 So it is a pleasure for me to introduce today's speaker, our very own Micah Prohaska, who is an assistant professor of medicine in the section of hospital medicine, associate director of the McLean Center. And since 2022, co-director of the McLean Center, ethics consultation service along with Dr. Will Parker. So a few things about Dr. Prohaska, he obtained his BA in history, philosophy and social studies of medicine from the University of Chicago. He went on to Wright State University for medical school residency back here at the University of Chicago. In addition to completing a clinical ethics fellowship here at the McLean Center, Micah received a master's degree in public health science from the University of Chicago. Dr. Prohaska is an active clinical investigator who is supported by grants from the National Heart, Lung and Blood Institute to study how red blood cell transfusion for hospitalized patients with anemia affects their fatigue activity and fatigability levels after they've been discharged. He is a co-investigator of the University of Chicago Hospitalist Project Research Infrastructure and the University of Chicago Translational Medicine Program and a program called Cultivating Health and Aging Researchers by Integrating Science, Medicine and Aging, which is the Charisma Program that some of you may know about. And I don't know who comes up with the abbreviations but they're great and so well done. Congratulations. I give all credit to Vinny for every acronym. So. So the acronyms are great. So this program, the Charisma Program trains undergraduate students in clinical and translational research. So in addition to his leadership role in the consult service in the center, Micah has really played a major role in our efforts to develop a master's degree program in clinical ethics that we hope to launch in the near future. And really, you know, this was an idea for many years but has moved to a concrete proposal on Micah's efforts. So Micah, it's a pleasure to have you present today. Thanks Peter. Thanks Peter for the kind words. Thanks for everything that you've done for me as well. So I really appreciate it. Bader told me if you scan this CME code you lose credit for being here and listening to me today. So I'll let you figure that out whether that's true or not. There we go. So Peter talked about though some of the work I do studying anemia and red blood cell transfusion. When I tell the medical students, sometimes the residents, they say, you know, I really wasn't that interested in anemia as a resident, red blood cell transfusion as a resident. You know, and I probably had a little bit of the view that they often have, which is like, so what? Like who cares? Everybody's anemic. We have the treatment for it as transfusion. We know all how to do it. And it's kind of boring. But I hope to tell you a little bit of a story today over the past, a little bit less than the decade where I've been studying this that makes it clear that it's not so boring. There's actually some nuance and that along the way and the development of our scientific base and practice of treating anemia and hospitalized patients we've made like some key ethical missteps, missteps in clinical research, missteps in our application of intervention as physicians. Here's the objectives, let me hide this. So to understand the prevalence of anemia and hospitalized patients and the scientific foundation for restrictive transfusion practices and policies to recognize the ethical principles neglected in the development of this evidence base and then recognize adverse consequences and the dogmatic adherence to restrictive transfusion practices. All right, there we go. Okay, so the background is gonna be a little bit on anemia and transfusion or restrictive transfusion practices. I think I recognize everybody here and most people have some clinical background but so that we're all on this but actually the clinical background is very so the fact that we'll get on the same page with the background and then we're gonna talk about transfusion and the evidence base for it and the ethical issues and we're gonna break it down into the relevant scientific evidence, patients values and preferences and clinician judgment. Okay, so anemia is just a reduction in red blood cells. I bet I could have, we probably would have gotten 90% if I asked everyone here what it is and it's measured by a reduction in hemoglobin or hematocrit. So at the University of Chicago, we measure hemoglobin. Many of you have been to institutions where they have measured hematocrit that's how it's clinically measured. The World Health Organization defines hemoglobin anemia in females as a hemoglobin less than 12 or less than 13 in males. Anyone know where that came from? A 1960 report on iron deficiency anemia which the World Health Organization never meant as a standard for defining anemia overall but it sort of just carried forward and it became the definition of anemia. The key thing about anemia, the key physiologic thing that matters is it results in reduced oxygen delivery to the tissues. So hemoglobin carries oxygen, hemoglobin's carried by red blood cells, you reduce red blood cells, you reduce hemoglobin, you reduce oxygen delivery to the tissues. And anemia is extremely common in the world. So 25% of the population outside the US, the majority of the people have iron deficiency anemia, much less common in the US. So if you're less than 65, it's about 10%. But I'm a hospitalist, so I take care of hospitalized patients and in hospitalized patients it's somewhere between 40 and 70%. Nursing home residents is about 50%. And when you're younger, or when people are younger, typically females have a higher prevalence of anemia but as we age, men have a higher prevalence of anemia. So the key is it's common, very, very common and increases with age. Prognosis. So these are two studies, one by Colleton in blood in 2006 and another by Penix in the American Journal of Medicine in 2003, but I could have swiped these figures out for like dozens and dozens and dozens of articles that show anemia affects mortality and quality of life and functional status. So here, this left side is mortality. And you'll notice like, if you move from right to left, the moment in men or female, you start to have an abnormal hemoglobin, there's an association with increased mortality. And on this table on the right, having anemia compared to being borderline anemic is associated with a double the risk of odds of having reduced functional status. Okay, so associated with mortality, functional decline and decreased quality of life. I was hoping Mike Rossi was gonna be here today, although I don't see him in the office. Anybody know where this, oh, actually I think I wrote it up here so I can't even ask you this. So this is William Harvey's Dumate Cordis, published in 1628. When I was an undergrad here in HIPPS, you actually had to read Dumate Cordis, so I don't know if you still have to read it, but Harvey discovered that blood flows through the arteries in the veins. He did this a couple of ways, but the plate on the right shows that he put tourniquets on and he realized that the veins had valves and they became engorged. And after 1628, about 20 years later was the first blood transfusion. This book by Holly Tucker, if you like history, you will definitely love this book. Even if you just like a good novel, that's true. This is totally worth the read because it's real and it talks about actually the first blood transfusions and how people actually ended up in jail for murder. So after the French and the English government imprisoned the people for transfusing blood in the 17th century, went a couple hundred years without blood transfusions. And then in the early 20th century, transfusions took off again. So we discovered antigens and antibodies and we learned how to give blood in the battlefield. And since then we've been treating anemia with red blood cell transfusion in the hospitalized patients since the early 20th century. It's the primary treatment. It's also the most common inpatient procedure utilized across the US. So it happens all the time in this hospital, it happens all the time across this country. And historically, I told you that 12 and 13 were the hemoglobin levels that defined anemia. Historically, we'd transfused patients using this 10 and 30 rule. So this meant you had a hemoglobin of 10 or a hematocrit of 30, but that was like this soft rule. And you sort of would say, well, their hemoglobin's nine five, but I think they're gonna go to rehab. And so if they're going to rehab, I wanna give them a little bit more hemoglobin, increase the oxygen delivery to the tissues, or you know what their hemoglobin is 10.5, but it was 13 yesterday. And so I should give blood because that's a big loss. So people, the doctors were making assessments using patients' clinical characteristics. Also, it also mattered what the patient wanted. Okay, and so the idea was we wanna treat anemia with transfusion. We wanna figure out why you're anemic. But there was no evidence base for this. It really was a little bit the Wild West. Doctors were doing it, and no one could really even describe where this 10 and 30 rule came from. So that should begin to change in 1999. This is not the first trial, but this is the first big transfusion trial. So those of you that practice, definitely in adult medicine and provide transfusions, you've heard of this trick trial. It was run by Paul E. Baer. It was in the ICU out of, I think University of Toronto is where he's at. And they compared a liberal transfusion strategy to a restrictive transfusion strategy. So they randomized several thousand people to get blood at a higher hemoglobin level, this liberal strategy, or a lower hemoglobin level, the restrictive strategy. And their primary outcome was mortality. And they actually hypothesized that the people that got lower blood or less blood in the restrictive strategy were gonna die at a higher rate. So they were thinking oxygen delivery to the tissues, critical care patients, they need blood. If we don't give them the blood, they're gonna die. And it turned out the result of the study was that there was no difference. Okay, so this actually was very fascinating. I'm gonna say this a couple of times. This trial set a methodologic precedent that's been carried forward for over 20 years and how we think about studying red blood cell transfusions. So Paul published that in 99. There's a few trials that were before that, but that was the big one. And then from there, these transfusion trials comparing restrictive or liberal transfusion thresholds just took off. So this I think is from 2021. It's a Cochrane review by Jeff Carson. And they added up, I think there's 30 some trials here all looking at liberal versus restrictive transfusion. Anyone that knows anything about the Cochrane review we'll see on the right side that risk of bias. So Jeff has actually rated these fairly high quality studies. Okay, and down at the bottom I've highlighted in red a risk ratio of one. So the conclusion from the meta analysis or anybody who has had basic biostatistics is that it doesn't matter if you get blood at a higher or a lower hemoglobin threshold, your mortality is gonna be the same. I say restrictive or liberal transfusion, both of them are safe. I got into this for a variety of reasons because I was working with David as a resident, but these first practice guidelines in 2012 came out. And they were the first set of guidelines to pick up on this accumulating set of trial data and say, whoa, we didn't have any evidence base. We now have an evidence base. And the evidence base suggests restrictive transfusion should be the standard for all hospitalized patients. 2012 guidelines actually had a whole bunch of other recommendations in them. And in 2016, they updated the guidelines because there was additional trials and they became even more concrete. There's some nuance, there's some variation, but transfused at a hemoglobin of seven. And most recently, last fall, the 2023 guidelines came out. Same thing, describe the nuance. So the trial or the committee's actually like fairly smart people, they can read the data. But the key here was that the key message was transfused at a hemoglobin of seven. So, and the nuance and the guidelines was completely lost. Busy clinicians saw the guidelines and they saw that primary recommendation to transfuse at seven. The nuance, the rest of the trials was all lost. And mostly what was delivered was, we gotta give people blood only when their hemoglobin equals seven. You'll see in the upper right-hand corner, Paulie Bear and Jeff Carson wrote this editorial, I think it's now six or seven years old and they said transfusion of seven is the new normal. 10 and 30 is gone, we're transfusing at seven. Now, if you read their editorial, they're talking about ICU patients with sepsis, but most docs didn't read the editorial. Most docs just said, oh, wow, okay, this is a new policy change, we should pay attention. There's been a really bizarre growth of an industry against blood. So, the AAABB, which is actually a legitimate scientific and clinician organization, up here in the left, they talk about patient blood management, but the patient blood management industry has been focused on getting doctors to use less blood. There's a bunch of reasons why some of them are economic. Sabham is an organization that I've been a part of that it's like dogmatically against blood. They're also heavily influenced by pharmaceutical industries that wanna push IV iron rather than transfusion. So you've ended up with major scientific and clinical organizations saying like, seven's the new normal, stop transfusing, transfusion's harmful. Anyone that was here in 2022 at the McLean Fellows Conference will remember that our guest speaker and award winner who was very influential in choosing wisely Canada made the argument that actually in the lower right-hand corner, if we stop giving blood, it will actually affect climate change because we're causing waste in the medical system by giving blood and that's leading to climate change. So if we just stop giving blood, climate change will get better, which is just a wild argument. In nature, transfusions are one of the most overused treatments in modern medicine. It's costing billions of dollars. Where did they come up with billions of dollars? Also for us in the McLean Center when we say like something's overused or unnecessary, that's often like should trigger, this is a value judgment, right? Like who overused to who? Unnecessary according to who? So there's been an entire focus industry saying like stop giving blood and doctors have responded. So we've basically stopped giving blood. We now use universal restrictive transfusion which means we don't give a patient a transfusion until their hemoglobin is less than seven. And this is monitored by the health system. So anyone that's transfused above seven or definitely above eight, you've probably heard from our blood bank, but this is national, okay? It's a quality and safety metric. I actually ended up on the guidelines panel from 2023 and the guy who's the second author, Simon Stanworth, he would say to me repeatedly like what do you want Micah? We've randomized 17,000 patients to liberal versus restrictive transfusion and we've seen no difference in mortality. How many more trials are you gonna demand before you are a believer in restrictive transfusion? He said, this is the best evidence-based medicine has to offer, right? So those of you that know the evidence-based medicine hierarchy you'll see here at the top, we have RCTs and systematic reviews which is what I've shown you. Trials and systematic reviews have sown no difference. It's been 17,000 patients and if I'm honest, when I dig into the outcomes other than mortality or small trials, there really hasn't been any difference in outcomes. Okay, so Simon says, what do you want? But I can't get past this idea that like, wait a second, all patients get a single strategy for treating their anemia. This just doesn't make any sense. And the trials, the trials tell us what the optimal transfusion strategy is. Like, are there other things in medicine we do where no matter who you are, what comorbidities you have, what your values and preferences are, we use the exact same treatment approach. So when Simon says to me like, well, what else do you want? This is the best of evidence-based medicine. And Gordon Gaia, who many of you also know is the evidence-based user author was at this meeting too. I said to them like, you know, Gordon and Simon, like, Gordon's the one that would tell you evidence-based medicine is more than just relevant scientific evidence includes clinical judgment and patients' values and preferences. And I don't think we've actually paid attention to clinical judgment and patients' values and preferences. And I'm not totally sure we've nailed it on the relevant scientific evidence too. So the goal of evidence-based medicine, Simon, is to support the patient by contextualizing the evidence with their preferences, concerns, and expectations. This results in shared decision-making in which patients' values, circumstances, and settings dictate the best care. But if we're just deciding that your hemoglobin value when you're hospitalized is the only thing that dictates whether you get a treatment for your anemia, which happens to have adverse consequences with it if we leave you untreated, is shared decision-making possible? You know, are patients' values and preferences even possible? So I'm gonna go through each of these domains and I'm gonna show you where I think we've messed up. Clinical research, clinicians, where we've made ethical missteps, they're small, they're minor, they're not the big things that you're gonna read about in a bioethics journal, but I think they've led our practice of treating anemia, common condition, and hospitalized patients astray. So the first thing I wanna do is I wanna take you back to this trick trial in 99, okay? And I said it's set a methodologic precedent. In trick, they're comparing transfusion strategies, this idea that you should get blood at a higher hemoglobin or a lower hemoglobin. They're not actually testing what the effect of a transfusion is. But I'm on service right now. David's on service, I think Nikki you're on service. Like when you sit in front of a patient, you say, should I give you blood? The question that you're asking in your head is what's the benefit of this single unit of transfusion at any given hemoglobin? It is not what is the strategy that's the best for this patient, okay? So they're related questions, but they're different. And why this matters is because this question that the trials have tested, that trick started out trusting, it doesn't answer our fundamental questions. How do we treat anemia or what's the optimal management? And very interestingly trick, but all of these trials, when they've been designed and conducted, they've captured the adverse events of transfusion, but not anemia. So this is actually biased the trials in favor of restrictive transfusion anyways, because we don't even know what the consequences of anemia are after the fact. So the scientific question is not the relevant clinical question for either patients or clinicians, the evidence base as good as it is wonderful or randomized in 17,000 patients is orthogonal to the primary clinical question. This methodological precedent also mattered not only because this trial has kept caring for it, but because that trick trial never included a standard care arm. Okay, so this is from a paper by Dean and Vox Sanguinis. It's really fascinating. But if you look back at trick, I told you that they use this 10 and 30 rule back in 99 and 2000 to transfuse people. And so they would look and say, does this patient have a reason for us to give them blood? Whether their hemoglobin is a little bit above or below 10. Well, one of the things that people were given blood for is if they had cardiac ischemia. Another thing that they were given blood for is if they were very sick. And a third thing is older people got more blood. So what Dean did is he went back and he looked and he analyzed subgroups of patients. So those with ischemia and not with ischemia and he looked at who got restrictive and who got liberal transfusion. And you'll see on the left here the restrictive group. So the lower hemoglobin group got had a higher mortality rate with less blood. And in the non ischemic group, the group that the standard of care at the time would not have been to give blood who got the blood ended up doing worse. Similarly over here, sicker patients actually, I'm sorry, less sick patients actually did worse with more blood. They would never have gotten blood under standard of care and younger patients did worse with more blood, but they would have never gotten blood with standard of care. The first issue is like a trial that doesn't include a standard of care arm raises questions of equipoise and safety. But what's really interesting about this is for 20 years, this figure on the left does a patient with cardiac ischemia. And remember, why does it matter to give blood with cardiac ischemia because you have reduced blockage of the coronary arteries, you have reduced oxygen delivery to the myocardium, that transfusion may increase the oxygen delivery and save the heart. This became sort of dogma that like it doesn't matter. It doesn't matter if you have cardiac ischemia. Why, who cares? We've already shown in trick it doesn't matter. So I put this up here because this is actually the recommendation from the most recent guidelines. You'll see the same recommendation in 2016 and 2012, but you'll see this odd remark here which I had to fight for this remark to get in there. And I convinced at least a majority of the group to include this remark. And then this remark is this weird, it says, yeah, we think you should give hemoglobin at seven, but you might use 7.5 if you're undergoing cardiac surgery. You might use eight if you're going on orthopedic surgery or pre-existing cardiovascular disease. Now I fought for the remark because the trials have never agreed or never used the same hemoglobin threshold. And I said, we have to be, we have to show fidelity to the trial data. And so we need to tell them, we need to tell the physicians that are going to read this, what the trial data actually says. The retort from Simon and Gordon for these people that were really kind of making fun of me was like, Micah, there's no difference between seven, 7.5, eight, these are clinically insignificant. And doctors actually cannot handle a complex recommendation like this remark. It's just confusing. And this is what leads to poor care. Okay, so their argument was it doesn't matter. Now everyone, and then I said, well, what about patients with coronary artery disease? And they pulled up this table that you'll find here. And they said to me, we had it on a big screen just like this. And they said, look, come on, you're making a mountain out of a molehill. The relative risk of restrictive transfusion in patients with myocardial schemas, it's the exact same as the wrong group. And a pretty narrow confidence interval. So I asked, I said, well, how did you summarize this data? And they said, well, we used a fixed effects model. They said, well, what does that even mean? You know, why are you using a fixed effects model? I don't wanna get too much into fixed or random effects. But those of you that know enough about it would say, wait a second, this is a strong assumption, a fixed effects model that every single patient with cardiac ischemia has the same outcome. And when you look at these across trials, we're looking for a single outcome. Very strong assumption. So I asked them to add a random effects model. And you'll see here down in the notes, the confidence interval changes. And what you see is that a restrictive transfusion strategy might result in four fewer deaths, but it also might result in 40 more deaths. Okay, so this happened about a year ago. We had this argument, they made fun of me. And what happened in this fall was the mint trial came out. And the mint trial was looking at this exact question. And they randomized 3,500 people with cardiac ischemia in the hospital to restrictive versus liberal transfusion. And the take home from this slide is this top line, which is basically people that got more blood did better. We probably could have figured that out in 1999 if we had a standard of care arm. But we've gone for 20 years and assumed it didn't matter and assumed that these patients with cardiac disease could get transfused at lower hemoglobin. In fact, one of the committee members said, I think this remark is unnecessary. You're making too big of a deal of this. We've already lowered our transfusion threshold given the blood shortages. And so she did that without actually checking what the outcomes would be. Okay, last few points under this trial data and then we'll move on to the next section. But the other ethical thing, other ethically interesting thing is when you look at these trials, these are some of the biggest ones. This is actually a few years old. I should update this table. But if you look at the biggest trials, I told you they've compared two transfusion strategies. Okay, well, in reality, they've not used a fixed number. They've used a range. So this is the trick trial. Their restrictive strategy was between seven and eight and a half and their liberal was between 10 and 10, seven. But there's wide variation in the lab result that we get on every morning on a hemoglobin. And if you look some of these trials, like this Villanueva trial, which actually ends up being very controversial, the two trial arms overlap. And the committee, the guidelines committee, these industries that have built up around reducing blood have never recognized that like in some of these, they use these trials to argue for less blood, but they've never recognized that if this is the case, if the two trial arms overlap, the intervention, which is which trial arm you assign to, can't even be tested. So Villanueva randomized 920 patients and they actually can't confidently walk away and say that their data means anything because there was trial arm contamination. And so those of you that do clinical research, like that's a ethical violation. You're exposing a patient to an intervention, 921 of them in Villanueva's case, without ever being able to say with statistical certainty or clinical certainty whether it mattered. Okay, and this is I think the final point with this, but this is the one that really drives me nuts. You're gonna just hear people and maybe many of you have said this, but they'll say like, well, listen, that's fine, but blood is bad. We, the trial showed us blood is bad. And I pointed this out earlier, but the risk ratio is one. That means blood is not bad. There's just no difference. It doesn't matter if we give it to you at a higher rate or a higher hemoglobin or a lower hemoglobin with respect to mortality. But I think many of our clinicians that are doing this don't recognize that point. And so they've come to assume blood is bad. So it's this misunderstanding of a statistical, of a statistic which is the relative risk. Even on this narrow confidence interval, right here, the absolute risk of a restrictive transfusion strategy is 1%. And people say, yeah, but blood is so harmful when you give it. Well, if the 1% absolute risk of death from a restrictive strategy scares you, like blood should not scare you because it's actually very, very safe. Almost nobody dies from a blood transfusion. Okay, so to sum up the ethical considerations in the trial data, the scientific question we've answered is not the relevant clinical question. Equal poise, safety, harm have been replicated over 20 years based on a single trial. We may have explicitly harmed patients with coronary artery disease, just a complete disregard of heterogeneity of clinical differences. We've exposed people to trials where we had absolutely no ability to determine the clear scientific conclusions, statistical bias, imprecise conclusions in our assessment and misunderstanding and misapplication of that data. And yet we continually apply restrictive transfusion policies to all patients without recognitions of these limitations. So what about patients' values and preferences? Like, are they involved in this restrictive transfusion world that we found ourselves in? There's a group that has taken that evidence-based hierarchy. Has anyone seen this before? I actually wasn't that familiar with it. They've tried to rewrite this evidence-based hierarchy to not focus on studies, but to actually focus on how we get the synthesis of studies out into the world. And so you'll see at the top, they put systems and they use this clinical decision support. So I'll show you what that means. Like, I think this is like absolutely insane. So we have studies and we've put them as the lowest level of medical evidence and instead we're gonna use synopsis of studies, synopsis of synthesis and say like, that's what's gonna dictate our clinical care. But for the point of this talk, I want you to focus on the fact that they think taking the sum of evidence, sum of trial data that we have and deploying it into a system is the top level of evidence. And it's really the result of this sort of long standing movement that sort of says like, standardization is the key part of improving quality and safety because everything that's not standardized is unnecessary, it's harmful, it costs money. So variation is bad, standardization is good. And so they cite these clinical decision supports, which there's sort of two main, there's a good more than this, but the two main that we see are best practice alerts. This is when you're on your computer, you're taking care of a patient, something pops up and says, hey, their hemoglobin is 75, do not give a transfusion. Or a computerized provider order entry that's when I go to write a transfusion order, it says, are you sure you want to do this because it shouldn't apply to your patient? Your patient's hemoglobin is too high. And it just happened to be that we were able to, as all this trial data and transfusion was accumulating and the electronic health record was coming online, we started to realize like, well, we can embed these clinical decision support tools in the EHR, will force people to follow restrictive transfusion, CMS is meaningful use criteria, we're updated in 2014 and 2016, right as the majority of these trials are being done. And this is really simple, like we want to standardize care, if we're trying to tell you that all you can do is transfuse when your hemoglobin gets to seven, like that's the simplest clinical algorithm we can build into a computer, like anyone can do that. So institutions started deploying these clinical decision support tools. And here's ours, I won't get into all the issues with it, but like it doesn't reflect anything that's in the trials, it doesn't reflect anything that's in the guidelines, it's like very haphazard and it was the result of like a group of people getting together and some people saying, I want to say this and I want others to say this. Yet, it dictates clinician behavior. And I'll talk about why that's important in a second. So I kind of was getting annoyed seeing all of these clinical decision support tools being published in the literature, just paper after paper, institution after institution, Hopkins, Harvard, I'm sorry, Stanford, Stanford published like six times saying, we built a clinical decision support tool to improve patient care and we stopped giving blood and look at how good we did. And so I asked my research coordinator, will you pull all of these studies? And we actually ended up, we've submitted this for review, but like we found 32 that we wanted to include, which is a fair number, okay? And all of these institutions wrote paper saying we deployed this clinical decision support, it improved outcomes, but I had her look and I said, like how many of those manuscripts that were published in the literature actually reflected the data in the trials or the guidelines? And it turns out it's about 12, about a little over a third. 50% of those just said, we just think blood is bad and we're seeing all this trial data come in. So we're gonna come up with a policy that just says to doctors, stop giving blood, but it wasn't based on any sort of trial data or guidelines, it was institutional policy. And then 12.5% just said, we just wanna do this because it's good and someone, some journal published it. But what's really interesting out of the 12 that cited transfusion guidelines or trial data was making the attempt to say, we want our doctors when they see these clinical decision supports to respond to it and we want the clinical decision support to be based in reality, it turns out that like 83% of them miscopied the guidelines, miscopied the trial data into their program, into their electronic health record. Maybe more minor, but 30% of them described any plan for updating the clinical decision support as new data emerges. So you think about something that's deployed in your electronic health record, it's influencing how doctors work and a new trial comes out and shows that it's not right and yet they have no plan for updating it. And only 50% described the patient population that they even deployed this in. This is I think the most interesting part. So of all these studies saying, look at how great we did in deploying a restrictive transfusion clinical decision support when their measure of why it was so good was they reduced red blood cell orders. They got different hemoglobin pre and post transfusion hemoglobin levels that doctors were transfusing at. And then the frequency of transfusion outside of that recommendation, whether it was trial based or not. My favorite one for both David's here is costs. Everyone said we reduce costs. And they did that by just saying we gave X number of red blood cell transfusions before the intervention, Y number after a red blood cell transfusion order costs or a red blood cell transfusion costs $250. So 250 times Y equals our cost savings. And for our economists in the room, like that's not how you measure cost effectiveness in a medical study, right? So all of them claim that we reduce costs, we stop giving blood. The patient outcomes, where is it? Almost none. So you have length of hospital stay, readmission made, mortality, nothing on quality of life. So like these, they're not even focusing on the things that matter to patients. They're focusing on things that are irrelevant to the patient at least. So this has just become dogma now. Health systems use clinical decision supports to promote unified restrictive transfusion. And these interventions direct care. Most clinicians don't spend as much time as I do reading these trials or thinking about this. Like they're busy, they have a million other things to do. So what are you gonna do when something pops up on your screen and says, the evidence says, don't give this, you're gonna follow it. At our last blood committee meeting here, there was a discussion about sending one of our faculty members a letter punishing him for overuse of transfusion. And I was like, wow guys, this seems a little unnecessary, right? But our institution's like, well, we don't want people giving blood above a hemoglobin of eight. So these clinical decision supports have allowed us to track physicians who are giving blood and punish them. The published literature has completely oversold the benefits. But what I think has also happened is that like there's a range of trial data, there's variation, there's some uncertainty. We all deal with that all the time. And the clinical decisions reports have sort of said, nah, none of that uncertainty exists. Like here's just an answer. And for those of you that are familiar with Bruno Latour's science and action, like to me, this makes us think of like, well, how did we get to a scientific fact? This is the black box of someone just saying, we have data, but now we've agreed and all decided that we should only get blood if your hemoglobin's less than seven. What I think it's also done, and this is how it leads back to patients' values and preferences is like, I think it's subconsciously taken the patient completely out of it. Like we just don't even have to think about what the patient, what matters to the patient. So you'll see in the guidelines, I'm gonna come back to this slide, but it's, we actually say it's good clinical practice to consider not only hemoglobin concentration with symptoms, signs, lab data, and patients' values and preferences. Like ask the patient, do you want a transfusion? But when you build these clinical decision supports and you drive physician behavior in a certain way and you make them convince that this is the right thing, and that right thing doesn't include any sense of patients' values or preferences, like we've completely removed the patient from the decision. So the transfusions are uniformly prescribed, they're promoted, irrespective of differences, and like we don't even actually agree on a framework for thinking about which preferences matter. So if the committee says preferences matter, which they did, I said to them like, what preferences should matter? What should be important? And they said, well, we don't really know. I said, well, we don't know because like we've actually removed preferences. So in thinking about that, one of the studies that we did was we looked at differences in hemoglobin level between males and females. So I showed you, I think on the first or second slide, that males and females have different standard hemoglobin levels that define anemia. That's been well described in the literature. I told you anemia changes over as you get older. And so men and women over time have different hemoglobin levels. And there's some really, really good clinical research showing that depending on the sex of the donor and the recipient, you get differential changes in your hemoglobin, okay? So this suggests that the oxygen delivery that you're getting, if you get blood from a male, you're gonna actually have a higher hemoglobin level, higher oxygen delivery, particularly if you're a female. Like this shows that these hemoglobin levels matter. And if we believe that hemoglobin drives things like fatigue, quality of life and symptoms, well, maybe these differences in hemoglobin are also associated with, or maybe the consequence of these differences in hemoglobin are such that patients have different symptoms. Okay, so we studied this, we collected data, we enrolled patients in this hospital, anyone that had a hemoglobin less than 10, and we asked them to how fatigued they were. We compared their fatigue levels by hemoglobin strata between sexes, used a simple linear regression model using fatigue as the dependent variable, and then two different models, which I'll go through here in a second. And what you see is that women, overall, and at every single hemoglobin strata have higher fatigue levels, anemia-related fatigue levels than men. This scale that we use is reverse-coded, so a lower score equals higher levels of fatigue. Okay, so this is unadjusted, and these are all statistically significant, except for between seven and 7.9, where the effect is still real, but didn't quite reach statistical significance. If you use an adjusted model, our regression models here, and you stratify, you'll see that females have an entire standard deviation, a higher level of anemia-related fatigue than males, and more importantly, as your hemoglobin drops, you get greater increases in fatigue across that drop than males. And if you just stratify simply by male and female, you'll see that males have a higher anemia-related fatigue level than males. So females have clinically and statistically significant higher levels of anemia-related fatigue within each stratum. The levels of fatigue are nearly a standard deviation. Greater fatigue changes. Anybody that was following that data, which I put up pretty quickly, might say, well, wait a second, females should have less fatigue because if they have a lower baseline, you might expect that if they're treated at restrictive transfusion, that's a smaller change, and that might be true. We don't quite know why that is, except that the effect that things like fatigue and genetic and biological variables affecting red blood cell structure may be driving that result. But the key for today is less that and more to say like, this calls into question this uniform restrictive transfusion policies because it's probably not optimal for either sex. And maybe fatigue, maybe symptoms are the framework by which we should be considering patients' values and preferences. How do you feel? How did your anemia made you feel? But we can't get there because we've been focused so much on standardization. How does the standardization work out for us? This idea that if we just give everybody the same, you know, critical decision support and we push restrictive transfusion, we're gonna get equity. That's really what some of this is pushing towards. So we collected data looking at transfusion differences by race. And you'll see here in the hemoglobin less than seven columns, so everyone here buy the guidelines and buy the trial data. If you're a restrictive transfusion advocate, should get red blood cells. And you'll see we have a disparity. So 80% of African-Americans, 86% of whites and 92% of patients of other race, okay? So this is under transfusion. This is guideline discordant care. What is very interesting though is in the seven to 7.9 area and the eight to 8.9 whites got actually a lot more blood. This is despite a clinical decision support telling everyone when they should get blood. And what I think is interesting about this is like, well, why are, this is guideline discordant too. If you're really a restrictive transfusion advocate, you'd say, well, whites got actually worst care in this category here. So I talked about this a little bit, I think a couple of years ago at the McLean Center Retreat. But what I think is happening here is if you're white and you're on the general medicine services at the University of Chicago, you probably had to drive past six or seven different hospitals to get here. And you probably have access to an iPhone with an Epic My Chart that shows a little red dot for an abnormal lab. And you get an abnormal lab and you say to your doc, why is my hemoglobin low? I want a transfusion, right? Compared to patients, are African-American patients who lower health literacy, lower socioeconomic status. So the association between driving here and paying attention to your lab values probably is also associated with your ability to advocate for your own values and preferences and get a transfusion. We don't have the data to show it, but that's one of my hypotheses. So clinical decision supports, promote clinicians algorithmically applying a uniform transfusion policy. This removes the patients, this removes consideration of the trade-offs of the risks, benefits and costs for individual patients and their preferences. Standardization is well-intentioned, like there's this movement to reduce overuse, which I understand. What it assumes trial data and average treatment effects constitute generalizable facts that can be applied to individual patients irrespective of context values. And those are the physician or the patients. And as I said before, I think this allows physicians to sit back and make a decision at the computer without actually having to go to the patient to see what they think is right. And then the consequence is suboptimal care and probably worsened health disparities. Oh, oh, there we go. This might have just died. So standardizing transfusion to improve care can result then in the false assumption that standardization is optimal for all patients and will result in a parity. If we assume that trial data or guidelines are neutral with respect to personal history, values and beliefs, parity will produce optimal outcomes, but this is where social determinants of health are then disregarded. And we know these social determinants of health actually end up driving the outcomes. And so paradoxically, the guidelines and the evidence meant to standardize care may be a source of exacerbate disparities. Okay, last few slides. What about clinical judgment? Where does clinical judgment fit in in the practice of restrictive transfusion now? So I think one of the things that happens is that we've now inadvertently deemphasize the significance of anemia because we've told people that it doesn't matter if you're hemoglobin's eight. It only matters when you're hemoglobin's less than seven. And reminder, anemia is associated with all of these adverse outcomes. It should trigger a diagnostic workup, whether you're in the hospital or in the outpatient setting. And so because it's safe to not give you, not do anything into your hemoglobin's less than seven, it's just not considered significant. So clinicians have a diminished appreciation of this. So we had around 1,000 patients in our database and I said like, how many of these patients are even getting worked up for anemia? And it turns out that almost a quarter of them with moderate to severe anemia during hospitalization don't even get a single lab beyond a hemoglobin level to actually work up their anemia. So this I think sort of speaks to this or suggests this idea that docs aren't even worried about it. They're not even considering it. And then what's interesting amongst that 25%, there's actually variation in the amount that get transfusion both by age and by the level of hemoglobin. So docs are now actually just saying, well, I might give a hemoglobin once it gets below seven or they're not actually thinking about what's the etiology of anemia, why am I doing this? It's very reactive. So then I had that I said, well, like what do doctors think? Cause like I talk about this a lot and any of you that know me will know like, they've heard parts of this like, what do docs actually think matters in a transfusion? Do they believe this restrictive transfusion? And I showed this slide earlier and the guidelines committee came up with this just like laundry list of variables which should guide transfusions. But then, you know, as I said in the committee, I was like, who made this list up? Like where did this come from? Well, it just makes clinical sense. So no evidence base for this list. So I thought, well, let's ask docs whether they believe that those things actually matter. So we enrolled 85 patients. Most of them were from hospital medicine, 25 of them were from trauma surgery. And we looked at 30 different clinical factors and said like, does this matter to you when you're making a transfusion decision? It turns out that there's no single factor that's universally agreed upon by all providers as being very important in making a transfusion decision. And only seven factors did more than two thirds of doctors agree are very important. But of those seven factors, five are related to hemoglobin level. So I think doctors think there should be variation that they don't totally believe this restrictive transfusions that's being pushed on them, but they can't actually agree on what factors matter. 68% of providers believe a patient's preference for a transfusion is very important. But you can't really say I believe restrictive transfusion and I believe patient preferences matter because there didn't disagreement with each other. We had 88% of docs said that restrictive transfusions practices are a standard of care and then said they're optimal for hospitalized patients. But then I also consider clinical context. So it's actually a mismatch in what doctors are saying they believe is important and what's even possible under restrictive transfusion policies. Most doctors believe that restrictive transfusion or that transfusion in general can improve health related quality of life and that the transfusions decision should involve shared decision making, but many of them are not able to do it. Only 38% agreed that a transfusion should be given for fatigue despite the fact that most thought it should be given for health related quality of life reasons. And then most doctors just thought that variation and this goes back to this long running movement in improving health outcomes that standardization is right. Like any variation in transfusion care increases quality and it increased costs. And like how can you actually personalize transfusion decisions? How can you have shared decision making if costs and quality go down when you individualize the decision? Okay, so to sum this up, Anemia is common. It's associated with adverse events. We've long treated it with transfusion. There's a large body of clinical trial data, meta-analysis, pushing restrictive transfusion or reduced use of transfusion. This has resulted in significant clinical shift. There's a belief among providers that this is the best we have of evidence made of medicine. Maybe cardiology has trials, maybe oncology has trials, probably 17,000 patients. But the trials and the data have limitations and uncertainty and that's normal. That's not something we run from. I mean, that's something you recognize with any single trial in the generation of knowledge. And patients values, preferences and individual clinical characteristics are now almost never considered in these transfusion decisions. And this has been reinforced by the electronic health system, systematically embedding or misembedding restrictive transfusion into the practice of Anemia or then into the care of patients with Anemia. The result is restrictive transfusion may be suboptimal for patients. It may exacerbate disparities. Clinicians now can't even agree on what factors should be relevant. And there's inconsistencies in their beliefs about how to optimally transfuse a patient. And so if the goal again of EBM is to support the patient by contextualizing the evidence with their preferences, concerns and expectation, through shared decision-making to resolve and best care, I'm not sure we're doing that. I thought Adam actually might show up. I don't know, has anybody read Adam's book? Anyone read this book? Yeah. So actually maybe it's good he didn't show up because I won't misquote him. But this is like a really interesting book that Adam wrote and he talks about this idea of medical reversal. And this is our assumption that some intervention that we do benefits patients only to find out later that the intervention doesn't benefit patients actually harms patients. So what are the ingredients for medical reversal? Uncertainty of data. So clinicians not totally understanding or misinterpreting data, variation in trial data. And I'm gonna add a third one which is the lack of considering patients' preferences and values. And so actually things that we think helps patients harm patients. And I think maybe I'll be dead by the time this comes around or there's a second version of this but I think restrictive transfusion might fall into this medical reversal. What does it have to do with clinical ethics? So Mark taught me and I think many of us here like clinical medical ethics, it's practical, it's routine, it's things we see every single day when we're caring for patients. It's intrinsic to what we do. The goal is to improve patient care and outcomes. Transfusion to treat anemia is an example of a common routine condition that just seems simple and straightforward. We think we have the evidence base. Why are we even talking about anemia? But the dogmatic adherence to restrictive transfusion is not the best care and it leads to harm in individuals. And so clinical medical ethics should have something to say about these things. And I think had we started back in 1999 calling into question some of the scientific evidence, some of the trial design, I think if we were to refocus on what patients wants in a transfusion and does it benefit them, we would have an evidence base or we can begin to develop an evidence base where clinicians could then even agree on what matters. These types of things are hard to get when you think about the world of bioethics. So bioethics, I love it. I love reading the interesting topics and stories and keeping up to date on it. But sometimes it can miss the clinically relevant things. It can miss the discussion about whether there's equipoise in a trial, whether there's safety harms and risk trade-offs in common interventions like a red blood cell transfusion. Like, what does it mean to standardize care? Why are we standardizing care? The statistical under certainty, clinically significant findings and variation in provider beliefs. So clinical ethics is what happens in these routine situations between doctors and patients in the hospital. And these situations don't always have some big rationalistic or universal norm that we can publish but they're nonetheless important for patient care. So thank you. I appreciate it and happy to take any questions. Scarce resource? Yeah, I think it's a second scientific problem or maybe I'll just say it's a policy problem. It's not, it's, you know, I've heard this before, actually, that's one of the retorts people I say to me. Like, well, we don't have blood. So why would you think that we should be giving more? It's like, well, okay. The scarcity is an issue that we should deal with but let's figure out first, what's the right treatment? What's the optimal treatment? What do patients want? So I don't mean to minimize it. Like, I think it's a problem. I wonder how much of the scarcity problem is also reinforced by our assessment that we just don't need it. You know, there's a little bit of this self-fulfilling prophecy. Dartmouth in particular has gone way over the top in transfusion at a hemoglobin of six and they're like, well, we solved our scarcity problem and then the next week they're coming back and they're like, well, actually we didn't have enough. Yeah, we're experiencing the blood crisis too. And so, yeah, I think it's a policy issue that's unrelated. Well, related but secondary to the clinical question. David, let me just say very quickly. So now it was probably 2022 or 2021 actually, the fall of 2021 when we first met with the 2023 guidelines, Jeff and Gordon and Simon sent out PICO questions and patient was hospitalized patients, intervention was transfusion. And it was like a Saturday, I remember it was in my car because I wrote it on my phone. I'm like, should I send this from my phone? And I was like, I want the intervention to say transfusion strategies. And they had spent like three or four months coming up with transfusion. I said, because that is what the trials have tested. So actually we did change it to transfusion strategies. And there was really not much pushback but I thought it was so interesting. Like they're the ones that have run the trials and they're making this confusion between transfusion and transfusion strategies. And what was your third comment? A series of recent stuff. Yeah. And you know, should we break that in the hope that you can save yourself as someone who's just an applicant? But, you know, maybe there's a different thing. Yeah, yeah, it's been interesting. I don't know that I've tried out any of these arguments in totality, but I've tried out most of them in some form. And I mean, just huge resistance, huge resistance. I would say that the people at the top of the chain, Jeff and some of the other people that, you know, Paul, you bear that run these trials, like they actually would agree. They feel, I was surprised at how incapacitated they felt by the whole thing. Like, well, you know, the NIH is not interested in funding additional trials. Like they believe the science is settled. So I think it's a good comment and a good question. I mean, maybe it needs to just be written out as a bigger, longer article and see who's interested in it. And maybe from the ethics perspective, maybe that's the door in to get this into literature. Thanks, David. Peter. So thanks, Mike. I thought that was great. It seems to me that some of what you illustrated by this talk is the extent to which it's much easier to, this is ubiquitous in medicine, especially when it comes to how they feel. Right. You know, I mean, that's just curious to your thoughts because it does seem as though the ability to say, imagine, you know, we have all this stuff. Yeah. You know, I'm not, I'm certainly not the expert in this, even in this room with patient reported outcomes. It's just extremely difficult to measure. I'm definitely not the psychometrician expert of patient reported outcomes, but those of you that know Robert Gibbons would, you know, if you sit him down and you ask him about the promise outcomes, which are the most widely used patient reported outcomes, he'll just be like, these are awful. They're terrible. And he'll give you, you know, a list of five to 10 reasons why they can't even capture the data that we say matters. So the subjectivity of patient reported outcomes is extremely difficult. And then designing a trial to capture those because there's wide variation in the outcome, you need huge numbers to detect a clinically significant effect and for it to be statistically significant. Well, you know, this is actually, the question you're actually is really where the focus of my work has been and the criticism with the trials is they say, well, look, we looked at patient reported outcomes and we saw no effect. Well, you can't enroll 30 people and ask about your fatigue level, your anemia related fatigue level and then say, we didn't see an effect. I mean, the underpowered nature of all of these trials that have looked at anything other than mortality. So that is the tidal wave. And, you know, the challenge is like, actually how do we design a cost effective trial to actually test some of these questions? So yeah, I agree. Yeah. Isn't it like kind of amazing that that like, like we have a, we have a, or we had one medical, we have a medical student, like this is like the thing you learn as a first year medical student. Like how do you feel this morning? You know, like, and yet like we don't routinely do this and we definitely do not capture data around. Yeah. Yeah. Well, so the trust thing is really interesting. That's one of the other hypotheses about the transfusion and race data. I haven't figured out if that, trust is so complicated. Like I feel like there's arguments that we have no trust, patients have no trust in us or they have lots of trust in us. So it's one avenue to consider. I do think that patients who either don't trust or they are super busy or have a whole bunch of other things going on who are part of this medical system. I think they're the least likely to get the blood and get blood in these instances. They're probably the ones that need it the most. But like, you know, if I understood kind of what you're saying, like, you got a mom who's hospitalized and like life's falling apart at home because you got four kids. Like the last thing you're gonna do is sit down and be like, yeah, keep me here for an extra couple hours for this unit of blood when you can't quantify even whether I'm gonna feel better or not. I've got to get home. I think, so I don't have an answer to that other than to say like, my ideal world would be one in which we say like your hemoglobin is 8.5. The chances you feel better are X. The chances you don't feel better are Y. I think you should take this. I think you should get the blood because feeling better by chance X is worth it compared to the risks of transfusion. And that's an empiric question. Like that is actually an empiric question. We can answer that question, but we've not yet gotten there. And there's just a huge tidal wave against anyone either studying it or even raising the issue. So, yeah, I mean, that's probably gonna be some AI chatbot will tell us how they're feeling. And we'll probably have a clinical decision support in place to tell us that. So yeah, smug. Yeah. Is this something that, and so, is this the right actual setting to judge whether this media is real? Is it just effective? Is it just... Yeah. ... ... So I'm ambivalent about this in terms of like, does it need to happen in the hospital or in the clinic? Like, now, I could, I think make a strong argument here that like the majority of patients who get hospitalized in general medicine service don't end up in our clinic later on. And so that's the argument to do it now. Now, if you had, I think we could talk about whether it's operational or clinical or scientific, we could design some study to figure out what's the right algorithm and when should people be worked up? And we don't have the answer to that. And I'm okay either way. Oh, what I was gonna say though, is interesting is so Andy Arts, who, those of you that have been around a number of years would know him. He's not a former oncologist here who left right before COVID. And Andy was working with David and I on some of this stuff and Andy was convinced and the literature supports him that like somewhere between two and 5% of these people have MDS. So if you have clear iron deficiency anemia, you have iron deficiency anemia. You know, if you have clear anemia of chronic kidney disease, you have anemia of chronic kidney disease, but chronic kidney disease is often misapplied. Like we just say, well, they've got a little bit of chronic kidney disease. They must have anemia of chronic kidney disease. And particularly in older adults, like a healthy dose of them have MDS. And so they're coming in, they're touching the health system. We think that it doesn't matter because, you know, their hemoglobin's above seven. We never draw a lab. We never send them to hematology. Primary care gets it and they have 20 different things that they're trying to evaluate this patient for. And the anemia just continually gets left over and lost. And some of them go on to end up having cancer. So another reason why, whether it happens in the clinic or the hospital, I don't know, but we shouldn't, we shouldn't let them get lost to follow up without thinking it's clinically significant. Okay. Thank you.