 On behalf of the McLean Center and the Center for Health and the Social Sciences, known as Chess, and the Bucksbaum Institute, David Meltzer and I, David, raise your hand. Oh yeah. Are delighted to welcome you to today's lecture in our 19, 19, 20 lecture series on the present and future of the doctor-patient relationship. Before I introduce our distinguished speaker today, I'd just like to call the group's attention to the 31st Annual McLean Conference that will take place this Friday and Saturday at the Law School on 60th Street across the Midway. The conference, as many of you know, is free and open to the public and will include presentations by local and nationally renowned ethics leaders who will address topics including health policy, the doctor-patient relationship, healthcare disparities, medical decision-making, end-of-life care, pediatric ethics and surgical ethics. When you leave the lecture today, you'll find a copy of the conference program on the table at the back of the room. It's now my great pleasure to introduce our speaker today, Dr. Doug White. Dr. White is vice chair and professor of critical care medicine at the University of Pittsburgh School of Medicine. He holds the University of Pittsburgh's medical centers in doubt chair for ethics in critical care medicine. He also directs the university's program on ethics in decision-making in critical illness. After receiving his medical degree from the University of California in San Francisco, Doug stayed on to complete his residency there in internal medicine and later a fellowship in their outstanding pulmonary and critical care program at UCSF. During his time in San Francisco, Doug also completed a master's degree in epidemiology and biostatistics as well as a fellowship in bioethics. Dr. White is an expert on end-of-life decision-making for patients with acute life-threatening illnesses. His great research has been continuously funded by the NIH since 2005 and has also received funding from many other foundations, including the Greenwall Foundation, the Gordon and Betty Moore Foundation and the Beckwith Institute. Dr. White has published widely on the complex topics of surrogate decision-making in intensive care units. In conducting this work, Doug collaborates with a multidisciplinary group of investigators, including faculty with expertise in bioethics, law, philosophy, sociology, biostatistics and health services research. His empirical research program has had two central aims. First, to identify factors that adversely affect surrogate decision-making for the critically ill patient. And second, to develop and test interventions to improve surrogate decision-making in the ICU. Dr. White is perhaps the leading scholar in the field that explores the ethical challenges that arise when clinicians have to work with surrogates for unrepresented patients. Dr. White's talk today is entitled The Doctor-Patient Relationship in the ICU When Patients Are Incapacitated. Please join me in giving a warm welcome to Dr. Doug White. Thank you, Mark. Appreciate that. Yes. Everybody can hear me okay? Good. Okay, well, it's really nice to see familiar and friendly faces in the audience. Mark, thank you for that lovely introduction and thank you very much for having me. It's a real honor to be here. Mark forgot to mention that I'm a downstater. So thank you for letting a downstater up into Chicago to talk to you. I was born and raised in Springfield, Illinois. So as Mark said, well, actually, let me just, by way of disclosures, this is my grant funding. I don't have any commercial conflicts of interest beyond the research funding. So, as Mark said, the topic today is centered conceptually on the doctor-patient relationship but when the patient is incapacitated and the locus of communication shifts from the doctor to the patient to the doctor and the family. And there are a variety of important aspects of that. What we'll focus today on the intensive care environment where many patients are incapacitated, but I do wanna bracket off and say that not all patients are incapacitated in ICU, so the doctor-patient relationship can and does exist in the intensive care unit above and beyond the doctor-surrogate relationship. And so when I was thinking about what topic to address with this fine crowd, one of the really important things that I like to talk about, I spend a lot of time thinking about, are the sort of tricky normative issues around surrogate decision-making in ICU's. So what is the moral authority of a surrogate decision-maker? What is the ethical motivation or what's ethically being achieved when we are asking surrogates to make substituted judgments? But as I looked at some of the really, I think, groundbreaking and important papers in the field, I realized that they're all by this guy who's sitting right there. So we're not gonna talk about the normative aspects today in great detail. I think because you have that already at your service in Dan. So instead, what I'd like to do is focus on, take an empirical turn that is normatively informed by the goals of surrogate decision-making in the conventional sense of advancing the patient's interest, manifesting respect for the patient as a person, et cetera. And what I'd like to do is to summarize empirically where things are going wrong with that process and intensive care units. And I'd also like to make the case that the way we've been thinking about how to improve outcomes is flawed. I'll mention more about that in a bit. And then finally, I'll present some clinical trial results that I think show a path forward for the field in terms of how we can strengthen the clinician family or surrogate relationship in ICU's and improve patient outcomes. All right, so for those of you who don't practice an intensive care unit, why are we even thinking about and talking about incapacity in ICU's? Well, there are a variety of reasons, but one is that a lot of Americans die in intensive care units. 15 years ago, the number was about one in five deaths occur in an ICU or shortly after the patient is discharged from an ICU. And more recently, Joan Tino and colleagues have redone these analyses and found that although the number of patients dying in hospitals is decreasing, the number of patients dying in ICU's is actually increasing. So if you're over 65, you have about a one in three chance of dying in an ICU. As we've talked about, patients are generally but not always too sick to make decisions about overall goals of care. So clinicians are turning to surrogate decision makers who are often but not always family members, sometimes close friends, sometimes legally appointed guardians who have no love relationship or close relationship with the patient. And over the last decade, a number of investigators have really looked into what is happening with communication processes in ICU's. And a series of studies have documented all kinds of breakdowns. So things like not having conversations with surrogates until it's time to withdraw life support or the clinic judges that it's time. Or when conversations do happen, shortcomings or failures to disclose, keep prognostic information, to elicit patients' values and preferences. And even thinks as simple as conveying that the option of comfort-focused care is a permissible option. And so all of these problems that I would consider breakdowns and care processes contribute to three interrelated health problems or public health problems. First, ICU care is frequently misaligned with patients' values. Critically old patients often receive more life-prolonging treatment than they would choose for themselves. This is both studies from 20, 25 years ago and as well as more recent studies. Second, surrogate decision makers in ICU's find the experience both distressing in the moment and then also are left often with long-term symptoms of depression and anxiety and PTSD that they temporarily relate to the experience. And then finally, the care that is provided in ICU's is incredibly expensive. The most recent estimates about costs are from 2005 and it was just under 1% of the US GDP goes to critical care, not hospital-based care, but care and intensive care units. So we can certainly disagree about what is the right allocation of healthcare dollars to critical care, but I think we can all agree that care that is delivered that is not concordant with patients' values and preferences is not a good allocation of healthcare dollars. So hopefully you're sitting there right now saying, well, why are we even talking about what's happening in the ICU? The work that should be happening is to keep people out of ICU's who shouldn't be in ICU's. And I want to acknowledge from the outset that I entirely agree that that is one very important stream of research to improve end-of-life care and end-of-life outcomes. And there are a variety of studies here, an IOM report, a National Academy of Medicine report and then a couple really valuable papers if you're interested in learning more about how people are now conceptualizing the advanced care planning process. So absolutely the case that we need better advanced care planning before patients cross the doors of the ICU. However, in reality, many patients would accept and would desire intensive treatment if a good outcome is possible. So certainly there are classes of patients, very, very end-stage cancer, very, very advanced Alzheimer's. The vast majority of them probably would not wish to receive intensive treatment, but that accounts for actually a very small proportion of patients in ICUs. And this is a really nice study by Teri Freed, now nearly 20 years old, in which she elicited from seriously ill patients, those who had limited life expectancy from cancer, CHF or COPD, their preferences for intensive treatment. And she did it in a really, I think, interesting way. That the line to focus on is the medium gray. I think this is before the New England Journal was published in color. So basically on the X-axis, you have the likelihood of an adverse outcome. And on the Y-axis, you have the number of people that were enrolled. And she gave these seriously ill, older adults treatment scenarios. And the one that's of interest here is the high burden scenario in which the likely outcome was death and they varied the probability of death. And they defined high burden treatment as I think two or three weeks in an ICU intubated on a ventilator and pressers and then a month, I think, in a skilled nursing facility. And they asked patients, would you accept that treatment if you had a 0% chance of an adverse outcome and the adverse outcome was death, a 10% chance, a 50% chance of death, an 80% chance of death and 100% chance of death? And what they found is that patients who have a 50, 50 chance, a 60% chance of dying, a 70% chance of dying, more than half of them would opt for intensive treatment in order to try to get through the acute critical illness. And what this study and others drive home for me is that there's absolutely a role for helping people avoid ICU admission if it's not something they would want. But there will also almost certainly continue to be a large cohort of patients who clinicians will have to help their surrogates transition from goals of care focused on cure or survival to goals of care focused on comfort. So what has been done to date? And what I'd like to do is take you through two or three studies that illustrate one conceptual approach to how investigators have been trying to improve surrogate decision-making and clinician surrogate conversations in ICUs. And then step back and talk a little bit about decision science and decision psychology as it relates to the surrogates experience in ICUs. And so I think a really valuable place to start is the support trial. Who here has heard of the support trial? Okay, good. The number is dwindling over the years. It's now 25 years old. This was front and center in my life when I was training. And I think has really, for many of the researchers in the field that it has been a seminal contribution and has really changed how people think about what needs to happen. So this was a 5,000 person cluster randomized trial that is by far the biggest RCT of anything related to end of life care. And at the time it was 5x bigger than anything that had happened. And it was an incredibly well done study by top health services researchers. And essentially they enrolled patients who many of whom were in ICUs, some were seriously ill and in the hospital but all had life limiting illnesses. And they randomized them to either usual care or what was called the support intervention. And the gist of the intervention is that physicians in the intervention group received from the support nurse estimates of the likelihood of six months survival, estimates of the likely outcomes of CPR and then estimates of the likelihood of functional disability at two months all from validated predictive models that the support investigators developed. And then the second part of the intervention was that the support nurse engaged with patients that they had capacity and surrogates of the patient's lack capacity to elicit the patient's values and preferences around life support and then shared this information with the physicians. So this intervention was very much about saying let's make sure that the doctors have good prognostic information or at least the best information we have, let's make sure that we get from families or patients what we know about their values and preferences and make sure that information gets to doctors. So conceptually that was what they were trying to do. And essentially what they found is that the intervention didn't impact any of the pre-specified outcomes of interest and it was kind of a funky set of outcome measures that they looked at but essentially no change in the time to changes in goals of care in these patients. No change in the proportion of care that was consistent with their code status. No change in the proportion of patients dying in an undesirable state or in pain and then no change in healthcare utilization. So this was the way I think of it is the most important negative trial in the field of end of life care and surrogate decision-making in the last 25 years. And this intervention, this is where we'll let's step back and talk just a little bit about theories of decision-making. This intervention was grounded in what's called expected utility theory but I think a simpler way for us to think about it is it's essentially the rational actor model. They assumed that giving people accurate information about the likely outcomes of treatment and giving good information about patient's perceived value of those outcomes will reliably lead to good decisions which is to say that the individuals making decisions will be logical, rational, and utility maximizing. That was the underlying premise of the intervention. And what that translates to, I think it maps on very nicely with how many clinicians think about what is required of them to foster good decisions either with patients or with surrogates which is something like this. I as the clinician need to make sure the family understands the patient's prognosis, that we've talked through their values and that they know which treatment options exist and are reasonable. That's a fairly standard view of, if the doctor's done that, that's pretty good. That's the standard view that was tested in the support trial and was essentially negative. You might say, well, listen, they gave the prognostic information to the doctors but it turned out that many of the doctors didn't share that prognostic information with family. So maybe this model, the standard view of what's required for good surrogate decision-making, maybe it's not that the model's wrong. It was just a ham-handed application of it. Yes? How are more information given when they realize they have a non-zero probability to survive before they want to? Yeah, well, so I think that the concern, the existing literature suggests that when you actually do either look back with families or you look at what patients say they want and what they get, that they're getting more than what they, at least at last preference documentation indicated that they wanted. It's a very tricky problem that you're referring to is that at the time that care is delivered in capacity to patients, there's no way to wake them up and ascertain now that you are really literally experiencing this thing, have your preferences changed. You're thinking of the decision-makers preferences which function on the family. Yes. I think the idea that the implicit way to describe it is that an irrational person would do less, but it's possible a rational person would do less. That was their assumption, and this was in 1995 when I think there was a large social view that over-treatment was rampant at the end of life. So that was their a priori hypothesis. There's rampant over-treatment. Yeah, entirely possible, but let's see. All right, so that was study number one in the rational actor of Vane. Study number two in the rational actor of Vane is one that we conducted, Chris Cox was the lead, he's at Duke, Shannon Carson at UNC. We developed using modern standards for decision aids, a decision aid for surrogate decision-makers patients receiving prolonged mechanical ventilation in ICUs, which is more than seven days of mechanical ventilation, in 13 different ICUs, medical ICUs, surgical ICUs, and neurologic ICUs. And essentially this was a web-based tool that on day seven in the ICU before or in preparation for conversations with doctors about whether to proceed with long-term mechanical ventilation, including tracheostomy, surrogate decision-makers viewed and worked through a decision aid that took about 35 minutes to complete. And it had all of the standard components of a modern decision aid. So it focused on making sure that the surrogate decision-makers had a clear understanding of what the patient's medical condition was in terms of prolonged mechanical ventilation and what are the likely outcomes of treatment. So again, in the decision aid prognostic information was given based on model-derived estimates. And then there was a section that laid out different possible treatment pathways. Full intensive treatment, comfort-focused treatment and a time-limited trial of intensive treatment and allowed surrogate decision-makers first to understand what the options are and then work through what's called the values clarification exercise to help them think through what might be the patient's values as it relates to these three treatment options. And then finally the last part of the decision aid was a variety of tips about how to communicate with the clinicians about this information. A summary sheet of the decision aid output was printed out, was given to the families, was given to the clinicians and then family meetings occurred. So this intervention overcame the concern that, oh, the families just aren't getting prognostic information and they're not being given the opportunity to really internally deliberate about the patient's values and preferences. And unfortunately, again, an entirely negative trial. The intervention had no effect on decisions to transition to comfort-focused care, no effect on healthcare utilization, no effect on surrogate's psychological outcomes, depression, PTSD and anxiety, measured at six months. And then also, concerningly, no effect on the accuracy of surrogate decision-makers' prognostic expectations. So giving them really clear risk information did not change their individual perceptions of their loved one's prognosis. And I think what was really nice about this study is that we had a number of secondary and process outcomes and one of the things that we looked at that was, I think, quite telling is that surrogate decision-makers, even after looking at the decision aid, were very optimistic compared to one-year survival models about the patient's likely survival to a year. The median estimate of these surrogates was that 90% chance of survival to a year when the model on average was a 56% chance of survival. So again, another what I would think of as information-focused intervention, foundering in the intensive care population. And then finally, one more study that I think is worth talking about briefly because it's a study that I think is often miscited or misunderstood in the field. So the title is effective, a palliative care-led meetings or palliative care-led meetings for families and patients with chronic critical illness. And what I think many people take from that title is, ah, this is an intervention that was a palliative care consult. And even some of the press releases from JAMIS framed this as palliative care in intensive care units. This was not a palliative care intervention. Again, a very similar cohort of patients, patients with chronic critical illness being on the vent for at least seven days. And the intervention arm compared to usual care, the intervention recipients received either one or two family meetings that were led by a palliative care physician. But that individual had to do a very prescribed thing in these meetings, which was disclosed to surrogates, model-derived prognostic estimates about survival, and then engage them in a conversation about what is the right course of action based on that prognostic information. And that had to happen on their first meeting with them. So that, if you think about, is that similar or different to what your palliative care colleagues do, I think many of you would look at that and say, that's really not what our palliative care docs do. They often start by getting to know the family, by establishing a rapport, meeting the family where they are, and then kind of walking with them through the experience, rather than starting with prognostic information. And the intervention, again, it's up here, had no change, essentially another negative trial. No change in treatment intensity, goal concordance of treatment, or quality of communication. But it did yield a concerning signal for increased PTSD symptoms among surrogates at six months. So yeah, you think about what may be going on here is what we think of as inflicted insight, right? They made more stark the reality of what the patient was going through without coming in with the suite of psychosocial support resources that may be needed to help surrogates through that experience. All right, so this feels like a situation in which we, myself included, have been barking up the wrong tree. You may think that we're fools for pursuing this approach, but I would just point out that there have been 105 studies with more than 30,000 participants of decision aids, which are very informational approaches to improving decision support, that have had, on average, a fairly important effect on a variety of outcome measures. And I'll just read this to you across a wide variety of decision contexts. People exposed to decision aids are better informed, clearer about their values, and have more accurate risk perceptions. There's growing evidence that decision aids may improve values congruent choices. So the question I have for you is why does this Cochrane review suggest they're so beneficial in some contexts and not in the ICU context? And I think to be fair, these are the kinds of decisions, to be fair to you, these are the kinds of decisions that were tested and fit into the Cochrane review. PSA testing for prostate cancer screening, colorectal cancer screening, and then different decisions about which birth control method is most appropriate. So. So if you can write this all, but that small chance just overpowers their rational thoughts. Okay. And the same thing in the left, like that small chance that you hear that your loved one may stay with you and have a positive outcome, they recognize the numbers, it's not a number issue. Yeah. So something overpowering rational thought. Any other comments or thoughts? Yeah. Very good. But she's still talking in big pictures and trying to let your family push it back. She let go. And I was the one that said, I know you'd like to be at your grandar's graduation in three years, but we need to think about what you want to do in the next 30 days. Mm-hmm. Mm-hmm. I just started looking like, who are you? I'm looking at, I don't know, I see you guys. But it's, I just, I feel like that the team, when the patient pushes back. Yeah. And when they're pushing back, I think it's really important just to unpack what we mean by pushing back. It's almost never that they're saying, I don't understand what you're saying. It's not, that's not the issue. It's not about rationality being challenged here. It's- We believe in God. Yes. We believe in one person. Right. You know, you told us that in the land of pediatrics, you started telling us this is a big deal and you're still alive. Yeah. Exactly. I think we're all on the same page here that there's something different in important ways about the kinds of decisions that decision aids seem to be valuable to address and something else that is going on in the care of patients facing death and their families are being asked to make decisions. And this is a series of three papers that I think just nicely get at what is going on. And we should just, I think, be explicit about it. What we're talking about here is that there is an extremely prominent emotional and psychological overlay to these decisions that is not well attended to in the traditional conceptions of the rational actor. That's the simplest way to say it, but I think it's different between the two. Yeah, that was it. So, drastically different than the decisions you're making in potentially so often in the fight, that the idea that you could take an instrument that doesn't work because people are humans. They trust, I think they trust other humans, but the problem is that the ICU doctor is just right here in the house. And as well as all the other people there and the big part of this is just the lack of trust that people are actually a telling of the truth that the actually looking out for the best possible outcome to their loved ones. I think that really marks, I think, the relationship of a lot of people with their health care systems. Yeah, well said. Well said. So as you're talking and we're talking, what I hope that you're constructing in your mind is a vision that there is certainly at the core of the surrogate physician relationship, there's something about information transfer that is probably necessary but not sufficient to ensure that good decisions are being made and that around that core or around that component, there are other components that also need to be attended to. The emotional distress that surrogates are experiencing. The psychological guilt, psychology and the guilt that they may be experiencing. And then just problems with healthcare delivery in terms of conversations not happening. So let me go on a little bit and just very slightly delve into what we know about the impact of strong emotions, particularly fear and anxiety on how people deliberate. And this is a great review by Michael Pham on the topic. I'll just read it to you so you can have the gist of what the science says now. Intense emotional states, such as anxiety, produce deficits in people's reasoning abilities. People have an intense emotional, anxious states have lower ability to recall information and organize this information in their memory. They scan alternatives in a more haphazard fashion which means they're less able to carefully deliberate and choose between existing options. And then they process persuasion arguments less thoroughly. Well, what on earth does this mean? Well, persuasion arguments are essentially what the physician and other clinicians does in family meetings when he or she is trying to talk with patients or families about what the best course of action is. And so this very strong body of evidence from the psychological sciences that we really have not been paying much attention to is pointing us, I think, in a substantially different direction than expected utility theory and the rational actor theory. And so when I teach my fellows in the intensive care unit about this, I use what I think of as a sort of a really dumbed down model of what FMRO studies of people in intense, who've been induced to be an intensely negative emotional states, how they deliberate and what is popping up on their FMRIs. And so you may not know this, but there is a whole genre of research happening in the psychological sciences in which they put individuals in scanners, in FMRI scanners, and then expose them to stimuli that have been proven to induce fear and anxiety. And so horrifying images, battle scenes, a host of things that are quite disturbing. And then once these individuals are in what we think of as a hot state, they put them through an FMRI scanner and they see what parts of their brain light up. And lo and behold, what they typically find is that when people are in hot states, their cortex is quite quiet and their amygdala is lighting up. And remember that the amygdala is the fear center, the fight or flight center of the brain. And then when they look at how people are actually making decisions and what attributes of their decision-making processes they observe, essentially what they see is that people are in these hot states are really not able to deliberate well. They tend to make black or white decisions and are insensitive to trade-offs between options. And when they study these same people in a cool state, put them back in the scanner, play nice, happy music, whatever is there, their stimulus to get them to a cool state and rerun them in the FMRI scanner, what they see is that the cortex is lighted up, the amygdala is quiet and the way they make decisions is quite different. It's much more along the lines of how we might think someone would behave as a rational actor. So they are able to be deliberative and appreciate trade-offs between options. They're able to weigh things in terms of risks and benefits. And so what I, this is, I think, a really scaled down way to think about it, but what I think a good physician does with surrogates and ICUs is either implicitly or explicitly is working to move them from a hot state to a cool state. And I would just say that when we think about, okay, information exchange is a key component of things, I would argue that everything that you just raised is sort of what makes surrogate decision-making hard can also be thought of through the lens of surrogate decision-makers being in hot states in ICUs and needing help to move them from an intense fear or anxiety state to something cooler. So there have been a few recent studies looking at this and I wanna take a bit of time and walk you through several of them. The first is a study by Randy Curtis in which he developed an intervention that he called the Communication Facilitator Intervention that essentially involved adding to the experience of the surrogate decision-makers a family support and mediation theory trained nurse or social worker who every day was touching base with the family and delivering tailored emotional support based on their coping styles. They filled out a coping styles inventory at the beginning. And then this individual also facilitated with communication with the team and not only ensure that good communication was happening but also intervene to identify and address conflicts before they really went from nascent to intractable conflicts. And essentially this intervention, they assessed three different types of outcomes. First, in terms of the surrogate's psychological distress, what they found is that the intervention substantially reduced surrogate's depression symptoms at six months here. It did not affect anxiety at three or six months and there was a trend towards decreases in their PTSD symptoms at six months. So something about this more psychologically attuned intervention did actually have a signal in terms of their psychological outcomes. When they looked at healthcare utilization, they also found a significant effect that was largely restricted to length of stay among decedents. So they looked at how long a patient was in an ICU before they died. They found a substantial reduction in length of stay among these individuals that was mediated by faster decisions to withdraw life support. They also found that decedents in the intervention arm compared to the control arm had substantially lower healthcare costs. So I think I would view this as at least the beginnings of proof of principle that intervening in ways that are not explicitly tied to information transfer may be a promising path forward for improving the surrogate experience in ICU's. So one story is that the presence of the psychologist who leaves the anxiety in some sense. The other possibility is that this person is infected in getting the family to make a decision to withdraw and because they make a decision to withdraw, they're spare all this traumatic stuff that's gonna happen in your words. Yeah. So it would also be super interesting to look at the people who kind of didn't change their decisions and understand whether it's alleviating negative outcomes even if the people decide to go all the way. Right, so yes, we'd have to restrict the, we'd have something to have to figure out what is your decision now and then determine whether you changed your decision over time. If nothing else, you know, full code all the way. Yeah. Yes. Is it better for that? Yeah. It might not be. Yeah. So one thing I will say is remember that this intervention did not change mortality rates. So the delta was whatever number of days sooner the decision happened. So you'd have to say that the active ingredient according to this theory is that there was just four or five fewer days that they were on a ventilator and the family was exposed to that trauma repeatedly. But potentially, Dan and then Mark. So I think forms of interventions help the family with their decision and the pool and how that's the decision that's really theirs to make. That's how they can be a little bit more at peace with us than they have. But it was also, it was me that what they were going towards in the park is how to tell what to do. And that the intervention functions that but as I'm trying to, this is the right thing to do these things. But they have such a lot of benefits but wouldn't count as facilitating their decision. Right. Yeah. Yes. Well, I think if the interventionists followed the protocol, these were non, each of the interventions, this one in particular but the other one I'll show you that we conducted were non directive interventions. So from a substantive standpoint the interventionists were trained to be non directive, to not make any normative statements about what should be happening. From a procedural standpoint, they were certainly directive in that the whole hope was to help surrogates have a framework to be thinking about how to make these decisions. And the framework was clearly sort of in general terms how we think decisions should be made for incapacitated patients, reflective of the patient's values duly considering their prognosis and options. That was the intention of the intervention. Success consists of family doing things with a team. Yeah, I guess I don't know why the addition of an interventionist would change anything about families between, across arms, anything about their interest in pleasing one team versus another. I'm just not, I'm not seeing that as a sort of a plausible mechanism. Mark. In many years before 2016, on this paper, one of our former fellows who worked for a community hospital in Chicago asked the president of the hospital if he could set up an ethics program. The president said he could have a six month trial to see if an ethics program would increase or decrease the cost of care. Dr. Meltzer and I wrote a paper with our former fellow as the lead author in which we showed essentially what you're showing in this 2016 paper. That in fact, ethics consultations substantially decrease patients' lengths of stay and cost in the hospital and generated for the hospital and what the president regarded as an adequate increased return to enable an ethics program to be established per. Yeah. So for this, David, I think it wouldn't have been 2000. Long time. Maybe even older than that. Maybe even older than that. Yeah, and so absolutely, a slide that I pulled just to make sure that we didn't go over was Larry Schneiderman's JAMA piece, a randomized trial of early ethics consultation which is a nice use, which only showed, the only outcomes they assessed were healthcare utilization. So I think only a partial picture of what's important to all of us, but they similarly showed a decrease in length of stay and hospitalization costs with ethics consultation. So what's important here is to ask yourself, so what's the ethics consultant doing? Is the active ingredient of an ethics consultation that you guys are that much better about giving crisp, clear prognostic information or is there something more going on? I would argue that there's something, if ethics consultation is done in a facilitative way that is longitudinal, that there's something more going on attending to the emotional and psychological experiences of the surrogate decision maker rather than just pushing towards a sort of a standard informed consent model. Well, without a question, and better interactions with the family and the immediate surrogates for the patient, yes. David, you're either suppressing a thought or... I'm suppressing lots of thoughts. What's the role of the ICU doctor in this, and by the way, this is among ICU doctors and their ability to sort of calm a situation. And then the other thing that I was thinking is, what's the role of the other members of the patient's care team? So in our work, we have the same doctor caring for patients in and out of the hospital. And when the patient's hospitalized, they may not be there, ICU attending. We have closed ICU, but they are physically there and much more available. One of the interests, and we're finding big decreases in expenditures and hospitalization rates, and we're outside here looking at end-of-life care around this, but one of the things that's a really interesting kid's result in that study is that one of the patient reported outcomes that really more than any others was patients telling us that they feel, what's the term, calm more of the time. And I think there's this idea that having this doctor who knows you, takes away the anxiety and maybe that feeling of a process that you're describing, which makes it easier. The other thing we do is we engage a lot with the family, even in advance of these end-of-life things. So that is your, for that is your role. Yeah, so Dave, thank you for that. That's a nice segue to one thing I would like to do. I'm gonna skip this and talk about just one more study and then we'll have some time to chat. And before that, so each of these interventions that I just talked about essentially relied on the external interventionist model, which is to say we bring in someone else to do the work that we think the clinical team is not doing well, whether it be an ethics consultant or a palliative care consultant or a nurse trained in mediation theory. And that's certainly one approach and maybe that's the best we can do, but there are certain risks of that. And David introduced me five or six years ago to the term coordination costs. And one of the problems with these models is that they're quite complex and you're trying to add one more person to an already complex ICU team. Problems come up with handoffs or role responsibilities. But really for me the concern as a clinician in an intensive care unit is I'm not sure that in the long term it will improve outcomes if clinicians in ICUs are outsourcing these conversations to other people. I absolutely think there are circumstances in which the conversations are so complex or the issues are so complex that consultants are needed. But I also think it's worth at least trying to see if we can have interventions that focus on the capacity of the clinicians in ICUs and see if we can improve outcomes that way. So with that, this is a study that does just that and I wanted to spend the last 10 or 15 minutes talking to you about this study that was published just this last year. A randomized trial of a family support intervention in intensive care units. This was all conducted in the UPMC health system in five different ICUs and was funded interestingly by UPMC. They came to me and said we feel like there are problems in our ICUs. We don't know what a good fix is but we are really willing to invest in trying to find solutions. So this is I think a nice example of a partnership between a health system and the academic side of the medical school. So the research questions that we sought to answer should look pretty familiar compared to usual care. Does a protocolized family support intervention that is delivered by the existing interprofessional team. So the doctors and nurses in the ICU decrease surrogates long-term psychological distress, improve the quality of decision making in clinician family communication and decrease the duration of intensive treatment among patients who ultimately don't survive. So this suite of outcome measures I think is an important way of getting at surrogate decision makers outcomes, patient level outcomes and then healthcare utilization outcomes. Each of which I think needs to be looked at if you're trying to fully appraise an intervention. We conducted what's called a stepped wedge randomized controlled trial among 1,400 patients in these five ICU's and I'll show you the sites in a second but for those of you who don't know what a step wedge trial is this is, it's an approach in which rather than randomizing individual patients you randomize units in our case it was ICU's to transition from a usual care phase to an intervention phase. And so this is ICU number one they start in the usual care phase and by the randomization order they step across to the intervention phase and the intervention is called the partner intervention. And so by the end of the study every ICU in a stepped fashion has moved from usual care to intervention. And this research method or clinical trial design is often helpful when an institution for example or when individual ICU's say we're unwilling to be in usual care for four years and that was actually the case for this trial. No ICU was willing to not essentially not do anything to improve end of life care in their ICU's. So rather than doing an intervention where we randomized individual ICU's to stay as usual care or intervention for the whole trial period this approach allowed every ICU to get the intervention by the end. The patients that we included lacked decision-making capacity and had at least they were either on a ventilator for 96 hours or the physician in charge predicted that they had at least a 40% risk of death or a 40% risk of severe long-term functional impairment. So these were just screening criteria to try to get at the sicker cohort in the ICU. These were the different ICU's in which we conducted the trial and you'll see that there was a neuro ICU, med surge ICU's a surgical transplant ICU and a medical ICU. And the intervention and I apologize for the tortured acronym is called the partner intervention pairing re-engineered ICU teams with nurse-driven emotional support and relationship building. And so this is a multi-component intervention as I said that's delivered by the ICU team. So we didn't bring in any external people instead we essentially re-trained and increased the skills of the existing nurses in the ICU. We didn't intervene on the docs at all except to orient them to the intervention. The intervention involved instituting a protocolized approach to family support that I'll show you on the next slide giving the nurses advanced communication skills training to deploy the intervention. And then finally from a healthcare delivery standpoint intensive implementation support in each ICU. And so what do I mean by each of those? Well, let me just say one more thing. We explicitly grounded the intervention in what's called the cognitive emotional decision making framework that very much is consistent with what we've been talking about today that good decisions arise out of both cognitive considerations and emotional and psychological considerations and that we need to attend to both in order to help surrogates in ICUs. So you're not meant to need to see what the individual components of the intervention are but essentially the family support pathway started on the patient's ICU admission and continued through their entire ICU stay and essentially involved daily encounters between the partner nurses and the family that were largely about attending to emotional distress and identifying nascent conflicts. So that was every day the partner nurse was checking in and then within 48 hours and then every five to seven days thereafter the partner nurse took responsibility to make sure that family meetings with the full clinical team were happening. And before each of those meetings so before the formal family meetings the partner nurse sat down with the family and essentially went through what's called a question prompt list a structured way to help them think through what are their main questions and concerns for the doctors and this is just a screenshot of it but it's organized according to the kinds of questions that family members often have about what's wrong with my loved one what happens to most people as sick as my loved one what are the different treatment options how can I tell if my loved one's getting better there's space where they can write in their own questions and then the other thing at the end of this the partner nurse talked with the family about the questions the doctors might be asking them and the kind of questions largely had to do with beginning the value solicitation process the doctors are likely to ask you has your loved one ever talked about what's important to them what kind of treatments they would want if they were too sick to speak for themselves and so this was the family support intervention and the rest of the intervention that I'm describing is really about how to operationalize that using existing clinical personnel and working it into the ICU's workflow. As I said it was 12 hours of in-person training for the nurses for four to six in each ICU they were certified at the end of it so they had to actually pass a test in which they interacted with simulated patients and demonstrate that they had acquired the skills and then in terms of implementation support this is a big lift to get ICU's to really change how they communicate with families and more importantly to get physicians from diverse backgrounds to say it's okay for nurses to take a much bigger role in how we support families and so a big part of the work that we did was working with hospital leadership to start with top down buy-in, engaging ICU directors, engaging champions in each unit on both the physician side and the nursing side and then spending lots of time in the ICU's helping the teams develop and implement the care pathways. The main outcome measures we've talked about sorry, it gets psychological distress. Two validated measures of the quality of communication around end of life care and then what's called the modified patient perceived patient centeredness of care scale which measures the degree to which care processes were reflective of the patient's values and preferences and then finally healthcare utilization outcomes. So here's what we found. This is just the sort of the skimmed slim down concert diagram 1400 patients enrolled. We had good long-term follow-up 72 and 74% which in ICU cohorts many of the families are bereaved and it's actually quite hard to retain them. So this is about as high as you can get in this cohort. The baseline characteristics of the patient there's nothing really interesting here except to point out that this was a very sick group. So 33% died in the hospital, almost 60% were dead by six months and 2% were living independently at six months. So even among the survivors, very few of them were living independently. A relatively small proportion were actually living at home at all. So outcome measures in each of these three groups. So first surrogate psychological distress that's measured at six months with the hospital anxiety and depression scale and then PTSD with the impact of events scale. No different, no effect on surrogate's long-term psychological distress. So these levels are actually quite high in both groups and the intervention did not lessen the long-term symptoms that these individuals had. However, every other outcome measure was positive in favor of the intervention. So surrogate decision makers ratings of the quality of communication was substantially higher in the intervention arm compared to the control arm that is about a six point difference and another way to look at it to say what's the highest score category that greater than 80% is considered the high range. There was a 10% increase from 30 to 41% of the patients in the intervention arm who were in the high range and then also a significant improvement in the patient centeredness of care scale and this is actually kind of a wonky scale but lower numbers are better on this scale. And then the last bit of the outcome measures of interest here were healthcare utilization and costs. So among survivors, there was no difference in length of stay among decedents. There was a two and a half day reduction in length of stay in the ICU. Overall, there was about a three day reduction in hospital length of stay, which corresponded to decrease in total and direct variable hospitalization costs of about $6,000 for total and just over $2,000 for direct variable. And then I think the other important part of doing an intervention like this is quantifying the costs of the intervention and what we found when we, I guess, inclusive of the nurse's time, the training costs and then the cost of all the implementation support throughout the trial, it was about $170 per patient. So a fairly large cost savings and cost reduction that came in parallel with perceptions that the patient centeredness of care improved and the quality of communication improved. So these are the, this is the summary of the trial results. Maybe I'll stop here before we finish up and take questions. Yeah. Thank you, sir. President, I want to ask one of you finished before. So you mentioned that when you talk about writing this trial that you'll even hear, what are the uses that we have with the notion versus something to extend that, how decisions should be made. So I assume that you can perhaps even better hear some talk about substitute judgment type ideas within surrogate decision making. But that also comes with a sort of baggage where you think the normal decision making processes or what the, in this case is a rational actor, but I think of surrogate decision makers as more culpability, more minimalism, right? So when you put psychosocial, although I'd like to change it, when you put psychosocial support, you're trying to help them either come to terms as someone's just earlier or here, trying to suggest that this is the optimal way to think about it. So from that trial and out here, how do you think the normative model surrogate decision making, or what that surrogate decision maker is, your decision's different? So that's what I wanted you to get out of the answer. This trial, my thing's interesting is that, again, and we always have to do this, right? There was sort of justifying cost effectiveness here, but it's just the fact that they felt, because we talked to them more, that they're burdened, right? Of more culpability was plus, right? That what you're doing by these interventions is actually helping them, not for the social measures, but they try to minimize their more culpability, then by the fact that these people are all sharing this decision, or they're talking about this decision, that I've lessened my burden of that, and therefore it's okay to do, again, what we think is right. That's entirely possible. I mean, mechanistically, it's very hard to unpack what are the either psychological causal mechanisms or emotional or otherwise about how this intervention worked. So I think your question about does this change my view or does it really inform the normative model? I don't think it really informs the normative model at all. We didn't, remember, this is a very slimmed down intervention to see can we focus on the interprofessional team trying to emotionally support families? We didn't train the nurses to go in depth in terms of talking about principles of surrogate decision making. Instead, we trained them to speak in very general terms about the goals, which largely, the way we asked them to talk about it, was to make decisions that are respectful of the patient as a person. So that's a very general thing, and we left it to the physicians to continue to do whatever counseling they typically do. So I think that's an inherent limitation with this kind of intervention is that it's not nearly as intensive or controlled as we might hope if we really wanted to get in there and begin to play with different normative models and what are the effects of framing the surrogate's role in this way or this way. Those interventions are very hard to scale. That's the issue, and that's why we wanted to try something that leveraged in a relatively scalable way the interprofessional team. But make sure you really understand it and think about how it relates to the team. Just think, heard it, I think, I'm very sure training the nurses, right? And that the doctors really weren't being treated. Correct. So, but obviously they knew what was going on, and there's something going on in the team as well. And I guess, I wonder if you could just tell each and every group, was it that the way the nurses were talking to the patients was sort of calming them, calming down their families, and helping them in a fairer place. Sure, what kind of doctors were doing something different because they knew this was happening. Yeah, let me tell you from... Can I ask a question? Sure, yeah. Did you study the nurses? In this trial, we did not. This was a very thinly funded trial, so we had to be quite parsimonious with our outcome assessment. So, David, again, what would have been valuable, we did not have the funding to do, was a really careful, qualitative process evaluation of actually the kind of things that you're talking about. You know, what really was changing on the ground? What were the experiences? But informally, the feedback that we got from both physicians and nurses were things like the doctor saying, you know, the family is just, they are more prepared for these conversations, and the conversations go better. I think to some of the doctors that was code for the conversations were faster and less conflict laden, it's really hard to tell in that regard. But the other thing that this intervention really was about was making sure that conversations were happening. You know, the old model of how communication plays out in ICU is essentially that the doctor decides when to have a family meeting. And this intervention flipped that and established a new default, which is family meetings are happening within 48 hours and then every five days thereafter. So that's the other thing the doctors noticed was, oh, you know, I'm meeting with these families in ways that I didn't used to meet with them in terms of, I'm talking to them when there's really no decision to be made. With both the training of the nurses and the schedule. Yes, that is the family support protocol, which is daily encounters between the nurse and then scheduling family meetings according to a preset interval. And doing a little, a small dose of decision support in the form of a question prompt list before the clinician family conversations. What are the nurses who work for the particular family? The same nurse that stays with them or did the nurses change? Yeah, so that's a great question. And it was something that was very eye-opening about this. The intention was that to the extent possible it would be the same nurse every day to have continuity in terms of the relationship. That turns out to be very hard to do with nurse-based staffing. Even though the nurses that we were training tended to be more senior nurses who were usually there during the day. So we had to really adjust the intervention at the very beginning to make sure that nurses were also trained in how to do handoffs from one partner nurse to the next day after day. That each ICU did a nice job of figuring out, given their own resources, how can we maximize the continuity from patient to patient or within a patient? But we couldn't guarantee it. So that was part of the work too, is how do you effectively do a handoff from one partner nurse to the next? And there are several other papers to be written about how each ICU took this intervention and tailored it to be feasible in their ICU. Yes? One of the outcomes that you reported is that the 2% of people with the intent of the response had this really question the intervention of intensive care as a therapeutic intervention. Thank you. And all this energy being placed on getting people out of the intensive care unit. I mean, I know you said putting not putting them in the first place, but I think the whole approach to how to take care of a very serious legal patient needs to be obviously reconsidered here. It's not just a matter of stopping it. It's a matter of what they're curing. Yeah. And I think the emphasis on people's values and goals is really misdirected. And I think most people would like intervention instead of the don't want intervention of bones. And I think not knowing what those are part of the problem, but also the incredible incentive to do this, these are the largest money-making units in the hospital. And so I think there's a real conflict here, especially with the hospital funding in terms of what is actually going on. Well, so I would respectfully push back on the notion that patients will take things that will help them and not take things that will hurt them. Having taken care of a lot of patients in ICUs who are frail and elderly, many of them say, even if this could help me, I don't want this. So that notion that it's too simplistic to talk about patients' values and preferences, I must just disagree with you. When you talk about helping them, is that really helpful? Well, that's where the conversation about values comes in. Defining what help means that they can see. Yeah, right. So then, yes, maybe I heard you, what I heard you say was conversations about values and preferences, they're not useful. I've been saying that I use thought. I think that they misconstrued what people actually are thinking. Because I think it's not a better that people have such drastically different values. I mean, some people want their heart just to be beating. That's good. But I think the vast majority of people really want to be active, involved, and living in the world. Again, this may be perhaps a separate conversation about plurality of views, about what constitutes a good life. Maybe that's a good conversation for afterwards. But again, my experience in working in a diverse hospital is that there is quite a lot of diversity around this issue about what is a life worth living? What is too much treatment to achieve certain outcomes versus not others? Well, so let me give you the counterfactual, which is 2% of people living independently doesn't tell you what's happening with other people. So there could be people with one ADL impairment living at home. They're not living at home independently. But they're still living a life that many would say, that's better than being dead. So if the question is, is intensive care doing something good, I'm not sure focusing on this 2% number is going to tell you the whole story. Because there are still people either living at home or living in an assisted living facility who, when you go talk to them, say, this is a life that I would accept. ICU management is changing across the country. There are two kinds of patients in ICU. One is that following operations, patients are there. And there are times that from injury, they come there. So some places they have ICU management. And there may be surgeons who are hardly involved in it. They're patient competent. And there are others that may be surgeons that are involved in the ICU and take care of their patient. So there is a lot of information gets lost. When you have taken care of a patient, after they've dealt with the patient, now goes to someone else. Do you know if there is any statistical evidence that makes any difference? In terms of the model of, I think what I hear you describe is a closed versus open ICU. Yeah, so the surgical issue is a tough nut. There are not great studies of whether open ICUs versus closed for surgical populations. To my knowledge, there are not studies that suggest that one is better than another. The studies in medical populations suggest that closed ICUs with what's called a mandatory intensivist staffing model are associated with better patient outcomes. That's the extent of the literature that I know on that topic. I was dealing with one ICU in another hospital, which is taking care of ICU persons. And the patients, I mean, the doctors said, we don't know what happened to our patient. And a lot of information is lost because when you examine the person, when you talk to the family, you prepare them for all these things, or communication, they can be lost after it goes to the ICU at the time that it's a serious management. Yeah, no, absolutely. And I think that the counterbalancing question is, does that same surgeon who did the procedure and is technically incredibly proficient with surgery, is he or she as proficient with intensive care medicine? And if not, then there's an empirical question about what's the right balance between those two specialties and improving outcomes. Yeah. Right. Thanks so much for that presentation. Do you have a sense informally from this study or from maybe some of the previous work where you taped conversations whether physicians actually do encourage surgeons to do substance to the judge? I mean, if I'm an ICU physician now, informally it doesn't seem like anybody actually does it in practice, physician encouraging the surgeons to do substance to the judge? So what's your sense of what's happening? So we actually did do one analysis on this topic. So Thomas Cunningham, who's a philosopher when he was a postdoc, worked in my lab. And we had a cohort of, I think, about 85 audio-recorded conversations in which the focus of the analysis was what normative guidance, if any, is the physician giving to surrogates during the process of deliberating. And so that was the question. We both looked to see, was there any normative guidance? And then if so, what type? And how often did it actually adhere to the sort of standard hierarchy of stated preferences substitute to judgment and best interest? And I don't have the exact numbers in front of me, but in, I think, more than half of cases there was no normative guidance about how to function as a surrogate. And then in the other half, it was very thin. And the kind of questions that may actually be enough were things like the doctor saying, what do you think your mom would want if she could magically sit up and see herself here? So that was just a bit. And then sometimes, very rarely did we see physicians appealing to other kinds of considerations that are not necessarily sort of tightly linked to the standard issues. For example, family regarding considerations. Some physicians said things like, you guys should make the decisions that you think are the right one for your mom without clarifying that the hope is that when you're making that evaluation, you're considering the patient's preferences, not just your own. So in essence, what we have found is that the actual normative guidance given to surrogates is quite thin in terms of how they should function in the role of surrogate decision maker. Before you think your intervention changed the normative guidance that was given by physicians? By physicians? I don't know. We did not audio record the conversation, so I would only be speculating. We know that in the pre-counseling period before each family meeting, the nurses were taught to make that general statement that your job as a surrogate is to make decisions that are respectful of the patient as a person. Thank you so much. Please join us, please. Thank you.