Add Spectator Initials with airSlate SignNow

Add spectator initials

thanks very much nick and thanks to all of you for coming this talk is a contribution to the emerging genre of philosophy of covid19 science the science and policy branch i suppose i'll talk for a moment about what i think we as philosophers of science can actually contribute to discussions around covet 19 that is distinctive and worthwhile we're not a public inquiry think questions of who is to blame for what or who should get credit or who should be rewarded or who should apologize or who should go to prison these are not our questions as philosophers we're also not journalists our job is not really to tell gripping compelling narratives of what has happened this year and we're also not at the table when decisions are made much as we might like to be none of the decisions i'm going to be talking about in this talk involve philosophers of science in the decision-making process but we can at least study the table we can study how those decisions were made why should we study the table well we might have various goals in view i think one goal we might have is to construct generalizable norms for effective scientific advising that hopefully go beyond this case to be useful in future emergencies and other contexts outside the uk we also might want a better understanding of how normal advising differs from advising in extremists and that's going to be one of the key themes of this talk and we might also want a better understanding of how the government advisor dynamic plays out how that relationship that subtle relationship between the government and its advisors plays out in extremis and how the science and values nexus or dynamic works out in extremis and i'll have all of those goals in view in this talk my focus will be on several resources that we can use as philosophers of science to try and make progress towards those goals one thing you can really say on behalf of the government in its favor this year has been the transparency of the scientific advising it's received and in particular sage the scientific advisory group for emergencies has been impressively transparent more than 60 meetings have occurred this year and the minutes are published very soon after the meetings 61 sets are currently publicly available sage feeds into cobra it's represented at cobra at least in the early stages of the covet 19 crisis cobra was the main decision-making body this is the committee cabinet office briefing rooms they call it that is traditionally but not always chaired by the prime minister the minutes from cobra meetings are not made available the fact that such meetings have happened has been made available but by looking at the stage minutes we can get quite a clear sense of the advice that was being fed into those meetings we also have the minutes from nerve tag the new and emerging respiratory viruses threat advisory group fantastic acronym 38 sets of minutes from that group are available as well now nerve tag feeds advice into sage they also have numerous papers and so-called consensus statements from spy m scientific uh pandemic influenza modelling group the at the start of the crisis was transformed into a subcommittee of sage and we have select committee testimony as well from the science and technology select committee in the health and social care select committee now i haven't uh had time to go through all of this in micro detail i'm really really pleased to hear from people who are interested in working with me on this this rather large volume of information i think for example that it could be a fantastic msc dissertation project to really mine some of these resources in relation to those goals above here i'm going to be focusing on a particular period the uk's initial response to covid19 22nd of january when sage met for the first time they called it precautionary stage through to 23rd of march when a national lockdown was brought in i'll be thinking about three key themes and i'll break for questions at the end of each discuss each discussion of each theme the normative force of advice reasonable worst case scenarios and their role in decision making and the difference between independence and neutrality there's more i could talk about but i'm going to focus on these three themes if there's time and i fear time maybe against me i'd also love to give an epilogue which is about september 2020. so first theme the normative force of advice we hear this slogan and chris whittie the chief medical officer has used this slogan in a select committee advisors advise and ministers decide it's a good slogan what does that mean in practice that is a very general statement of how the people on stage see the relationship between advisors and ministers doesn't give us a lot of detail i think compatible with that slogan are three broad approaches to thinking about advisor government relationship and the sort of advice advisors should be giving i think it's useful to introduce three categories at this point you can think that the role of an advisory group like sage is to not give any unconditional recommendations at all but rather to provide conditionals conditionals of the form well if your goal is this we think this will be effective if on the other hand your goal is this you may find this alternative option effective you present various means to end relationships without providing a normative recommendation for any one of those actions second approach you could take is what i'm going to call the disjunctive unconditional recommendation where you don't just say well if your goal is this this will be effective but rather you do give a normative recommendation but it's disjunctive in character that it's do this or this or this and then a third approach is the single unconditional recommendation that simply says do this i think it's interesting to think about these three categories in relation to the sage minutes because i think there's a broad trend that sage at the beginning of the crisis sees its role as one that falls under the nur category the no unconditional recommendations category then there is a perceptible shift a shift towards stir towards disjunctive recommendations and finally a shift towards the single unconditional recommendation is some evidence in support of that trend there's a document a draft advice document that's drawn up by the sage secretariat on 26th of february that i think captures the received wisdom of sage at that time and that would have been i think widely circulated within government it includes various statements that are clearly avoiding any unconditional recommendations they talk for example about the various options one might take to try and mitigate a pandemic including social distancing including shielding and so on and they say implementing a subset of measures would be expected to have a more moderate impact than implementing all of them still substantially reducing peak incidence making a second wave of infection in autumn less likely this might be the preferred outcome for the nhs now the substance of what they're saying there is interesting because you can see at the forefront of their thinking is this concern that if you suppress the virus too aggressively there will inevitably be a very severe wave in the autumn and i'll come back to that point later but i want to draw attention to the form the advice takes they don't say well the nhs should prioritize they just say this might be the preferred outcome and if it is it's taking a subset of measures but not all of them is the most effective way to achieve that outcome and they say it's a political decision to consider whether it's preferable to enact measures at first lifting them gradually is required or to start with fewer measures and add further measures if required that's striking because it is clearly not a purely political decision this is an example of what we might call a mixed political scientific judgment there's obviously a scientific element to the question of what the effects on the pandemic are of introducing all the measures at once and then relaxing them if you've gone too far versus introducing as few measures as you can and building them up as they become necessary it is not just a political decision but they don't want to make that call they want politicians to make that call and even more explicit is this comment from 4th of march where they say as if it wasn't clear enough sage has not provided a recommendation of which interventions or package of interventions that government may choose to apply so the start of the crisis they believe their role is not to make unconditional recommendations as i see it but then there's a change there's a change by 9th of march where we now get different kinds of language there's the claim that measures relating to individual and household isolation likely need to be enacted within the next two weeks to be fully effective and those concerning social distancing of the elderly and vulnerable two to three weeks after this that's perhaps still compatible with with no but drifting in the direction of something normatively stronger and then here's the claim that a combination of these measures is expected to have a greater impact implementing a subset of measures would be ideal so we no longer have this claim that if the nhs's priority is to avoid a second wave in the autumn then implementing a subset of measures would be an effective means rather the claim is that that would be ideal normatively that is what your aim should be whilst this was had more moderate impact it would be much less likely to result in a second wave so now that's disjunctive because of course if you're saying what you should do is implement a subset of the measures we've suggested there are many ways to implement that there are many possible subsets pragmatically i don't think they're considering the empty set as one of the things but there's many possible uh sets of measures ranging from nearly all of them to just one or two that could be taken that's 9th of march now a big change happens on the 16th of march you might remember that day there's a paper published by imperial college and given to to sage with a very striking conclusion that is the first example in all of these published documents that i found of a single unconditional recommendation they say they contrast various mitigation strategies they contrast that with the strategy of suppressing transmission as aggressively as possible by doing everything at once including school closures and they conclude that epidemic suppression is the only viable strategy at the current time it's the first instance i found of anyone in the uk government advisory network making a claim of that type and it's simply because two things had clicked for them as captured in this famous graph that was widely circulated at the time they already had models for some time showing that the demand on critical care beds would be enormous and that you can see in the in the do nothing case this requirement for almost 300 critical care beds per 100 000 people but it was only in this paper on 16th of march that they put that together with accurate information about the number of critical care beds that actually existed in the uk and they plotted that on the graph as a red line along the bottom and that was what created a great deal of concern one might say because the mismatch between the reality of where demand would be and where the surge capacity was perhaps had not been fully appreciated now interestingly they say in response to this that suppressing the epidemic as aggressively as possible is the only viable strategy at the current time that clearly involves a political judgment it is not just a scientific judgment there's a political element to it and in fact if you read the paper what they are endorsing under the heading of suppression is something that is not considered politically acceptable even now because what they describe as the suppression strategy is closing schools whenever cases start to rise sharply and they predict in the model that this will have to be done for two thirds of the time indefinitely until a vaccine is developed and there's this graph here in the paper saying well until about november 2021 let's say schools will be closed two-thirds of the time the blue blocks there represent periods of school closure to say that's the only viable strategy that it's the strategy that must be pursued is a political judgment it's one that even now politicians don't endorse the war might well argue that that unconditional recommendation at the time was crucial to inducing the government to take steps that were that paper by imperial feeds into the advice given by spy m on 16th of march or spy m o they add the o to indicate that it's operational subcommittee of sage they agree that this strategy should be followed as soon as practical at least in the first instance strikingly though remember this is 16th of march the week before a legally enforced lockdown sage at that point still hangs on to du to the strategy of disjunctive recommendations what it says is that there's clear evidence to support additional social distancing measures be introduced as soon as possible grammatically a bit strange like the previous statement but you can see what they're saying they want to leave the government's options open they want to make their recommendation disjunctive by saying take additional measures but the decision of what those additional measures should be is up to you and they don't at that time endorse school closures they do two days later two days later sage advises that available evidence now supports implementing school closures on a national level as soon as practicable to prevent nhs intensive care capacity being exceeded and that in fact happened two days later schools closed on 20th of march that's the first and from what i found only case of sage making a single unconditional recommendation not menu of options not if then but rather do this do this now so what can we learn from this story this slide from nur to sir here's a case i suggest that shows the limitations of nur and er in extremist i think they're potentially sensible approaches for advisors to take in normal circumstances but what's the problem here well in this case as i think is usually the case there's no clean separation of evaluative or political and scientific judgments they are thickly entangled a decision about whether to close schools for example it's a clear case of a mixed political scientific judgment what happens in that sort of situation if you pursue no unconditional recommendations well you just get back demands for a recommendation i don't know what form those demands took but i would imagine them getting increasingly fraught as early march progresses government clearly wants advice here about what to do they're saying in public we're following the science i can imagine them screaming out to sage saying tell us what to do tell us what the science says of course i don't know that with speculation the problem if you take doer as sage slips into further on is that it's this rather awkward compromise there's unintentional outsourcing of some scientific judgments to ministers because you're still trying to avoid making those mixed calls and in particular those decisions about what social distancing measures to bring in those decisions about what order measures should be introduced in and whether you should go all at once or whether you should try and do them slowly and incrementally the all at once option being epidemiologically more effective but the incrementally option being perhaps more intuitive that judgment is left to ministers so this seems to me like a case in which sir was the right approach of course there are reasonable concerns one might have about sir to make a single encourage unconditional recommendation like that close schools do it now requires value judgments as i've been emphasizing now those value judgments if you take this approach are being made by unelected advisors they're being made by people like neil ferguson patrick valance chris witte and there's a real worry that there's no democratic accountability for those value judgments the people in those positions have not been elected there's no particular reason to think they reflect the values of voters that's a really important concern in normal times it's what we might call an accountability gap no one is democratically accountable for these normative calls i just don't think that's a decisive consideration in this setting it leads me to my first proposal but i think there's a normative difference between normal scientific advising and scientific advising in extremism what do i mean by in extremis it's hard to be precise about this but what we were looking at early march is a clear case i think we might say using the american terminology that there's a clear and present danger to public health that requires rapid action i'm influenced here by michael walzer and his work on the so-called problem of dirty hands where the idea in that literature is that there is a normative difference between the norms of political leadership in normal times versus in extremeness that in normal times a political leader should adhere to the moral norms of the community they lead but warsaw suggests in extremis which is when the moral community itself is in danger of being destroyed it may be reas different norms apply and it may be reasonable for a political leader to do things that violate the moral norms of the community in normal times in a way i'm extending that point to the scientific advisor and suggesting that different norms apply to scientific advisors in extremis and that worry about accountability gaps that is completely reasonable in normal times can be suspended reasonably in situations like this in extremis the usual concerns about sir may reasonably be set aside some accountability gaps may be tolerated some examples of mixed political scientific judgments may simply be handed over to the advisors so the adviser is asked to just make a single unconditional recommendation compromise is that i think ideally the government rather than the advisors should decide when that shift has occurred there's obviously a meta call to make there there's not just the mixed scientific political judgment but there's also that metacall about whether we're in extremists when those decisions should be handed over to the advisors in effect and their advice simply implemented or whether we're in normal times where the advisor should stay away from making political calls and simply present menus of options and ideally it would be the government not the advisors that decide when that's occurred now that is not what happened in this actual case in this actual case there was no top-down signal saying we think this is a dire situation we now want single unconditional recommendations please rather there was a bottom-up process in which the scientists at imperial in effect took it on themselves to issue a single unconditional recommendation now i think that may well have been the right thing to do in those circumstances but i don't think it was ideal and that ideally it would be the government deciding that the situation is one in which single unconditional recommendations are required so before moving on to the next issue i'll pause there for any questions about proposal one okay so my second topic is the use of reasonable worst case scenarios this is absolutely at the core of sages approaches to the pandemic from the beginning you hear this term all the time implicitly there is a principle that i'm going to call a reasonable worst case scenario principle they don't explicitly state that but it must be what's driving it that if you prioritize planning for the reasonable worst case scenario then you will also be prepared as well as possible for less severe scenarios that must be the logic of it otherwise why the huge emphasis on reasonable worst cases now what was the reasonable worst case scenario that they were using at the time well it was in some cases highly pessimistic you could even use the word apocalyptic to describe the assumptions they were making there was an assumption that in the reasonable worst case eighty percent of the population gets covered and the fatality rate is one percent so you're looking at around half a million excess deaths a figure in the almost 10 times the figure we actually had in the wave in the spring there's also a pessimistic background assumption not formally part of the reasonable worst case scenario but in the background partly due to public health england the contact tracing is assumed to break down once you have 50 cases per week and another background assumption that's it in the background is that interventions can realistically be sustained for 13 weeks so when interventions are recommended they are described as being in place for 13 weeks moreover the modeling they're doing at the time of the reasonable worst case scenario assumes partial compliance so it's assuming for example 50 will comply with household quarantine this is initially described in the 4th of march memo as high levels of compliance so they say we're assuming high levels of compliance but it's only 50 percent the modeling at that time robustly indicates that maximally aggressive suppression merely postpones the epidemic to a time when measures are relaxed and that's crucial to understanding sage's thinking in this period this is a graph from a memo the memos i've been talking about that were drawn up by the sage secretariat to capture the received wisdom in sage at that time and this figure features heavily that if you suppress too much that's the green line you make things worse later so don't suppress too much now some transmission now so that you flatten the curve as it were and your peak is not as peaky what this graph doesn't show is that all of these scenarios are completely apocalyptic in that nhs capacity is vastly exceeded in all of them with the pos the only possible exception being the green line up to summer but then the the pessimistic assumption is that once it gets to autumn nhs will inevitably be overwhelmed completely inevitable all the models at that time show in lots of different ways that this always happens in the worst case scenario that nhs capacity is inevitably exceeded regardless of mitigation strategies doesn't matter what you do nhs is going down there's this line in the 26th of february memo that in the event of a severe epidemic without action the nhs will be unable to meet all demands placed on it in the reasonable worst-case scenario demand for beds is likely to overtake supply well before the peak is reached the words without action are deleted between the 26th of february draft and the second of march draft to take account of the modeling that's been done in between and become available on 2nd of march showing that this is with or without action the without action qualification is pointless so they're incredibly pessimistic but in other respects they're excessively optimistic they assume that our naught this is the reproductive rate of the virus in the absence of mitigation is 2.4 now estimates vary a great deal but that's right at the lower end of estimates for r naught the the more pessimistic extreme is something like seven are not somewhere between two and seven they stream 2.4 they assume doubling time of four to six days even though the the data at that time was suggesting a doubling time of about three days so even people who had nothing to go on except the published figures about cases would have said four to six days is optimistic and there's an optimistic background assumption introduced on the 25th of february that surveillance is good and that surveillance will give about 9 to 11 weeks warning of a major epidemic they're writing this less than a month before the lockdown comes into force on march the 23rd so reasonable worst case scenario is this mixture of incredible pessimism and excessive optimism what are the consequences of those choices well the optimistic elements of the reasonable worst case scenario when combined with that background assumption that realistically the maximum duration of any measure is 13 weeks led to incredibly costly delays and i think the scientists involved with themselves admit that that because they thought the epidemic is not here yet we've got at least nine to 11 weeks and we need to time the measures so that they fall in the 13 weeks across the peak we should delay action that logic was based on optimistic assumptions that turned out to be false and secondly maximally aggressive suppression was recommended against as we saw in some of those earlier quotations don't do everything do a set there was a period of time in which maximally aggressive suppression was recommended against before 16th of march because it was regarded as mere postponement of the inevitable it was regarded as putting us on the green line in that graph where we have lower cases now but come the autumn nhs capacity is completely overwhelmed many times over so some reflections on this then on the role of the reasonable worst case scenario let's think again about that principle if you prioritize planning for the reasonable worst-case scenario then you'll also be prepared for less severe scenarios that's quite seductive at first glance but we can see circumstances in which it is not true it's not true if your reasonable worst case is not in fact globally pessimistic i'm introducing this term globally pessimistic for pessimistic in in all relevant respects but is actually optimistic in some respects and then reality catches you out by being worse in those optimistic respects it's exactly what happened arnold was almost certainly larger than 2.4 the doubling time was almost certainly shorter than four days it's also not true if your reasonable worst case is one in which the pessimistic assumptions you're making turn out to justify some actions delays or omissions that will be far from optimal in a less severe case now that's the sort of thing a lot a lockdown skeptic might say i'm not uh saying it to motivate skepticism about lockdowns but rather i think what we can see is in this case that delays and omissions that would have been reasonable in the in the worst case where you needed to avoid that green line don't suppress too hard because you'll end up on the green line heading for healthcare system collapse in the autumn the delays and emissions that were motivated by that assumption were far from optimal in the situation we were actually in which is not one in which healthcare system collapsed was completely inevitable in the autumn but we're in the autumn now and i think we can confirm that it was not completely inevitable so we see here the limitations of the reasonable worst case scenario principle and that leads me to my second proposal that reasonable worst case scenarios if they're used at all they should be globally pessimistic which is to say at the pessimistic end of scientific opinion in all potentially relevant respects not just some because in doing that you make the scenario less likely so you invite the criticism but why are you basing all your planning on an incredibly unlikely scenario in which everything falls out badly unluckily the answer to that is that reasonable worst case scenario should not dominate planning in the way that arguably they have done other scenarios need to be considered too and in particular if there are some steps that are favored in scenarios that are more optimistic in some respects then they should at least be considered and that was the case for maximally aggressive suppression in this case now they should be considered i don't mean they should always be tried because of course doing nothing is something that's favored in in in an optimistic scenario or trying a sweden style approach with no legally forced lockdown is something that be favored in a more optimistic scenario so they shouldn't always be implemented but they should be considered and a crucial proviso is that provided we have the option of reversing at minimal cost if the evidence tells against optimism now i would suggest that was actually the case for the maximally aggressive suppression strategy it could have been implemented in early march could even have been implemented in february but let's think about early march if it had been but then it had turned out that that reasonable worst case scenario was the actual scenario in which releasing lockdown measures that are very aggressive inevitably leads to a devastating second wave evidence would have come bad news would have come so to speak from other countries in the world like china taiwan singapore korea and so there would have been an opportunity to reverse that and say let's gradually loosen these measures because we know we're heading for total disaster if we hold them in place and then then release them all at once so there's the option to reverse that at minimal cost and this ties in with some of the things brian roberts was talking about in his talk earlier this year and that suzanne bury has written about the importance of treating options differently depending on how easily they can be reversed and the fact that an early lockdown could have been reversed easily if the assumptions had turned out to be too pessimistic is an important consideration pause there for any questions about reasonable worst case scenarios and proposal two so we've talked about normative force of advice and we talked about the reasonable worst case scenarios my third topic is independence and neutrality what do i mean by these terms well by neutrality i mean that the advice remains neutral on politically contested value judgments and that can be distinguished from independence which is that the advice is formulated without undue constraint or influence from government now these are rough definitions i'm not going to attempt to provide a precise definition of undue because that is a place for debate and i think it's through examining real cases that we can try to develop norms for what is undue constraint or influence obviously some influence is needed the government has to provide questions in a way it has to to some extent set the terms and reference terms of reference for advisory bodies it can pose questions and it can present constraints as well it can say we can or cannot do this the debate is about when constraint or influence becomes undue to the point of compromising independence so in sage's 68-page guidance document weirdly neither issue is raised so neither neutrality nor the independent of stage is discussed even though there is some discussion of how subordinate bodies like spy m relate to sage regarding neutrality it's clear enough from the minutes that sage has this general aim of maintaining neutrality if you're trying to see explain the reluctance in early march to make a recommendation i think that has to be part of the story that the that sage and presumably in particular chris whittier and patrick valance do not want to be drawn into making political value judgments along the lines of which measures have unacceptable costs attached to them should measures be brought in all at once or should they be brought in gradually they don't want to go there it's also clear that in spite of that general ambition they don't always succeed there's a clear example of this on 11th of february if you can remember back then we were all outdoors back then we were going to work in workplaces on 11th february in the minutes then they said it's not possible for the uk to accelerate diagnostic capacity to include covid testing alongside regular flu testing in time for the onset of winter flu season 2020 to 21. there's clearly a political judgment involved in that use of the term not possible because of course with the political will if you put in 12 billion pounds and if you say it's the number one priority of the government nothing else is as important as expanding testing capacity you can expand it and the department of health did expand it and we now can get covid tests so what they said was not possible was in fact possible and i think it is clear that they're not talking about logical possibility or physical possibility they're implicitly making a political judgment about what sort of hurling of resources at this effort would be realistic so they aspire to neutrality but i don't think they always succeed but in any case it's not even clear to me that neutrality is even desirable in extremis and that's an implication of my proposal one that i was suggesting that when the imperial researchers took it upon themselves to make some political judgments about what could be tolerated and what could not and they explicitly judged that a verting healthcare system collapse was more important than keeping schools open they made that judgment they said this has to be done now it's the only viable strategy i suggested that they acted reasonably in doing that so i don't think that neutrality is even desirable in extremis what about independence though well strikingly from the minutes government officials were and are present at the vast majority of sage meetings from the 20th of february ben warner from 10 downing street becomes a regular attendee dominic cummings is not he's registered as only attending two meetings but that was then nicked to the press and became a source of great controversy when it was found out that the prime minister's chief advisor had attended these meetings i was criticized by sir david king a former chief scientific advisor who founded the independent sage as it's come to be known an independent shadow advisory group in response to this that is explicitly criticizing sage for not being independent by calling itself the independent sage an agenda have been added to the published minutes to distinguish scientific experts from observers and government officials they were not that distinction was not drawn in the original minutes but agenda have been added so is david king right is this a violation of sage's independence is it an example of undue constraint and influence well sage has a clear defense which is that the government officials are not there to be scientific experts the scientific experts give their advice on scientific matters and the government officials are there to provide input from the point of view of government departments and to listen that is their official view but i actually think david king is right that i think there is still a problem there in that it's virtually impossible to assure it ensure that there's no undue influence and i think that's the case even if the officials say nothing there's no evidence that they did say nothing that the minutes don't drill down into that kind of detail presumably they did say some things but even if they said nothing there would still be an issue in my view because i think there are some points that a scientific advisor may reasonably want to make salient in a meeting of an advisory group in discussion with other advisors without thereby making these points salient to the government for example they may want to highlight sources of uncertainty in the evidence i already said that the the subgroup feeding interstage spy m is better at this than sage itself potentially people in the room may may have wanted to say more about the degrees of uncertainty in the evidence we'd want to do so without providing the government with reasons not to act for all you know the government officials in the room are looking for excuses not to do what you think they should do that gives you a reason to not make salient sources of uncertainty particularly given that levels of uncertainty are not always well understood by people who are not scientific experts and they may want to give a frank opinion on how government has interpreted past advice and may interpret future advice they may want to say that this or that was not acted on properly so let's word this in a different way in the hope of getting better results and they may want to make salient some of the questions i've raised questions about what form the advice should take and how much normative force it should have they might want to raise the issue of services to versus no and all of that is very difficult when government advisors are in the room listening and are able to into you are able to react to whatever moves are made in effect they become part of a game part of a strategic interaction which should clarify don't mean game in the ordinary sense of the word the part of a strategic interaction where issues that are made salient by the advisors may be used in ways the advisors don't want and they know that and so they have an incentive to remain quiet about certain things so even if the government officials says nothing by changing the character of the interaction that goes on in the meeting room where the real or virtual independence i think is compromised and that leads me to two more proposals the proposal three scientific advising and extremists may reasonably set aside political neutrality in the interest of giving advice on mixed scientific political questions and i think spy m in the example i talked about in the first part of the talk acted correctly in doing that they set aside political neutrality by recommending aggressive suppression they they did but it was unproblematic that they did my proposal four is that independence by contrast remains extremely important even for scientific advising and extremists that's not a norm that we should suspend and moreover that even subtle borderline complements compromises of independence such as the mere presence of government officials in the room should be avoided so pause there for any questions about proposals three and four so i'm going to do the epilogue now that really completes the talk in a way i've been talking up to this point about the period of january to march 2020 the epilogue but i've suggested that my proposals actually capture potentially generalizable norms that are more broadly useful the epilogue is about trying to illustrate that using an example from more recent history september 2020 but of course what we've seen in the uk is that a second wave has to some extent materialized and we are indeed living through a second national lockdown how did that come about and how does it relate to those three themes from this talk well let's think first about the normative force of advice in relation to september there's an element of history repeating where again in september sage is very reluctant to use the single unconditional recommendation what it prefers quite clearly is the disjunctive unconditional recommendation in the minutes from 21st of september that have been discussed quite a lot sage offers a short list of non-pharmaceutical interventions including a two-week circuit breaker lockdown more than a month before the current knockdown was actually instigated and recommends that some consistent package of these interventions be adopted there's no single unconditional recommendation for a lockdown or anything else rather the shortlist is presented this seems to be a conscious approach patrick vance has also talked about this in a select committee that the aim is to provide a menu of options so that it's clear that ministers are deciding and that advisors are not deciding contrary to the sort of relationship between advisor and government i was recommending earlier consistent package was indeed implemented a consistent package of one item from the shortlist you can get into the semantics of what's meant by the word package i suppose but one of the items on the list was advising people to work from home and that was indeed implemented but everything else on the list including universities moving teaching online and the two-week circuit breaker lockdown was not implemented that's an epilogue about the normative force of advice you could argue that again there should have been a shift to a single unconditional recommendation in this case it was not made the measures taken fell well short of what it would appear as though sage felt would realistically be required and cases continue to rise what about reasonable worst case scenarios what's the epilogue there well they continue to dominate planning a new reasonable worst case scenario was drawn up on 30th of july that was confidential despite the transparency that i praised praised earlier they didn't release this until it was leaked to the spectator on 29th of october there's a significant change which can be seen in the spy m consensus statement of 16th of september the now the reasonable worst case scenario is not drawn up independently by sage it is agreed with ministers to quote directly from spy m the ministers now um they're no longer meeting in cobra there hasn't been a cobra meeting for on this issue for a long time as far as i understand but rather a separate committee has been created called covey s that has taken over the functions of cobra um and as far as we know is still normally chaired by the prime minister but this is this is very opaque no one knows any minutes from covides but the point is that the reasonable worst case scenario is now agreed there that the politicians have taken it partly on themselves to in effect perhaps negotiate the reasonable worst case scenario the new reasonable risk case scenario introduces an assumption that cases will rise in the autumn but then highly effective measures will be taken in mid-september and that the effect of those measures will be to reduce are to warn by october we've just seen what measures are actually taken advising people to work from home of course the rule of six as well and things like that it was part of the reasonable worst case scenario that these measures would be highly effective that i don't think is coming from the scientists in all honesty i think that is part of what was agreed with ministers so the new reasonable worst case scenario is once again over optimistic in important respects the actual scenario is once again worse than the reasonable worst case scenario in important respects the spy m consensus statements start to get very interesting because it becomes fairly clear that they are trying to point this out very explicitly i can only speculate about the tone of the scientists in the room when they draw up these consensus statements but my speculation is that frustration is building 23rd of september spyen reports that the actual scenario is already on track to be worse than the agreed reasonable or worst case 8th of october sage agrees with this and reports that the actual scenario is in fact already worse than the agreed reasonable worst case you might well think what on earth has happened here how did this happen again how was the reasonable worst case scenario more optimistic than reality twice there was a very interesting comment from sir patrick valance chief scientific advisor on 3rd of november to the uh health well forget which select committee i think it was the science and technology select committee or health and social care one of those he said we model what the civil contingency secretariat sees as a reasonable worst case and that is then modeled by the spy-n modelers when he says civil contingency secretariat i assume he means covides i think he is uh i think he's possibly been told not to use that term even though the term is public knowledge has been reported in the daily mail that is a dramatic reconception to some extent a disfigurement of the government advisor relationship it is quite extraordinary really that the reasonable worst case scenarios are seen as subject to negotiation and regarded as a political judgment in such a way as to build in assumptions that to any scientist looking at this seem incredibly optimistic this was a point made by aaron vale mp in this in this select committee he said you know in what way is assuming extremely effective measures coming in mid-september in what way is that either worst case or reasonable that's a sad epilogue i suppose it's a sense that some of those proposals i was putting forward earlier remain just as important now as they were in the spring there's still this over reliance on reasonable worst case scenarios combined with here problematic compromises of independence further reinforcing i think the point that independence of scientific advice is still important even in extremists what i hope is that in the future i might be able to give the further epilogue that has a happy ending that is in some way about how planning was better the next time but we'll see in general i hope that some of these norms for scientific advising that i've suggested based on study of this particular case do have more general relevance and that can be generally useful in shaping the way advisory structures are formed for dealing with future pandemics and future crises on that sort of scale thanks very much for your attention you