Skip to content

Amber Barnato is an expert in simulation studies.  A health services researcher and palliative care physician, Amber lauds the ability of simulation studies to isolate one variable in a study.  For example, we spend the first half talking about a RCT simulation study of clinician verbal and non-verbal communication with a seriously ill patient with cancer. In one room the physician under study interacts with a white patient-actor, and in another room interacts with a Black patient-actor.  They found no differences in verbal communication, but clear differences in non-verbal rapport building communication: physicians stood farther away, crossed their arms, didn’t touch the Black patient as frequently.  Amber tells the moving story of how these findings led a clinical colleague, her chief, to question and change his behavior. Of note, we talked about implicit bias in depth in this podcast with Kimberely Courseen. 

As we’ve written about on GeriPal when we were a blog (a decade ago!) these simulation studies can be used to study language, such as patient or surrogate choices when we use the terms “allow natural death” vs “do-not-resuscitate.”  This change in framing is a nudge, more evidence that the choices we make to use one phrase or another, or the order in which we present options, are all nudges that influence patient choice – listen to our podcast on the ethics of nudging with Jenny Blumenthal-Barby and Scott Halpern for more. 


Additional links to simulation studies:

https://www.atsjournals.org/doi/full/10.1513/AnnalsATS.201411-495OC

https://journals.lww.com/ccmjournal/Abstract/2011/07000/A_randomized_trial_of_the_effect_of_patient_race.9.aspx

https://www.liebertpub.com/doi/full/10.1089/jpm.2015.0089

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3687021/

https://journals.sagepub.com/doi/pdf/10.1177/0272989X14522099


Theoretical underpinnings:

https://home.csulb.edu/~cwallis/382/readings/482/nisbett%20saying%20more.pdf


Eric: Welcome to the GeriPal Podcast. This is Eric Widera.

Alex: This is Alex Smith.

Eric: And Alex, who do we have with us today?

Alex: We are delighted to welcome Amber Barnato, who’s a palliative care physician and health service researcher, and she’s director of the Dartmouth Institute for Health Policy and Clinical Practice. Welcome to the GeriPal Podcast, Amber.

Amber: Thank you guys. I’m a huge fan.

Alex: That’s good to hear.

Eric: So we’re going to be talking about the topic of simulation. We’re going to define what that is and how it applies to the work that we do. But before we go into that, Amber, I think you got a song request for Alex.

Amber: I do. I do. Being a palliative care physician, the one that came immediately to mind was Knocking on Heaven’s Door. You know what? It has that literal meaning of you may be coming near death, but when I was rereading the lyrics, there’s also something in there about losing hope. And I think the crisis that we’re in right now in healthcare delivery, the idea of laying down my badge, I think that’s maybe a metaphor for being a sheriff, but what about being a healthcare provider? So I just can’t keep going. And so I think that’s why it has its moment now even in 2023.

Alex: That’s great. I love this song. Well, for those of you who are close listeners, we have done this one before, but I couldn’t help it because I love it so much. And I wanted to redo it with bass drum track for those of you are listening and not watching on YouTube. So here we go. Here’s a little bit of the song.

(Singing).

Eric: Very nice Alex. Okay, Amber, I’m going to jump right into it. What the heck is simulation?

Amber: Well, simulation is pretending, so I think of the way I use simulation is in trying to simulate clinical care. We use simulation all the time in training our physicians. So most of our medical schools have simulation centers. You remember practicing and learning how to do ACLS with Annie. And so that’s sort of it’s based what simulation is. We’re trying to create a safe environment for people to practice skills in clinical delivery. They use simulation in other settings. For example, airline pilots and astronauts. They even use simulation in war games, diplomacy? In my research, I use simulation in order to try to simulate the kind of encounters, complex encounters with patients you’ve never met before, that hospitalist, emergency physicians and intensivists experience all the time walking into a room, meeting someone for the first time and having to solve a clinical diagnostic problem and make some decisions. But the reason that I use simulation, and I’ll tell you more about the details of my simulations, is to basically try to understand how clinicians think.

Eric: And before we get exactly what you did, why did you decide to do this in the first place?

Amber: Well, I got interested in variation and physician decision making really early on in my career. I think the first time I noticed it was, like as a medical student when you would rotate on one service with one attending and they would make decisions about how to treat a case one way. And then the new attending would come on and they’d have a different way that they did it. And this is before we coined the term evidence-based practice. But even within the constraints of evidence-based practice, clinicians still have to make that initial decision, is this a case of X, Y, or Z?

There’s that recognition piece. And when you and I talk about, what’s the key medical decision or diagnosis that we’re making all the time, we’re making a decision about like, “Is this patient sick enough to die? And if this patient is sick enough to die, what kinds of things do I need to think about in order to ensure goal concordant care?” And so in order to be thinking about goal concordant care and not just evidence-based practice, you have to first call into your consciousness that this might be a situation where those skills are needed. And so that’s the particular thing that I’m interested in is when do doctors think it’s time to have the conversation?

Eric: Yeah. It also feels like, I also grew up in med school in a three hospital system. There’s the variation between the hospitals, feels like there’s cultural issues that play a role. And certainly we’ve learned a lot over the last many years about variation based on geography. A lot of this stuff coming out of Dartmouth around variation. Why use simulation for that instead of just looking at recording clinical practice encounters, like real life stuff instead of, I’m going to use the word you used, fake encounters?

Amber: Well, so I think they all need to be taken together. So just as a physician makes a decision about whether or not there’s a particular, let’s say it’s appendicitis. They look at the signs and symptoms, they do a physical exam, maybe some lab tests or some imaging. You’re basically taking data from different sources to say, corroborate what’s going on. Try to really understand what’s happening. So as a researcher, that’s the way we think too. So we have all this data from the Dartmouth Atlas that shows if you go from region to region, there’s differences in spending. So the amount of dollars per Medicare beneficiary in the last six months of life varies more than twofold from one region of the country to the other region of the country.

So that suggests there’s some variation happening. So then you drill down further, just like you said, Eric, when you were going back between different hospitals. We find the same thing when you actually measure at the level of the hospital. Spending or ICU U days in the last six months of life, varying widely from hospital to hospital. My research has included ethnography, which is the kind of research where you actually follow around medical practitioners and take notes and watch their clinical conversations and medical decisions. And then when you debrief them about why they made decisions, they can tell you all kinds of stories about why they did.

And that the beautiful thing about humans is that we’re terrific at coming up with rationalizations. “Oh, I did that because this patient, really concerned he might not be compliant with his medication. So I decided I was going to do it this way instead of this way.” Physicians are constantly both making decisions that might be influenced by biases and heuristics, but they’re also unaware of some of the things that might be leading them to make certain decisions that they can’t rush, they can’t sort of observe in themselves so they don’t see it. And that’s kind of that implicit or unconscious bias. So when I talk to clinicians, sometimes they’ll give me reasons. And so the question becomes, well, are those reasons real?

Is that really why they’re making those decisions? And so the way that you test it is to basically take the doctors into a simulation. So with actors and a clinical scenario where you hold everything else standard. So every single doctor that sees the same patient gets the same stimuli. And if the doctors still come up with different diagnoses or perceptions of what’s happening with the patient and or treatment plans, that’s all on them. They can’t say, “Oh, it’s because the patient did this.” Because the patient did that for Dr. Jones and Dr. Smith and Dr. Widera. And if the three of you make different choices, then those choices emanate from the perspectives that you bring into the room. So I tried to triangulate data from those different sources.

Eric: So this is interesting because it sounds like part of this is standardization. Everybody gets the exact same thing and potentially you can change even a variable to see if that variable influences anything. I just want to go to your work around the effect of patient race, that you’ve done some work around that with simulation, right?

Amber: I have. And so have some of my mentees. It’s a great example. So you can basically take otherwise similar clinical scenario, but you can make one of the actors and their family members Black and one of the actors and their family members white. And you can run the same doctor in our case, in the simulation, in the rooms from one simulation room to another. And you can see if there’s a difference in the way that they behave.

Eric: And before we jump into that, why did you decide to do this study?

Amber: The studies between Black and white patients? So my area of interest broadly is in this idea of trying to understand that variation we talked about at the top of the conversation between regions, between hospitals. Why is it that there’s so much difference? Okay, are there norms in the organization that are different or culture, I think was the word you used. Are there things different about the physicians in terms of things like their prior experiences that influence the way they think about patients, their beliefs, their values, what else? What could be bringing about different decisions? And so that was my main interest.

The interest in race largely arose out of this conundrum or this curiosity that we see, which is that you may be aware that in the United States, Black patients, they get less of everything. They get less curative therapy, preventive therapy, all the stuff that we have worked assiduously as a field to develop, great technologies to improve quality of life and extend life expectancy. Black patients get less of that, but what they do get more of is aggressive life sustaining treatment at the end of life. They’re more likely to die in an ICU on a breathing machine. And this is a curiosity, like why would that be? And so that’s one of my hypotheses is, what is it that the physician is bringing to the table that might be explaining these differences?

Eric: Because it’s interesting. Preferences rather a specific patient population group, or it could be what the physician is bringing into the conversation and you’re looking into that physician component.

Amber: Well, that’s right. There’ve been tons of studies that have tried to drill down on the question of whether it’s patient preferences and there’s some mixed results. These are surveys if you tell two patients, one who’s Black and who’s white, if you had six months live and you got a pneumonia would you want to go on life support? And what you find when you do those studies, and I’ve done one of them, is that Black patients are slightly more likely, about 20% of them will say, “Yeah, I’ve got limited time. I’d be willing to take some time on a mechanical ventilation machine to live longer.” Rather, 30% would say that, whereas 20% of white patients will say that.

So there’s a slight difference, but if you look at the converse, 70% of Black patients and 80% of white patients don’t want that. So there are less different than you would otherwise think. And there have been smaller studies that have found differences as well. But my question is not so much whether Black patients might carry different preferences and people have all kinds of hypotheses about why they might carry different preferences for healthcare, but preferences are constructed. So you guys know that most of your patients don’t walk around carrying a preference for this type of anticoagulation versus this one. They have a preference maybe for being able to walk two blocks to the grocery store. They have preferences or goals.

Eric: It’s not about the intervention, it’s about the goal, the outcome.

Amber: Exactly. And your job is to say, “Okay, what’s medically achievable? I think I can get you to a place where you can walk those two blocks, yes or no. And if so, which of these strategies is more likely to get you there?” You’re the expert in whether or not it’s this treatment or this treatment. And so the idea that patients are walking around with these on their shoulder like, “Hey, I got the mechanical ventilation preference, just want to make sure.” This is what our field is trying to do, move away from treatments to goals. And so I’m already suspicious of someone who’s saying, “Oh, Black people want to be put on life support.” Because I’m like, “Yeah, to what end?” Black people probably have goals just like white people do, and they can understand what’s medically achievable and not. So that’s where we should be talking.

So I’m not even really interested as much in the preference hypothesis because I think it’s invalid. I’m interested in the goal hypothesis. If Black people on average are more vitalist and prolonging life, no matter what is important to them, that might lead to them wishing life support more or rather less for being more goal aligned. But I don’t know that anyone’s really had those conversations because they’ve been pretty much at the preference level, but backing up. So yes, it could be the patient preferences, which again, I would say goals and values is what I’m mostly interested in.

Eric: But it could also be how we interact or think biases as clinicians.

Amber: That’s my thinking. And again, these are things that are unconscious to us.

Eric: So you actually did a randomized controlled trial on this. So you looked at, it was an ICU patient, right?

I’ve seen a lot of 78 year old with metastatic gastric cancer. Is that what you used in this study?

Amber: I did.

Eric: In 78 year old, I’m just going to summarize. Metastatic gastric cancer, life-threatening hypoxia, thought to be cancer progression, stable preferences to avoid ICU admission intubation. Is that the general theme for this one too?

Amber: Yeah. That’s the general theme. That’s my archetype patient. And I bet you guys see this patient a lot.

Alex: Yes. I’ve seen that patient.

Amber: Yeah.

Eric: So in this study, the big difference, the case was exactly the same, but the difference was just the patient race.

Amber: Well, just for your knowledge to the geeky details, they had different types of cancer and one looked like it was lymphangitis spread causing hypoxia. The other one looked like it was cholangitis, but both of them were sort of on a septic pathway kind of fixing to die. But the thing that I did is that sometimes the patient with gastric cancer was Black and the pancreatic patient was white, and then sometimes it was the opposite. So we did a perfect cross so that the race and actual primary diagnosis was perfectly matched across groups. So half of the time the white patient had the gastric cancer and half of the time the Black patient did. We created these using mortality models such that both, if you actually put them into a mortality calculator, they both had the same risk of death in the hospital.

Alex: So did you give the two different cancers so that a given clinician wouldn’t see the exact same clinical scenario other than the patient’s race twice?

Amber: Right, exactly. Otherwise, they would be like, “Hmm, this patient has the exact same chart as the last patient.”

Eric: They’d figure out the question.

Amber: I’m suspicious that this study isn’t about cancer, but it might be about race. And I’ll tell you more about that because we didn’t tell the patient-

Eric: Yeah, interesting.

Amber: We didn’t tell the doctors the study was about race and I’ll tell you more about how he…

Eric: So what’d you find?

Amber: Yeah, so this was amazing. So the first thing we found… So I just want to tell a story about this study, which was so fun. So I don’t know if you guys know Deepika Mohan, she’s a trauma surgeon and intensivist at the University of Pittsburgh. And she was at that time my research mentee. She’s now a terrific researcher in her own, and you guys should probably have her on the show. But she and I were running the simulation rooms. They’re side to side. And so the control room is a contiguous room, but the individual rooms where the patients, the actors and the doctors are separated by a wall. And so, I don’t know if you can see my hands for those of you on the podcast, not going to be able to see all my great hand work.

But basically a doctor would start in each room. So one would see the white patient and one would see the Black patient, and then they would go out and they’d walk around and go to the opposite room. So they each saw some doctor saw the Black patient first and the white patient second and vice versa. The other one saw the white patient first and the Black patient second. So Deepika and I made sure to stay with our doctor. Let’s say I’m over here with the white patient. I’m running the room, and as soon as my doctor’s done, I sort of jog over to the other side to meet my doctor as they’re coming into the simulation. I’m looking through the smokey glass window. They can’t see me. And she runs through-

Alex: So those of you who are listening, Amber just stood up and jogged in place… [laughter]

Amber: Well, the funny thing, this is why I want to tell you this. So we’re going back and forth like this. And we start saying to each other as we’re jogging about, we’re like, “Are you seeing this? Oh my God, is this really happening? Are you seeing this too?” It was so visible to the naked eye. What we were noticing is that the doctors, when they went into the Black patient’s room, they were standing farther away from the patient. They were keeping their arms crossed. They were not touching the patient, and they were more likely to be looking at the nurse or the vital signs tracing than the patient or the family member. And it was really obvious to the naked eye. And so of course that wasn’t enough because we want to make sure that we had did valid measurement. So we as any good professor, we handed it off to a resident without telling her what the hypothesis of the study was.

But we had her code, each doctor’s interaction with the Black patient and the white patient, and she coded it with video using detailed information about how many times they touched the patient, how close they got to the patient’s bed, where their body was positioned, and whether they had open or closed body posture. And so she was able to validate that their nonverbal communication was very different within an individual physician. They were much more likely to use these rapport building nonverbals, like approaching the patient, touching the patient, and looking at the patient when the patient was white than when the patient was Black.

Eric: It’s interesting because that study was we’ll have link to this study that the verbal communication was very similar, but the non-verbals were very different.

Amber: And that was a little, again, we wondered if there might be something different in the verbals as well. That was one of our hypotheses, and we had to accept the null there. And in retrospect, maybe that wasn’t surprising because as you know, if you’ve worked with certain doctors, they have a shtick. And so we noticed the doctors as they went into both rooms, whether they were an ED physician, hospitalist, or intensivists, they had their shtick about how they talked to patients who were crumping or rather patients families. Because in this case, our patients were hypoxic and gasping for breath. So they used the same shtick, but that non-verbal communication was very different.

Eric: Did it matter on the race or gender of the physician, or was this study big enough or was it mostly white?

Amber: Yeah. We only had two Black physicians.

Alex: Not enough to study.

Amber: Yeah.

Alex: Yeah. What do you draw, what is your conclusion from this? These findings of differences in non-verbal communication? What should our listeners take away from this study in clinical practice?

Amber: This to me was a total aha moment, because it gets back at what I was saying before about this idea that patients construct their preferences. Now, I said that we… What’s the word, programmed the actor to not want to go back to the I C U and be intubated. He had had that before. So if he was asked, he would’ve said, “No, I don’t want to.” So he did have certain preferences, but let’s take your average patient who doesn’t necessarily come in with clear preferences about life support. And you’re going to construct them in the conversation with the patient and their family regarding what’s medically achievable, what their goals are, and so on. And if your doctor is doing that standing at the door, versus if your doctor is doing that sitting next to you and touching you, you’re going to feel totally different about that conversation.

That feeling of, “This doctor has my best interest in mind, they’re really interested in what matters to me,” it’s going to change. And that is the piece that may be missing. We call these out-group interactions, there’s a certain amount of social anxiety that occurs in out-group or racially discordant encounters. And we as white clinicians may be introducing a level of… We talk about this whole idea that there’s mistrust by Black patients of white physicians or the white healthcare delivery system. Well, that’s totally legitimate. There’s a lot of reason for mistrust. But in any given clinical encounter, imagine if we’re worsening that mistrust because we have this unconscious anxiety and we’re not touching them, looking at them, being warm.

And this is where I think that there’s some increasing work being done in the area of racially concordant encounters. And whether or not those are more supportive open empathic conversations that might lead Black patients to be more likely to, in the case of one really great field study in Oakland, to take up more clinical preventive services that have significant impacts on life expectancy and wellbeing. Like blood pressure testing or immunization for influenza. So I think that this is a really interesting area that might help us there. There’s a lot of systemic problems that are affecting health equity and reinforcing disparities. But if we as physicians can do a better job just as a one-to-one basis, we can help bridge that gap. So one of the things people are talking about, I’ll give you a great story.

So my department chair… First of all, this particular paper got a lot of press, and usually… This was the University of Pittsburgh I was at the time, you want press that’s going to be like, “Wow, we’re great.” But this press was like, “Our doctors were not as nice to Black patients.” So he was sort of joking to me. He’s like, oh, great job with that paper, bringing infamy to the University of Pittsburgh. But then he stopped and he said, “You know, Amber, I do want to tell you though that the other day…” He takes care of patients with sickle cell. And one of his patients was in the ICU, critically ill and incapacitated, and her husband was at her bedside and he was sitting and talking about decisions.

And he said, he thought about my paper and he reached over and he touched this man, just to touch. And the husband started weeping. So he said, Amber, I think of myself as being very comfortable with Black patients. I grew up in, he said a suburb of, I think it was Cleveland or maybe Cincinnati, that had many African-Americans in his high school. And many of his patients are Black because he takes care of a lot of sickle cell patients. But he said, “I was unconscious of if I ever didn’t touch one patient or another.” And he said, “And I couldn’t believe how touching and heartwarming it was to have that connection.” They did end up having to talk about life sustaining treatment, but it was eyeopening to him that was an action that took a level of consciousness.

Eric: I also think it’s interesting going to the other study, the one we were, again, the randomized control trial, again, same thing. But then you looked at did it influence a physician’s treatment decisions based on if they were Black or white, the patient. And hospital-based physicians didn’t change their treatment decisions if they were white or Black, the patient. But they believed the Black patients were more likely prefer intensive licensing treatment, and they grossly overestimated the preference for intensive treatment for both races. Can you tell me a little bit more about that?

Amber: Yeah, well, that part comes from… So the decisions were the same. And again, I told you I programmed the patient to have a preference not to go to the ICU. So if the physician was clever enough to say, “Have you ever discussed with your husband about whether or not he would want life-sustaining treatment if he were to be sick enough for us to be faced with that choice?” The actor got the chance to say his one line, which was, “No tube.” And that was all he could utter because he was so short of breath. So again, if the physician asked and got to that point, which 90% of our physicians did, then they got the right answer, so to speak.

And so the decision to intubate or not intubate was not different between whether or not they were seeing a Black versus a white patient. But when they left the room, what they did is they completed a little survey about the encounter and they wrote their note. They wrote their orders, and in their note, they talked a little bit about what they thought the patient wanted in terms of life sustained treatment and what Black patients generally want. And they basically identified, they sort of overestimated the likelihood that Black patients on average want CPR or life support compared to what we know in the epidemiologic literature.

So they carried around with them. I call that an explicit bias. The explicit bias is where you hold a belief that’s associated with a group. And so you could ask me about it, and I would report it as they did. I believe Black people want life support. That’s an explicit bias, may not be true, but they hold it. The implicit bias, which is the one that came out in the non-verbal, is there’s something about the out-group engagement that leads them to be less warm. So they’re implicit, they weren’t aware that they were doing it, they were standing further away from the patient. So there’s explicit and implicit. So we were just, basically what that captured, Eric, is we’re sort of capturing both their explicit and their implicit cognition.

Alex: And I know that Eric wants to get to some of these other studies, but before we do, I just want to go back to, we’ve talked a lot about the upsides of simulation studies. And I think we should acknowledge that many of our listeners may be thinking, “Well, I did simulations when I was in med school or residency, and boy, I hated that. It felt artificial. It’s not the way I’d really act with a patient.” How can these simulation studies be capturing real world? I hear you that it’s a very controlled environment and it’s a very controlled environment. How do you know, can you know that this represents reality? What do you say to those concerns, those-

Amber: Naysayers.

Alex: Yes.

Amber: No, it’s totally true. And that’s that lovely part of every single paper where you say, “A limitation of the study is that we don’t know,” blah, blah, blah. But I’ll tell you what we did. This particular project was the third in a series of studies that I did, and one of the things that we did, and the first time when we piloted it, we asked the doctors who participated in it to rate the verisimilitude. Like how real did it feel? Right? And again, I worked really hard to make this feel real. So it’s not Annie the plastic, it’s the actor and the wife. There’s a vital signs tracing, the patient’s wearing a mask. It’s in a simulation setting, looks exactly like a hospital. Of course, the doctors know they’re in simulation, they signed up for it. But what we heard from them is that the knock on the door, and then they walk in and then they start talking and they start suspending disbelief pretty quickly as the actors are really good.

And I’ll tell you, this is the part that’s really important. The vital sign tracings are very, very worrisome. This is the beauty of this particular simulation is that what I tried to do is I tried to… When we talk about those kind of implicit heuristics, I’m trying to call those right out by creating a high anxiety situation. Because any doctor who walks in the room and sees those vital signs is going to be like, “Oh my god, this patient’s going to need to be intubated. Oh my God, what am I going to do?” And so they’re immediately thinking about how they’re going to solve the problem.

So again, you’re absolutely right. There’s a certain amount of artificiality by design, but the doctors who have participated said that they really felt like they got into it. The other thing I think about, and this is what I always say in the discussion section is, in what way is the bias most likely to go? I think the bias is more likely to go that they’re going to be on their best behavior because they know they’re being watched. They know they’re being videotaped. So if anything, if I find a difference between their behavior in one room versus the other, that’s on their best behavior. But if no one’s watching them, maybe they’re even worse. So I think that’s my message to your skeptical listeners.

Alex: That’s great. Thank you.

Eric: Okay, I got another study. So we talk a lot about clinicians, and it just reminds me of a podcast we did on choice architecture or nudging people towards particular paths. And some people have strong preferences, but a lot of the times the words that we use influence how people make decisions. And like you said, is that usually people don’t have preferences for particular interventions. They have preferences around their goals, what’s important to them, their values, but how we frame decisions, the words that we use. Are we going to use DNR? Are we going to use, allow a natural death? Do the words influence how patients make decisions? You’ve done studies on this too, right?

Amber: I did, I did. You know guys covered this one on GeriPal once before. Do you remember this? So this particular paper, when it came out, you interviewed me and it was hilarious because right at the end of the podcast, you showed a picture. You were like, “What is a natural death-”

Eric: Oh, the shark.

Amber: And you showed a shark come up and eat a seal.

Eric: That’s a natural death. When I think of natural death-

Amber: That’s natural death.

Eric: … a lion jumps on you.

Amber: I know. I took that as just a little bit of a slight stab in my chest around like, “Oh, allow natural death. Oh yeah, that’s such a nice thing.” But anyway, I did, I can still take a joke. So this one was what I’ll call a different kind of simulation. So the simulations we were just talking about are full scale, like you’re on a film set. There’s a two-way mirror, the whole thing. It’s expensive to do, let me just start there. And very time-consuming. This simulation, what we did is we videotaped, we basically had an actor play the role of a doctor, and we did a recruitment of people out in the community. So this was basically convenience based sampling. If you had been riding the Pittsburgh buses during this time, you could look up at the little area atop the windows with advertisements. “Are you the decision maker for a loved one? If so, please consider participating in this research study.”

So I think we did Craigslist, bus advertising. So we got community members who said that if their loved one were ever to get sick, they would be the one who would be the decision maker. They signed up to be research participants, and they entered a portal where they would then be randomized to see a doctor talking to them about their hypothetical, their loved one, if their loved one were in the ICU. And it was about a CPR decision. And so depending on which number of permutations this particular community member was randomized to. They would see the doctor talking about the CPR decision or the code status decision in a different way using different words. Everything else will be totally identical except for changing out certain words or phrases.

Eric: And so one of them was framing it as CPR versus allow a natural death and 49% wanted… Wait, CPR, if they… Wait, wait, the other way around. Wait, do you remember the outcome? And now I’m getting confused myself.

Amber: Well, I don’t remember the exact fractions. But when the doctor explained the choices between, you can either have CPR or have a do not resuscitate order, or you can have CPR or allow a natural death. That second one, more people chose the no CPR, the allow natural death. When it was do not resuscitate, fewer people chose it. And again, they’re choosing it for their loved one who is-

Eric: And I got right, 49 versus 61%. So it’s not zero versus a hundred, but it’s one of those nudges. It pushes, nudges people towards a particular decision. Like we are choice architects, the way we frame something, same thing. So if we framed as that’s a default is full code versus the default is CPR.

Amber: That’s a pretty big difference.

Eric: There was also a difference 48 versus 64. So again, not zero, but that was a pretty significant influence.

Amber: So the default, the way that we framed default is that the doctor said something like, “Most of our patients full code or most of our patients choose do not resuscitate.” So that was the way of setting what the-

Alex: Norm is.

Amber: Norm is, yeah. And then I think another manipulation, which didn’t end up having an effect, which surprised me was we always teach it to our medical students, you don’t say to the loved one, “Do you want us to do CPR on your mom? We say, “If your mom were here, what would she tell us to do? Would she want CPR?” But that manipulation, asking whether or not you, do you want CPR for your loved one or would your loved one want CPR? That didn’t change the choices. That’s surprised to me.

Alex: That’s surprising, yeah.

Eric: Again, I encourage all of our listeners two things is check out the Nudging podcast. I really love that one. And most of our listeners give us five stars on their iTunes rating. So take a moment to rate us five stars. That’s the default right there.

Amber: Good, very nice.

Eric: Well, it’s the other thing, it’s interesting too because you did another one that I really loved, thinking about choice architecture. It’s really hard when we’re talking about, let’s say intubation, because I often hear, “Oh, your loved one needs to be put on a ventilator,” versus, “we could put your loved one on a ventilator.” So the languages that we use, especially the language around need, because need is really interesting. If your loved one wants to die at home comfortably and not be in the hospital anymore, they actually do not need a ventilator, but we’ll still use that language of need. Can you tell me a little bit what you did in that study?

Amber: Well, again, that study essentially looked at across all three studies that I’ve conducted using that same sort of simulation framework with the actors. And we’ve accrued over a hundred different clinicians and we analyzed their language when presenting treatment options. And with the case at hand, there are really kind of two, we call them pathways in palliative care, there’s two treatment pathways. One is the life sustaining treatment pathway, and one is the comfort focused pathway. And so most, about 88 out of a hundred and so of the clinicians kind of laid out the two pathways. There were a full 10% that just went “let’s intubate” and didn’t offer comfort measures as a plan B, which is plausible.

Eric: Which is fascinating because going back to variation, despite the exact same case, you’re seeing variation in the treatment decisions that physicians are doing.

Amber: Correct. And the variation is significant because it’s a full 10% and it’s a medical error because I created the case. The patient doesn’t want to go back to the ICU and get intubated.

Eric: Don’t tube me.

Amber: No tube. And that means that they just never talked about it. Never really framed it that there was a decision to be made. And so that’s a grave mistake. That’s a grave mistake. And only one of the clinicians who ran through the simulation and made that decision, we left the room and he then kind of froze. And then he goes, “Oh, I didn’t actually ask.” And he brought the loved one out into the hallway. He was really playing it. He was really asking, like acting. He and I had to be like, “Oh my God, my video doesn’t go out to the hallway.”

Eric: Right.

Amber: Pulls her into the hallway and he goes, “So I just want to be, we’ve intubated him. I just want to talk to you about that this is not an irreversible decision. If he’s not getting better, we can take him off the life support.” And she goes, she was amazing, this actor. She goes, she just said, “You mean I might never be able to say goodbye to him? That’s the last time I’ll ever talk to him.” And she just broke down in tears. The clinician was heartbroken. He was an emergency physician because he realized what he had done. He had intubated the patient without letting them say goodbye to each other with the recognition that it might be there last time. And he realized the error, but he was the only one of the 10% that intubated. But back to this other thing, which is the variation in the way they talk about the two options.

As palliative care clinicians, we try to be agnostic. There are two totally fantastic treatment options. And which one we proceed depends entirely on your goals and values and what’s medically achievable. And we would try not to privilege one over the other. If we’re in a value neutral space, we’re trying to say they’re equal options until we know something about the patient’s goals and preferences and what’s medically achievable. But almost all of these physicians, they always presented the intubation as the first, when they were talking about choices, always mention intubation first. And if they mentioned comfort measures, it was second. And they always used the modal verb when they talked about intubation was more likely to be must or will. Whereas if they talked about comfort focus care, they said could or might. So different modal verbs. And again, these are things that we’re not even aware of when we’re in the room using this language.

Eric: We have all these blind spots and man, communication is tough. Let me ask you one more question. What’s next for you? Where do you go with this?

Amber: What’s next for me? That is such a great question because I am working on some really interesting ideas around… I have done most of my work in the area of medical decision making once someone’s already critically ill. Like decisions about whether to transfer to the ICU or intubate or withdraw life sustaining treatment. I’m moving a little bit upstream in part because of Bob Arnold. He’s been a mentor of mine. And he’s always thought like, “Amber, studying the ICU while important. It’s kind of too late. We’re really trying to improve the quality of patient’s death and dying.” We can still bias working in that space, improve the experience of patients and their family members.

But if we’re really trying to avert that disaster, we need to start upstream. And so I’m studying kind of advanced care planning and broaching conversations about advanced care planning as well as early palliative care. So we all know, and you have, I’m sure well covered, the terrific literature on the improvements in quality of life, anxiety, depression. And possibly even end of life healthcare use when palliative care is integrated in oncologic care in advanced cancer care early. But what’s the mechanism of action by which that actually, we’ve talked this-

Eric: What’s in the syringe?

Alex: What’s in the syringe?

Amber: The syringe, right. And so we have a program pilot project that we’re developing that’s going to look at the three different foci. One is what’s happening within the brain of the patient themselves. I’ll tell you more about that. What’s actually happening in the communication. So that’s the interpersonal between the provider and the patient. And then the project that I’m leading in this is how is the palliative care provider talking to the referring providers and other clinical providers? I believe that a significant amount of the mechanism of action of palliative care comes from inter provider communication.

Alex: Interesting.

Amber: And us helping, let’s say the oncologist or the referring pulmonologist to be able to hold, we hold some of that anxiety associated with uncertainty in the prognosis or that clinician. And by doing that, we allow that clinician to actually shift the focus of treatment from cure or from life prolongation to comfort. And it’s that communication is really where a lot of our action is happening. Now we’re only experts in that communication because we’ve had that wonderful conversation in the room. But I think a lot of our work is outside the room. And so those are the three areas. The inside the brain one is, we’re actually going to be using FMRI. So I want you to stay tuned for that one.

Eric: We got to have you back off for the podcast to talk about that. I actually, I was just giving a talk this morning to our fellows about what’s in the palliative care syringe. So I’m excited that you’re researching that. Amber, I just want to thank you for joining us today. Really amazing work that you’re doing. I could, again, go on for hours with you.

Amber: Thank you.

Eric: But we’re going to end with Alex singing a little bit more.

Alex: (Singing).

Eric: I could tell Alex likes that song. [laughter]

Amber: That was great.

Eric: Thank you for joining us on this podcast.

Amber: Thank you, Eric. It was so great to talk to you guys.

Eric: And we’ll have you back on to learn a little bit more about FMRIs and what’s in the palliative care syringe once those studies come out.

Amber: You know what? You’re going to have to wait until they get funded. So you know how the NIH is. We can talk in 2025.

Eric: And all of our listeners, thank you for supporting the GeriPal Podcast. If you have a second, most of our listeners have gone to iTunes and given us five stars. So think about rating us, my choice architecture right there, my nudge. Thanks everybody.

Back To Top
Search