Professor Sunetra Gupta the epidemic is on its way out

hello and welcome you're watching lockdown TV the news and discussion program from unheard com in recent weeks we have been talking to a lot of epidemiologists and scientists trying to understand one big question is the corona virus pandemic a very deadly disease that has only passed through a small fraction of our populations and we need to reorganize our whole societies around protecting from in the future or is it a much milder disease that has passed through big chunks of the population and may have done as much damage as it's going to already in sections of those populations this is the question and in a sense the author of that first view is professor nob Neil Ferguson we spoke to him about his Imperial College study the figurehead of this other view is Professor Sunita Gupta from the University of Oxford and I'm delighted that she joins us down the line right now hi professor Gupta hello so it was March the 16th that Neil Ferguson came out with his earth-shattering model that defined government responses to this pandemic around about a week later your group out of Oxford University published a very different model that speculated as much as 50% of the population may have been exposed to the virus already and that the infection fatality rate may be as low as Northpoint 1% this caused a big news story and people sort of started lining up either they were University of Oxford people or imperial college people on this big question of the day we haven't heard much from you since then but we have learned some new evidence since March so what is now your view about which of these two polls the truth is working to be closer to so the purpose of the exercise that we published in March was to show that from the available data on deaths we couldn't really tell just how lethal this disease was so if you use a very simple model an SAR model simple compartmental model that divides the population to those who accept well those who are infected those who are recovered and try and fit that to the available data there are actually a range of scenarios that are compatible with the available data on the rise in deaths I'm just referring to the rise in deaths because I think those are the only data one can use the rise in cases is completely contingent on how much testing there's done so I don't believe that's a good source of data to fit a model to any model to so if you simply if you take the deaths and you try and fit the SAR model to it you can get one scenario which is essentially at the kernel of the Imperial model which is a scenario in which the infection has only just arrived and it's killing lots people or and it's going to kill it even more and the other scenario which is equally compatible is one where the pathogen does not have a very high risk of death upon infection and arrived approximately a month earlier and has already spread through the population so these are two extremes and they are both compatible with the data as well as anything in between so what so that was as you say sort of end of March from what we've seen since both in the UK with the way that the deaths are gradually reducing and but also from serological studies we've seen in other countries and do you have an updated estimate of where which to which of those two poles you think the truth is closer to so we're only just starting to see some reports using methods that are validated often sort of in house antibody tests and they vary widely I've just been looking at a study from Birmingham which of health care workers which says the overall rate is higher conversion is twenty four point four percent and per tick hi in people who are working in housekeeping or acute medicine up to 34 33 percent so that's on the high end of the things I've seen and then I've seen measures that are much more modest than that but the problem is that unless you have a very good statistical framework in place to make an inference from these observations what that means to the general population it's a bit difficult to conclusively say determine what what the SERP evidence is furthermore some people may actually be resistant to infection so to actually get a handle on total exposure might be harder than we think that said I believe my own personal view is that the while the jury is still out I would say that it's more likely that the pathogen arrived earlier than we think it did that it had already substantially spread through the population by the time lockdown was put in place so will these when these seroprevalence numbers come out we were expecting them last week we may get them imminently or next week from the first wave of the ons Oxford ongoing study will that settle the argument I mean if it shows let's say 10 percent for the population and nearer 20 percent for London just to pick a random number will that be a case of our professor get to was wrong and it was less than that but it was maybe a bit more than professor vergessen thought or is there grounds for ongoing discussion I think there are grants for ongoing discussion because as I said some people may not actually produce these antibodies we certainly know of cases you know people who have tested positive who don't events an antibody response or one that's detectable and there are various reasons for this one is that if we're lucky enough we may be able to deal with the virus basics of fundamental level with our innate immune responses we may also be able to fend off the virus with pre-existing responses against other coronaviruses which I think has is very likely to play a role in protection particularly against severity of disease and death so it while the antibody tests give us an indication of how many people have been exposed they may not tell the whole story and so the the number that comes out of an antibody test in your view is there's kind of low bound of the percentage who are likely to have been exposed and the truth lies somewhere upwards it about that's a very good way of putting it so does that mean that when we've got this study from Spain I'm sure you've seen that talks about under five percent of the total population sharing antibodies and they had a really quite bad death total and that's really frightened everyone so they think well if any folks then have had it and they've had thirty thousand people dying then you know what would happen if you let the disease spread you would approach that result with with more caution or how should we interpret a result like that I think one has to be aware that a lot of these antibody tests are extremely unreliable so and also as I said if the studies are done in a particular group that may have been less exposed than others one really has to be quite careful about interpreting results but whether they're low or high you may also you know chance upon a group that has for whatever reason been heavily exposed so I think what we need to do is be patient which is difficult given the costs of lockdown we need to be patient and analyze these all the results we get using a proper statistical framework that will give us a better indication of what this lower bound is so does that mean that you stand by your earlier model and still think the best though since model was wrong so with the data that were available up to the time of knockdown in this country on deaths either the model where it had already arrived and swept through and not killed too many people or the one Neil Ferguson's model where it had arrived only quite recently and was set to kill a lot of people both were compatible with that those pathetic rise in deaths so his was the worst-case scenario which still remains compatible with the rise in deaths up to that point and then from that point onwards what happened would either have to be a result of lockdown if his model is correct or if it is still compatible what happened after lockdown with the model where the infection fatality rate is low I think this is partly in part an answer to the early question else what do we know now that we didn't at the end of March so in almost every context we've seen the epidemic grow turn around and die away almost like clockwork and various countries have had different different countries have had different lockdown policies and yet what we've observed is almost a uniform kind of pattern of behavior which is highly consistent with the SAR model to me that suggests that much of the driving force here was due to the build-up of immunity I think that's a more parsimonious explanation than one which requires in every country for lockdown or varying degrees of lockdown including no lockdown to have had the same effect so so those kind of immunities so that would then be in other words I guess in different populations there might be different degrees of underlying immunity or or would it be that that you talked about these it is too similar but slightly different coronaviruses or is it that there would be basically a big chunk of the population that was never vulnerable to this virus because it had some sort of other kind of immunity that might not show up in an antibody test so yes that would be one factor so there two factors here one is resistance to infection at all or certainly the untoward effects of infection which could accrue either from some genetic factor that stops you from being infected ball or from previous exposure to related pathogens which creates some kind of barrier to infection or to disease or death the three things being different but there's also the issue of the vulnerable fraction of the population varying between population between different settings so a very simple one is just the demographics so the age structure in this country is very different from the age structure in South Africa hmm so on just on the basis of what we know now about the dependence of mortality on age and the age structures of different populations we can expect the burden of deaths to be very different so how can we get to the bottom of this then it's it's feels no exaggeration to say that it's the most important question in the world right now because if you're right and in a lot of these countries the epidemic has passed through and we're actually the other side of it these are ongoing enormously destructive lockdown policies are a waste and they're a self-inflicted wound in many places so how come if we get serological numbers and now we're still not sure whether they represent the true levels of the immunity what steps can we do to actually get a definitive answer that's a very good question I think first of all we do need to actually do the serological survey they're only just taking off they need to be done properly they need we need to have a representative section of the population and we need we need to get those numbers that's the starting point as you say it's a little bit worrying that that may still not give us the true picture but then we need we also need to rely on other indicators so first of all the structure of deaths so who is actually vulnerable and we do have more information on that now than we had in March and it's clearer and clearer that the burden of deaths is unfortunately being borne by the elderly and those who have comorbidities or other predisposing conditions that has certain implications for infection fatality rate just how nasty this virus is so we need to pay attention to that and then it also has implications for how to protect the vulnerable and how to move out of lockdown in a way that protects those who are vulnerable to cover it but also protects those who are vulnerable to lockdown hmm so you mentioned the impact fatality rate there do you have a kind of estimate of what you think that currently what we should be expecting that to be I think that the epidemic has largely come and is on its way out in this country so I think it'd be less definitely less than one in a thousand and probably closer to one in 10,000 so that would be no point naught 1 percent yes probably not not quite not one bit naught point naught 5 so if we're talking about like policy now we've still got this huge sort of area of uncertainty and clearly when people talk about the science scientists as you're demonstrating now disick so what is a what is a responsible government approach look like I mean they will say we should err on the most cautious on the most conservative side and kind of be most cautious about coming out of lockdown what do you think the approach should be as we go forward well exactly so the government's defense even now is that this was that the plausible worst-case scenario I agree was a plausible or a possible at least worst-case scenario the question is do we or should we act on a possible worst-case scenario given the costs of lockdown and it seems to me that given that the costs have locked down and mounting that that case is becoming more and more fragile so what would that mean for policy next week or in two weeks time or does that mean a more rapid exit from lockdown is that what that means practically yes I think I would have to say it means a more rapid exit from lockdown based perhaps more on certain heuristics like who is dying what's happening to the death rates and even though we cannot prove I mean it's true scientifically it's difficult for me to say I now have proof that the epidemic is largely on its way out and many of us are likely immune especially if not all of us are going to share the correct sort of you know if that signature isn't going to be present in everybody but we we can see who has died from the epidemic we can see what's happened in a number of countries as I said that they've shown this as a clockwork behavior that mimics the SAR very nicely and you know you can take together these pieces of evidence and and I think one can come up with a strategy that takes us out of this situation when balancing those pieces of evidence against the enormous costs of lockdown so that what is it specifically for example that the focus on the are rate I'd love to get your views on that because that seems to be what the government is sort of looking at in order to decide whether to put the brakes back on or whether to come out faster and so on it's not only deaths they're looking yeah they're trying to model the are rate as they go do you think that's a mistake or is that the right way to to measure things the our rate is principally dependent on how many people are immune so the reason we talk about an R naught for a pathogen is because that's its maximum transmission potential and that maximum transmission potential can only be realized in a population that is completely susceptible so once it starts to spread through a population its are declines from its maximum potential of our nought and the reason it does is because it's using up its resources it's using up the susceptibles and those susceptibles are becoming immune at least temporarily so I don't think you can calculate R in the absence of a knowledge of how many people are immune you can construct an expression for R I imagine it would be extremely difficult to come to a sensible assessment of what that number would be just from measuring the elements within that expression so in other words there's a danger then that we're looking with with modeling are wrong as we go as we could be doing the really coming out of lockdown incredibly slowly well mistakenly modeling our when in fact if we just looked at deaths as the only kind of reliable source of data or maybe hospitalizations and death you think that would be a more certain way to decide the speed of coming out of lockdown I think so I think measuring km in cases you can't rely on at all because it's so completely dependent on the extent to which the testing is done so I mean I don't understand her cases how there has been any kind of emphasis on cases at all I don't think data on cases should be presented at all in the should enter into this discussion so there's another question that has popped into my head that I have just got the numbers for which is I know that when people speculate about very low infection fatality rates a lot of people come back with New York City as the counter point there and they say well more than 15,000 people have died I've covered in New York City or the population of around about 8 million which makes it no point more than naught point 1 percent of the population has perished and what do we make of that how do how can you match that with these very low projected IFRS I suppose and I say this with some trepidation but I suppose one would have to look at the population structure to see what fraction of New York City constitutes the vulnerable fraction I mean I'm not saying this as a way of getting out both an idea the the assertion that the overall infection fatality rate is low but one must remember that overall infection fatality rate is a linear combination of the very high charity rates in a vulnerable class and a very low one in those who are not vulnerable so one can justify or explain higher rates in a set population by saying that the fraction bondable might be higher the extreme example which is the Diamond Princess cruise liner from which the original kind of infection fatality rates were obtained it also seems like when there are kind of perfectly bad high dosage fare ups in places like hospitals that might make a difference is that something we should think about viral load being a kind of factor certainly that's something I did think about when some health care workers were sort of shown to be affected who might not didn't seem to be part of the traditionally vulnerable classes that maybe the actual infection load was higher in those people that's possible it's also possible that you know when you have pockets of vulnerable people it might rip through those pockets in a way that it wouldn't if the vulnerable people were more sort of scattered within the general population so that their risk of exposure was not so high as being if they were sequestered with other vulnerable people so there are elements as the differences in population structure which along with the sort of innate resistance and innate vulnerability can also make a difference to the outcome so you can have a really high risk group which is perhaps somewhat isolated and then nothing happens to them and then suddenly it hits them and everyone dies well everyone is an exaggeration but you know what I'm thinking else so do you think it raises the really sort of worrying prospect that might actually have made things worse in western countries like the UK by reacting as strongly as we did and you know one the goal for this example is that we essentially cleared out the hospitals in expectation of this huge wave of incoming patients and ended up accidentally ceding infections in half the care homes in the country you know do you think there's any chance that we might actually have done better had we done nothing at all yes I think there's a there's a chance we might have done better by doing nothing at all or at least doing something different which would have been to protect the vulnerable to pay attention to protecting the vulnerable to have thought about protecting the vulnerable 3040 years ago and we started cutting hospital beds so the roots of this go a long long way back if we'd had a well-funded National Health Service which had capacity for this sort of event we I think we would have done a lot better you know the other thing to say is that remaining in a state of lockdown is also one which I think is extremely dangerous from the point of view of the vulnerability from the point of view of the vulnerability of the entire population to new pathogens so effectively I think we used to live in a state approximating lockdown a hundred years ago and that was I believe what created the conditions for the Spanish flu to come in and kill 50 million people and that is because what used to happen in that setting the lockdown setting that we had with very little international travel with not as many contacts between people low density populations was one where a virus like corona virus would come in and sweep through and die out which is what we're trying to get it to do and what that meant is every if you take influenza for example it would come back every 30 years or so rip through and disappear but what that means is that every 30 years there would be a sort of juicy fragment of the population under the age of 30 for the virus to attack and that is exactly what we want to avoid so actually people being exposed more being out out and about more is better for Public Health I think so yes so this the sort of follow-on question for is there a chance that things might have been better had we done nothing is is there a chance if we go back to normal tomorrow and say rites lockdown is cancels reopen the pubs we hope the nightclubs everybody as you were things might be fine is that a possibility and how can we eliminate or prove that possibility well I think it is a very strong possibility but as you say it's not something that one can easily prove obviously that the whole drive towards getting the serum epidemiological data was to provide that information on which one could hang such an argument or if that proved to be wrong to take alternative measures but that is proving a little bit difficult so well so that is a strong possibility so what do we do I think we've weighed that strong possibility against the costs of lockdown I think it's very dangerous to talk about lockdown without recognizing the enormous costs that it has on other vulnerable sectors in the population so I know there's a sort of libertarian argument for the release of lockdown and I think it's unfortunate that those of us who feel we should think differently about lock down have had our voices added to that and the libertarian Harang the truth is that lockdown is a luxury and it's a luxury that the middle classes are enjoying and the higher income countries are enjoying at the expense of the poor the vulnerable and the less developed countries and I think it's a very serious crisis and to think of it simply in terms of is this epidemic going to be over or not is really unconscionable thank you so much for that that was Professor phonetic Gupta of the University of Oxford who is professor of theoretical everdene and she was giving a very different account of the way this pandemic might be taking place and certainly what she had to say should give our policymakers pause before making big decisions about continuing lock down for indefinite periods so thank you to her and we'll be back in a couple of days time you