So thank you so much for the invitation to speak. Again, my name is George Nisi from the University of Pennsylvania. I'm a medical intensivist and clinical epidemiologist and health services researcher. And I'm a bit of the kind of the odd speaker out in today's session, as you might notice. And what I was asked to talk about was not post-COVID syndrome, like post-COVID clinical syndrome on a patient level, but instead the impact of the stress and strain of acute COVID surges on acute care delivery and outcomes. And so kind of the aftershocks of COVID. And so that's what we'll talk about today. And then I think our third speaker will bring us back to kind of the core goals of the session. Here are my funding disclosures. My research is funded by federal and foundation sources. I have a few financial relationships, none of which I think pose a conflict for today's talk. So I have three learning objectives. The first is to understand the operations concept of hospital stress, or also called healthcare capacity strain. The second is to understand two related phenomena. One is hospital adaptation, one is hospital resiliency during acute surge events. And the third is to understand the now measured or ongoing measurement of the impact of COVID specific capacity strain on acute care delivery and outcomes. And so we'll start with this concept of hospital stress or healthcare capacity strain. This is kind of the concept that anchors my work. And this is an operations concept that says that there are limits placed on a care team, a hospital unit, any unit of care delivery on their ability to provide high quality care to all patients needed at a given time. And we know that there are numerous patient system factors that contribute to capacity strain, things like occupancy and turnover. So prevalent incident patients, more patients, how sick those patients are, what kind of care demands they require. So certainly the same number of patients, but who require elevated levels of isolation and equipment and specialized personnel and so forth is gonna be a higher capacity strain load. And then even in the face of static demand, if there's a reduction supply, anything that reduces resources, that's a source of capacity strain. And in the preparedness literature, we talk about the S's, space, staff, and stuff. So where to take care of patients, hospital beds, the appropriately trained personnel deliver that care and then stuff, the tangibles, medicines, equipment that's required to deliver that care adequately. And we've tried to argue that healthcare capacity strain exists along a spectrum. And so in the middle of the spectrum, under the umbrella of what you might call dynamic strain is kind of routine changes. Anyone who's been in a busy hospital or busy emergency department, semi-predictable seasonal changes. And then if you swing to all the way on one end of the spectrum, you get acute surge events, which is kind of what we're talking about today. And obviously what we've all experienced over the last couple of years, a dramatic increase in the demand for care and or a concomitant reduction in resources. And in addition to epidemics and pandemics, certainly things like natural disasters and attacks on the public account for acute surge events. And then just for completeness, all the way on the other end of the spectrum, a different pillar of my work, which we won't talk about today, but is under the umbrella of static strain, which is care delivery and persistently resource limited settings, where there's more of a chronic, a fixed difference between supply and demand, a mismatch of supply and demand. And certainly acute on chronic exists in that as well. So capacity strain has been studied well before the pandemic and certainly well before I started studying it. And it has been over the course of a good amount of research been associated with lots of deviations in care delivery and processes of care. And so this is a few bullets of ICU capacity strain in particular and how it's related to a number of different process outcomes under pre-pandemic periods under what we might consider kind of routine strain. And so things like reductions in ICU length of stay and increased ICU readmissions, but no changes in overall hospital length of stay, changes in end of life care planning, shorter rounding time, loss of some adherence to evidence-based guidelines. And then my contribution, which is reduction in the likelihood of ICU admission. And the relationship between ICU capacity strain, for example, and outcomes has been a lot more difficult, a lot more nuanced, a lot more subtle to tease out. And so a number of studies have attempted to do this, I'll highlight just one or two that show a much more muted relationship. And so for example, this is a paper in the Blue Journal now a number of years ago by a former colleague of mine, Dr. Nicole Gabler, who found that there was in fact a relationship between ICU capacity strain and ICU mortality or in hospital mortality, but only at the extremes of strain, the highest deciles of strain and only in certain scenarios. For example, closed ICU staffing models, but not an open ICU staffing model. So much more nuanced relationship between capacity strain and hard clinical outcomes. And we may kind of come back to that a couple of times. But everything I've talked about so far is under pre-pandemic conditions, under what you might call routine strain. Certainly there are intense periods, influenza seasons and so forth in that timeframe, but nothing like what we've experienced over the last couple of years. And the COVID pandemic is not a single strain event or it's at least many strain events within a larger, bigger parent strain event. And so this brings up this important concept of hospital adaptation. And I'll argue this in two parts. And the first part is the kind of the nice side of the sword which is the ability to improve care and outcomes for infected or primarily affected patients, that is to say COVID patients during a COVID pandemic, during that surge event by implementing new care processes based on accumulated experience. So this is the learning process, the first part of hospital adaptation. And so early on in the pandemic, my center, like many centers around the country and the world published their initial experience. And one of the findings that really stood out to us, again, over that early, earliest part of the pandemic was a very swift improvement in the outcomes of our patients. So over, this is over 15 day period. So total of 60 days that you're seeing on this figure, our ICU, COVID ICU mortality decreased from 45% down to just over 20%. So a very, very, very rapid reduction kind of in real time. And that was robust to adjustment for lots of patient level factors. So what could account for that? We certainly observed very big differences in how we were treating these patients over a similarly very short period of time. And so on the left, you saw our quick adoption and then very quick de-adoption of mechanically ventilating a lot of these patients as we've now experienced and has been reported by many centers around the country. But even that was not enough to explain this outcome. So on the right, you see the same exact analysis from before, but restricted to patients who are mechanically ventilated, who presumably were getting more complicated, sicker over the course of this period as we were innovating fewer, smaller percentage, but still had an improved mortality. So this is the schematic from our COVID leadership group that I'm sure every hospital in the world really can produce. And this was those chaotic early weeks and what we did to try to combat this novel disease. And it details our team resources, the surge planning, the clinical practice guidelines, the modeling and the iterative change that we went through to kind of try to improve our care. And so certainly one, almost certainly more than one thing in this schematic, as well as things I've almost certainly left out, account for that improved outcomes. And this is hospital adaptation in real time. But there is a negative side to the hospital adaptation definition, which is whether you can continue to do this in the face of the capacity stream that comes with a novel threat. And so this was a really elegant paper by Khajiri and colleagues in Alice in General Medicine last year that showed both of these phenomena. So part one on the left, again, shows that improved outcomes very quickly over time. But when stratified by time period, so looking at what should be relatively stable outcomes, now showing that as a surge index, what they call the surge index, a composite measure of hospital strain, increased the adjusted odds ratio of mortality for these patients also increased. So worse outcomes as the strain was increasing in the setting of an acute surge event that's happening in parallel to our overall learning of this novel threat. And then this result has now been replicated in a number of very high quality studies. Bravada and Gemma Network Open used two capacity strain measures, a COVID ICU load, defined as the mean COVID ICU census divided by their pre-pandemic ICU beds, and then a COVID ICU demand, which was the mean ICU census divided by their maximum COVID numbers. So how much of their maximum COVID kind of capacity they were using at that time. And there's a dose response here. So as both load and demand increase, the likelihood of mortality increases. And this is concentrated within those patients who experienced that strain in the ICU. And then a final study, Matt Chirpek, a dear friend who has very sophisticated methods for teasing out lots of data. And I always respect his results, had a very interesting study looking at the attributable contributions to COVID mortality. And so certainly, you know, dominant by acute physiology and demographics and comorbidities, but for our purposes today, a very notably significant 9% from capacity strain. So what's going on around the patient impacting patient level outcomes, independent of patient level factors. So everything I've talked about so far is what happens, the stress being placed by COVID on the system and how it impacts COVID patients. But COVID patients are not the only patients we have to take care of during a pandemic. And in fact, probably more the largest, you know, mass number of patients we have to take care of don't have COVID during a pandemic and wouldn't have the primary affliction of a future event. And so how do we account for those? And so this is the second concept that I'll introduce, which is hospital resiliency. And this is the ability to continue to deliver high quality care to all or specifically to uninfected, what we might call bystander patients, despite the presence of a surge event. So how do we continue to deliver good care to everyone else who needs it while this acute surge event is going on? And while we're trying to learn about this novel threat and adapt, we also have to be resilient. And hospital resiliency is a concept that we've known about before. And I'll point to one very elegant, even kind of microscopic, so to speak, example, a local effect. So this is a really elegant research letter in JAMA from a number of years ago that showed that after there is a critical illness event on a medical surgical ward and hospital, so a code, a rapid response, an ICU transfer, the likelihood that a separate, totally unrelated patient, otherwise unrelated patient will have their own critical illness event is elevated. Okay, so two independent patients have different physiology. They're not identical twins. One has an event and over the next six hours, and in fact, longer than that, all the other patients, in fact, have an elevated risk of their own critical illness event. And this persisted over a number of adjustment strategies that has a dose response, that is to say two critical illness events is gonna even further increase the odds of a third and so forth. And this persists over time, decays over time, but persists so that even at the 12-hour mark, there's still an increased event. So a shift later, there's still an increased likelihood that another patient will have a critical illness event if a first one did. And so a really elegant example of what you might consider shunting of resources, right, towards an acute surge event, in this case, local, but an acute surge event at the expense of what's going on around them. So a microscopic example, for example, of a COVID pandemic and all the non-COVID patients around them that still require our attention and care. And so we've now started to see this bear out, I believe, in the literature coming during pandemic times. So this is data from Brazil showing their ICU outcomes restricted to non-COVID patients. And you can see that in the years preceding the pandemic, they had a year-over-year steady improvement in their ICU outcomes. This mirrors the continued year-over-year progress pre-pandemic that we had, for example, sepsis and ARDS outcomes. So a nice improvement over time as we continue to kind of get our care better and better. And then a loss of years worth of progress during the COVID pandemic. And this is all non-COVID outcomes. Okay, so their COVID outcomes are separate from this. Now, of course, the non-COVID population in the ICU changed during the COVID pandemic. So this is a nuanced analysis that has to be approached with care, but certainly a hint that there's this shunting of resources going on. Perhaps a more dramatic example, and one closer to home, this is data from the U.S. Department of Homeland Security, their COVID task force published in MMWR that looked at the relationship between what they called hospital strain and excess deaths. And here, hospital strain is defined actually pretty simply as ICU occupancy, kind of an all-comer American hospital ICU occupancy. And the excess deaths are all deaths everywhere, in the hospital, out of the hospital, all comers. And what they show is that as ICU occupancy increases, two weeks later, excess deaths are increasing all around us. And certainly at that 60, 70, 80% mark, there's a real takeoff. And we certainly existed in that degree of capacity strain at various points during the pandemic. And so this would suggest, and again, these are relationships that need to be, that have many variables, that need to be addressed with nuance. But if this were to hold up, that's a profound relationship between an in-hospital surge event and kind of our, essentially our public health around us. And I'll leave you with, I think, a last paper, which is a beautiful paper, very pretty recent in critical care medicine by Dr. Wilcox, that I think unified these. Looking at ICU capacity strain as a standardized ICU census, standardized to baseline. And looking as a capacity strain increases, you again see a step up in the adjusted hospital mortality. You see it as a dose response, and you see it across COVID and non-COVID patients. And in fact, perhaps amplified in those non-COVID patients. So even the brunt of the surge occurring as researchers are shunted to what's a cure, what is most dramatic in front of us and at the expense of the patients around us. And I would put together an entire talk of this same duration or longer to talk about all of the preventative medicine and age appropriate cancer screening and all the things that we do not have time for today, but that are gonna, we have yet to scratch the surface of how that will kind of tentacle out and we're gonna recover from that for a long time. So I'll end with a few parting words. The COVID pandemic provided, I believe, kind of both sides of that hospital adaptation sword. The really incredible side, which is the improved patient care and outcomes based on accumulated experience. And I should mention all the data we've shown you here is before the vaccination errors, early pandemic. So the amount of improvements that we've made is just incredible. And I don't mind toot our own horn for that. But on the other side, threatened patient care and outcomes based on the acute severe capacity strain of the pandemic for COVID patients themselves, and then turning to resiliency to non-COVID patients. So the pandemic provided a real challenge to hospital resiliency and to bystander patient care. That is to say all the patient care we wanna deliver. And so we have much to learn about preparing our healthcare delivery systems for acute surges in demand of which there will be more. So I will stop there and thank you very much for the opportunity and I look forward to discussion.

George Nisi

medical intensivist

acute COVID surges

hospital stress

healthcare capacity strain