 Let's get started on today's Health Policy and Bioethics Consortium. I want to welcome everybody to the first Health Policy and Bioethics Consortium of the Year. My name is Aaron Kesselheim. I'm a professor of medicine at Brigham Women's Hospital and Harvard Medical School. And I am the director of the program on regulation therapeutics and law, which has the honor of helping organize and co-hosting these Health Policy and Bioethics Consortium with the Harvard Medical School Center for Bioethics. And if you want to tweet about anything that you hear about or see on this talk, you use the hashtag, hashtag policy ethics that you can see there. And you can have a conversation going over Twitter. I just want to give a little bit of background on the organization of this portal, the group that I run is a research center focused on studying the intersection of law therapeutics and regulation in public health. We are an interdisciplinary group with experts in medicine and business, law, ethics, epidemiology. And we run courses at the Harvard Medical School Center for Bioethics, as well as the School of Public Health and Down at Yale also at their law school and school of public health. If you want to learn more about portal, you can go to our website at portalresearch.org. But you can see there at the bottom recent articles and blogs that we've authored. And we also try to get involved in a lot of policymaking and policy discussions around the issues relating to the pharmaceutical market and other regulated products. You can also follow us on Twitter at portalunderscoreresearch. And you can join our mailing list by checking us out on our website or go into this short website. So one of the things that we've done at portal for the last five years is organize these consortia, which occur monthly on the second Friday of every month during the academic year. The goals of the consortium are to articulate key issues in the health care system and public health that involve ethically challenging policies or practices to bring together experts with different perspectives or experiences to consider and propose solutions to these issues and to stimulate conversation and further academic study that will advance the field. And just to give you a little bit of a taste of the policy and ethics consortia to come, in October we'll be talking about drug shortages in November, concussions in youth sports, and then in December, accessing experimental drugs and stem cell treatments. And so for today's session, which is a fantastic session, couldn't be more timely and we couldn't have better speakers to share with you, I want to first introduce our moderator, Leah Rand, who is a postdoctoral fellow in the portal group. Leah got her doctorate and applied ethics from the University of Oxford and studies intersections between bioethics and health care with the focus on justice, fairness, and health care access. And apropos to this topic in particular, before she was at portal, she worked at the National Academies of Science, Engineering, and Medicine and helped write a book about public health emergency preparedness. And so Leah is going to introduce the topics for today and introduce our guest experts as well. So take it away, Leah. Great, thank you, Erin. So our topic today needs little introduction since the SARS-CoV-2 virus and the pandemic have affected all of us, changing how we live day to day, interact, and affecting the health of millions of people around the world. So as of this morning, there have been over 20 million confirmed cases and 900,000 deaths globally. Now, the many issues and topics relevant to COVID-19 today we're focusing on the challenge of making decisions when there is uncertainty, uncertainty about the pandemic and its future impacts, and moral uncertainty about what actions are right to take. We can characterize the pandemic context as one of great uncertainty and rapidly changing information. Since December, what we know about how the virus is spread, who is at greatest risk, and how to reduce transmission has changed and improved. Meanwhile, there's ongoing research to develop pharmaceutical interventions, including treatments and vaccines. And against the backdrop of this uncertainty is an urgent need for action at the federal and state levels to protect health. In different countries in US states, we've seen different policies and responses to whether and when to close borders, restrict travel, require masks, and when and how to reopen businesses and schools. So in public health ethics, there's the so-called precautionary principle, which is often invoked to highlight that uncertainty cannot justify inaction when there are risks of serious harms or put another way when there is a serious threat of harm. Reasonable cost-effective interventions to prevent that harm should be enacted, even if we are not confident in the available information. But with COVID-19, we've seen a divergence on what people think are reasonable interventions and even on what the greatest harm is, whether it's the direct health loss, economic impacts, or restrictions on individuals' liberties that could result from the virus or the actions taken to prevent its spread. So two ways to approach these challenges and decisions will be the focus of our discussion today, how we know and think about what will happen and how we decide what actions ethically should be taken. Guiding policy decisions so far has been the work of epidemiologists, analyzing data and using models to predict the future trajectory of the pandemic and the likely effectiveness of interventions to reduce transmission. Models can provide valuable predictions of what will happen, but there have also been notable examples of models that under-overestimated the impact of the virus early on, bringing the resulting policy decisions into question. Models will continue to play an important role in managing the response to COVID-19 and informing future actions. So how should policymakers make decisions on the basis of model predictions? In addition to needing to make decisions when information is limited, policy decisions like whether to require quarantines or close businesses face moral uncertainty. What are the acceptable trade-offs between liberty, livelihoods and protecting the public's health? And if that wasn't a big enough question to ask, the pandemic has made this debate how to allocate healthcare resources with new guidelines for who should be prioritized for ICU beds and ventilators if hospitals are overwhelmed with cases. Now we're beginning to consider who should be vaccinated first once a safe and effective vaccine is developed. These are big questions about what we owe each other that have been asked for centuries, but they are especially urgent right now. So I'm glad that here to share their thoughts and expertise are doctors Mark Lipsitch and Matt Winnie. Dr. Lipsitch is professor of epidemiology at the Harvard University T.H. Chan School of Public Health with a joint appointment in the Department of Immunology and Infectious Diseases. He directs the Center for Communical Disease Dynamics. His research concerns the effect of naturally acquired host immunity, vaccine induced immunity and other public health interventions like antimicrobial use on the population biology of pathogens and the consequences of changing pathogen populations for human health. He is an author of more than 300 peer reviewed publications on antimicrobial resistance, epidemiologic methods, mathematical modeling of infectious disease transmission, pathogen population genomics and immunoepidemiology of streptococcus pneumoniae. Dr. Lipsitch is a prominent voice in the novel coronavirus research community with six peer reviewed publications and four manuscripts currently under review on the topic. He has also been active in science communication on the subject with a dozen op-eds published since the start of the outbreak. Dr. Lipsitch received his BA summa cum laude in philosophy from Yale and his D. Phil from Oxford as a Rhodes Scholar. He did postdoctoral work at Emory University and the CDC. Honors include mentoring awards from Harvard Chan, the Robert Austrian Lectureship and the election to the American Academy of Microbiology. Following him will be Dr. Wynia, who is director of the Center for Bioethics and Humanities at the University of Colorado and professor of medicine at the University of Colorado School of Medicine. His training is in internal medicine, infectious diseases, public health and health services research. From 1997 to 2015, Dr. Wynia worked at the American Medical Association, AMA, where he developed a research institute and training programs focused on bioethics, professionalism and policy issues at the AMA Institute for Ethics and founded the AMA Center for Patient Safety. He also practiced the University of Chicago. His research has focused on understanding and improving the practical management of ethical issues in medicine and public health. He has led projects on a wide variety of issues related to ethics and professionalism. He has served on committees and expert panels and as a reviewer for the Health and Medicine Division of the National Academies, the Joint Commission, federal agencies, the Hastings Center, the American Board of Medical Specialties and other organizations. He has delivered and held more than two dozen named lectures and visiting professorships nationally and internationally. Dr. Wynia is the author of more than 140 published articles, chapters and essays. His work has been published in numerous leading medical and ethics journals and he is a contributing editor of the American Journal of Bioethics. He is past president of the American Society for Bioethics and Humanities and has chaired the ethics forum of the American Public Health Association and the ethics committee of the Society of General Internal Medicine. He holds current board certifications in internal medicine and infectious diseases. So I'm delighted to welcome both of you to speak with us today. Drs. Lipsitch and Wynia will each present and that will be followed by a moderated Q&A session. So at the bottom of your screens you can see a button for Q&A and please type your questions into the box and we look forward to hearing what you have to ask and hearing from Drs. Lipsitch and Wynia. Shall I begin? No, I can't hear. Yes, you can begin. Okay, sorry, little zoom hiccup there. You think by now. Well, thank you, Aaron and Leah and colleagues for the opportunity to be here. I think ethical issues and the sort of decision quandaries that we face are under appreciated dimensions of this pandemic. I think ethical dilemmas arise a lot in any crisis. Better decision making can make certain kinds of ethical dilemmas disappear because they arise sometimes from the bad decisions that we've made in the past. And I think that's a theme that I'm not sure if I'll directly touch on but has been very striking that the decisions become harder as the policy responses become worse and conversely the better decisions under uncertainty not only are themselves an ethical contribution but they also help to prevent their sort of preventive medicine for ethics. Their better decisions mean that there are fewer hard problems to solve. I would propose and we'll see if people agree with that by the end. So after the 2009 flu pandemic a large group of us held a conference at Harvard School of Public Health where we talked about the lessons learned from that pandemic. And I'm gonna talk a lot about 2009 in this talk as well as about COVID in part because we have the classic problem of fighting the last war in thinking about all pandemics and making the lessons learned those that were relevant to the previous one. And we can't do better. We can't know what the future will hold but we can at least learn from multiple past pandemics and from the contrasts and maybe in those ways fill in some of the possibilities that we haven't yet experienced. So when we talked after 2009 we discussed what sorts of decisions at the top have to be made in pandemics and what sorts of evidence in the middle might inform those decisions in an ideal setting and what sorts of inputs from basic epidemiology and surveillance might inform those types of evidence. And we made a really ugly diagram which then was improved in a subsequent publication and this is the improved version. And what it shows is that there are a number of basic pieces of information, number of case counts, various forms of surveillance and outbreak investigations. There are a number of approaches to understanding the basic data which can be synthesized into evidence and then help to inform decisions. And in a well functioning public health system something like this obviously modified by circumstance but something like this would be how we would react. That has not been the case in this pandemic in the United States as we're all aware. If we weren't aware then we're certainly aware after the recent news coverage. But I think it's important to realize that it's not just a failure of leadership and that would be a topic of its own specific to the United States and certain other like minded countries. And I don't really wanna focus on that although it's crucial because I think as we think about the future and of this pandemic and of other pandemics it's important to understand the sort of structural reasons why even good decision making well intention decision making is hard. And the reasons largely are listed on this slide which is importantly that every decision has a long lead time before you see its results. So that means that we need to make them early and we therefore need to make them before we have good evidence about how well they'll work. And these are decisions like vaccine development and purchase lockdowns and other control measures, contact tracing. And this has been true for a long time. The classic book by Newstatt and Feinberg on the swine flu, the epidemic that never was as that book has also been titled in other editions. There were many of these issues and we wrote about them also in 2009 as that flu pandemic was emerging. So this is not unique here and these challenges even with good leadership would have been real. So in such a setting we need as many types of inputs as possible to inform our decision making weighted appropriately and we could talk about how that might be done. And I think important and often forgotten in the scientific discourse is reasoning from historical analogies. In the case of flu, the previous pandemics in the case of this coronavirus previous the previous epidemic of SARS non-pandemic versions of the same viruses including seasonal coronaviruses and even other infectious diseases. And these have a lot of lessons. They need to be interpreted carefully but we should not forget what we know about the past because sometimes our models are a little too clever and a little too historically unrooted. Clinical and public health experience of course and political considerations and the need to do something and be seen to do something whether or not it's known to be effective. Everyone carries an implicit mental model of what's going on and what might be the response to interventions and then explicit mathematical models can help to sharpen and hopefully improve certainly increased confidence but not always even increased confidence just to let's say improve in the best case and we'll see an example at least. I think as people who are mostly working as scientists many of us are used to a sort of approach to uncertainty where if we say that our knowledge is inadequate then the answer is to do more science and be agnostic until we know and scientists are sometimes known and even caricatured for being very cautious and what caution means for scientists in normal times is to avoid type one errors that is accepting something as true for which the evidence is inadequate and so we default to agnosticism but we and we require a very high standard of proof. When we're doing decision making where it's not simply a decision a question of do we believe something is true but how do we act then there's no wheel then the default to agnosticism doesn't work the consequence of judging a model or a knowledge of a topic to be inadequate is that we have to rely on something else to make the decision because there is no such thing as just remaining neutral you have to decide how to respond even if that is to do nothing and so caution dictates protecting people and maybe saving money and saving the economy, et cetera and therefore agnosticism is not really an option it's just paralysis and I've contrasted these two views I think in the interesting exchange in the Boston review that I participated in with Jonathan Fuller and John Ioannidis an interesting discussion for those more in medical ethics and public health ethics is how these types of thinking map on to medical decision making but I'll leave that aside for now what I wanna talk about mainly though is the challenges the specific challenges in epidemics and some of the things that we do to try to get around those or find out if they're really relevant as relevant as we think and these are some themes I'm gonna talk about why the data are usually terrible how those limitations can influence our conclusions I'll talk about a paradox that I it's a big word, fancy word for it's maybe not a paradox but a surprising thing that I discovered in the last two months and then I'll talk about evaluation metrics and other considerations outside of models that can sometimes help us move forward even when we're not sure so physical scientists often don't realize how bad and how unstandardized data from epidemiological models are so if you do physics or even economics or meteorology you measure the temperature and everybody has a pretty well calibrated instrument they may be slightly different but they're pretty much in space and time and you're getting a consistent reading that means the same thing when you see it as it meant the day before and whoops in a crisis especially but even not in a crisis epidemiology is a bit different and it's really the analogy I've given which is a little harder to put on a slide it's as if everybody was every county health department was designing and recalibrating their thermometers every day and you were trying to figure out what the temperature was we're all using different instruments at every level of geography and over time as the testing and et cetera changes even putting aside the variation and what it is we're measuring also the problem of delays and the timeline goes for many infections including this one from infection to symptoms and then to sometimes to hospitalization sometimes to ICU admission and death and each of these has its challenges which are listed below and in particular the events that we observe better because they're more severe and more important such as deaths tend to be lagged several weeks from the time of infection and since we're trying to control infections by our countermeasures a lagged but pretty reliable indicator like death or hospitalization is good in that it's reliable but as bad in that it's lagged and makes it harder to control and many of us I think have sort of settled on hospital admissions as perhaps the best compromise for many purposes what's infection activity but it's not far from perfect the delay distribution is not even constant these are the delays from a particular health reporting system from symptom onset to self reported to the time of diagnosed to the date that the test was performed and you can see widely varying and sometimes very long delays that has consequences for control and this is the first data that I'm showing from work that we actually did during this pandemic this is just a display of the data from Wuhan where the demand for critical care beds or intensive care beds is shown in orange the deaths by day are shown in black so this is a prevalence measure this is a new deaths measure so different types of measures but the peak demand for critical care intensive care peaked right around the total capacity per capita of the United States for intensive care in Wuhan but it peaked exactly a month after the lockdown of January 23rd in Wuhan so for various reasons some of which we understand and some of which we don't the draconian measure to control the epidemic was here and at a time when things looked pretty good I mean obviously they were bad enough to lock the city down but there was not a health care crisis yet but by the time it peaked there was a health care crisis so this makes control much more challenging it also makes it much harder to analyze data and here I'll use an example from the H1N1 epidemic pandemic of 2009 so on May 4th in Mexico this disease emerged in late April early May of 2009 the one of the earliest reports was that there had been about 500 cases of 19 deaths or 4% case fatality rate in the US on that same day there had been about a thousand cases and one death and because of our experience from the SARS epidemic we knew that there were two countervailing biases in these data one was the notion of censoring bias that we are missing deaths that haven't happened yet meaning some of these people were confirmed yesterday and they're going to die but they haven't died yet and on the other hand especially in Mexico probably the mild cases were not being detected and so we were overestimating severity by underestimating the total number of cases when we saw only the severe ones and so you got in the literature within 12 days of one another some colleagues in New Zealand guessed that the true case fatality rate was somewhere well south of 0.1% maybe considerably further below that while the imperial college group estimated that it was probably somewhere between 0.2% and 1.2% the New Zealand estimate turned out to be correct and it turned out essentially that this bias of not detecting the mild cases was swamping the more moderate bias in the other direction in that case opposite of what had happened in SARS and that's why it was so confusing the same issue arose at the beginning of the COVID-19 pandemic I'm not going to go through this graph in detail but this is our paper which we wrote during the Ebola outbreak in West Africa to try to discuss some of these issues of bias in the case fatality risk and this was an op-ed that the Washington Post asked me to write in May, I think it was May maybe even earlier than May to explain essentially these issues because the numbers were coming out they were not being very well contextualized and it was very strange to essentially adapt an article from Plosnicoactive Tropical Diseases to a Washington Post op-ed but that's more or less what I did the there are various ways to try to address these challenges of estimating severity one that we tried in 2009 that ended up providing a fairly definitive and I think lasting estimate was this sort of idea of gluing different estimates together of severity so in that pandemic the problem was that there was no it was so mild that there was no jurisdiction that had large enough numbers of deaths to do statistics on and also had really counted very well the lower levels of this so-called severity pyramid so what we did was to use different data sources and statistically glue them together in an evidence synthesis framework to try to make estimates of the risk of dying which are here and under different assumptions get essentially the same range that the New Zealand colleagues had guessed at earlier on this also makes these delays also make decision-making harder in that in that as we saw with the Wuhan case there is a delay between when you know about the problem and respond to it and when you and when the problem stops being so bad so this is some of the early modeling that we did on repeated rounds of social distancing in COVID-19 and the blue bars are periods of having social distancing on and the white is periods of having it off and what you see is that the red is the intensive care demand and the black is the number of cases and so essentially you have to turn back on social distancing here at a time when the intensive care demand is not too bad because it's going to keep growing before it gets better so it's essentially the same problem I mentioned with Wuhan but in a model now and it also complicates analysis of transmissibility and this is another example from 2009 where we looked at different ways of augmenting the data to try to correct for these delays and the details are unimportant but when we were trying to estimate the reproduction number the R0 of that flew what we found was tens of percent of variation depending on how you modified the data to account for our best estimate of the biases and these may or may not look like big differences to you but given that this is a question of how much control is required to stop the epidemic a 40% difference is a really big difference so these turned out to be quite important one part of the solution that we've contributed to and others have contributed other approaches to is this notion of now casting which is essentially a formal way of asking, answering the question if we know about a certain number of cases today and in the recent past how many cases will we eventually know about that occurred today and in the recent past so it's essentially trying to fill in these missing data of cases that have not yet been reported and as it happens we were working on this software approach or this statistical approach and software to do it before COVID came Sarah Magoo had graduated and gone off to an industry job and her bosses at Genentech very generously gave her a week of pro bono time to put all of this into our package that could be used by public health departments around the world as it is now being used so there are a lot of examples but I think you get the point the meaning of data changes the next point I want to make is that sometimes the most important contributions to policy are just expressions of uncertainty and this is a quotation from ProPublica which did a big story on how New York City and New York State have responded to the pandemic and they describe this thing they call it the Lipschitz model that's very grand let me show you what it actually is we had at the request of New York City created a shiny app a web base tool that people could use to make estimates of what this is going to work of what of how many cases there might be in the United States given the number that we knew about I'm not seeing how I can share it but what we did was to let people set to let people set what they thought was true of the epidemic spread lots of uncertainty at that time and still but especially at that time about how many introductions there might have been what proportion and so on various randomly generated scenarios in which those assumptions were made and so there was not really this was not deep science this was saying we don't know almost any of these numbers but if you assume certain numbers and assume that the epidemic progresses in certain ways then you might have very few cases in a few months from now under the same circumstances with bad luck might have a lot more cases and for New York City that turned out to be a very valuable statement not because it was scientifically novel but because it but because it said that not having a problem yet that they could see should not be reassuring that there may be a problem you don't see or absence of evidence is not evidence of absence and so only in retrospect was it at all clear to me that this was a useful thing to have done we did it because we were asked to and it was pretty straightforward to do but it turned out that this was one of the more useful things I think we did simply to give the warning that things might be worse than they look because of lack of detection then the last area I want to talk about is the importance of decision rules and here I want to go back to the question that didn't happen happily but was much debated in the early 2000s which is how we would respond to or prepare for and respond to the risk of a bioterrorist smallpox attack so the notion was a terrorist may attack us with smallpox and the three choices were that we could vaccinate the whole country against smallpox now we could do that only after an attack happens or we could use so-called targeted vaccination or ring vaccination sort of test trace vaccination after an attack to vaccinate the contacts of infected people the way it was done in the eradication campaign and for political reasons the first option seemed implausible although Vice President Cheney was pushing for it it never really became reasonable so two and three were the two options and so the assumptions about smallpox were that without vaccination there would be millions of cases and deaths if there was an attack but the mass vaccination would lead to almost certain death of a small fraction of vaccinated people due to adverse reactions to the smallpox vaccine that was available then if we mass vaccinated it was assumed only few secondary cases would occur because we would have herd immunity a term that's now in common parlance but then was still specialized but the more vaccines we used the greater the cost so with trace vaccination we might do fine but there's a risk larger than the risk with mass vaccination that it could fail leading to a large epidemic that killed millions infected in millions and killed many the tradeoff then is between the certainty of needing more vaccines with mass vaccination and more cost and adverse events but a lower risk of a large epidemic and there were dueling models that were proposed during that time and there were all sorts of debates about which parts of the models were right and wrong and uncertain etc but I think despite some serious problems with the model that was used by by Kaplan and colleagues from Yale I think the most important diagram that was ever made for this question was this one and I've added some pieces to reflect what's described in the text so what they showed is that if you think that there is likely to be a small enough initial attack and that the transmission is likely to be low enough then the optimal policy is trace vaccination and on the other hand if you think that it's likely to be a large or highly contagious attack then the optimal policy is mass vaccination but what they pointed out was that the difference that if you're wrong the risk of being wrong if you do trace vaccination when you should have done mass vaccination is that you get millions of deaths because mass vaccination would have prevented runaway epidemic but trace vaccination is inadequate and the risk if you do trace mass vaccination when you should have done trace vaccination is that you have 10 to the 2 adverse events we could talk about there's also a probability in there no, there's not a probability this is all after the attack so it seems to me that one of the clear messages of this is that there's a lot of uncertainty about parameters and the optimal decision when you think about the possibilities and integrate over those possible uncertainties it's very hard to argue for trace vaccination because the risk of being wrong is far, far greater another analysis that was published after that in science made a case for trace vaccination but using a very weird metric in my opinion which is cases prevented per dose so it's a sort of efficiency metric but the question is why would we care that much about using a lot of doses if this was stopping a large attack so that seemed to me to be an answer to a question that we probably shouldn't need to ask so I think it was otherwise a better model in the science paper so with this problem the capital model had lousy assumptions but I think a better metric of trying to minimize the maximum badness or minimize the chance of the worst outcome and I think the decision rule of Howarin et al was less relevant so I think some takeaways are that early action is often necessary despite uncertainty the data are not what we would like but we have to use them and that we should use decision analytic principles simply scientific principles to make decisions and that really means taking into account the possibility that you what happens if you're wrong some ways I think we can improve rapid decisions under uncertainty in the short term is to recognize that no decision is permanent fast does not mean unchangeable and we can build in points to re-examine and off-ramps I think that term was coined by this book on the swine flu pandemic that never was in 1976 we can invest in lower cost lower risk preparations now to preserve options for the future for example we've advocated preparing for human challenge trials and manufacturing others have pointed out the need to manufacture vaccines when we're not sure and we need to respect science but not always wait for it to be perfect in the longer term we need to improve our surveillance capacity we need to improve data sharing across jurisdictions and then bridge the educational gap between the sort of decision theory that we teach in somewhat mathematical and specialized classes and public health decision making more broadly so I will stop there and thank you for your attention thank you and next we'll be hearing from Dr. Winia and meanwhile during the presentations if you have questions Reiner just put them into that Q&A box what are you seeing on the screen am I sharing my slide or am I sharing yes I can see great so I'm going to reflect a little on some of what Mark has said there are some areas that I might have touched on but he did such a nice job of describing the uncertainty associated for example with estimates of morbidity and mortality in pandemics that I'm going to skip over some of that I'm going to frame this around a few issues but I'm going to start with a poll question so if we can get the poll question up and I'll just acknowledge number one this is a question that other people have already used in the past and that's why I'm using it it's not a perfect question and it has an obvious sort of false dichotomy baked in because I think one of the real sort of points of frustration for many people today is that these were supposed to go hand in hand and we made enormous sacrifices as a nation and many individuals made personal sacrifices economic sacrifices health sacrifices health sacrifices in order to slow the pandemic so that we could reopen sooner and the fact that we bought time we bent the curve we flattened the curve and then failed to take advantage of that to be able to safely reopen the economy is probably one of the places where people are most most frustrated today looking back on the last months people feel like you know what we did this back in March and April we did what was necessary as a nation and then and then we sort of didn't follow through in the way that we might have I will say the results here are maybe not unexpected given that this is something like a public health crowd but it's also not that far off from what you get from the general public well over 80% of people when given this forced dichotomous choice will choose slowing the pandemic over slowing the economic decline so let me get started here I don't have any disclosures to make I am going to be referring to the two reports one of which Erin mentioned at the beginning that Leah was very much involved in came out at the end of August on the evidence basis for public health emergency preparedness and response and the second which I probably won't talk about because Mark did such a nice job of this but this just came out on Wednesday of this week and it's a framework for assessing morbidity and mortality after large-scale disasters and this was actually a FEMA funded report that looked at how do you tally morbidity and mortality came out of the very discrepant assessments of mortality from Hurricane Maria where the direct mortality from flying debris and drownings during the hurricane were something like 64 people but when teams at Harvard and Georgetown did more comprehensive assessments it looked like closer to 4,000 people had died so that sparked the need for some assessment of how should we best count up the number of people who died during disasters I'm going to use some historical examples to explore what I call the three Rs of ethics and epidemics and I'm going to avoid the two historical examples that everyone talks about which is the 1918 pandemic flu and the 2009 pandemic flu and instead talk about maybe some more obscure examples and I'm going to spend a little bit of time focusing on a few guiding values these as I will say are not the only guiding values but they're the ones that seem to have gotten the most attention recently and they are certainly very important I could speak a little bit about research which I suppose would be a fourth R of ethical considerations during pandemics but we'll save that for the Q and A so let me start with an old pandemic this is a pandemic that arose in the 10th through the 13th century it swept through Europe many people were affected mostly in overcrowded poor communities people developed patches of skin which would grow and become entirely anesthetic so they could not feel pain in these areas they would then get an infection end up losing their limbs the bridge of the nose would often collapse eating along with build up of this bacillus under the skin to what was called a leonine face or a lion face and people with this illness were often segregated into colonies because they were seen to be the word that was often used was unclean so Hansen's disease or leprosy was an early example of a pandemic and a terrific example of the ways in which people with a contagious illness are stigmatized are feared, are segregated out and some of the sort of moral panic that can arise around a pandemic it also is an excellent example of the poor being hit hardest we had a number of early outbreaks in Colorado that arose in Aspen in Breckenridge and Vale in Tony Resort Ski Towns but that's not where it stays and I heard this actually from a reporter from the LA Times who was tracking down the wealthy residents from Mexico City who came to Vale got infected and took it home to Mexico City where of course that's not where it stays it then goes into the neighborhoods where people can't afford to travel to Vale and she said the very similar thing arose in Los Angeles where the first few cases in the Hollywood Hills because those are the people who can travel around the world and pick up a virus that they can then bring home and transmit to the housekeeper and then it's in a different neighborhood and that's of course where the majority of the illness and deaths have arisen this is not new and the idea that pandemics will affect the poor and will lead to stigma, fear, blame, ostracization the great pox you may not be familiar with this term but small pox is in contradiction to great pox great pox was syphilis syphilis in its first iteration across Europe caused a very aggressive disease many people died from syphilis in a fairly rapid fashion in the secondary stage when giant boils would form on the skin it has obviously changed in its nature since then but great pox was in some circles blamed on Columbus but the reality is every place that had great pox blamed it on someone else so the Italians called it the Spanish pox and the French called it the Italian or Neapolitan pox and the English called it the French pox the English people called it the German pox and the Russians called it the Polish pox so everyone wants to blame a contagious illness on someone else and this is again a strategy a human response and it is something we're seeing again today the idea that you can isolate this into islands just like Hansen's disease also very similar and there was what was framed as a medical controversy in the US over whether it was appropriate to treat people with syphilis and the argument against came largely from the religious community who felt that you were interfering with God's will that people who developed syphilis deserved it and that if we saved them we were just bringing people back from the appropriate punishment from their sins this level of fear and blame generates of course a great deal of interest in protecting oneself and my favorite story in this regard is the very famous Doctor's robe which was used during the great plagues of Europe in the dark ages essentially and I love this quote because the Doctor's robe the most famous part of it is the beak of course which was a hollow beak that would be filled with either herbs or spices or sometimes rags that had been dipped in cat urine the idea was that something very pungent would prevent the humors from the house getting into you and so because the humoral theory of disease was the prominent theory at that time and this priest in Italy complained that the robe which was associated with this which was a muslin robe which was made of wax so it's a very hot sticky heavy waxy thing that would drape all the way down to the floor ostensibly to prevent the vapors from getting in from the room into the person inside and this priest said this was very uncomfortable and useless against the plague the only thing it did was to protect from the fleas and this is where I really miss having a live audience because you would be cracking up right now those of you who realize plague of course is transmitted by fleas but I can't see or hear any of you so let me talk about self-protection which thank you Mark so self-protection blends quickly into questions about restrictions on liberties and how much are we allowed to restrict the liberties of ourselves and others in a democratic community in order to protect the larger community and this of course is an old question in 1984 there had been a series of countries around the world that had implemented martial law and other forms of very strict civil liberty restrictions in response to ostensible threats from the outside and so leaders who were essentially authoritarian leaders were using public health and military threats as a way to implement martial law so a group of lawyers for the United Nations got together and produced what are now called the Syracuse of Principles saying that when you're going to use coercive public health measures they have to be legitimate necessary legal, non-discriminatory and the least restrictive means and that phraseology has become very widely used this sits in contradiction in some ways to Mark's earlier point which is really about the precautionary principle right and so these sit in balance with each other the idea that you want to do everything necessary to prevent really terrible outcomes the more terrible outcome you can envision the more restrictions you can justify and at the same time you have to be sure that what you're doing is not unnecessary that what you're doing is respectful respectful of individual liberties and rights just by the way very similar to the principles in the US Constitution right the 5th amendment 14th amendment due process so the idea that a state can only impose restrictions on its on its constituent individuals if the state has a compelling interest if they're using the least restrictive means if there's a due process for appeals and so on is very American in fact the Syracusa principles are this principle of proportionality turns out to be the one of the more commonly discussed and deliberated and debated in the whole arena of pandemic response it comes up with regard to restricting liberties but it also comes up questions around resource allocation so for example when you're performing triage and you've got a limited supply of whatever resource the principle of proportionality requires that you do repeated assessments based on a repeated understanding of what your resource limits are right now because God forbid you tell someone we can't give you what you need right now because we don't have enough and 10 minutes later or two hours later a new shipment comes in and you don't go back and reassess that person because they've already got a black tag on their toe and this is a very real consideration in disasters so this is Louis Armstrong airport after Katrina and the DMAT team there actually black tagged a fair number of people as being triaged to receive comfort measures only but very few of those people actually because they kept going back to them and saying we can now evacuate you or we've received more resources we can actually take that black tag off your toe and put a red tag on your toe so being willing and able and recognizing the necessity of repeated assessment is one of the principle principles of resource allocations let me put up this second question because I know Mark has worked on this I'm sure Erin's worked on this a number of people have worked on this question of who gets to the front of the line if there are not enough ventilators to go around and this became a very hot topic back in March, April, May even June about whether or not healthcare workers and keep that broad about physicians here but people taking care of sick patients in a hospital should they get some kind of preference in prioritization and this is a hard call question that we're doing in an episode of our podcast hard call on this exact question because it turns out to be quite controversial as many folks around Boston and Massachusetts have learned the hard way these questions are not without some level of controversy and it looks like within this audience we're looking at a little over 50% who say yes another third say no I guess closer to 60% or so are saying yes and another third or so are saying no so let's go to the next slide here one of the mistakes I'm trying to close this sorry one of the mistakes that I think folks sometimes make is to oversimplify the ethics in disasters when it comes to resource allocation and triage and there's a sense sometimes that in normal medical ethics we've got these four principles we've got balancing act that's going on but in a disaster in a catastrophic disaster all of that goes out the window and the only thing that matters is just save as many people as you can and that unfortunately doesn't work and is morally fallacious reasoning in my view for several different reasons starting with the fact that it's really hard to save the most lives there are not great metrics to know when you are saving the most lives and it almost immediately devolves into questions about well are we really talking about saving the most lives or are we talking about saving the most life years so do you give priority to people who have a longer life expectancy in front of them versus people with a shorter life or a contentious issue of well what about the quality of those life years what about other principles like women and children first which is really reflective of the idea of saving your community so that the community can live on beyond just saving the most lives and what about things like the people who show up first and what about the ability to pay which is in fact how we distribute healthcare resources often in the U.S. and the point I want to make is not that one of these is right or wrong I do think some are more right and some are much more wrong the point is that there are a bunch of principles that come into play some are substantive type principles where you might be able to develop a set of guidance around those principles that would tell you exactly what to do but I think much more important is what we have used when making allocation decisions and that's where the issue of proportionality equity transparency accountability and so on come into play and in reality multiple values are important even in a disaster so yes you would like to try and save the most lives but the way we put it in the Colorado state guidance on crisis standards of care is while sustaining social cohesion while retaining some level of trust in our healthcare system and while preserving the ability of our community to come together and heal in the wake of the disaster and those types of values matter as well as the desire to try and save the most lives we put up the last question here which is about who's lending us away from rationing and into questions of individual liberty so people who choose today not to wear a mask and they get infected should they be deprioritized in the resource allocation protocol and this is something that I think Art Kaplan actually wrote a little essay about this suggesting that this might be a reasonable approach but I'll say consistent with what we're seeing in the results here about 60-40 most folks in the healthcare world in the medical care world say you know what people do stupid stuff all the time and we don't punish them for it when they show up in our emergency department so I think in the medical ethics arena the answer to this is relatively clear in the public health arena in a disaster in a resource shortage situation I can imagine people trying to make this argument I feel like personally I feel like this would be a very difficult cell for clinicians who are very used to the notion that we don't punish people for the bad decisions that they make we don't use healthcare as a tool of punishment so let me go on to the last a little bit here about personal liberties and public safety Ron Bayer sort of famously back in the 90s during the sort of height of the AIDS epidemic said that the ethos of public health and that of civil liberties are radically distinct and yet I think even Ron agrees now that there are a number of ways in which attempts to restrict personal liberties can actually backfire and lead to worse outcomes in terms of public health we talked about this in the evidence based practices for public health emergency preparedness and response report which came out a couple weeks ago and we described the fact that people who think they're about to be placed in quarantine or if there's some other sort of public health intervention that is going to force them to do something may actually get their hackles up and become increasingly defensive and flee the area and then entirely you know sort of hypothetical response during the SARS epidemic there were a couple of very good examples of this there was a rumor early on that all of Beijing would be quarantined and 245,000 migrant workers fled the city now if some number of those people had had SARS when they left the city that would have been a quarantine effort that backfired because it would have spread SARS all over the country we got lucky in that instance Hong Kong's Amoy Gardens apartment similarly the site of the first outbreak in Hong Kong it was in fact placed under quarantine and when the police arrived to enforce the quarantine half of the homes there were empty so people left rather than stick around for quarantine and without being stereotypical I will note that these are ostensibly communitarian cultures where people obey the state authorities you can imagine what might happen in the United States if a military style quarantine were attempted to be enforced say here in Colorado where most people have both an SUV and a gun and this same dynamic by the way played out in Wuhan when Wuhan was locked down about 5 million people were not there normally there some of those had been traveling anyways because it was around the Chinese New Year but some of them presumably fled the city to avoid the lockdown so could something like a mask mandate backfire in this same way and I'm not going to say one way yes or no I will acknowledge at the outset by the way I'm not arguing against mask mandates they are absolutely protective these are my two favorite pieces of evidence in that regard this choir practice in Skaggett Washington where one person standing way up in the back row infected almost every other person in the choir over a two hour choir practice none of them wearing masks of course and on the contrast are the two hairstylists who saw and cut the hair of 140 clients which by the way puts them in very close contact obviously with all of these clients and not a single one of the clients became infected because they were both wearing masks they were not wearing N95 masks these were paper masks and cloth masks so masks are absolutely are absolutely effective but forcing people to wear a mask could actually backfire because it's a very visible symbol of your agreement with the government's recommendation or mandate and unfortunately that turned out to be played out for a while at least is that people were saying well the CDC didn't say it now the CDC does say it why are they giving us mixed messages and this is a cabal and you're just trying to use this public health emergency to infringe on my personal liberties and I thought Angela Duckworth Z can really nice piece for the New York Times about how to talk about mask mandates with people who initially raised those kinds of objections so I'll just refer you to that and finish with a reflection on the thing that Mark also ended with which is the possibility of politicization of public health information smallpox vaccine program is an excellent warning of this. It was a really nice IOM report about this back in 2005 or six and I wrote an essay about it for the American Journal of Bioethics sort of summarizing what they said which was that the effort to do broad based vaccination for smallpox was largely driven by a political effort to drum up support for the war and the CDC and a number of scientists were unwitting accomplices in the attempt to drum up support for the war and the minute the war got started the whole CDC the whole vaccination program essentially stopped with no explanation and the CDC's reputation was according to this IOM report harmed as a result because the public and health care communities started to wonder is the CDC part of this political apparatus rather than being the independent scientific and public health entity it needs to be and I'll just I'll close with that because I think that is very much at the forefront today people today are looking at the FDA they're looking at the CDC and saying you know can we salvage the reputations of these really historically important and critical organizations for national public health and what will it take to salvage and to rebuild the reputation of trust that both of those organizations really deserve a need in order to be effective right well thank you thank you Dr. Winnie and Dr. Lipschitz for great introduction to these topics and raising lots of good questions and so as a reminder if you're in the audience you can type up your questions in the Q&A box and we'll try to get to as many of those as possible in the remaining time but I am going to take the moderators privilege and start by asking a question I have especially listening to your presentation mark which is that with all of this uncertainty about decisions and you've both brought up the cautionary principle and trying to weigh up harms on whose side does the burden of proof fall so if a model is predicting certain harms in the future is it to a state governor to refute it and show why they think the economic harms might be greater for example or is it on a public health researcher to show that the health impacts will be greater due to the disease than something else I think there are at least two layers to the question one is sort of the the decision maker versus the academic and the other is kind of the known versus the more known versus the less known and I think I don't have a clear and crisp answer I think what decision makers should be getting and I think it's been quite chaotic in this country partly due to the sort of near disappearance of the networks of academic groups that were that were providing advice on infectious diseases to government and partly due to the fact that there just hasn't ever been a really good integration between those two groups even when there was more funding the the path of information has been chaotic and it would be very much to ask of any government or even the federal government especially state governments to fully digest all the ideas and all the calculations that are being made and so I think the process of sort of scientific consensus and bringing in multiple scientific voices from from each discipline and then across disciplines is important you know in the discussion about lockdowns versus not which Matt nicely dichotomized in a way with his first poll question but reflecting a dichotomy that neither of us believes in and that I think much of the Twitterverse does believe in you know I think that's a case where the the successful control of the virus it appears examples in other countries leads to a better economic outcome and where the leaving it to spread uncontrolled is bad for health and bad for the for the economy and what we seem to have done is to do ineffectual lockdowns so we've suffered the economic losses due to the virus the health losses due to the virus and the health losses due to the lockdown so we've in a way had because of our chaotic response the worst of all worlds so I don't have a really good answer except that when there was a debate in March about should we begin lockdowns I tried to frame it very much as we see a train coming at us and we should get out of the way as best we can that does not mean we should be in lockdown until we understand better how to respond and by the same token we should have been the minute that the intelligence came that this was a problem which was early January or at least late January we should have been making preparations low relatively low cost insurance like preparations so that we could then have more options on the table and that's really what I mean by what I said at the beginning with better decision making leaves us hopefully with fewer ethical dilemmas because now we're really in a pickle and I don't even know how to take it apart thank you I'm straying a little away from your original question but Mark is raising this issue of coordination and the sort of chaotic thoughts and I think the illustration of this is really even back in February and March when the lockdowns were happening they were happening erratically in different states so Colorado implemented an early lockdown but we were surrounded by states that never implemented any kind of meaningful lockdown and the metaphor that I thought best exemplified this was it's like having certain areas of the pool where it's it's not you know you can't do that in a in a nation that is geographically all contiguous and where there's lots of traffic across state borders having this be a state by state response was just astonishing and the fact that we were relatively effective even with that initial sporadic response and then failed to take advantage of the fact that we did curb spread quite effectively to use that time to get ready to ramp up testing to ramp up you know mass production to all of the things that you know were in the playbook for how to deal with the pandemic right we have a playbook for how to manage a pandemic and we just ignored you know a number of those things and if I say one other thing about the the known unknowns and unknown unknowns and stuff because I think we actually the biggest risk here is not the one Donald Rumsfeld you know famously called to our attention which is you know there are things you know that you know and there are things you know that you don't know that you don't know and then there are these unknown unknowns and he worried very much about those I actually worry a little less about those and a little more about the fourth box in that two by two square right Epictetus actually made two by two square table the fourth box is things you know that are wrong right the things you think you know but they are wrong and that's where we make the very worst mistakes right so I think that is how we made the you know worst mistakes going into a rock but it's also some of the worst mistakes we've made here is by making assumptions about what we knew and I really loved early on mark you said something about sometimes the most value that a model will bring is to demonstrate the level of uncertainty under which we are operating and and I think that is really true models and public health people in general should be pretty good at conveying uncertainty and teaching humility and teaching the need to go back and relook at things as we learn more that's again part of the proportionality principle is willingness to revisit that decision willingness to revisit those data that's a good point and a question that's come up and touches on something you've both raised is that the recommendations have changed and we wasted some of the time we bought ourselves through these early shutdowns to implement stronger public health measures and so either as a modeler in your work in public health ethics and actions what could be done to facilitate public trust in the information that we do have and then the recommendations that come out of it Mark you want to go first I'll start I think the first thing is the source if the recommendations come from a source that is trusted and it doesn't seem to have a political agenda and we just heard a few caveats to that at the end of Matt's talk but the CDC had regained a lot of credibility and in 2009 really did I think an almost perfect job there's obviously not perfect but a really very good job in 2009 of saying every day from the same podium this is what we know this is what we don't know this is what we're trying to figure out and you know people are not dumb they can understand that that data come in and that ideas change I think that that it's the changing messages on masks have exemplified the problem you know the sort of what can go really wrong because there is a coherent explanation of why the message has changed which is initially we didn't know if they were effective for people and they were needed in hospitals and all things considered it was better for people not to wear them and so that hospitals could get them as it became clearer that the supply could be extended and that we were getting benefit the risk benefit changed I think some people were sloppy and they're with important consequences in saying oh no it's bad to wear masks they're not explaining or saying they don't work rather than we don't know if they work and having the work personally experience the worst health event of my life due to my own doctor not knowing the difference between absence of evidence and evidence of absence I try to be very very careful about that also for reputational reasons because things change and if you say something with too much confidence you look like an idiot so I think I think we just have to that one distinction of evidence of absence versus absence of evidence gets you a pretty long way it takes longer to explain and it's harder to explain to everyone but people can understand it and there are people who do risk communication well who can help us to explain that better the only thing I would add is that honesty about the reasons for the recommendations as well as the reasons why they're changing I think you can explain to people why things changed it is much harder to explain if you didn't tell them the reason you were making the decision to begin with in an honest way we made recommendations about mask usage that were crisis standards of care recommendations that we knew the recommendation was not optimal it was optimal given the resource limitation right so the idea that masks aren't useful or can't possibly be useful that should have been really not said because we didn't know what we knew was that we had a shortage and we were going to need N95 masks in hospitals and we didn't want the public to use them because there's a shortage and we need them in the hospitals no one said that and if we if we've been honest about it at the outset and acknowledged that we were making you know a sub optimal decision but it's the best we can do given the shortage it would have been easier to change track later and say okay we've got more now you guys should start using them and unfortunately it's just a lost opportunity to acknowledge that we were in crisis standards of care we still by the way still today we are in crisis standards of care with regard to mask usage with regard to mask supplies we are not today in our hospitals using masks in the way we know to be best we're using it in contingency at best if not crisis mode we're asking people to carry that mask with you for a week keep it in a plastic bag overnight put it on again the next day that's not optimal that's not how we normally use masks but we're still doing that because we still have shortages and we're afraid that we haven't ramped up production to optimal levels thank you and there are a number of questions raised about how to explain what is complex uncertain changing data and evidence to the public in ways that we can all understand and affect our behavior and I think you've both touched on that and your answers about being transparent and honest about the reason for decision making given such uncertainty something else that has come up and I want to ask is that throughout this pandemic there have been disparities in which communities have been affected by COVID and not that's something you touched on we've seen the people of color suffering a disproportionate number of cases so for both of you how are these types of equity concerns are they incorporated into evidence or trying to figure out what might happen next and how are those playing out in the public health actions you were discussing your first map oh okay I mean that's huge right we could easily have spent two hours on that set of topics alone I think I'll just sort of list the places where they're showing up there's an enormous amount of interest right now in understanding why we are seeing these large disparities both in the rates at which people acquire the illness and the relative severity of the illness and the number of people dying those are contentious conversations because I think many people tend to go straight to sort of biological explanations for why people with dark skin might have vitamin D changes or genetic differences or some kind of physiologic explanation and yet you know it strikes me that it will be extremely unlikely that that's the primary reason you know when you think about the level of exposure in communities of color which you know tend to be folks who come from more crowded neighborhoods who have to go to work can't work from home you know are working in public service right all of the structural factors that lead to racially disparate outcomes in every domain of medical care those are the same factors that are probably leading to these disparate outcomes and the fact that you can do you know a multivariable model and show that even after accounting for socioeconomic status and housing all of these things that you still see differences between races to me that is less likely due to genetic difference between races and more likely due to racism and that's the fact of living in a community that is racist right so if you're seeing persistent differences between racial groups after adjusting for all of these socioeconomics factors there are two possibilities right one is that one group is genetically different than the other and I'm not saying that never happens it certainly does happen there are examples of that but it's a lot more common and I think it's going to be more likely that when we sort of tease all of this apart two generations from now after thousands of dissertations are written on this pandemic what we will have learned is that most of that difference race based difference was explainable by non biological non physiologic factors the second way this is really playing out in policy discussions is around triage and resource allocation protocols and whether those protocols should divert resources to communities that are particularly hard hit by this pandemic and I would say that consensus is leaning towards the answer yes that equity is an important value and that while saving the most lives is critical it is important to save lives in a way that doesn't rip our nation apart and that doesn't discriminate against people who are already systematically discriminated against and at higher risk so diverting resources into communities that are hardest hit by this pandemic make sense from a values standpoint for us as a nation I would agree with all of that and I would maybe make two further points I think one thing about disparities is that not only are they an issue of injustice but they also show us that one can do better meaning if there is a group of people with better outcomes and a group of people with worse outcomes who are allegedly members of the same society then the existence of that better off group means that we could be doing better for the worse off with the worse off group and that is almost a tautology but it should remind us that these causes of disparities are ones that are potentially fixable and that relates to my second point that I've been really struck by an absence of research on trying to tease apart really what are the especially the determinants of greater mortality and bad outcomes after infection by race after adjusting for other predictors I agree is very very compelling evidence of racism as a cause although there are other explanations that have to be tested against it but what struck me is that not much work has been done to figure out what are the activities and settings in which transmission is a great risk the CDC just published something recently that identified going to restaurants and bars and that was based on 400 or so cases in a case control setting that's a pretty paltry study to be a major thing but I think there's a lot of evidence that we have more than six months into a pandemic or more than six months into a pandemic the fact that we have still very limited evidence to say whether the disproportionate exposure of people of color and low income people is due to public transport or denser housing or job or public transport and that's a great example of how we can actually publish a questionnaire to be used alongside serologic studies as one tool to be used to try to figure that out but that seems like it should be at the top of the research agenda and I find it strange that there's national news stories about a 400 person case control study done? we're at the end of our time inoculum size thank you both because this has been an excellent thought provoking discussion and thank everyone who's joined us today and for your excellent questions since we could keep going but we are out of time I want to remind everyone the recording of this session will be available on the portal website portalresearch.org and we hope you will join us next month for a discussion on drug shortages and ethical issues with that and policies to prevent them thanks everybody very much thanks Mark thanks Matt thank you both