 Welcome everyone. I am thrilled to go ahead and start. So welcome everyone. This is our to the seminar on retrospective contact tracing co-hosted by the Berkman Klein Center on the Internet and Society, Harvard Medical Schools Program and Global Public Policy, the National Governors Association and Partners in Health. I'm Dr. Margaret Bordeaux and the Research Director of the Program and Global Public Policy at Harvard Medical School and I co-chair the Berkman Klein Center digital pandemic response. I'm thrilled to welcome all of you to our first seminar and what I hope will be a series that focuses on practitioners and implementers of the COVID response. I'd like to welcome with today Dr. Hitoshi Oshitani who is a member of Japan's novel coronavirus disease control subcommittee whose pioneering work has helped develop the retrospective tracing methodology. He will present today on this methodology how it was developed and how it is being implemented in Japan. I think that before I turn it over to him, I should note that we are really happy to have him here because as I think many listeners know, we are in a very difficult position here in the United States when it comes to the COVID epidemic. We yesterday had more cases reported in the country I think overall if not at least since its first peak in April and we are at a moment where we really need to think about what we're going to do over the next couple of months. So I've always maintained and I know Berkman Klein Center and Harvard Medical School has always maintained that we are not helpless before this epidemic, that there are things that we can do and public health strategies that we can develop and implement and that will leave us in a much, much better place. And so many of you have heard me talk about sort of the three-legged stool approach to the public health response with one leg being these environmental modifications that we can make ventilation, air filtration in our buildings to make transmission less likely. The other leg being the public health population-based strategies of asking people to adopt behaviors that will protect them like mask wearing or wearing PPE in appropriate contexts. And then the third leg of contact tracing which is where you try to interrupt individual chains of transmission to decrease forward transmission. But the thing we haven't talked about so much is the seat of that stool of that three-legged stool which is really health intelligence and how you understand the epidemic where it is spreading, how it is spreading so that you can refine each of those three legs of the stool. And in some sense retrospective contact tracing is one approach to helping build more robust health intelligence during the epidemic and not only helping you with your basic contact tracing approach as well. So without further ado I'll turn it over to Dr. Oshitami. Thank you, Margaret, for kind introduction. It's my great pleasure to be at this seminar to share our experience in Japan. I'm going to share my slide. I hope you can see my slide. So I'm going to talk about the cluster-based approach in Japan with the focusing on retrospective contact tracing. So these early findings of COVID-19 in Japan, among close contact, we had a very low positive rate, the 1.3% in early cases, the close contact. And similar findings were obtained in China and also the chain of transmission patterns were analyzed in the very early cluster in February in Japan which showed that the one individual infected many others and many other infected individuals did not pass the virus to anybody else. These are the early findings. And also Dr. Nishiura, Hiroshi Nishiura had some preliminary data which suggested that the majority of infected individuals did not pass the virus to anybody else. And the small proportion of the infected person infected many others. And this over dispersion characteristics of COVID-19 was also shown in other countries, other places like Hong Kong. And so that was a basis of our cluster-based approach. And if there is no the cluster or super-spreading event, there is no sustained outbreak. But the only one, the chain of cluster is established. That's the time. That's the situation where we have sustained outbreak of COVID-19. And the prospective contact tracing are done like this. And if you define confirmed cases that you identify the contacts and among contacts, you are trying to identify the cases, COVID-19 positive cases. But as I mentioned, the positive rate is quite low. And probably the one of the two contacts that may be identified by the prospective contact tracing, but they know the cluster that can be found by identifying the 10 or so confirmed cases. You need to find many more confirmed cases to identify the clusters. And the also the another tricky part of this virus, the COVID-19, is the invisible nature of this virus. For SARS, the majority of infected individuals develop the very severe pneumonia, viral pneumonia. That's why we could identify the almost all chains of transmissions. And we managed to interrupt the all chains of transmission. And but for COVID-19, there are many mild cases or even asymptomatic cases and which makes this virus more difficult to identify. And this invisible nature is another challenge for COVID-19. And also there is a significant difference between COVID-19 and SARS in terms of infectivity. For SARS, only when the patient developed very severe symptoms, that's when they had infectivity. But for COVID-19, the peak of infectivity is believed to be before the onset of the illness. So this also makes this virus more difficult to control. And in retrospective contact tracing, if when we find the one confirmed cases, we try to identify the source of infection. As I mentioned, if there is no cluster or super spreading event, there is no sustained the chain of transmission. So there must be the cluster somewhere around the each case. So we try to identify the source of infection of which some in many cases, which is a cluster or associated with cluster. So this is a retrospective contact tracing. So in Japan, we are also doing prospect tracing, but we are more focusing on retrospect contact tracing, which identify the common source of infection. And the contact tracing in Japan is mainly done by the public health nurses. There are over the 460 public health centers all over Japan. And there are more than 28,000 officers stationed in these public health centers, including the more than 8000 public health nurses. And these public health nurses are mainly in charge of the contact tracing. And the retrospective contact tracing has been done as their routine work for investigation of particularly the tuberculosis cases. As you probably know, we are still having many TB cases in Japan. For TB, identifying the source of infection is very important. So that's why they are used to these retrospective contact tracing. And actually, from the beginning, the public health nurses have been doing these retrospective contact tracing. And this is the diagram for our contact tracing. In addition to the to identify the secondary cases from the confirmed cases, we try to identify the source of infection by asking their activities in the past 14 days. And probably 14 days is too long. And in most cases, we identify the source within the five or seven days before the onset. But in some cases, we identify the source nine or 11 days before the onset. And for many cases, we identified the source the three to five days before the onset. And by doing this approach, the one objective of this approach is to interrupt the chain of clusters. And it is usually when we identify the super spreading event, there are some secondary transmission from the cluster. But we still can interrupt the chain of transmission from the cluster. And we also try to identify common characteristics of clusters. We have identified the many clusters in Japan. And so these are the most important common characteristics of the clusters, the closed space, crowded places, and the closed contact settings. And so these are now known as some myths in Japanese. And there were three cities in English. And even the primary school kids know about this concept. And the public was asked to avoid the risky environment to reduce the number of the clusters. And the further analysis of clusters identified some additional risk factors, such as the exercises, talking in loud voice, the singing, and also the nightlife settings. And also we identified who are likely to transmit the virus to the others. So these are the characteristics of the primary cases of the cluster. The primary cases that tend to be young in 20s and 30s. And although the average age of the total cases were much older, so they are probably more infectious because they are more active. And the majority of primary cases, more than 40% of the primary cases, they were the pre-symptomatic phases. And we are analyzing the more cases in Japan. And over the 17,000 cases have been analyzed. And the pattern is still exactly the same. The more than three-quarter of the cases did not pass the virus to anybody else. But the small proportion of the infected individual, the infected many others, sometimes more than 10 people. And so the secondary cases is generated by the relatively older people. And from the children, the percent of the cases generating the secondary transmission was quite low. And this is the epidemic curve of Japan as of November 10. And we had the first outbreak, the first wave of outbreak from February to May. And the second one started at the beginning of the June. And this wave is still continuing. And now we are seeing some increasing trend. And so we are analyzing the clusters. And many clusters have been analyzed. And the setting where the clusters are occurring is changing. And so it is important to monitor the clusters so that we can implement the specific measures for the specific setting. And especially in June and July, we had many clusters in the large nightlife entertainment district in the major cities, especially in Tokyo. And from there, we had many community transmission, including household schools, restaurant parties, and then at the end, the virus reached to the nursing homes and hospitals. And where the majority of CBI cases occurred. So it's important to interrupt the transmission somewhere before the virus reached to the hospital and the nursing homes. And these are the transmission patterns in the community. And even there is a medium to large size clusters. We may see some household clusters and the clusters in the workplaces and eventually in the hospital. But if there is only small cluster, there might be some household cluster, but there is no further transmission. And many of the transmission chain is interrupted in the community. So now we are discussing the prioritization of cluster investigations, including the retrospective contact tracing. And partly because we are seeing the increasing trend of the cases and the public health nurses and the public health centers are overwhelmed now. And in such settings, the high priority should be given to the hospital and the long-term care facilities, including the nursing homes. And the nightlife entertainment is also the risky setting because from these settings, the community transmission, they can be the starting. And also the large social gathering like parties and the theater live music event and so on. But the schools, university colleges, and the workplaces, these are the settings from where the community transmission is less likely to occur. Unlike in the US, we haven't seen the very large clusters in universities. And so the further information that we set up our website, and you can see some of the information about our approach, including the English guidance for the epidemiological investigation, including the retrospective contact tracing. And so this site was actually established by our colleagues, including Dr. Jindai and the Miyeko Kikuchi. And so I appreciate their effort to set up this website. Thank you very much for your attention. Fantastic. Thank you so much for going through that with us. I think we can all start to see how this approach can really change how we think about super-spreading events being not necessarily about an individual, but where that individual goes and how they're relating to others and which environments. So just to say, I'm going to turn it over to Dr. Sung. And then afterwards, I see some questions coming in. We'll have plenty of time for questions after Dr. Sung's remarks. So keep them coming. So to introduce my colleague, Dr. K. J. Sung is a total honor. He is the Chief of Strategy and Policy at Partners in Health of Massachusetts COVID-19 Response. He's also an associate physician at the Division of Global Health Equity at Brigham and Women's Hospital and assistant professor at the Department of Global Health and Social Medicine at Harvard Medical School. And he's going to focus on sort of, okay, yes, this is a cool, amazing technique, but how do you do it? How do you execute on it? So, K. J., I'll go ahead and turn it over to you. Good morning, everybody. I think for a lot of you, it's the first time you've heard Professor Oshitani. For us, it was actually, I think, it was back in July when we heard, you know, Professor Oshitani was very gracious to get on a quick Zoom call with our project. And it was really a different way of doing things. And so we want to explain exactly how we restructured our project to incorporate a lot of these techniques. And just a little background on our project here in Massachusetts. This is a state 7 million people. The project, this is the community tracing collaborative. And you can see up there on the left, there are 351 local boards of health in the state in the Commonwealth of Massachusetts. So each one of those has jurisdiction over its town, has its own public health department. The, and so the collaborative is really, is really meant to incorporate all of these. So it's really the, those local health departments that are doing a lot of the contact tracing. And there is another surge workforce of contact tracers that is meant to backstop those in case those, those public health nurses really who are very closely aligned with what's happening in the field start to get overly high caseloads. And so a lot of these concepts and strategies we have incorporated into this surge workforce, which tends to, you know, which can really float across the state and is not restricted to work in any one jurisdiction. So it's really set up as a large call center. And a lot of other states have done this as well. And it's a unified force. It's a, it's got a single database. There might be a few things in Massachusetts that are different from other states. For us, the CTC contact tracing units have, have case investigators and contact tracers. The case investigators do the initial case investigation. The contact tracers are then making calls to close contacts. In reality, those two roles are very, very similar. And in fact, those people really do both things depending on the need. We also have a third cadre called the care resource coordinators. And we know that people need support. They need nutritional support. They need medicines. They need transportation. They need all of these things to stay in isolation and quarantine. So, so there's a third cadre, a smaller cadre that is also floating throughout the, the contact tracing units. And for those cases and contacts that are flagged, the CRCs will, will be able to provide that support. So the first problem, you know, I think as you listened to Dr. Oshitani and probably had the same reaction that I did, which was that what he is really explaining with respect to retrospective contact tracing is different. It is not actually what we are doing. It is not part of the WTO protocols or CDC protocols. There is, you know, our, our protocols are meant to do perspective contact tracing. And that means from 48 hours prior to symptom onset, that is anybody, you're really looking for the people who that case could have infected. But to go back retrospectively, you know, in all of these protocols, there is some, there's some data collection about, about possible exposures. But really, I think if you listen to, to the professor, you can, you can, you can see that it's actually much more involved. And so what we have done is if you really think about as level one and level two, instead of, and instead of really pushing out that level two, which is a retrospective component into the entire contact tracing workforce of over a thousand, a thousand people, we chose to, to develop a smaller unit. So really a unit within the overall contact tracing team. And we called it the Epidemic Intelligence Unit, the EIU. And it's really less than 20 people at this point. But you can see here, though, that, that if you're going to take this approach, the, the larger contact tracing workforce, so level one clearly has to involve some aspects of level two. So in some sense, level one is the first, the first contact, the first communication with the case, that initial first call. And so if the person, the case investigator is not really sensitized, not really thinking about where the case was, was infected, then that hint or that tip about a possible cluster at a workplace, at, at a restaurant, at a yoga studio, or hockey team, that cannot be given to the EIU and level two can't start. So level one, definitely here, I think it's, you know, I think it shows the different roles and the differences between the level one and the level two. But really that level one, even if you do have an EIU, a smaller unit that can really handle the retrospective component, you can't completely, you know, there's still a lot of training that has to be done to your larger content tracing workforce. So really over, in our case, it's, it's over a thousand people. So that's one issue here is that, is that how do you expand your protocols and, and how do you expand your, your workflows to get around the first problem, which is that it's not really what Professor Oshizani is describing, is not contained in current content tracing protocols for COVID. So the other problem is, is this one, which is that the, you know, the, the cluster, and I think that in the previous presentation, there was some discussion of this is that clusters are not obvious. They are the cluster analysis the cluster analysis and the investigation is, is fragmented. And so you really, you know, I know that Professor Oshizani has talked a lot about the danger of the isolated case. So when you're seeing a lot of isolated cases, so cases that say, you know, I don't know where I got infected. I have no, I don't, I don't know, I know who I could have infected because I'm living maybe with my household. I don't know where I got infected. That's an isolated case. It's not connected to a cluster. That's dangerous. And, you know, the other thing what I would expand on that is that it's really the isolated household, right? So we know the easiest part of this is, is, is the household transmission. I'm not saying that it's easy to prevent, but it's certainly easy to, to find everybody is living with, with people, their close contacts that are, they're a high risk for infection. So you, what you see is isolated households. You have high households, there are household clusters are all over the place, but you don't seem to be connected to any, any cluster. So when you start to do this retrospective analysis, what you find is that, you know, each household and, you know, in Massachusetts is that we consolidate those household contacts and those cases into one contact tracer. So really, those follow-up calls can be made in a much more efficient manner. But, you know, as your retrospective contact tracing continues, you may find that this household is actually connected to a larger cluster. So suppose actually three households went to a play date for, for one of the, one of the addresses. So now, you know, now these aren't actually isolated households. This is a small cluster. But the way that you have set up your content, or the way that we set up our content tracing program is that there is one contact tracer. So there are actually three contact tracers who are following this relatively small cluster because there are three households that are involved. If you continue to do your retrospective contact tracing and hypothetically, you find actually this is part of a larger, not a gigantic cluster, but still a larger cluster because one of the children went to a pool party where there were some teenagers involved who had actually gone to a Halloween party. Well, again, this is a, this is really not a gigantic cluster. But you can see here that there are seven different contact tracers involved because there are seven different households. So it's very difficult, in fact, to see for each of these contact tracers, they may, part of the training in that phase, in that phase one is to at least get a hint is to really ask further than the 48 hours. So they get an idea of how, where the exposure happened, whether they're played, those are play dates or Halloween parties. But really, it's impossible for all of them. And it's not efficient for them to try to figure out the extent of this cluster and to trace back even for this relatively small size cluster to the Halloween party. So the EIU really has to be, it's really a problem of communication. So you have fragmentation because the, you know, people aren't getting tested. You cannot test 100% of the people in the community every day. You may only pick up certain cases in the cluster. But then those cases are followed by different people. And you really have to have a central unit. For us, that's the EIU. And there has to be just free communication with the entire contact tracing workforce. So, you know, this cannot be done with a database. You know, there's no way to do this. There's no way to do this from the outside. The EIU has to be able to communicate with the people who are directly interviewing the cases in context. Because what you find, and, you know, this is one of my pet peas, is that when people are talking about why contact tracing is not being effective, and why people are not listening, or people are not complying with contact tracing, what I usually say, and I tell them is that it's because, in fact, you're not, you're not listening to what they are saying. And in fact, in our experience, for example, retrospective contact tracing cases are extremely interested in how they got infected. They want to talk about it for hours. They, and what we've in fact trained our contact tracers to do is to ignore that information that the cases want to tell them. So it's really a systems failure rather than an individual failure, a failure of the community to comply, quote, unquote, with contact tracing. So, just the last slide here, you know, because there's this other question that's, and that's the third problem is that it's really what one of my colleagues calls the so what problem, what are you going to do with this information? So we have here a case, we have a case investigator, we have, you know, the follow up, the support, there is resource coordination, so this case can stay in isolation. But actually this case, let's say is connected to a cluster. This is not an isolated case. This is perhaps it's a yoga instructor that was it was a yoga retreat that was held at a weekend retreat that was held at a hotel or it's a it's a hockey team with a coach that's part of a league or it's a business, a large, it's a large meat processing plant with a manager and multiple shifts. So there is a larger social background behind this cluster that this case is in. You know, the case investigator, you know, that's phase one can get a can get a tip and can provide a tip to the EIU, but it's really the EIU that's responsible for looking out and really delineating the the boundaries, the size of of of where this this cluster took place, this particular cluster took place. There, we have a lot of tools for that. You know, that I've shown in previous slides, for example, those cluster maps. But the in the end, you know, the question is really is this so what so how does this actually prevent future clusters. And that's really with, you know, in our case, with our local health departments, right? So that they have really jurisdiction, they have responsibility. They can those public health nurses are able to talk to coaches, to talk to leagues, to talk to hotels, to talk to businesses, and to have very much more in depth discussions about how they're trying to defend against COVID to prevent outbreaks in their in their settings, and can even, you know, as a blunt instrument, implement more other sorts of restrictions on hours or operations that will cross out, you know, that will encompass an entire sector. So the EIU here is, you know, the idea is that it's supposed to provide actionable intelligence. So you're, you're, you're collecting information from all of these case investigators, thousands of contact tracers, you're trying to put them into a format that's useful, not not a bunch of names and numbers and dates, but to put them into a cluster map. And we've developed some of those tools that I showed you previously. But those can be taken by local health departments, can be used to understand, you know, very specific small or large clusters can be used to discuss with business owners, with places of worship, with, with restaurant owners, and even, you know, with, with mayors, with, with city councils can, can be used to, to change policy surrounding COVID in, in those jurisdictions. Thank you very much. If, you know, the US PIH's public health accompaniment unit is available to help states. And, you know, I hope that we'll be able to talk more in the future. Fantastic. Right. So a lot of, a lot of learning has happened here at the contact chasing program in Massachusetts. And thanks to Dr. Ishitani for, for coaching us, us through it. I want to quickly give the first, the right of first question to Dr. Tefeki, who is a techno sociologist, who has written about specifically retrospective contact tracing and the nature of COVID is something that spreads in these clusters. And so I know that ever since your article in the Atlantic, you've also been getting lots of questions, Dr. Tefeki, about, about this. So I'll turn first to you and then we'll go to our Q&A in the Q&A panel. Thank you. I do want to start by thanking, especially Dr. Ishitani, both for the presentation and how willing he has been throughout the process to share from Japan all these important key points. And as you've noted, I have written about some aspects of the pandemic, including potential for airborne transmission and its relationship to clusters and this backward cluster tracing that from, as we learned from Dr. Ishitani's example. So the first question I have is what percent of the cases you're finding are coming from this retrospective cluster busting? Or are you not doing like the traditional prospective tracing? Because as we heard from the United States, that is not at all in our protocol. We just sort of mostly in most places do forward tracing. And one question I got a lot since writing this has been, well, if we do do this backward tracing, what do we expect? Do we have some numbers? Some, I mean, the theory is easy to explain and a lot of people are convinced, but it's kind of hard to get the practical side moving. And so people would like to hear what's the payoff? Do you have some percentages? Do you know what percent of your cases are found? Maybe first backward, then forward. Or if you could enlighten us a little bit more, that would be very interesting, I think, to many states that are considering these protocols. Okay. Can I answer this question? That I don't have exact numbers, but it also depends on the situation that when we define the cluster by doing the retrospective contact tracing, we usually can find the number of cases sometimes the cluster is more than 50 or even more than 100. So by identifying the many clusters, we can find more cases. And I don't have an exact number, but the probably the nearly half of our cases were found by identifying the clusters. And also, we also identify many unlinked cases, the cases without any epidemiological links. But the many of these cases, many of these isolated cases, the probably they do not generate any secondary cases or the very few secondary cases. It's more important to identify the cluster than identifying the many isolated cases. And that's our concept. And that's why we are putting more effort to identify the clusters. Fantastic. KJ, do you want to follow up on that? And maybe you can include a little bit of, so okay, so you think that the cluster was at a yoga class or at a Halloween party, then what do you do? Do you go and test everybody that was at the Halloween party or at the yoga class or how does that practically work? Yeah, I think that we're really getting to that question, which is the so what? So what are you going to do? What will you do with this information? And it really depends on the type of cluster. So I think that what we're seeing now in Massachusetts anyway, are we're not seeing, well, I can't really say that we're not seeing any very large clusters. Certainly we do see some of them. But what we see more of are lots of little clusters. And whether this is happening in the workplace, which is relatively controlled or at schools or social gatherings, which I don't think we're having the same sorts of large social gatherings that we did in the spring, when we were really taken aware. But these are harder to do. And I think that certainly in certain cases, you can increase, you can target your testing. So if you have an idea that there's an outbreak at a church or a workplace, it's quite common in fact for business owners to say, well, now I'm going to test the entire workforce. So that, you know, at that point, you're not trying to characterize the cluster, you're actually trying to stop transmission there. Social clusters, you know, this is the very, very difficult. These are not huge gatherings. These are gatherings of 15 people, maybe four households, maybe six people, but with attack rates of 60, 70%, you know, pretty much everybody with it at an indoor gathering, at an indoor dinner party getting infected. And there, you know, it's really about, you can certainly inform your community education, but also just community advisories and restrictions of the size of social gatherings, it can inform those things as well. So I don't think that it is, I think really for every type of cluster there is, there are certain actions that can be taken on a local level. And I, you know, I would also agree with the professor that it is a clustering disease. So when you find these, you know, we certainly are not able to find 100% of the clusters. We may not be able to connect 100% or even 80% of all cases to clusters, but I don't think that's necessary. You know, this is, if you have characterized well, and certainly when you're having a huge caseload, it's really impractical. You cannot, you cannot do retrospective contact tracing on every single person, but I don't think that's the point. The idea is to get a sense of where the transmission is taking place, what sorts of social gatherings, you know, is it happening in colleges, in universities? Is it happening in churches? We know overall there's quite good evidence of outbreaks at all of those places, but this will give you on a local level where it's tending to happen in in your area. And it doesn't, so you don't need to find 100% of the clusters. You just have to get a sense of where they are happening by essentially doing a sampling and then you can take action. Fantastic. Yeah, so I just to underscore those points, retrospective contact tracing helps you both identify cases so that you can do prospective, well, prospective contact tracing and help isolate or stop onward transmission. But maybe more importantly is you sort of learn about the environment in which transmission is happening so that you can modify that environment and so that the transmission doesn't occur there anymore. So the, just to humor you with a little bit of an example of this, KJ has watched as my family, which is a very strong hockey playing family. I have four daughters, they all play hockey, my husband coaches, four hockey teams. And hockey was, you know, it was discovered that it was, had led to a couple of clusters. It may be more than a couple in Massachusetts. And much to the heartbreak of my whole family, hockey in Massachusetts was stopped, youth hockey was stopped. But really what was able to happen was they started looking at, well, why was it happening at hockey games? You know, why were they finding this? And they were able to at least try to modify the practices of the youth hockey teams such that such that they could restart youth hockey. And you know, that's the kind of battle rhythm that we all need to need to get into here. If we're going to figure out how to both drive down the transmission and get a hold of the epidemic, but also, you know, keep living, keep living life. And so I think there's a number of questions around that in the Q&A, really about the issues around the, well, I'll pick a few. Why are universities in Japan and schools less, why are you having fewer clusters there than in the United States, do you think? Is it something about how you've changed your school, something about the age of the students, or something about the nature of the epidemic there? Thank you for the question. Actually, we are also seeing some clusters in the university, but most of the clusters in university have been occurring in dormitory settings. And or the outside of the lecture, the drinking party after the lecture, and some other activities. And particularly, we do not have the many dormitories in Japan. And the compared to the university in the US. And so the dormitory is the probably the most important setting. And for the sports clubs, they are living in the dormitory. And we've been seeing the many clusters, the large clusters in such settings. But the usual student, not the many, most of the usual students are not staying in the dormitory. And also that we are still doing online lectures in most of the universities. And but probably the dormitory is the most important factor. Interesting. So that's interesting in terms of how you sort of think about clusters. I think that another sort of bulk of a cluster of questions in our Q&A has to do with really the US context. And we've been struggling with testing and having very sort of big limitations in our testing capability. And also, we are, you know, seeing the surge in cases. So some questions like Dr. Sung, maybe you can answer, does this work in a low, where you have low testing capability? And, you know, what about in with rising caseloads? Is it kind of worth it, worth the effort? I do think it's worth it. I think that it gives you, you know, with the rising caseloads, you know, you cannot contact your trace, contact trace yourself out of an epidemic, not a COVID epidemic. But the retrospective contact tracing and the prospective contact tracing is not going to catch every case and catch every cluster. It's not, but I think that for retrospective content tracing, that's not the goal. The goal is not to find everything. The goal is to understand better where the transmission is happening, and also to guide your efforts. So targeted testing, targeted outbreak response. And but just going back to the first one, you know, there's another question there about really this rigid definition of a close contact. And I think this is really a great example of how it changes, you know, doing retrospective contact tracing changes your understanding. This is a really a, you know, the definition of close contact that's meant to facilitate contact tracing, but it has been so driven into the minds of the community that they think that it's actually protective. And, you know, we've seen many, many, you know, as you do, as you talk to churches, as you talk to businesses, we can see, you know, it is obvious, right? If you have, you know, if you have 12% of your workforce in a large warehouse that is infected with COVID documented infection, you know, these and none of them actually fit the definition of close contact. Okay. So what you realize is that in certain situations in the three C's, for instance, you know, there is infection that can happen actually quite easily, even for non-close contacts. And no, it's not difficult to figure out that there is an outbreak going on that all of these are linked. You know, if 12% of your workforce is infected with COVID is PCR positive and is symptomatic, obviously they are linked in some ways. These are not independent infections. So this is, you know, even on the, you know, even on the very basic understanding of how COVID is transmitted within a place that fits the three C's and for broadening your perspective and giving you the impetus to go a little bit beyond these very, very rigid protocols that we're using, I think retrospective contact tracing is extremely valuable. Fantastic. Well, you mentioned sort of the idea of six feet being kind of drilled, you know, just like the mantra now is like this protective, invisible warp shield as long as you're staying six feet away from someone. And we, you know, we just know that that is just not the case in tight spaces with bad ventilation and you're there for a period of time, you're going to get it, whether you're six feet apart from other people or not. And Natalie Dean asked a nice question here about how do we communicate the need for retrospective contact tracing with the public? How do we handle some of the sensitivities that might come up? You know, I know you said people like to talk a lot about where they got infected, which is definitely chimes with my experience with patients as well. So for some, that's not going to be an issue, but for others, you know, they're not going to want to reveal, you know, that it was their hairstyling business, you know, or, you know, there's a lot of difficulties there. So how do we explain this to the public and how do we address some of the sensitivities around doing this kind of investigation? Well, you know, I will say that I think that in the United States, we are really behind in this aspect. I think that people, you know, and I'll say what I, again, I'll repeat that I don't think the problem is with with the community. People are quite happy to talk about these things. They are not worried about talking about this. They, you know, and like I said, you know, obviously, is there some reluctance sometimes to give up names and phone numbers? Of course. But, you know, in some sense, in retrospective contact tracing, that's not even necessary, right? Because you can certainly, you know, if somebody says I went to a Halloween party, if there if there if if Halloween parties are are coming up again and again and again, 10 days after Halloween, you don't have to know exactly who went there. You don't have to know exact guest list of every single one of those parties to know that this is a major major problem going along, popping up, and that in Thanksgiving and Christmas is going to happen again. So I think that, you know, we do have information here that we can gather. We have to be we have to figure out how to convey that to the public in terms of community education. But, you know, even more broadly, like I think that because we haven't figured out how to explain COVID outbreaks to the American public, there is a lot of of of both stigma and discrimination, which I think is ridiculous for for what is relatively, you know, which really by now, you know, well into the epidemic should be routine. There should not be any shame. There should not be any discrimination about being about exposing your church or your business while you you know while you were not symptomatic, which which is really the most common situation. And people shouldn't be scared about getting email saying you've been exposed, you need to go get tested out of quote unquote, in abundance of caution. We have not been able to, you know, we haven't done very good communication on this, we have, we have a lot of media that jumps on clusters and likes to make a huge media event out of it. And that I think perpetuates the problem. But I definitely would like to hear from Professor Oshitani, how Japan has dealt with this problem. So thank you. In Japan, we also have some issue regarding discrimination and so on. And sometimes people do not want to talk the especially that we have this problem the with the cluster being the host club or hostess clubs, that's kind of the nighttime nightlife in entertainment settings. And the people do not want to talk to where they were. And the both customers and the the the people working in these settings. And that was the one of the reason or probably the main reason why the these clusters were difficult to control the in June, July. And we are also seeing the some increasing of the clusters in the foreigners community. And so, so the communication is quite difficult. And the people do not usually go to the clinics for testing. So the we are now trying to find a better way to communicate with them. And but usually the people are the willing to collaborate with the local health authorities, particularly with the public health nurses. But in certain setting, it's more difficult to investigate all around the world, people are people and have certain things they like to keep private. Dr. Tafaki, do you want to wrap up with any other any other question and then we'll and then we'll have to continue our conversation in the offline and the in the coming days. I would love for Dr. Oshitani to address the close contact question a little more, because I think that's a part of the puzzle that's not very well understood in the United States, we just kind of talked about it. But there was a lot of debate to try to get airborne transmission, aerosol transmission accepted as a method of transmission. And CDC has acknowledged it. But there's some question on like, how prominent is this method of transmission? And how does this connect with contact tracing? So if you could address how like what Japan's epidemiologists see as the primary mode of transmission or secondary mode of transmission, if you have any data on how much of which one you're seeing and how that links to how you define close contact, what role ventilation plays and how that links to the kind of cluster busting you do, I think that would be very illuminating for our audience. Thank you very much for the question. And we still do not know the exact the proportion of the different mode of transmissions. But from the beginning, we were considering the possibility of the aerosol transmission, especially the short distance aerosol transmission. Because from early findings, we saw many clusters from the pre-symptomatic individual. And so they did not, of course, have any cough or sneezing, but still the transmission occurred in these clusters. And so and also we had the many clusters in close contact settings, just having the conversation between the people in these settings, many the transmission that were occurring. And so I think at least the short distance aerosol transmission is probably the one of the important mode of transmission. But we still do not know how the is the proportion of such transmission. And that we are considering the two, they implement some control measures that we are considering this possibility. And we are using the supercomputer to simulate the different settings and like the theater, the live music plays, and so on. And in this simulation, we are considering at least the short distance the aerosol transmission. And we also highly recommend the good ventilation because the many clusters, the most of clusters occurring in the closed environment. So the ventilation is probably the important. And that also suggests that the importance of the aerosol transmission in these settings. Thank you. So fantastic. I feel bad that we have to wrap up. I feel like we could go on for another couple of hours on the issues, everything from how you set up your data. Are there any tools that we can use to help investigators put information together in a way that is more effective or faster all the way to the issues around building trust with communities and thinking about privacy and protections once you are doing this type of really deep dives into how people interact with one another. So lots, my mind is all of a sudden full of ideas for follow-up seminars. But we will stop here and just a huge thank you to all of the folks that spoke today and to the folks that tuned in.