t really describe how many people would go to a nursing home or how many would go home, after the hospitalization, but in terms of in-hospital mortality, it’s fairly similar. And that’s one of the important findings that this paper brought up. Because those prior large studies excluded people who died in emergency department. But actually, those are actually rare. And the numbers are very similar to people who are intubated after hospitalization.
Alex: Is there a specific clinical case that comes to mind, when you think about this issue?
Kei: So, I had one patient, maybe a month ago, who, let’s say, his name is Mr. C. Who was 85, had metastatic lung cancer with pleural effusions who are coming in with hypoxia and hypotension. And the patient was actually on home hospice service and the daughter called 9-1-1 because he didn’t look very well. Of course he arrived in the emergency department. Of course, the patient has no ability to speak for himself at all, because he’s huffing and puffing and he’s on BiPAP and he can’t really make this decisions. But, now this decision is made with the surrogate, who’s the daughter, who arrived with the patient. And she is crying and yelling, “Why can’t you put my dad on a breathing machine?”. So this is actually a common, it’s not a common, but it happens frequently in emergency department. And then, when I think, back to this paper, that’s really not the time for you to bring up this decision aide. Actually, these items can be used, but in a different settings.
So, for example, I watching my trainee go through this conversation, next to him. And when she says, “Why can’t you put my dad on a breathing machine?”, he starts to talk about, “Well, the probability of surviving this hospitalization, if he were to go on the breathing machine is, I don’t know what it is,” which I know, “I don’t know what it is… Blah, blah, blah”. He starts to explain about the numbers of why this is important. That’s a perfect example of when this decision aid should not be used. Because I think, in my mind, what that daughter was asking is, “Can you please help my dad?”. So, this is an important point. The numbers are very important for clinicians to know, but you have to choose the right moment to use this type of decision aid.
Eric: So, let’s talk for a second, when you are in that situation with a resident and maybe they’re starting to talk about numbers. And you’re feeling that this is not the right place. Because it sounds like, you’re thinking maybe this happens even before the emergency room. But this is still important in the emergency room, at least it sounds as well. So obviously we want this to occur as early as possible. But when you’re in that situation with a resident, they’re talking about numbers, what do you do in that instance?
Kei: I struggle with that, too. And I asked people who are a lot senior than me, “What do you think I should be doing?”. And I tried variations of it, to see which one might work. But the last example I had is, I had to stop them and I just said, “Hey, I want to add something. Would it be all right if I interject right now?”. And the trainees won’t keep going because I’m the attending. And then I kind of start to take over the conversation. So, it’s a really hard balance of allowing the trainees to do as much as they want, as well as keeping the patient safe. As the lead clinician, it’s always a hard balance, especially in this hyper acute setting.
Eric: And then do you reverse the conversation to go back to big picture goals, values? You’re not actually talking about the numbers, but you’re talking mainly making recommendations based on those numbers, is that right?
Kei: That’s correct. So, what I usually do is that, I say things like, “It sounds like a very important topic and that number is also important. And I also hear you…” the patient’s daughter, “…that you really want to learn more about this. Now, I don’t want to ignore that, but I want to put that on hold. And I want to learn a little bit more about your dad, before I can talk more about what I would recommend in this situation”.
Eric: Well, I really want to thank you for joining us today and having a great discussion. It was a really phenomenal JAGS article and I loved the color picture, which we will have on our GeriPal website. So, thank you for joining us.
Kei: Thank you so much for having me, I really appreciate it.
Eric: So Alex, do you want to, since we’re talking about intubation, do you want to give us a little bit more of “Just Breath”, by Eddie Vedder?
Alex: Yes. [Singing]
Eric: I want to thank all our listeners for joining us this week. Especially live from the AAHPM, HPNA, and SWPHN meeting in Boston. And we will talk with you next week.
Alex: Thanks folks.