Prognosis Superspecial: A Podcast with Kara Bischoff, James Deardorff, and Elizabeth Lilley

Eric 00:00

This episode of the GeriPal podcast is CME eligible.

Alex 00:05

To claim credit, please go to the CME tab on GeriPal.org.

Eric 00:09

Welcome to the GeriPal podcast. This is Eric Widera.

Alex 00:12

This is Alex Smith.

Eric 00:13

And Alex, we have a super special today, three different articles.

Alex 00:18

Prognosis, super special today. And we have some great guests. First time on the GeriPal podcast, Liz Lilley, who’s a surgeon and faculty in surgery at the Brigham Women’s Hospital and Harvard Medical School. Liz, welcome to the GeriPal podcast.

Liz 00:32

Thanks. Good to be here.

Alex 00:33

And we have two returning guests. First, we have James Deardorff, who’s a geriatrician and assistant professor at UCSF in the division of Geriatrics. James, welcome back to GeriPal.

James 00:42

Thanks for having me.

Alex 00:43

And we have returning guest Kara Bischoff, who’s a palliative care doc and associate professor at UCSF in the Division of Palliative Medicine. Welcome back to GeriPal.

Kara 00:51

Thanks so much, Alex.

Eric 00:53

So we have three articles to talk about, about prognostication and prognosis. But before we do, we always ask for a song request. Liz, I think you got the song.

Liz 01:04

I do have the song. The song request is “What was I made for?” by Billie Eilish, from the very fabulous Barbie movie.

Eric 01:13

Why did you pick this song?

Liz 01:15

I think that when patients think about, or when we think about prognostication, there’s a lot of uncertainty and some existential distress sometimes. And I think that this song kind of captures some of that distress.

Alex 01:28

In addition to wanting to torture me with female singer songwriter Diva, which I secretly enjoy the challenge of. Here’s a little bit. Yeah, I didn’t see the Barbie movie, but…okay, here we go.

Alex 01:47

(singing)

Eric 03:20

Alex says you’re torturing him, but I know he loves the song because he continues to play the extended intro version. [laughter]

Alex 03:29

Thank you for the opportunity, Liz.

Liz 03:32

Well done.

Eric 03:34

So, we’ve got a lot to cover on prognosis. We’re going to be talking about three different articles, one on the PPS. We’ll talk about what that is, another one on thinking about prognostication in nursing home patients with dementia, or whether or not they’re going to be admitted to a nursing home. And the third one on prognostic alignment. So I was thinking about starting off with the PPS, the Palliative Performance Scale. Is that right, Cara?

Kara 04:03

That’s right. You got it.

Eric 04:06

Why is the PPS important?

Kara 04:09

Sure. Yeah. So the palliative performance scale, or PPS, is essentially a tool to help clinicians with prognostication for seriously ill or palliative care populations. It’s essentially a measure of functional status. It’s modeled off the Karnofsky performance scale, but it was adapted for palliative care populations, and it was developed in canadian palliative care units in the 1990s, about 30 years ago. But it’s still the prognostic tool used most often by palliative care clinicians.

Eric 04:39

Yeah, I see it used on inpatient, side on consult clinics in hospices. A lot of people use the PPS when they’re thinking about function and then aligning that with prognosis.

Alex 04:51

Most used prognostic index on e prognosis.

Eric 04:55

Oh, really?

Alex 04:56

Most used?

Eric 04:57

Most used. Out of all the ones, out of.

Alex 04:59

All those indices, why do you think that is?

Kara 05:02

About 12,000 views per month? I think it’s just. It is appropriate for all patient populations, and it is developed specifically for the palliative care and hospice populations. And it’s pretty straightforward. It’s pretty easy to use by clinicians in the course of care.

Alex 05:19

I suspect that eprognosis gets a lot of hits because we have simple, easy to use dropdown menus and there are all these skip patterns in estimating the PPS. And so we make it really easy for clinicians to use it. That’s a suspicion. My other suspicion is that hospices are using it for recertification because it predicts such a short term mortality and they have data to back that up. And as Kara will discuss, that may be problematic.

Eric 05:47

And the PPS we may discuss.

Alex 05:48

I guess we didn’t do hospice specifically.

Eric 05:50

Yeah. What’s in it again? It’s a measure of function. Right?

Kara 05:54

Yeah, exactly. So how physically active the patient is, how much time they spend awake, how much they’re eating, how much care they need, that type of thing.

Eric 06:03

There is a question about disease extent, but it’s wrapped in with another question.

Alex 06:08

Right.

Eric 06:09

It’s pretty basic.

Kara 06:10

Yeah. And so these are easy things that we usually are aware of when we’re seeing patients in palliative care.

Eric 06:15

Yeah.

Kara 06:16

Doesn’t require a lot of advanced measurements.

Eric 06:18

Yeah. And then I could imagine. I guess the question is, so you did this paper on validate. Validating the PPs in the populations at UCSF. Is that right?

Kara 06:29

Yeah, it’s exactly right. It’s an external validation study.

Eric 06:32

So why. Why did you do that?

Kara 06:35

Yeah, of course. So, essentially, I was seeing patients in primarily outpatient palliative care settings, and I was noticing that the patients I was seeing didn’t have the prognosis that their PPS scores would predict based on classic literature. For example, if I saw a patient with ALS with a PPS score of 50%, classic studies would predict that that patient had about 50 days to live. But that just didn’t seem right.

Eric 07:03

And those classic studies, were those done in the hospital and PALP care units and clinic?

Kara 07:07

Yeah, they were done over ten years ago, and mostly in inpatient palliative care units and hospice settings, primarily in cancer populations. So we really wondered, you know, in modern palliative care practice, where we’re seeing patients a lot farther upstream than we used to, and we’re seeing patients with a wider range of diagnoses, we wanted to update, sort of. What are the prognoses associated with various PPS scores?

Eric 07:31

Yeah, and I can imagine, I. The patients that we were seeing, a, first of all, in the 2000 2010s, there were very few clinics out there, and the patients that we were seeing in palliative care were mostly inpatients. The folks that we were seeing either in clinic or inpatient, were probably a lot sicker than the ones that we’re seeing now. Cause there’s been a big push for seeing them upstream, which also. Right. Would change their prognosis. That’s the hypothesis, absolutely.

Kara 08:00

Yeah. We hypothesize that we’re seeing patients with significant disability, so lower PPS scores, but are living for a lot longer than they used to.

Eric 08:10

Did you find what you thought you would find?

Kara 08:13

Yeah, indeed we did. So, using data from our UCSF inpatient and outpatient palliative care practices, we found that the prognoses associated with each PPS score are two to twelve fold longer than they were per previous literature. So, coming back to the example I gave a moment ago, in classic literature, an outpatient with a PPS of 50% would live about 50 days. But we found that outpatient with a PPS of 50% was actually living 598 days.

Alex 08:44

I like the way that CarA is referring to the literature as classic, which is not as pejorative, as old or dated. Very generous to the.

Eric 08:55

Alex and I are classic attendings.

Alex 09:00

I remember, like, the 2010s, when, like, I wrote a paper with Mike Raybo, who is a director of the cancer focused outpatient clinic at UCSF, about the emerging landscape of outpatient palliative care. So you’re absolutely right. Like, those clinics just weren’t around now, and they were just emerging in cancer, much less what you do primarily, which is see non cancer folks.

Kara 09:21

Yeah, it’s kind of interesting to remember how far we’ve come in just ten or 20 years. As you alluded to the non cancer space, we also found that patients with non cancer serious illnesses, on average, live longer at each PPS score. And we also, similarly, using our data from our outpatient palliative care clinics, we looked at whether the PPS scores determined through video visits were associated with the same prognoses as PPS scores determined in, in person visits, and we actually found that they were the same. So that was reassuring that pts can be accurately determined through a video visit.

Eric 10:03

So you’re worried that maybe if you’re doing telemedicine, you wouldn’t be able to accurately engage functional status.

Kara 10:08

Yep.

Eric 10:09

So you’re reassured on that concern there.

Kara 10:12

Yep.

Eric 10:13

You know, I guess one question for me is a ways back, I looked at our. God, it must have been like 2015, 2016. We started a clinic here in 2008 at the VA, and what we found was that, you know, our median time from consult to death before we started clinic was like 18 days, somewhere around there, like somewhere around two weeks. And since starting the clinic, median time for consult to death was. It’s now like over a year.

Alex 10:40

Wow.

Eric 10:40

Yeah.

Kara 10:43

Those are inpatient consults?

Eric 10:44

No, all consults. Just if you look at all consults. But if you break it down from inpatient to outpatient, really the biggest growth and time consult to death is really those people who’ve seen in clinic, those people who’ve seen an inpatient, there is a growth, but it’s still a lot smaller, shorter than a month. And I wonder, are you seeing the same thing when you look at that data from the PPS? Is that the folks that you see in clinic are the ones that are really extending that time from consult to death versus the inpatients?

Kara 11:15

Yes, absolutely. So there were really significant changes in the, the prognosis predicted by each PPS score in the outpatient setting, particularly at higher PPS scores, or when predicting a longer term mortality, six month mortality, twelve month mortality compared to one month mortality.

Alex 11:33

Yeah. Okay, I have a question. Is a PPS a good prognostic index for mortality?

Kara 11:41

So that’s a great question. It actually is decent. We looked at the PPS ability to discriminate between those who live and those who died. And we found that it is good. The PPS has good ability to discriminate between people who live and people who die. When predicting short term mortality, like one month mortality, both in the inpatient and outpatient setting, but particularly in the outpatient setting, when you’re predicting longer term mortality, like six month mortality, twelve month mortality, or median survival, that’s where the discriminative ability of the PPS is less good. And that’s where I might recommend that clinician consider using a more disease specific prognostic tool, like the Bonnite index for COPD or Seattle heart failure model. Or, as a shout out to my colleague here, the deardor model for dementia as examples, though those need to probably be explored further in palliative care populations.

Eric 12:37

So let me get this straight. Short term mortality, PPS does pretty good. Inpatient, outpatient?

Kara 12:42

Yep.

Eric 12:43

Longer term mortality, maybe not so good. Why do you think that is?

Kara 12:49

Yeah, I think that there’s a whole lot else that can impact how people do over time, including the disease state. So especially in the outpatient setting, people can live with a large amount of debility for a long time in many disease states, like dementia, Parkinson’s, COPD, etcetera, ALS. Yeah. In those disease states, we really do have to be more thoughtful, think beyond just the PPS, gather additional information to pull together in order to give the best prognostic estimate possible for a patient.

Eric 13:29

I love that. It reminds me of one of my favorite articles. I think it was Tom Gill who wrote about the functional trajectory in the last year of life, and there was no disease, cancer, COPD, heart failure, that had a universal trajectory of functional status in the last year of life. Largely because cancer is a mixed bag. There’s a lot of different types of cancer out there. It’s hard to unified. The only disease that came really close was dementia, where pretty much everybody had severe functional disability in the last year of life. Which would argue is that I feel like it’s like age. One single prognostic factor, age, functional status, whether or not you have cancer or not, is never going to be that great. And predicting the future.

Kara 14:21

Yes, and it may be that a trend in PPS scores, or another way to say that would be a decline in functional status is more meaningful than any single PPS score. Fortunately, we didn’t have enough patients with repeated PPS scores over time in our data set to be able to look at that. But I think that would be a really interesting next step to see if the trend in PPS scores adds information about the patient’s prognosis, which I expect it would.

Eric 14:50

So, Alex, is eprognosis going to have the new data in it?

Alex 14:54

Eprognosis is going to have the update. Carl helped work with our programmer to make sure it looks good. And we are holding this podcast. We’re recording it in what is this, first week of May, but we’re holding it until the article drops. And then we will also have access to the new updated eprognosis using the UCSF outcome rates.

Eric 15:18

Is it going to split it up inpatient and outpatient, too, like you have? Yeah, that’s right.

Kara 15:22

Clinicians can find a prognostic estimate associated with their patient’s PPS score based on the setting in which they’re seeing the patient inpatient versus outpatient care, and based on the patient’s disease category, cancer or other serious illness. And I also wanted to mention that clinicians working in hospice can still access the older version of the PPS tool on e prognosis.

Eric 15:46

Yeah.

Kara 15:46

To find estimates that were generated in hospice populations.

Eric 15:50

That’s fabulous.

James 15:52

Thank you.

Eric 15:53

I guess a question for you, James, is so PPS, we’re looking at mortality when we think about prognosis and prognostication, is mortality the only thing we should think about?

James 16:05

Yeah. I mean, I think patients care about a whole host of different factors, survival being one of them. But another important one is, like, prognosticating someone’s future level of function. They want to know, how many years does it take before I might expect to see some difficulties with getting around, difficulty walking, difficulty driving, things like that. And so those types of daily activities are also another important outcome that we should think about.

Eric 16:36

Yeah. And that kind of leads to your paper. So you got a paper published. There’s a jama. Jama. I am. Jama. I am.

James 16:44

Jama.

Eric 16:44

I am.

James 16:44

Yep.

Eric 16:45

Jama, I am.

Alex 16:46

On time to nursing, needing nursing home level of care. It doesn’t just roll off.

Eric 16:51

No, it doesn’t. You are not. You were the senior author, Alex, right? Yeah. So you’re kind of the guest host today. Why did you do that paper, James?

James 17:03

Yeah, I think so. This is a particularly interesting topic because we know that individuals with dementia have a high rate of entering nursing home. This is often due to disease progression that leads to things like functional impairments, needing help with bathing or dressing, or even like behavioral issues like wandering or need for supervision. That just leads to issues that can make home care very difficult for these individuals. And so we know that the decision to enter a nursing home is incredibly challenging for families, especially because it often comes with thoughts that they’re, like, abandoning their loved one, really.

Our goal with this study is to provide clinicians, patients and families with an estimate of when a person with dementia might require the level of care typical for individuals at a nursing home. And we would hope that these estimates could help with future planning on when a time might be when you would need to either hire an extra caregiver, provide 24/7 level of care at home, or think about actually entering a nursing home. And so that was kind of our goal in this study.

Eric 18:15

So ultimately, the outcome that you were looking at was, what was the final outcome? It was the need for nursing home level of care.

James 18:25

Yeah. So initially our study, we wanted to just kind of straight up predict need for nursing home admission. We ultimately decided against using that outcome because you can imagine that outcome varies a lot by socioeconomic status. And just like cultural background, there are some individuals and families that would really never have their loved one enter a nursing home, choosing instead to take care of them. And so many individuals with dementia remain at home, despite requiring a high level of care that you would typically see at a nursing home. And so really, the metric that we chose, the need for nursing home level of care, is a more standardized metric that can be applicable across any cultural background or socioeconomic status.

Eric 19:11

Yeah, I love that, too, because I always think about healthcare systems, too. Like, the entire role for a pace program for all inclusive care for the elderly is to take nursing home eligible patients and keep them at home. So it’s not just like the patient or the family, but it’s also the healthcare system. And what does society help people with at home so they can stay at home.

James 19:32

Yeah, exactly. And it’s kind of tricky. So how you define need for nursing home level of care is challenging. We compromised on a definition that, based on previous guidance, we said, essentially you have to meet one of three criteria. One of the criteria was needing help with three or more activities of daily living, things like bathing, getting in or out of bed, toileting, walking across the room, or needing help with two or more AdL’s, plus the presence of wandering or like constant need for supervision. And then the third criteria could have been needing help with eating, which we felt like was a pretty high level of care needs.

Eric 20:13

Yeah, that seems like a higher bar. Like, it seems like most people with three ADL impairments, that’s often what we’re thinking about nursing home, or at least really 24/7 care at home or close to that.

James 20:30

Yeah, I think it’s pretty typical of what you would see, like, from, like, the Medicaid criteria for paying for nursing home care, is that you really have to have a lot of functional impairments. And we wanted to make sure that we added some other behavioral issues, like wandering or need for supervision, that are unique to individuals with dementia, because that’s another important thing that can cause a lot of caregiver burden at home.

Eric 20:57

So you created a prognostic model. You validated that model. What factors were included in that model?

James 21:05

Yeah. So we included. We ended up creating two separate models, one that included just self responses, so responses from the person with dementia, and another from proxy responses, which is essentially a proxy report. So, information from, like, a family member, spouse, or a son or daughter. And these two models included some of the factors that you typically see in these types of studies. So, for both models, we included age, baseline functional status, as measured by AdL’s and IADLs, and whether the individual was still driving. The proxy model had body mass index and falls history.

And then the self response model also had female sex incontinence, and then a simple test of orientation, like whether the person can recall the month, day of the week, in the year. And so I think we wanted these models, really to have good face validity, be very pretty parsimonious, and also include things that most clinicians would ask about during a visit. I think the final models are successful in the sense that they do have a lot of face validity. You can imagine that someone who’s driving to the farmer’s market, shopping for things like carrots and cabbage, and cooking a meal at home, is going to have several years before they would require nursing level of care compared to someone who is already needing assistance with one ADL bathing, they’re going to be a lot farther along, and it might not be as long before they require that level of care.

Alex 22:46

We should title this podcast. It’s in the cabbage.

Liz 22:50

I was just wondering. I was just wondering what vegetables from the farmers market would make someone a little bit worse off.

Eric 22:58

Exactly, exactly. Core of it is not included in this model. I think, I guess cognition, which is.

James 23:08

You have a little bit cognitive test. Uh, this model didn’t have much in the way of comorbidities, and I think that just attests to the fact that functional status is really, like, the end impact of a lot of different comorbidities. And so someone’s baseline functional status is just such a powerful predictor of their future functional status. And since this model is so, is predicting future functional status and really their baseline functional status is a very important one.

Eric 23:36

And I don’t like to get wonky, but I’m going to get wonky as the wonky as I can even get, which is pretty like, you’re probably going to laugh at me as I ask these questions, because it’s not wonky. But when you create these models, do you actually look at a bunch of things like comorbidities, and if they’re not that influential as compared to others, do they just fall out or how does that work?

James 23:57

Yeah, so essentially we have like a pool of kind of candidate predictors. You include things like comorbidities, age, functional status falls, some other variables. And then usually what we try and do is in order to make this more feasible, rather than clicking through 2030, predictors for busy clinicians, is perform some type of variable selection so that we can come up with a more simplified model. It still performs pretty well and is just a lot easier for use, for clinicians to. To use basically, within 30 seconds or a minute, they can punch in all these numbers, all these factors, and then get a prognostic estimate.

Eric 24:40

How do I punch in all those numbers? Because I looked at the supplement and the example and it was beyond my math ability of what I was reading.

James 24:52

Yeah. So to do it on your own would be pretty challenging and require a lot of math.

Eric 24:56

But I am not alone. I just could not.

James 25:01

Luckily, this model is also hosted on eprognosis.

Eric 25:04

I’ve heard of that site.

Alex 25:05

You’ve heard of that one?

James 25:06

If you’ve heard of it.

Eric 25:08

Super special.

James 25:10

Exactly.

Alex 25:11

It’s our stealth way of getting it.

Eric 25:12

Is it up and running?

James 25:14

It is up and running. You can just go to the website, find the model, input the patient characteristics, and then it’ll give you a certain few numbers. It’ll provide you with like a percentage estimate of probability of nursing home level of care at two, five and ten years, and also the median predicted time to nursing home care. So, like, it would essentially say that this person will likely need a lot of help in, let’s say, two to five years as an estimate.

Eric 25:46

And that Cara brought up discrimination, which in this case is a good thing if it can discriminate real quickly. James, what is discrimination? And was yours any good at it?

James 25:58

Yeah. So discrimination is one aspect of prognostic modeling performance, and it really indicates, like, how well the model can differentiate between someone who does and does not have the outcome. So a good model predicts that, let’s say someone gives them a high risk of mortality or something compared to someone else, that a well, discriminated model would assign a higher probability for someone who actually had an outcome of mortality. And so this model that we did discriminates fairly well. It’s certainly not as good as other prognostic models that we’ve developed in the past, and I think that just is a testament to how difficult this outcome is to predict.

Eric 26:43

There’s a longer term outcome, too, right? It’s not just in the next couple weeks or months.

James 26:49

Exactly. Yeah. And also, a lot of it is kind of challenging to predict, like, if you. Someone’s function in the future can depend on, like, a crisis event, like a fall and a hip fracture. That is really challenging to predict. So I think the estimates that you’ll find on ePrognosis maybe have a little bit wider range than some of the other calculators. But this information can really still be used as a starting point for enhancing conversations with patients and families.

Eric 27:20

How do you expect people to use it?

James 27:22

Yeah, I think it can be used essentially a lot with future planning and counseling. I think when you’re talking with patients and families with dementia, unfortunately, we know that it is a progressive disease, and most individuals with dementia will be admitted to a nursing home or require just a high level of care. And so putting an approximate number on when you might expect that to happen is helpful in just preparing for the future, when you may need to hire more help, get more family supervision and things like that, so that people don’t feel as blindsided when that time comes.

Eric 28:07

Yeah. Which gets us to our third part of this discussion. So we’ve talked about the PPS. We’ve talked about James’s model around dementia and nursing home placement. So there’s the how do we estimate prognosis and using our clinical stalt versus using these prognostic tools? You know, the other really important part when we’re talking about prognostication is how do we communicate prognosis amongst our interdisciplinary healthcare teams? And then finally to the patient, which is like the other really big part of prognostication. And, Liz, this is why I really liked your paper. Was it annals of surgery?

Liz 28:48

That’s right.

Eric 28:48

Yeah, annals of surgery. We’ll have link to it on our show notes as we think about moving into that second part of prognostication, the communication part. You talk about prognostic alignment as a really important part of that piece. What do you mean by prognostic alignment?

Liz 29:08

So, prognostic alignment, which I kind of think of as both a noun and an. And a verb, but it’s this process that we take each of our individual prognostic awarenesses or prognostic ideas and we communicate them and we understand where we’re each coming from and we make sure that we’re starting from the same place. And we hopefully are doing that before we’re talking to patients and family members. So they’re not getting different prognostic understandings from each individual member of the team. This idea of prognostic alignment actually came up during my palliative care fellowship, where I noticed that palliative care, I spent a lot of time making phone calls and emails and trying to get people to talk to each other, which was just a very different.

Eric 30:00

And people. You mean by team members or patients and family members?

Liz 30:05

Team members.

Kara 30:07

So I’m going to talk to each other.

Liz 30:12

It’s just, it’s a wild concept, but I found that I was spending a lot of time sending emails and trying to get the person in the ICU to talk to the surgeon who operated on the patient and the nephrologist and the cardiologist and all the who were taking care of pieces of a person.

Eric 30:31

Yeah, I wish we were a palliative care ologists. That would be pretty cool. And I really encourage everybody to read Liz’s article. But you actually gave a case that included some ologists and going back to the farmer’s market, a cabbage case. Do you remember what you describe that case? Do you remember it?

Liz 30:53

Yeah. So this was actually a case that Ricky Lieder, who came up with the word prognostic, prognostic alignment. So snaps to Ricky, but he presented this case to us of a person who had had a heart operation. And as things sometimes happen in surgery, patients can experience complications and one complication, particularly in someone who’s frail, who’s not buying the cabbage, maybe they’re buying the chard at the farmer’s market.

They can end up with more than one thing happening, and all of a sudden the kidneys aren’t working and the heart isn’t waking up the way that we hoped it would. And so this case that we presented was based on a real case where there were different members of the team, really smart people with good intentions, who were telling the patient and the family just wildly different things.

Eric 31:47

And this patient had infection and delirium, and nephrologist was saying that they’re not gonna, you know, they’re not gonna get better. From a yemenite, a renal standpoint, the cardiac surgeon said, we’ve seen people get better from these really bad cases. And then you got the ICU team, who’s probably thinking like, oh, my God, this is such a train wreck. And the palliative care doctor kind of somewhere stuck in the middle.

Liz 32:12

Yeah. Yeah. And I’m sure that everyone in palliative care has been that person stuck in the middle, where it’s like you’re getting called because no one likes each other right now.

Eric 32:26

Go ahead.

Alex 32:26

Did this come out in the meeting?

Liz 32:29

This does come up in the meetings. And what’s so funny is I remember one case, not this case, but a very similar one, where I spent very little time in direct contact with a patient or the family member. Our intervention was a team meeting where everyone talks and says, here’s what I’m seeing, and here’s what I’m thinking and feeling. And sometimes by having those conversations, we can reach alignment, and sometimes we can’t, and we can recognize that we, with good intentions and a lot of thoughtfulness, are never going to meet eye to eye. So how do we share that with a patient and a family member so that they can feel supported by everyone that cares for them?

Alex 33:13

Because many of us, I’m sure others on this call, our listeners, have had that experience of, you’re in the family meeting and, shoot, you forgot to discuss prognosis before, beforehand. And, you know, one doc says one prognosis, another team member says a different prognosis, and all of a sudden, you’re trying to hash it out in front of the patient and family. Like, that is the worst situation. Right?

Liz 33:38

It’s like you wish that you had, like, a panic button, and then the meeting ends. Yeah, it happens. I’m sure it happens. It’s happened to all of us.

Alex 33:49

This is one of the reasons when I’m teaching, like, how to run a family meeting to trainees, I say the pre meeting is as important and sometimes more important than the meeting itself. You know, just like you said, Liz, getting everybody on the same page is vital before you go into that meeting. And sometimes that’s the most important outcome, is trying to get people on the same page about prognosis and then all of the different options that fall out once you’ve agreed on prognosis.

Eric 34:19

How do you ask that question of all these different disciplines? Because I’m assuming it’s like talking with patients and family members. How you ask the question is really important. How do you ask that, Alex?

Alex 34:30

I usually anchor at death because most people, if you leave it to their own devices, they will be optimistic, because that’s just who we are. We hope for the best for our patients. I like, to anchor at the other end of the spectrum and say, is anybody worried that this patient is dying? And then sometimes people will shake their heads, whereas if you left, and then I say, well, can we use those words with the family?

Eric 34:56

So is dying. That’s very different than maybe dying or may die or the odds of this person dying is dying.

Liz 35:04

Yeah, that’s such a, that’s a really. I’m going to try that. That’s such an interesting, because the thing is, you know, I’m a surgeon. I’ve been the person who had that unrealistic optimism because my patient’s going to pull through, and, and sometimes I’m right and sometimes I’m wrong, but I’m always worried that the patient is dying. And I can recognize that. And maybe by vocalizing that and saying, yeah, guys, I realize that things are really bad. I’m just not sure. I’m not sure how wide that uncertainty is. And it’s so hard for all of us to make decisions amidst that uncertainty, let alone to be a family member dealing with that.

Kara 35:47

Yeah, I’m thinking about the outpatient palliative care setting as well. And this conversation really resonates there, too, where often it’s the patient and family going to different visits, hearing wildly different things from different doctors that aren’t talking to each other, and they’re just going on this rollercoaster ride that can be pretty brutal, getting whiplash in the process, and they’re having to be the conduit between the different clinicians and just wondering how we can improve this prognostic alignment in the outpatient setting as well.

Liz 36:20

Yeah, that’s such a good point. I think in the outpatient setting, again, getting everyone into a room is that much harder sometimes, to Alex’s point, sometimes it would be helpful if we could all just agree that this is really worrisome and that we’re all a little bit scared that the patient is dying. We will all worry that. And I think that sometimes this misalignment happens because some people, most people, I would say naturally, as you said, are inclined to focus on the positive and try to avoid using words like dying. And maybe by getting that out there and also having room, holding space for that uncertainty allows us to work as a team, a functional team.

Alex 37:11

Liz brings this really unique perspective as a surgeon and palliative care doc. Liz, I wonder if you could talk about one of your other research interests, which is around pancreatic surgery and the uncertainty inherent in treatment and options and prognosis in particular.

Liz 37:26

Yeah. So I’m working with an amazing research resident, Orly Farber, who will be a future guest at some point, and she’s doing work with me and George Molina, looking at pancreatic cancer uncertainty. Patients who have borderline resectable pancreatic cancer, and just understanding what it’s like to go through treatment from that period of time where you have a diagnosis and a treatment plan that’s only partially finished, and you won’t find out for a couple of months whether or not you can have a resection that is potentially curative, and that if you are one of the people who gets to have your pancreatic cancer removed, that there’s a 20% chance that that is a durable cure, and an 80% chance that the cancer is coming back.

And so you have people with this huge degree of uncertainty. Personally, I’m a sarcoma surgeon. I take care of patients who have multiply recurrent disease. The cancer just keeps coming back, and we don’t know when, and we don’t know how quickly it will grow if we try to ignore it for a few months, and we don’t know if the chemo will help you have anything from. We could take it out and it will never come back, and you’ll live the rest of your life feeling great. We take it out and everything goes wrong, and you suffer, and you’re functionally disabled, and you end up not buying cabbage and go into the nursing home. And by the way, the cancer’s back when you get your next scan in three months. And to even try and hold all of that uncertainty as, even as a clinician, that is just wild. I don’t know how our patients manage that.

Eric 39:13

Yeah, I mean, just even the topic of uncertainty, too. I feel like we often don’t talk about it a whole lot, but even hearing from Cara and from James, like, everything we do has a component of some uncertainty, you know, with. With appropriate guards on that. Like, we do know that, like, everybody here will die. So there’s a certain degree of certainty is at some point we’re all gonna.

Liz 39:41

Die, and what really not me.

Eric 39:46

And, like, everybody on this call is probably gonna make it through this podcast. Like, we have degrees of stuff I.

Alex 39:53

Don’T know about me, see how I do on the end of that.

Eric 39:59

But I wonder how you actually end up talking about it. I’d love to talk each of you about it. When James is thinking about. We have this uncertainty about when somebody will need nursing home placement, but we’re actually fairly certain that they’re going to get to the point with dementia, that everybody’s going to have that need for nursing home level of care.

James 40:22

I mean, I think this is where a lot of the palliative care hope, worry kind of statements come in, in the sense that, like, I think we all hope that we will maintain our function for long periods of time and also worry that there may come a time that things do progress and we need more assistance. And so, like you mentioned, there, there’s always boundaries around uncertainty, but there’s some, some, we have a pretty good sense of how things are going to progress, and there’s always, not always framing it as a negative, but things thinking about what we can do proactively to help prepare for that is important.

Eric 41:00

How do you think about it, Cara?

Kara 41:02

Yeah, I feel like in the outpatient care setting, a lot of what we do is just supporting people in the face of what is hard, which is oftentimes the uncertainty, and just offering them a space to verbalize about that, to be heard, to have it normalized to a certain degree, that this just is really hard, that there’s no way around that, and to acknowledge that different people and different families might have this, hit them in different ways or have different needs, different needs for information about what’s to come, but also different needs to process those feelings.

And so offering to talk to family members separately if they have different needs than the patient, for example. And I’m just so grateful that we have our interdisciplinary team, including a social worker and spiritual care provider, to help do a lot of that support for patients and families.

Alex 41:56

Yeah, I feel like it’s incredibly important, as you all said, to acknowledge that uncertainty, because that’s where the patients and families are living. They’re in that emotional space of that difficulty of prognosis being so uncertain for various outcomes, for cure, for prolonged life, for disability, for whatever. And so we need to acknowledge it. And on the other hand, we can’t use it as a crutch and run from giving a prognosis because it’s inherently uncertain. And I’ve seen that danger as well, of people saying it’s, you know, who knows? You know, I don’t have a crush.

Eric 42:34

Doctors are terrible. We don’t have a crystal ball, you.

Alex 42:37

Know, who knows what it is? So you have to find that middle ground where you do have information to offer patients based on the experiences of other patients, based on your knowledge as a clinician, and that information is important to share with patients and their families, and based.

Kara 42:55

On that, one more time.

Alex 42:58

Good. Good. Plug.

Eric 43:00

Well, Liz, I want to end with you, because, again, you’re doing this daily with patients. Maybe not daily, but pretty close with patients with these sarcomas. How do you both acknowledge you actually do have a crystal ball? Like your experience, your knowledge about the procedure and evidence base, like, you have an idea of what to expect, but there’s a level of uncertainty of exactly what’s going to play out. How do you bring that up with patients?

Liz 43:29

So my feelings on this are rapidly evolving as I’m trying different things. I started by trying to hold that uncertainty with patients and just acknowledge, but I think I overshared, and I think it was. I was adding to the uncertainty, and I started to feel a little bit of. I was giving them more existential, barbie level suffering. And so I’ve started to use some of the palliative care skills of asking permission and trying to get a sense of who someone is and what they want. I’m more and more owning that at the end of the day, I have to make a decision of whether or not I think an operation is right and helpful to this person.

And so I get to know what I can about who they are and what they want and package my recommendation in that and then talk about the other options. But I’ve started leading the conversation with, who are you and how can I help you? And then making a quick move to here’s what I think is the right next step, and tell me if this sounds like it’s right for you.

Alex 44:45

Mm hmm.

Eric 44:46

That’s also why I like, you know, talking with surgeons, too, because, like internal medicine doctors, we often like just to give a bunch of options, and then we get really upset if they choose the one that we think is really bad for them. But like surgeons, they’re much more like, no, like, we’re not going to offer this thing or, you shouldn’t get this thing. Much more on the paternalism, paternalism kind of standpoint as. As physicians. Does that feel right, Liz?

Liz 45:15

I’ll tell you, I think you’re completely right. And it feels so hard. And, like, I think there’s a part of me that wants to say, like, I’ll give you all the options and you pick, but that’s just, it’s unfair to the patients and unreasonable for me to expect that someone can understand how to, like, manage their third recurrence of their differentiated lifeless or, like, they can’t even pronounce it sometimes. And so I think my job is to hold that uncertainty for them and with them, which is incredibly challenging. And I agree. I’m so happy that we have social workers and spiritual care and all of those other people who can help patients with this.

Eric 46:00

Well, I want to acknowledge your time. I want to thank all three of you for joining us today. Before we end, I think Alex is gonna do a little bit more of the Barbie song. I forget the name of the Barbie song.

Alex 46:13

What was I made for?

Eric 46:14

How did you not see the Barbie musical movie, Alex?

Alex 46:19

I know I’m the only one, I guess. Is it good? Did you all see it? It’s great.

Kara 46:23

Oh, yeah.

Alex 46:24

Gotta see it. Gotta see it. I should have worn, what, pink today? Or baggy clothes like Billie Eilish, right?

Speaker 6 46:40

(singing)

Eric 47:48

Liz. James Carr, thanks for joining us on this podcast.

Kara 47:52

Thanks so much for having us.

Liz 47:54

Thank you.

Eric 47:55

And thank you to all of our listeners for joining us on this prognosis…

Alex 48:00

Super special ePrognosis.

Eric 48:05

The tie-in right there. Thank you.