Y Pwyllgor Iechyd, Gofal Cymdeithasol a Chwaraeon

Health, Social Care and Sport Committee

11/11/2020

Aelodau'r Pwyllgor a oedd yn bresennol

Committee Members in Attendance

Andrew R.T. Davies MS
Dai Lloyd MS Cadeirydd y Pwyllgor
Committee Chair
David Rees MS
Jayne Bryant MS
Lynne Neagle MS
Rhun ap Iorwerth MS

Y rhai eraill a oedd yn bresennol

Others in Attendance

Craiger Solomons Llywodraeth Cymru
Welsh Government
Dr Andrew Goodall Llywodraeth Cymru
Welsh Government
Dr Brendan Collins Llywodraeth Cymru
Welsh Government
Dr Rob Orford Llywodraeth Cymru
Welsh Government
Fliss Bennee Llywodraeth Cymru
Welsh Government

Swyddogion y Senedd a oedd yn bresennol

Senedd Officials in Attendance

Claire Morris Ail Glerc
Second Clerk
Dr Paul Worthington Ymchwilydd
Researcher
Lowri Jones Dirprwy Glerc
Deputy Clerk
Lucy Morgan Ymchwilydd
Researcher
Sarah Beasley Clerc
Clerk

Cofnodir y trafodion yn yr iaith y llefarwyd hwy ynddi yn y pwyllgor. Yn ogystal, cynhwysir trawsgrifiad o’r cyfieithu ar y pryd. Lle mae cyfranwyr wedi darparu cywiriadau i’w tystiolaeth, nodir y rheini yn y trawsgrifiad.

The proceedings are reported in the language in which they were spoken in the committee. In addition, a transcription of the simultaneous interpretation is included. Where contributors have supplied corrections to their evidence, these are noted in the transcript.

Cyfarfu'r pwyllgor drwy gynhadledd fideo.

Dechreuodd y cyfarfod am 09:30.

The committee met by video-conference.

The meeting began at 09:30. 

1. Cyflwyniad, ymddiheuriadau, dirprwyon a datgan buddiannau
1. Introductions, apologies, substitutions and declarations of interest

Croeso i bawb i gyfarfod diweddaraf y Pwyllgor Iechyd, Gofal Cymdeithasol a Chwaraeon yma yn rhithiol drwy gyfrwng Zoom yn Senedd Cymru.

O dan eitem 1, cyflwyniadau, ymddiheuriadau, dirprwyon a datgan buddiannau, a allaf i groesawu yn y lle cyntaf fy nghyd-Aelodau i'r sgrin? Yn naturiol, nodir taw cyfarfod rhithwir ydy hwn gyda'r Aelodau a'r tystion yn cymryd rhan trwy fideo-gynadledda. A allaf i ymhellach esbonio bod y cyfarfod yma'n naturiol ddwyieithog a bod gwasanaeth cyfieithu ar y pryd ar gael o'r Gymraeg i'r Saesneg? Weithiau mae yna ychydig bach o oedi ar ôl i rywun fod yn siarad yn Gymraeg cyn i lefel y sain ddod nôl yn uchel i'r cyfrannwr nesaf, felly ychydig bach o amynedd. Mae'r meiciau yn cael eu rheoli'n ganolog tu ôl i'r llenni, ac efallai bydd yna neges fach yn ymddangos ar eich sgrin i ddad-fudo pan fydd angen.

Er mwyn y Cofnod, os byddaf i fel y Cadeirydd yn colli fy nghyswllt rhyngrwyd, fel sydd yn gallu digwydd o bryd i'w gilydd, yna rydym ni wedi cytuno fel pwyllgor cyn hyn y bydd Rhun ap Iorwerth yn camu i mewn i'r bwlch fel Cadeirydd nes bydd fy system rhyngrwyd i nôl yn gweithio.

Yn naturiol, Dydd y Cadoediad yw hi heddiw, felly mae gen i fwriad am 11 o'r gloch i fynd i mewn i ddau funud o dawelwch i gofio'r sawl a gwympodd mewn rhyfel. Gyda chymaint â hynny o ragymadrodd, a allaf i ofyn am unrhyw ddatganiad o fuddiannau gan unrhyw un o'r Aelodau? Dwi'n gweld nad oes.

Welcome, everyone, to this latest meeting of the Health, Social Care and Sport Committee here in a virtual capacity via Zoom at Senedd Cymru. 

Under item 1, we have introductions, apologies, substitutions and declarations of interest. May I welcome in the first instance my fellow Members to the screen and note that this is a virtual meeting with Members and witnesses participating via video-conference? The meeting is, of course, bilingual, and simultaneous translation is available from Welsh to English. There may sometimes be a slight delay between a contribution in Welsh and the next speaker coming back up to full volume, so do please be patient. The microphones are being controlled centrally behind the scenes, as it were. You might receive a prompt on your screen asking you to unmute as required.

And for the Record, if I as Chair were to lose my internet connection, as can happen on occasion, then we have agreed ahead of time as a committee that Rhun ap Iorwerth will step into the breach as interim Chair until my internet connection can be re-established.

Of course, it is Armistice Day today, so we do intend at 11 o'clock to have a two-minute silence to remember all of the fallen. And with those few words of introduction, may I ask for any declarations of interest from any of the Members? I see that there are none.

2. COVID-19: Sesiwn dystiolaeth gyda Chell Cyngor Technegol Llywodraeth Cymru
2. COVID-19: Evidence session with the Welsh Government's Technical Advisory Cell

Felly awn ni'n syth ymlaen i eitem 2, sesiwn dystiolaeth COVID-19. Mae ein cynulleidfa fyd-eang yn amlwg yn gwybod beth rydym ni'n gwneud fel pwyllgor bob wythnos rŵan yn y pwyllgor iechyd, yn craffu ar ymateb y Llywodraeth—a phawb arall, a dweud y gwir—i'r pandemig COVID. Felly, heddiw, mae gyda ni sesiwn dystiolaeth gyda chell cyngor technegol Llywodraeth Cymru. Ac i'r perwyl yna, dwi'n falch iawn i groesawu i'n sgrin Dr Rob Orford, cyd-gadeirydd y gell cyngor technegol a phrif gynghorydd gwyddonol iechyd Llywodraeth Cymru; Fliss Bennee, cyd-gadeirydd y gell cyngor technegol Llywodraeth Cymru; Dr Brendan Collins, cadeirydd is-grŵp modelu polisïau, aelod o'r gell cyngor technegol, pennaeth economeg iechyd Llywodraeth Cymru; Craiger Solomons, cadeirydd fforwm modelu cenedlaethol Cymru gyfan, aelod o'r gell cynghori technegol, dadansoddwr arweiniol cyfarwyddiaeth gofal sylfaenol a gwyddor iechyd Llywodraeth Cymru; a hefyd Dr Andrew Goodall, cyfarwyddwr cyffredinol iechyd a gwasanaethau cymdeithasol a phrif weithredwr y gwasanaeth iechyd yng Nghymru.

Croeso i chi i gyd. Mae gyda ni nifer helaeth o gwestiynau ar fodelu ac ati'r bore yma, ac yn naturiol, mae gyda ni bump o dystion hefyd, felly fe gawn ni weld sut aiff hi. Mater hynod bwysig. Rhun ap Iorwerth i ddechrau.

So, we'll move straight on to item 2, and an evidence session with regard to COVID-19. Our global audience all know what we're doing at the moment as a committee on a weekly basis during this pandemic. We are scrutinising the response of the Welsh Government, and all other agencies, to the COVID-19 pandemic. So, today, we have an evidence session with the Welsh Government's technical advisory cell, and to that end, I'm very pleased to welcome to our screens Dr Rob Orford, co-chair of the technical advisory cell and chief scientific adviser for health with the Welsh Government; Fliss Bennee, co-chair of the technical advisory cell of the Welsh Government; Dr Brendan Collins, chair of the policy modelling sub-group, a member of the technical advisory cell, head of health economics at the Welsh Government; Craiger Solomons, chair of the all-Wales national modelling forum, a member of the technical advisory cell, and lead analyst, primary care and health service directorate at the Welsh Government; and also Dr Andrew Goodall, director general of health and social services and chief executive of NHS Wales.

A very warm welcome to all of you. We have a vast number of questions to ask you on modelling and so on this morning. And we have five witnesses too, so we'll see how it goes. It's a very important issue. Rhun ap Iorwerth to begin with.

Roedd hwnna'n ddechrau da, onid oedd? Rydw i am ddechrau, os caf i, efo'r cwestiynau rydw i'n meddwl bod y rhan fwyaf o bobl yn mynd i fod eisiau chwilio am yr ateb iddyn nhw rŵan, sef pryd ydyn ni'n mynd i wybod ydy'r firebreak wedi gweithio. Mae hynna'n gwestiwn eithaf amlwg. Mi rydym ni'n aros am ddiweddariad ar senarios, reasonable worst-case scenario Prifysgol Abertawe. Pryd gawn ni hwnnw, yn gyntaf, er mwyn gwybod beth oedd impact y clo dros dro?

That was a good start, wasn't it? I want to start, if I may, with the questions that everyone will want the answer to, namely when we'll know whether the firebreak has succeeded. It's an obvious question, isn't it? We are waiting for an update on the reasonable worst-case scenario modelling by Swansea University. So, when will we receive that, so we can know what the impact of the firebreak was?

Pwy sydd eisiau dechrau yn fanna? Dr Collins.

Who wants to start? Dr Collins.

Yes. Thank you, Chair. So, the firebreak will be successful if it slows down transmission of the virus and if this feeds through into reduced harms in terms of hospital admissions and deaths. So, in terms of the data that we've been looking at over the last few weeks, we have seen a bit of an early signal in terms of the reduced number of confirmed cases over the last week, and, crucially, we've seen reduced positivity, so this is the proportion of people who get tested who test positive. So, both of these seem to show a signal of having fallen over the last week or so. So, that's a good sign.

In terms of hospital admissions and deaths, because these are a bit more lagged, so these happen a few weeks after people are infected with the virus, we don't expect to see a signal right now, but maybe in the next couple of weeks, we should see a signal around this.

In terms of the reasonable worst case, we keep this under constant review. So, every week, we look at the data and see how it compares to where we are with the reasonable worst case now. The reasonable worst case is a long-term scenario. We don't necessarily expect it to be accurate from week to week, it's more about representing a challenging scenario for what we might see through the winter, but we do have it under constant review, and we have scenarios where we've incorporated the potential impact of the firebreak, which we're now refining, based on actual data on what has changed since the firebreak. So, we are reviewing the reasonable worst-case scenario, and if we think that it's time for a new version of it, then it's likely that there will be a new version in the next few weeks.

09:35

So, we can expect in the next few weeks, and you would maybe expect an update every fortnight or so. Is that the way you're, as a rough guide, expecting things to go?

I think Rob might wish to comment on this.

Thank you. So, the reasonable worst-case scenario has a set of assumptions in there about what the future could look like, not what the future will look like. So, we've got to be careful that we're not confusing the reasonable worst case with the short-term projections and what we're actually observing, and that's the important thing, what we're observing and what we might think comes next in the next couple of days, the next two weeks, the next six weeks. So, the reasonable worst-case scenario sits there in the background, to say, 'This is reasonably what "bad" could look like for us.'

Now, we can change that scenario. Say, for example, policy colleagues say, 'Look, we'd really like you to think about a projected cold snap, or we're really concerned about flow, and we need to see what that would do to the overall dynamics.' Then, that's something we can go back and do. So, we've got the reasonable worst case in the background, and then we'll track where we are compared to that over time. So, it's important—it's not what the future will be, it's what it could be with these assumptions, and it's very complex, all of the different—[Inaudible.]—that are included in the reasonable worst case.

There is clearly a link between the modelling work and the work of TAC through to the policy oversight, and then, ultimately, review mechanisms that Ministers have in place. So, there is a review mechanism of the firebreak, as Rob has outlined, but, inevitably, in a fast-moving environment where we're learning more about the virus, there'll continue to be mechanisms that we need to adapt, including some of the operational experience from the NHS, for example, and, clearly, announcements like through this week about vaccine potential and any scheduling that would happen, should that vaccine actually become available to us, would have to be drawn into that modelling approach as well, because the modelling is about trying to understand the mitigation actions that are taking place in the system and seeing how much they have an impact on reducing the chain of transmission.

Talk us through the kind of data and information that are fed into that Swansea University process.

Thank you. So, we are quite lucky in Wales to have a great working relationship across the NHS and wider public service through the all-Wales national modelling forum and policy modelling group. Swansea University, as part of the policy modelling group, is provided with data on cases, and that's from Public Health Wales, and additional deaths data from the Office for National Statistics, and hospital activity data from NHS Wales—[Inaudible.]

In addition to this, Swansea University are part of the SPI-M modelling group, a sub-group of the Scientific Advisory Group on Emergencies, who provide data from all four nations. As part of the group, they have the most recent planning assumptions that have been used by other modelling groups, and that input into the SPI-M and SAGE planning scenarios.

We also see the proportion of cases that are asymptomatic, from that, infectivity, the proportion of cases ending up in hospital, and other modelling parameters. They are also provided with demographics data, such as the age structure of the population from the ONS and Welsh Government sources. The majority of these data are made available to approved researchers as well through—[Inaudible.]—which is the Secure Anonymised Information Linkage database at Swansea University, and that's somewhere that Wales is really thriving. SAIL enables data to be anonymously linked to enable our approved rate of researchers access to datasets to answer research questions that have been identified, and a really good example of where that's been used as well, quite recently, is in the Welsh Government publication on shielded individuals.

09:40

You suggest that Wales is pretty well-placed to bring all this information together, but to what extent do you compare worst-case scenarios here with the worst-case scenarios for other nations in the United Kingdom and compare the way that you compile the data in order to see if we genuinely do have like-for-like comparisons? Fliss Bennee's nodding there—did you want to come in there?

Diolch yn fawr. So, I'll try not to be as loquacious as last time, Chair. My apologies for that.

Hopefully it won't have escaped select committee members' attention that we actually have a slightly different peak and at a slightly different time in our reasonable worst case for Wales than for the UK as a whole. Now obviously, because, as Dr Orford has said, the reasonable worst case is based on a policy scenario, our reasonable worst case has a slightly lower peak in cases as a proportion of the UK's one, but actually we also predicted our peak in December—this was before the firebreak was considered. The UK SAGE scenario was commissioned by Cabinet Office in central Westminster, and that had a peak in early 2021. It's difficult because we can't necessarily compare them directly because the policy scenarios are commissioned by different national Governments, and if the policy scenarios are different, then the reasonable worst cases that come from them are going to be considering different effects. That's because we're reflecting different demographics, different geography, and of course our different real-world data. So, the situation that England, Scotland and Northern Ireland are in is necessarily different, but we try to find consensus by coming together in SPI-M to discuss our reasonable worst cases and then bringing them to SAGE and to the technical advisory group.

There's a bit of a chicken-and-egg question coming here now. You've repeatedly said there that your worst-case scenarios model the policy direction at the time. To what extent, though, does your worst-case scenario planning then colour the decisions taken by Government on its policies? Which way around does it happen? Does Government give you a policy direction and ask you to predict what will happen? Or do you genuinely influence what happens in policy by showing them, 'Listen, this is the worst-case scenario if you do x, y and z'?

Craiger, obviously, is the chair of policy modelling, and is probably best placed to give the specific answer. My answer, of course, frustratingly, would be 'both', because as a scientific group, we need to consider all of the different things, and indeed how to model them. But of course, it is to Ministers and to Dr Goodall as the head of the NHS to say, 'What would happen if this occurred?', especially with regards to other pressures and other harms coming built into COVID. We take those commissions and model them. But I'd ask to go to Craiger for a fuller answer.

Thank you. I guess, to try and answer the question, our approach is typically guided by policy in that we always have a policy commission to be able to establish analysis, and there's always a scenario that's depicted to us to be able to evaluate with modelling. So, whilst there is obviously a conversation as part of that, we do have iterations of models and we do discuss underlying assumptions and things that we should be considering as part of developing the model with our policy colleagues. But typically it does come from a policy scenario, and anything that we identify from the wider scientific evidence that we should be taking into account.

Yes, just for completeness, it's obviously important to recognise that the scenario planning is there to allow us to determine the actions that we want to take, and certainly as we look at the way in which we would be responding from a Welsh Government perspective, but also to understand where the NHS was, it's not a set of scenarios to observe and to think that we'd ultimately get to the end point. We would make some assumptions that there would be actions that would be taken, for example, by the NHS in respect of its response, the balance of activities that we would take. And at the Welsh Government level, given that we have the coronavirus control plan in place, one would make assumptions that Government itself would act as it has through the firebreak mechanism to come in to try to mitigate some of the potential extremes of the scenarios that have been set out in the reasonable worst-case assumption.

09:45

Is it too simplistic to say that what you’re trying to do is do better than the worst-case scenario at all times? Actually, going back over—. Yes, come in there, Dr Collins.

I think you’re right about that. The reasonable worst case is just that—it’s a worst case, it’s not what we expect to happen or what we want to happen, and it’s meant to represent a challenging scenario. We look at what we need to plan in terms of NHS capacity for different scenarios, but, as Dr Goodall said, we don’t want to hit that capacity because we know that, sadly, a proportion of people who get the virus will sadly die from it. So it’s not just a case of having enough NHS capacity to process people through the system; it’s a case of that we don’t want as many people to get the virus, because somewhere between one in 100 and one in 200 people who have the virus will sadly die from it, so we need to think about how we mitigate those kinds of harms as well.

But leading up to the firebreak, Wales was actually tracking the worst-case scenario. Does that just tell us that policy isn’t working, and that’s when a different policy direction is triggered—in this case, the firebreak? Dr Orford.

Yes, thank you. These are models, these are forward looks at what could be. What we have now, both from our country and others, is data showing what actually is happening. And what we’re doing is we’re very, very closely evaluating what’s happening in Wales, whether that’s through the local interventions that we saw a few weeks ago or the national firebreak that we’re observing now. So, that’s the most powerful thing that we can do—actually evaluate what we’re doing and compare that with others. So, for example, in Northern Ireland, they’ve had a longer national intervention that’s slightly different in that the schools were closed for two weeks and then they’re back on for two weeks. So, it’s important that we work with Northern Ireland, work with Scotland on their local interventions, work with England on their tiers to understand what has the maximum impact and creates a minimum amount of disruption and harm.

It’s really, really difficult to assign causality in what we’re seeing in those trends. We look at lots of different things, to say, ‘What is the impact?’, whether that’s movement, whether that’s per cent positives, whether that’s hospital admissions. Some things we can evaluate really quickly, and other things we’ll have to wait and think about. But just in Wales alone, we are very different across local authorities, whether that’s demographics or population density, the types of industry that are in areas and types of communities that we have, whether they’re very connected, whether they’re rural, whether they’re urban. And so we need to take all this into account, so it’s very, very complex, and colleagues like Brendan and Craiger and colleagues that work with them are constantly churning the data, rerunning it. They’ll see new data and they’ll rerun it and keep asking these questions. So, going forward, it’s really important that we’re evaluating actually what we’re seeing and using that to say, ‘Okay, what will come next? What would we reasonably expect to come next?’ and then go again. And so it’s an iterative process.

Sorry to interrupt, but what would be your evaluation, then, for example, of the kind of local restrictions—because there could be other kinds of local restrictions, of course—that were put in place, starting in Caerphilly and then rolling out over the past weeks? Because the data that we have in front of us suggests that, actually, with those restrictions in place, we were still on track for the worst-case scenario. Does that mean in future you don’t advise going down the route of introducing those kinds of restrictions again, or local restrictions at all?

We’re really carefully evaluating them to see the difference between local authorities, how they behaved at the time, why they may be behaving differently over time, what would have happened if we hadn’t had an intervention there, because that would have looked different from what we actually observed. In terms of the analysis, Brendan or Craiger will be able to give you examples of the types of statistical analysis that’s ongoing. I’ll hand over to Brendan.

Yes, what we’re looking for is basically whether there’s a change in the trends curve. So, there are methods like interrupted time series using segmented regression analysis, where we’re looking at the trends before an intervention was introduced. We know that the incubation time for COVID is around five days, so we might expect to see some difference more than five days after an intervention was introduced. We’re also looking at methods like autoregressive integrated moving average, which is another method used for looking at interrupted time series. With these kinds of interventions, we don't have a control group to compare against. We can try and look at synthetic controls, but what we're actually looking at is interventions that are being introduced into very complex systems, with a lot of different dynamics in those systems. So, people might change their behaviour a little bit before interventions are introduced; they might change their behaviour after the interventions were introduced. It can be tricky sometimes to isolate the effect of these interventions that are introduced, but we are doing our best with a range of methods, and working with our colleagues in Public Health Wales and groups like SPI-M to try to use the best methods to understand whether we see the kind of discontinuity in the curve at the point at which interventions are introduced. 

09:50

Diolch. I also just wanted to highlight that the policy scenarios for a reasonable worst case are, to a certain extent, based on a commissioner's appetite for harm. So, we knew that there was going to be a much higher need for the NHS to be available in the winter, and much less likelihood that we'd be able to, for instance, clear out wards and so forth. So, the reasonable worst case for this winter was not based just on the direct harm of COVID, but based on how much COVID could be dealt with, managed and looked after before we started to see a threat to other services. There are four different harms from COVID, only one of which is direct. So, tracking against the reasonable worst case early, once we'd started to try local interventions, was a very good indicator to feed back that the Government needed to consider something much more serious, which they did. 

One last question from me. Again, just looking at the impact of local restrictions as an example of policy decisions that are made, I think the last time the technical advisory committee came in to speak to us, cases per 100,000 were key drivers in relation to justifying local lockdowns. And in a number of those areas, the case numbers per 100,000 were still way above the thresholds that were set at that time. How do you now assess things differently, in a way that justifies the lifting of those local lockdowns, given that—. Talk me through why it's safe not to have those local lockdowns now, when they were certainly justified in your minds a few weeks ago. 

Thank you. I think there may be a misunderstanding in terms of it not being considered safer. The local lockdowns have not been lifted; they've been brought to be national restrictions. So, our advice is, and always will be, that interventions should keep the effective reproductive number below 1 if possible, and otherwise as close to 1 as possible, so that we have the slowest exponential growth as we can manage. The advice we've given, and the evidence that we've based that advice on, was that restrictions will be more effective if they're simpler. And that includes having one set of regulations for the whole of Wales instead of multiple, up to 22, regulations for different regions. And alongside that, there is this overriding need for the public to have a more balanced share of responsibility for COVID risk reduction. Individual decisions about what's right for the individual, about what's too risky—. And they're harder to make; if there's a lot of regulation, then people tend to lean on those regulations to see them sometimes as targets instead of limits, and really not to change their behaviour to start to learn and understand what can reduce their risk, and how they can live sustainably. And until there's evidence of a better understanding of the risks, and until there's evidence of a real change in behaviour in the public, it's not really responsible for Government to remove too much regulation. The ideal situation is one in which there is no need for central regulation because people are able to understand and to make those risk decisions for themselves, not only about their personal risks for COVID but taking that shared responsibility for community well-being. 

Yes, as few restrictions as possible, but properly enforced by Government or by people self-enforcing. That's certainly the way I see it. Okay, thank you. 

Definitely not lifting restrictions, and not safer, you're right. The incidence level remains high, and the infection and positivity levels remain high, and they're still very dangerous. 

Dr Goodall, you were indicating earlier. Sorry, we missed you. I think you want to come in here. 

09:55

Yes, it was only to comment, again, on recognising the actions that we put in place. And there is a danger with modelling work—was it right or was it wrong? If we look at the first-wave experience, actually we were able to mitigate the number of hospital admissions that had originally been forecast, and certainly the number of deaths, materially so. But obviously, irrespective of our preparation on the NHS side, it was the actions that took place from a society and from a community perspective that really did contain the transmission of the virus and brought things down. And there is a difficulty sometimes in that people say, 'Well, was the modelling right or not at that time?' What happened was an extraordinary set of actions that was able to bring Wales together, the UK together, to actually respond and act. And we need to make sure that we're continually able to bring in those assumptions because I would never want the reasonable worst-case assumption to happen.

I was at the Public Accounts Committee, speaking to Members there, on Monday and ended up with a discussion on field hospitals. So, is it a sign of success that we were able to fully utilise all the capacity within a field hospital environment, or is it a sign of success that we didn't need to use it? But at least it was available if needed. And I think that's a good, balancing view of how we should be using it to respond on a society basis, but also from the NHS.

Excellent. Thank you for that. Moving on to the next section of questions—David Rees.

Diolch, Cadeirydd. Good morning, all. In your answers so far, I think we've got a reasonable understanding, to an extent, of policy modelling and its impact, but I suppose I just want to get more clarity for my own purposes. Now, you've indicated that, in policy modelling, sometimes you model policies that are given to you by Government and, at other times, you may be creating a scenario that then creates policies, as such. I suppose what I want to try and work out is, in the situation of the firebreak—Swansea University's modelling was such that there was a two-week or three-week option, based upon the evidence and the modelling you did—what advice do you give to Government in those scenarios as to which one, perhaps, would have a bigger impact? Because one was two weeks, and a three-week possible knock-on effect; the other one was three weeks, with a five-week effect. And particularly as we were seeing the peak come closer and the wave, almost, closer to us than to Christmas time—the last time we talked, it was around about December we were talking about a peak—and where we are now. So, what's your advice to Government in those situations? Do you actually make a recommendation on one or the other, or do you simply say, 'These are the two scenarios we have', and leave it to them?

Thank you so much. So, just to start, obviously we're extremely grateful to our colleagues at Swansea University for their work in modelling the different policy scenarios. It's really important to note that the modelling is one part of the evidence that's provided to policy officials to then advise on options. So, we provided the summary of the impact of the two-week and three-week firebreak, as you mentioned, and that was then compared to the current trajectory, using the most timely data available. So, there were a few particularly useful elements of that modelling that we wanted to highlight, which you've also mentioned. So, notably, at the height of the peak of the curve and the delay to where the peak would be, and then that was able to feed into the wider evidence that was provided around the practicalities, and the wider harms as well. So, essentially, looking at data and trends was a big element of the decision-making process. But, obviously, there's the practical elements that also need to be considered by policy officials to be able to implement the advice.

Can I ask—[Inaudible.]—policy officials? Because clearly, Andrew is a senior player in policy officials in one sense. I know that's in operation, but you will be giving guidance to the Welsh Government on this. So, how does that work? If you were advising, and Andrew's involved with you, surely that's the recommendation that should be going to Ministers.

It's not just a single area of Welsh Government that would make these sorts of decisions. It's Cabinet as a whole that has the opportunity to discuss and evaluate each of the options, and there is a lot of evidence that comes from across the Welsh Government. I guess from our perspective, that specific piece of modelling was focusing on the direct COVID impacts. And that is one piece of the equation that needs to be brought together with a wide set of advice from colleagues across Welsh Government, including legal and other policy areas, because obviously, something like the firebreak has a much bigger impact than just health.

10:00

Rob, did you want to raise a point? You put your hand up.

Yes. I think Craiger made the point that this is part of the decision-making process. These options and recommendations come as evidence from across the Welsh Government, so it's just a part. Thank you.

Does your model include economic impacts and not just health impacts?

Diolch. So, yes, the modelling itself is looking at the direct—. The model for the firebreak, for example, was based on the number of infections and how that would lead to a reduction in the number of deaths per day at the peak, which is a very unpleasant thing to have to consider in any circumstance. But obviously, whilst Craiger and Brendan's modelling groups feed in what the specific effect on the infection can be, and whilst our advice to health contains not only advice on the infection itself for our public health colleagues, but also to Dr Goodall for the capacity that the NHS will need to have, we also take in information from the socioeconomic sub-group, which is chaired by the chief economist. We take in information from the behaviour and risk group about the sorts of things that are going to affect behaviour.

The rounded advice that we try to give to Ministers just on the science side, and separate from the legal or from the economic profession and others, is that in order for the firebreak to be effective, obviously, the longer you can do it, the more effect you will have in terms of reducing COVID. But at the same time, if it's not affordable, then there will be socioeconomic harms that will increase. If it's not possible to staff or if it's not possible to run schools, then other educational harms will come into play. And what we have identified is—and this has come through very strongly in some of our behavioural groups' work—if you find out how long you can have the firebreak for, or once you've chosen how long a firebreak would be, you need to announce what the end point will be in order to reduce both the economic harm and the mental harm from thinking whether or not something will continue. So, once they had made the decision, it was vitally important that it was then stuck to, because anything else would reduce trust significantly among the Welsh people and potentially reduce the effect of further restrictions. But, from a modelling point of view, it's just about the harms and then we feed that into a much more rounded scientific assessment.

And do you therefore look—I'm sorry, Andrew, I'll come back to you, don't worry—at the behavioural—? You said you work on behavioural models. Because clearly compliance is a big question we have on the whole approach to the restrictions. How does the behavioural model really link in very closely? Do you make alterations to your health model as a consequence of some of your behavioural modelling scenarios?

Yes, absolutely. The reasonable worst case for the first wave was based not on how many people would react under certain circumstances, but based on whether or not we had a high or a low compliance. So, the reasonable worst-case scenario for the first wave was that only 40 per cent of people were able to bring the behaviours that the lockdown required to the table, because we knew that that behaviour was going to drive the most infections. Sorry, Dr Goodall.

Thank you very much. Just to allow some understanding of the balanced views that need to feature around the table. And obviously, from an official's perspective, the importance of ensuring that the advice is available and can feed into that collective process, but to allow for Ministers to make decisions in respect of their individual portfolios or obviously a collective discussion around the Cabinet table with the First Minister overseeing those decisions as well. And as much as we draw the modelling in and need to have mechanics like the reasonable worst-case scenarios, we need to draw in a sense of the resilience of the NHS, we have to have that economic assessment that is taking place along that.

But one of the features of both the setting of the local restrictions and also as we've gone into the firebreak and out of it has been the opportunity in Wales to use the more intimate relationships that we have with stakeholders for Ministers and officials to reach out to the sectors themselves so that they are participating in the process of decisions, allowing them also to influence and inform. So, the data and the modelling tells you one thing, but also you would need to know, practically, how would the hospitality sector implement a set of changes, how would it work from an education or a university perspective as well. So, I think that's been a strength as well—to add those opinions and views alongside the data that we have available and the evidence base for it. 

10:05

Can I ask, then—? Brendan, I'll come back to you in a second. Can I ask the question, then—? Clearly, I understand the issues, and Rhun was talking about the local restrictions, for example. What kind of timescales do you operate on to assess whether your model, actually, is working, based upon the data being collected, so that you can advise on policies in future down the line? Because people are frustrated when they hear that it's two weeks, three weeks. So, there's a type of timescale you have to collect data and produce your models to see whether this actually is delivering or not. So, on a local restriction, what time would you be expected to say, 'Well, we can now decide whether these policies are working or not'?

Schoolboy error. So, I'll speak for my colleagues, but I know Brendan and colleagues were up until 4 o'clock on Saturday morning, maybe, working through outputs and models, and they are continually turning the handle and evaluating, so it's a non-stop process for us.

That was back on behavioural science—

Just to say that we do factor in behavioural science to understand whether people are adhering to guidelines, and if not, why this might be the case as well. We've got people from our technical advisory cell who attend the scientific pandemic influenza group on behaviours, which is the behavioural sub-group of the Scientific Advisory Group for Emergencies, and this is used to help develop interventions with the greatest chance of having a positive impact—so, trying to understand individual behaviours and trying to meet people where they are. And so far, the collective effort of individuals to really try to reduce their contacts and things like that has saved many lives.

Two final questions from me. One is: on the behavioural science and the behavioural modelling, are you seeing a change in people's attitudes and behaviours now compared to, perhaps, when you were initially modelling maybe in May, maybe in June, ready for the summer period?

I think, from the data that we're seeing emerging from the national lockdown, we've had a really, really good response from the public. Overwhelmingly, people have got involved, they recognise the need for us to do this, and there's been a really high level of compliance, at least from the travel data that we're seeing. People really did stay at home. It's a huge imposition on people, the national interventions. They're a very blunt tool, but the Welsh public have done an enormous job, really, and that's been reflected in some of the gains that we're seeing in the early indicators coming through now.

But we're not out of the woods. We need to make sure that we're all continuing to think about our individual choices and that we're saying, 'It's not what I can and can't do—it's what I should and shouldn't do, in terms of the number of contacts that I have.' And this will take us through. If more people have fewer contacts, then we'll see less growth in the epidemic. So, it's all about mixing, and we just need to keep people engaged and aware and informed of what sensible choices look like.

Chair, I've got one final question, but I saw that Rhun put his hand up and wanted to ask a supplementary here.

Just on the behavioural science of it. This is a question, perhaps, for the policy makers—for the Minister himself—but I'm going to chance it anyway. Is enough being done with the results of the behavioural science work to share that with the public by means of holding the mirror up to people, to say, 'Listen, this is your behaviour, this is what you need to change'? Because, generally, I'm just hearing a lot of the same messages time and time again, going back to March, really, of 'Stay at home, do as little as you can.' Is there more, specifically, that could come out of that behavioural science that could help, in messaging terms, drive a change in that behaviour?

Obviously, Ministers will have their perspectives on this in terms of how they approach it on the decisions themselves, but I appreciate that there is some frustration that some of our messaging has been pretty consistent throughout, and that's because they remain the things that make a difference, and it's really important to emphasise that whilst there may be fatigue on some of the messages, it's really important that we do know that social distancing, that hand hygiene et cetera do make a difference in terms of breaking the chain of transmission at this stage. But it's really important to make sure that we continue to focus on communication and engagement. That is a big feature of, again, collective decisions that are taken, the importance of underpinning things. And one of the pieces of advice that came through from the TAC team that Ministers needed to reflect on was the importance of the coming out of the firebreak arrangements with a sense of simplifying the rules in order to be able to convey them and communicate that. So, that was a very strong characteristic, I thought, of both the advice and then featured as part of that very balanced discussion around the table. 

10:10

I've swapped to another microphone, so hopefully you can still hear me without the crackle. The behaviour and risk communication group actually identified that one of the risks to the technical advisory group's work was the fact that we weren't working closely enough with a variety of policy groups both in the central and in the wider public sector. So, we've been working very closely not only with the communications group now—we've started publishing for transparency—but also we've started trying to do explanatory videos with members of the advisory group, with academics and with Ministers as well, and anything we can do and anything you could help to do as Members of the Senedd to get some of those really clear messages across.

Overwhelmingly, what we see is that the majority of people in Wales really do want to do the right thing, they're keen to help and sometimes there's a confusing message, but anything we can do to help them understand some of those more complex things—the fact that it's not about choosing which six people you want to go to the pub with today, but thinking about, 'Are there a few people that I could try to see without swapping between groups?'—that would be less risky and it would still allow you to have some social life. So, things that are sustainable in the long term. We need to get over the feeling that this is very short term, because COVID's going to be with us even after the vaccine comes—and there will be several vaccines, hopefully, that have a beneficial effect. We're still going to need to work out how to survive with this without causing extra harm, and that's going to need people to change the way they think about what they do.

I agree, because with things that have been effective and really have been taken on board by people—you know, washing your hands—you're building there on something that we've been told to do since we were toddlers, 'Wash your hands because it's a good thing to do', except we're having to do a lot more of it now. There are things like keeping out of people's houses, keeping away from your friends, which need much cleverer messaging and maybe that's what we've seen not getting through in the close-knit communities of the south Wales Valleys, for example.

Part of that, I think, is perhaps because we've tried to give a really simple message, but we haven't necessary explained. The reason why washing your hands with soap and water is useful is because with this virus, the envelope of lipid, of fat, that protects it so that it can be in the air and alive and then come into your respiratory tract, your throat, and settle and start to grow—the combination of rubbing, that abrasion, with soap, which breaks down fat, takes away the shield, the shell of the virus, and kills it, and that's why it's important. And perhaps if we explained a bit more why it's important to do a certain thing, then people would be able to understand what it was that was driving those messages.

The final question from me, then. Obviously, the behavioural modelling is important, and therefore it has an impact on what your policy modelling looks like, what does each policy do and how it impacts upon it, but we've been focusing very much on the COVID work. What about the non-COVID work as well? How does that play into your modelling agenda? Because, clearly, in the first lockdown, the NHS decided, or the Welsh Government decided, not to actually undertake any basic elective work or other type of services whilst they dealt with COVID. How does the non-COVID work fit into your modelling of how services are working?

It may be worth me just giving an overview, because that's how the NHS is responding, and then maybe colleagues may want to fill in some of the data side of things. The modelling work, from my perspective, tells us how many hospital admissions we should expect, and therefore we can make assumptions on the level of capacity that needs to be taken. I see it as less about how we should create the capacity. I think that's an NHS operational role that we engage with the health boards and the health organisations across Wales. Certainly, as we needed to respond to the first wave, one of the reasons why we needed to step away from routine activities at that point and to step up very significant opportunities about capacity was because we knew that we needed to create a very high level of capacity that wouldn't have been possible if the NHS had just continued with its day-to-day work at that time. So, I know it was an extraordinary decision, and it was part of our preparation, but it was necessary, and, of course, that was a consistent action that was taken across the UK in respect of the different NHS services.

I think where we are now is about trying to ensure that we're able to, given that we know more about the virus—how we can focus on the other harms that colleagues have described. So, it isn't only about the response to the harm caused by COVID, it's the recognition that we also have an ongoing responsibility to accommodate patients within our system, and we have therefore seen a level of activity being restored over these recent weeks and months and we've had to learn and adapt to those environments. The one thing unfortunately that we've learnt is particularly in respect of our more planned activities happening in any setting, that we are slower in our ability to see and treat patients, simply because of the way in which we have to accommodate a different environment that is safer. Of course, the use of PPE by staff just slows things down in those sorts of environments. But we know that we need to maintain more non-COVID work. We need to increase that. We need to develop our plans for that, even at the most difficult times, and our current approach within the NHS is constantly on a daily basis looking at our available capacity, to ensure that we can keep as many activities going in as many settings for as long as is possible.

10:15

Let me ask a question, then, because I understand what you're saying and we've had this message time and time again; I fully understand it. I suppose my question is: does your modelling also look at excess deaths, not just COVID deaths but excess deaths, and some of which may be the result of non-COVID work not being available, for example?

Yes, we do track that. It's part of the ONS reporting. Again, colleagues will feel that they want to come in on some of the data areas. We've actually tracked that very carefully in Wales over these recent weeks and months and we do draw it into there, because we know that the Welsh population has more significant underlying health characteristics, and we also know that we have an older population proportionately in Wales, compared with other countries as well. The current excess mortality that ONS are reporting, even with this week's figures, is showing that Wales is operating at around a 12 per cent excess mortality. I think England's figures at the moment are around 19 per cent for the March to October period of time. We probably would have expected that to have been higher and hopefully some of the actions that we've put in place have been able to mitigate that at this stage. But, clearly, there is a focus on excess mortality.

In the longer term, we will need to continue to work through those, because if there are individuals who are not accessing services, or feel that the system is not able to support them through, we do need to make sure that we can continue to focus on those, and that's why in March, April, when we were giving guidance to the NHS, we had a particular focus on essential services. It was in line with the WHO guidance, but it was important that we wanted to keep a focus on other outcomes and other harms, inevitably, even though we were going through a very exceptional response. And at the moment, the NHS has twice as many non-COVID patients in hospital beds as we did in March, April, for example, and we're still maintaining some of those activities. But we're not at normal activity levels at this moment in time. There's been a recovery and a restoration of activity, but it's very difficult for the NHS to operate at the normal level of activity that we were experiencing before. Colleagues may want to comment on the data.

I think Dr Goodall said quite a lot of it, but the initial response to the virus, the pandemic, was an emergency response to try to get the virus under control, but I think since then, we've tried to have a more fine-tuned approach in terms of our advice, in terms of choosing non-pharmaceutical interventions that have the best trade-off between controlling the virus but trying to minimise harms to individual well-being, harms to the economy as well as indirect health harms.

The focus of our modelling work has really been on direct COVID harms, but there is also a socioeconomic harms sub-group of the technical advisory cell looking in particular at the long-term scarring effects of a COVID-related recession, and these will particularly affect young people. We've done work at looking at health inequalities, so we know that COVID isn't a great leveller. We know that people in deprived areas have suffered both directly from the virus, but also from the economic consequences of the virus and the response. And there's also a sub-group of HSS recovery looking at indirect health harms as well, so we're looking at these kinds of things. And I think there's an essential services group within Welsh Government as well that looks at how to maintain essential services.

So, there is a lot of work looking at indirect harms and looking at socioeconomic harms and how we mitigate these harms and trying to look at the long-term picture as well as the short-term picture of trying to keep the virus under control.

Thank you. There's a heck of a lot of research that's going on as well associated with the technical advisory group. We've got a research sub-group that's chaired by Dr Kieran Walshe, director of Health and Care Research Wales. We work really closely with SAIL—the Secure Anonymised Information Linkage Databank. That pulls in lots of different nested assets, and enables us to look at the impact of cohorts of people that may have had COVID or not had COVID, or different other diseases and morbidities. That's remarkably powerful, the use of those assets in Wales, and building on clearly what is a significant strength in Wales when you look across the UK. The dynamics of research around COVID really has changed. It's very rapid research. Just enough information for us to make decisions and to inform policy is of fundamental importance, and so understanding what's happening quickly, accurately, comparing that and then asking questions—'What are the impacts here? What are the likely impacts there?'—is really important in informing us going forward.

10:20

Thank you very much. It was from the excess deaths points, I just wanted to highlight specifically that we—. As part of the analysis on reasonable worst case and other models that are produced, the Office for National Statistics does produce an adjustment measure for us to take into account excess deaths. So, that work does happen. Interestingly for Wales, it has been quite a lot lower, the change, than other nations, so the adjustment we need to make on the models actually is quite negligible.

Reit, amser i symud ymlaen nawr i gwestiynau gan Lynne Neagle. Lynne.

It's time to move on now to questions from Lynne Neagle. Lynne.

Thank you, Chair. Obviously, it's early days in terms of the impact of the firebreak, but are there any early indications that it's been successful in moving the peak later in the year?

As you'd expect, we're very much monitoring the impacts of the firebreak and, as I said at the start, I think we are seeing a signal in terms of a reduced number of cases over the last week or so and, crucially, reduced positivity, so a smaller proportion of people testing positive. We've also seen very positive signs in terms of reduced mobility and more people staying at home. So, we've got data from a range of sources in terms of people making trips outside, and whether people are at home as well. So, we've seen quite a lot of positive signals, and then in the future we're looking at a range of indicators like 111 calls and GP sentinel surveillance and, obviously, things like hospital admissions and deaths, which are more time-lagged indicators because they occur some weeks after infection. But, so far, we think the signs look good in terms of the firebreak at least having some kind of effect.

I think Brendan's covered the majority of it there. Essentially, the indicators—. There's the additional indication of the ZOE prevalence—that's also showed some decreases as well. So, there are the signs there.

Does the same go, then, for the impact on hospital admissions and the critical care capacity? Is it still too early to say?

Yes, I think it is too early, but we review this every day; we have contact with the system. It's really important that we allow ourselves a few days to see the implications because, really, it's going to take two to three weeks to see the full impact flow into the NHS, certainly in respect of hospital admissions and in respect of critical care. There are some signs of a slowdown in hospital admissions. I can't say that that yet feels like it's a trend, but it's enough to—again, if it's encouraging signs, there are some early encouraging signs at the moment about that. But the reality is that it will still be really, I think, the next 10 to 14 days that will allow us to be more confident about that. Clearly, irrespective of the community prevalence, it will be the impact on the NHS, because the more we can protect the NHS, the more we are able to make sure we can carry on with that balance of other activities, and not just simply revert to a COVID-19-only emergency response.

Okay, thank you. In terms of where we are, we've not really hit what would normally be the peak winter pressure time yet, have we, in Wales? How are you factoring in the normal winter pressures into the modelling that you're doing?

10:25

Perhaps if I give the general statement for the NHS, because I'm bringing the operational perspective to the table here, and then the modelling side of it can be picked up. If you think about our normal winter experience and the level of work we do on modelling and numbers to try to mitigate some of those pressures, normally in a winter you'll see the NHS looking to put about 500 additional beds into the system. We've made plans and contingency plans that would allow us to expand with additional capacity physically up to about 5,000 beds, and I think that shows the scale of how we're trying to balance normal winter pressures alongside what we need to do here. There are some other factors we need to account for. If we were going into a winter where it looked like there was going to be high flu prevalence, then clearly that would be an extra dynamic within our system that would lead to hospital admissions, and there is some hope from perhaps some of the southern hemisphere experiences, and given the way society is behaving in terms of social distancing and hand hygiene, that perhaps we won't see that conversion in the same way. To reassure you, we look at this from an operational perspective within the NHS and I'm able to fall back, therefore, beyond the modelling on just some of our normal day-to-day experiences and contacts, so the reviewing of the escalation levels within the system, seeing how the responses are working around the different areas, looking at individual communities, perhaps, where prevalence is higher to see whether that is having a bigger impact on some of the local health systems.

Okay, thank you. In terms of people who've had COVID being discharged, is there any information and modelling on those patients?

Again, colleagues may want to draw in some of the data and the evidence around it, but we estimate that there's probably been around 16,000 discharges of patients that came into hospitals on a COVID pathway since March, and there will be different experiences from that, and there is still growing evidence around it. So, there will be a number of individuals who, despite a hospital stay, will have been able to get reasonably back on a normal track. It probably would have taken them some time. You will have seen the development of the concept of long COVID over these recent weeks and months, seeing individuals that are really struggling to get back to normal lives, with much longer term effects. Positively, we've introduced some work around rehabilitation frameworks in work even back in May, and have been able to use some of that to make sure that we're able to track these patients coming through, but these are lived experiences and we need to learn from how patients are recovering at this stage.

Certainly, we can see some impact in some of our areas. So, as we go through this second wave, we are seeing some perhaps different critical care experiences in respect of treatments and interventions that can make a difference, and that should be impacting on the number of patients who we're seeing now in critical care areas. Originally, the modelling of the first wave was that about 30 per cent of patients would end up in critical care. I think in our overall experience, probably, that's been about 12 per cent, which obviously changes the assumption on some of the critical care numbers that we would see. I think, even in the recent weeks, as we're building up to a second wave, we've actually seen areas of Wales that have been below that as well, even to 7 or 8 per cent as well, and that is very helpful because, ultimately, the fewer patients who will end up in the critical care environment, we would therefore expect mortality to be much improved, and therefore we are saving lives. So, there are a lot of important aspects about what we're assuming, and if critical care can have a better experience through the winter, then that would benefit Wales as well as the NHS system. What I should say, however, is that critical care is busier than it normally is, so whilst we have COVID patients passing through, currently there are over 180 patients in our critical care beds. They will be busier just dealing with normal, urgent elective patients coming into our system as well, and I think we need to respect that level of workload that is on the shoulders of our critical care staff and our intensivists across Wales.

Andrew R.T., you had a supplementary on this point, and we'll come back to Dr Collins then.

Thank you for that, Chair, and my apologies to the witnesses and the committee for being late this morning. Just on the point of demand and the service demand in particular that Dr Goodall was touching on there, as a regional Member, I cover two health boards and I get a briefing at the end of each week, on the Thursday from Cwm Taf, and from Cardiff and Vale on the Friday. I was amazed at the difference in how those two health boards, which are cheek by jowl, are experiencing demand in their hospitals. On the Thursday, a briefing that said that every bed practically available to Cwm Taf in critical care was full, whether they were COVID patients or non-COVID patients. In Cardiff and Vale, 80 critical care beds available, eight patients—eight patients—in those beds on the Friday. Is there evidence, and I look to the wider witness base here as well, that shows that the Cardiff and Vale area was harder hit in the first wave, and then, with the second wave, obviously, other areas now are having a greater infection rate and a greater pressure on the health services within their areas? Or is this solely down to the hospital-acquired infections that, sadly, the Cwm Taf area seems to be the hardest hit—from what I can see, anyway—health board area in Wales?

10:30

Yes, that's a really excellent question, and it would be really good to look at those trends over time of what happened in wave 1 and wave 2. Certainly, in some countries—there seems to be a difference between countries, so, eastern Europe looks to be hit differently to western Europe now. I think—and I'm speculating here, so be very careful—really, the force of the epidemic in those communities where we have areas of much higher community incidence—it will force its way through the population and it will find its way to older people who are more vulnerable. And we think that's what we're seeing in those areas that are notably higher than Cardiff and Vale.

Now, the Cardiff and Vale area seems to be higher, probably being driven by younger people, younger adults—you know, university age. People are in university, they're probably not living in extended households; they're probably in halls of residence. So, that was probably reflected in some of the Cardiff figures, and yet, when you look at the numbers of over-60s in, say, the RCT area, it's much, much higher. And so we're carefully monitoring this; we don't want to see Cardiff go that way, and we really want to see those areas come down. Ultimately, it's the force of the community epidemic that's forcing its way into the more sensitive areas of the community and finding its way into hospitals. Thank you.

Dr Collins, sorry, did you want to make a point at this stage? Then I'll come back to Dr Goodall.

Yes. I was coming in on the question before, but, on that question as well, I think Cwm Taf was being quite hard hit in both waves of the virus as well. So, I wouldn't necessarily say that areas that have been hit hard the first time will have been spared the second time around. We know that it's likely that only a small proportion of the population have actually had the virus. So, we've got the prevalence of antibodies from the Office for National Statistics COVID infection survey, which indicated that around 4.2 per cent of a sample of the community actually tested positive for antibodies. Now, it may be that antibodies wane over time, so it might be a little bit higher than that, but it's certainly not the case that we're at a point where we're reaching any population immunity type effect. So, that was just on that point.

I think, as the Member outlined in his question, there inevitably will be a combination of factors. A couple of reflections from Cwm Taf Morgannwg will be that, inevitably, there is a higher level of community prevalence. It'll find its way of feeding into both the population, but to the broader system, and it will reflect that, whether that is within a healthcare environment, whether it's in a care home environment, inevitably, because it will reflect the nature of its communities. I think, also, this is a virus that targets vulnerabilities, so, where there are areas with underlying health characteristics, inevitably, the virus will—it gets traction there.

In respect of the balance between areas, it's local community pressures that will feed into those hospital environments. A difference in critical care, however, between Cardiff and Vale and where it is on its provision, as opposed to Cwm Taf Morgannwg, would also reflect the nature and the function of the hospital. So, of course, the University Hospital of Wales, with its more regional role, with a much larger critical care facility and the more specialist work that it's undertaking, will have more flexibility on its available capacity as well. So, that will be part of it.

But we have seen the virus operate differently at different times in Wales. So, it's interesting, at the moment, looking at Betsi Cadwaladr in north Wales, for example. Clearly, it has prevalence within its communities, but there is a lower level of capacity being taken up at the moment in north Wales, for example, and there are clearly connections around travel routes to the north-west. Perhaps one would have expected that maybe that could've been higher at this stage, but they are definitely at a lower level of capacity use than they were in the first wave. But that can change very quickly, as we know, over a matter of weeks. 

A final question from me: what modelling has been done, or is possible, then, on COVID transmission within hospitals? 

So, our current models include nosocomial cases, so hospital-acquired cases, in the data that we use. Modelling the dynamics of hospital transmission is really tricky to do. We know, as the number of people in hospital with the virus increases, then the risk of transmission to other patients increases. There's a SAGE sub-group around nosocomial transmission that Wales is a part of, but, to our knowledge, there isn't a model that you can easily pull off the shelf that allows us to model what can be quite random but quite high-impact events, I guess, in terms of transmission in hospital. There's also a nosocomial group within Welsh Government, and, like I say, people from Public Health Wales engage with the SAGE nosocomial group. So, it's included in the model, but the complex dynamics of transmission within hospitals, I would say, is not completely catered for within the model.

10:35

Thank you, Chair. Good morning. How are you monitoring cases and hospital admissions and deaths internationally?

Thank you. There's a great deal of activity that goes on, actually, in this area, right across UK Government, and in Wales as well. So, we've got the international intelligence sub-group, as part of the TAC group, and daily we monitor information coming in from different governmental sources, like the Joint Biosecurity Centre, or places like the Foreign, Commonwealth and Development Office, to see what's happening in other countries. We also rely on places like the European Centre for Disease Prevention and Control, which very carefully monitor situations in other countries, World Health Organization information, and then also sources from within countries—so, the Robert Koch Institute in Germany is very good at sharing information.

So, we look at lots of different sources and we ask questions as to why things are happening in those countries. We also can go to our regional offices and ask questions about what might be being observed in those countries—so, you know, what's happening in France, why is that happening, could that happen here? These are the questions we ask all the time—you know, what's happening in Sweden? Why is it different from its neighbours? What's happening in eastern Europe? So, we are regularly reviewing, on a daily basis, what's happening in other countries and asking those questions about what it means for us and trying to learn lessons as well, albeit that not all countries are created the same—in the same way that different local authorities in Wales are very different, other countries are very different from us for different reasons.

Are there any particular countries that you look at in terms of evidence?

So, for example, Israel had a short national intervention. So, recently, we were looking at Israel, at how that happened over time, and the indicators, how they were shown over time. We've looked at places like South America, you know, earlier in the summer when they were colder, what was happening in those areas, any that were similar to us. So, we are carefully looking at those countries. And, of course, earlier in the summer, when our prevalence was lower and other countries were higher, we were very closely monitoring, as we are now, the potential for importing infections and seeding different areas for Wales. But now we have a fairly well-seeded population, with infection right across Wales but at different levels.

Okay, thank you. In June, TAC published a report on the impact of lockdown measures across Europe on the R number. Has the work been updated since June?

The work is constantly going on, but we haven't published another document, I don't think, since June. But that's certainly something that we could do, if Members would find that helpful. So, that analysis is—. We have regular TAC updates from that sub-group, and it's something that we share internally with colleagues and Ministers.

Okay, thank you. And in that report in June, it looked at different countries across the UK, lockdown measures, and across Europe. What lessons have been learnt that can be applied now we're facing going into the winter here?

I think there are lots of lessons that we are learning—learning from ourselves and learning from others. As I mentioned previously, there's work ongoing to look at the interventions within the UK and saying which ones have the greatest effect in terms of reducing community transmission, but have the least harm in terms of the other indirect harms associated with these national interventions.

So, we're doing that, and, in all of these things that we're doing, they need to be very clear and communicated effectively, and they need to be achievable. So, the work of the risk, communication and behavioural insights group is very helpful in informing us on what might work, because what might work in Wales may not work in England, what works in France may not work in Wales, and so on and so forth. We need to be really careful. There are no magic bullets. Everything, every decision, is difficult, and every decision has negative consequences.

So, we just need to be really mindful that sometimes it's comparing apples with oranges, and I suspect these snapshots of what we're seeing—for example, we could have looked at eastern Europe and compared us with western Europe and eastern Europe in wave 1, and that would have looked different, quite different to looking at eastern Europe right now, which seems to be faring worse than western Europe right now. There may be a myriad of reasons why that's happening. I suspect the only true reflection, the accurate one, will be a retrospective analysis of age-adjusted cumulative deaths, and, unfortunately, that will be too late for us. So, again, as the research is very much real time, asking questions, and then using that to inform our thinking—so, it's a different way of working, for sure.

10:40

So, how do you balance that, because, obviously, some things might be effective in certain countries? I have somebody who contacts me and tells me about how China are handling the virus and the outbreak now and how that is more effective, but, obviously, there are differences between Wales and China. How do you balance some of the good ideas with, actually, the behavioural side of things and compliance, that perhaps people in this country might feel a bit different?

I think there are a number of variables that we'd need to consider: the trust in Government, how that society thinks about public health measures as a community versus their individual freedoms—that's very different from western Europe versus the far east—what capability a country has, so, if it can bring x million tests on a day with all the staffing that goes with it. So, we just need to keep learning from each other and asking those questions, 'Would it work here and, if not, why not?' So, it's not straightforward, unfortunately.

In terms of things that the World Health Organization—they've had a big campaign around the three Cs, I think, in terms of avoiding areas full of people. There's quite a big campaign on that. Are there lessons from that that you can learn, or how other countries, perhaps, have been able to encourage the population to behave in a particular way?

Yes, absolutely, and it would be good to have my colleagues from the risk, communication and behaviours group to talk about lessons learnt with other countries. But there is a lot of international learning that goes on; there are a lot of conversations through the European centre for disease control and through bilateral relationships. So, there's a lot of learning going on. So, nothing is off the table, really, in terms of learning.

I just want to say the three Cs is a really good example of how we're all learning from each other. The three Cs has been taken up by the World Health Organization, but it actually started in Japan—I assume you're referring to the avoid closed places and crowded spaces and—[Inaudible.] The thing that the World Health Organization can do in those situations is take a message that's worked really well in one country and help other countries to analyse how to distil the message so that it'll work in different places. China deals in a certain way, but has a very different approach to civil liberties, which means that some things that they would consider we would never consider, and we're bound always not only by the need to reduce harm, but also by the underlying ethical principles that allow us to be a part of civil society here in the UK and in Wales.

Right, the final group of questions, then, from Andrew R.T., and, if we haven't finished, I do intend to break for a two-minute silence at 11 a.m. So, Andrew R.T.

Thank you, Chair. I'll quickly run through these, then, about presentation of data. I think most people, at the start of the year, if we'd said that we'd all be so fixated on data—most people would have looked at us stupid, to be honest with you, but it does now dominate our lives and it informs the decisions that govern us. What confidence have you got in the data that is presented by Public Health Wales, about its accuracy and reliability, because that has evolved as the pandemic has evolved, obviously, so there have been changes in that. So, what confidence have you got about the reliability and accuracy of it?

Thank you, Chair. So, I guess we were extremely grateful to our colleagues at Public Health Wales in developing such a rapid piece of evidence provision. It's very rare to see daily counts of any statistic available, and it's incredible the work that they've done—the Office for National Statistics as well—in bringing together two different sorts of data sets. I guess we consider these data to be reasonably reliable. Obviously, there is a lag in some testing data, for example, so we do have to wait a couple of days to make sure that we do have a full estimate of cases sometimes, but the data that is made available we think is sufficiently reliable to use.

Just to explain the differences between some of the data that are available, Public Health Wales are very clear that their data is rapid surveillance data, and that data brings together both Welsh NHS labs data in addition to UK Government lighthouse labs data. So, bringing that together into one coherent data set is extremely useful for anyone that's looking at the numbers and trying to identify a trend. The Office for National Statistics, however, looks at the data in a slightly different way, and they collect their data using a stratified sample. That ensures that there is essentially a robust mechanism for evaluating the population, and that sample is sufficiently robust to be able to reflect the cases within the community of Wales. So, we use different data for different purposes, just to try and understand the situation. I think that's the main point.

10:45

Is it solely the time lag that allows you to be reasonably confident rather than very confident, or are there other anomalies in the way the ONS, for example, collect their data, as opposed to Public Health Wales, that give you that cautionary approach? Because there's a difference between being reasonably confident and very confident, stating the blindingly obvious. So, is it just the time lag that is the issue about collecting the data?

Yes, I think essentially the time lag is the biggest issue. I guess, from Public Health Wales's perspective, obviously their data uses people who have had a test carried out. We know that the total number of infections is likely to be a greater number than the number of cases, because we know that not everyone with COVID is getting a test. That's specifically the asymptomatic population. From the ONS side, it's a sample survey, so obviously we don't have a census at the moment to understand case data. Essentially, as that sample size increases, we have more and more confidence in that data.

I think it's to build on the point about data being used for different purposes, and to really clarify that surveillance mechanism that allows us a more rapid view of potential actions we need to take or not. Just beyond the Public Health Wales data and its use on a surveillance basis, which means that it's better to get it as quickly as possible in order to make some different judgments, if I look at some of the feeders that we have in on the NHS operational data, which is looked at on a daily basis, is it a precise overview of every single patient in all of our beds across Wales? It's probably not absolutely precise, but it's absolutely good enough to allow us to make the kind of decisions that we would need to make on the capacity side. So, ultimately, all data will follow through to be really confirmed through all of the normal information sources, but I would rather have the daily surveillance and operational data available to make the right judgments on behalf of the system, whether it's in the NHS or actually in our oversight role.

Just on behaviour—and I appreciate I joined the meeting as I heard some conversation around behavioural patterns, so if we have covered it, forgive me, but I only picked up the tail end of it—what is the data showing about the behavioural adherence to the rules and regulations and, importantly, the age bracketing as well? I think we're seeing at the moment some stories coming out of universities—Swansea University, for example. I heard yesterday they've expelled students, but I think it's more suspension rather than being expelled, because of, obviously, falling foul of the rules that were put in place during the firebreak. So, there’s definitely information there that supports that different age groups have greater adherence to the rules, but are you able to enlighten the committee on your thinking and understanding of that?

10:50

We did cover this ever so slightly earlier. We’ve got to be very careful that we’re being objective and that we’re able to measure those trends across the population. It’s really easy to see examples and then extrapolate that and say that’s what’s happening everywhere. There are lots of young adults that have been very, very sensible throughout this period. Younger adults have more contacts, or tend to have more contacts, than older adults, and so as a group we’re different—I say ‘we’; I am an older adult. I’m not trying to put myself in the younger adult group there. [Laughter.] But certainly from the data that we look at, in terms of mobility data, movement data, whether people are staying at home, over the last few weeks, people have made a tremendous effort across Wales and there’s been a significant drop-off of movement and activity.

And this is just one indicator, really. The true test will come if we start seeing test positivity going down and, fingers crossed, hospital admissions and deaths reducing over time. But the Welsh public have made a tremendous effort, and not just in the last two weeks but before as well. But, if we were as we were as we entered into this rather insidious affair earlier this year, R was at three, and now it’s not been near three for quite some time, and that’s because of the collective efforts of people making the right decisions, and we just need to encourage people and support people to continue to make the right decisions and recognise these interventions are a significant imposition, as is COVID. And to get this right and to make sure the NHS is protected and as few people as possible become ill, then we need to keep going and keep working with the public.

One final point, as I’m conscious that we’re getting close to 11 o'clock. Will firebreak options be part of the suite of measures still that you’ll be recommending to Government as we progress through the winter, if the evidence points that way, or would you be, given the more detailed data that we’re able to deal with now, able to focus more locally on interventions and make those local recommendations, rather than a firebreak-type intervention again before the end of the winter?

Brendan may want to come in here, but for us as scientists, we need to keep an active open mind, we need to interrogate the data that we’ve got and say with confidence and provide some confidence to policy makers and decision makers what we think we’ve seen, what we think works, what we think doesn’t work, and then present those options for discussion. So, that’s our job.

At the moment, as well as the reasonable worst case that we’ve talked about, we’ve got a range of forecasts that we look at in terms of medium-term and slightly longer term forecasts of the number of cases of the virus, the number of people ending up in hospital, the number of people in critical care, which, as Dr Goodall has said, looks like it’s a bit lower this time around as a proportion of hospital cases than it was in the first wave, which is good news, if it continues like that. So, we monitor a range of measures and we try to produce forecasts and look at other forecasts that are out there in terms of what the situation will be in the coming months and to understand at what point any action might be needed. But we’re only providing advice; we’re not making the policy, so we can only look at policies that might be enacted at different points in time and then look at the impact in terms of transmission of the virus, which then is propagated in terms of hospital admissions and people, sadly, dying from the virus.

I wanted to add to that point, and come back to the behaviour thing. We know that most people want to do the right thing. When they’re told to self-isolate because they’ve tested positive or they've been contact-traced, we know that most people will try to isolate, but we also know that a surprising number of people will go for a final shop rather than trying to do it, or will break their quarantine in order to go and care for somebody that they have caring responsibilities for. And we need to work out how to support communities to cover that so that somebody can self-isolate until they're not dangerous. But the other thing is, with any recommendation, whether local or national, we have to bear in mind that people don't live comfortably within local authority boundaries, and something that is done in theory to control the virus at a smaller level may actually have disproportionate harms to people whose parents live five miles away but across a local authority border.

The final thing that I wanted to say is that when younger adults, as well as older adults, when anybody, says, 'Well, the likelihood of me dying of COVID is low'—Brendan said it was between one in 100 and one in 200—that increases significantly when you get older. If you're over 60, over 65, your chance of catching COVID and eventually dying raises to about one in six, I suppose. It's really, really, really dangerous. I'll have to check that figure. But the other thing is, even if COVID doesn't hurt you, even if you don't die of it, even if you don't end up with severe lung scarring that affects your health for the rest of your life, if too many people get what is a very preventable disease, and pass it on, and fill up the health services, and stop social care being able to function because too many people have it, then they won't be able to get help when they have an accident, or when they have a heart attack, or other things. So, getting people to understand that it's not just about whether or not it's going to hurt you, but, actually, wider society and how they can help with that is the most important thing right now. 

10:55

We've got a few minutes, haven't we? When do you start feeding into your modelling of the likely impact of the vaccine? What's the kind of timescale that you're hoping you might be able to work to?

Thank you. In terms of reasonable worst-case modelling, the reasonable worst case is that there will not be a vaccine. So, we have not fed a workable vaccine into our reasonable worst case. In terms of short-term forecasting, as soon as a vaccine has passed trials and begins to be available, then we expect that we will be commissioned by Dr Goodall, or by Ministers, to consider it in our policy modelling. 

Will you be doing a more informal look at what the likely impact would be, so that you can then formalise that? We know what Pfizer say their vaccine can do, and you could model what would be likely to happen if what Pfizer say is going to happen actually turns out to be true.

Yes, and more importantly than how effective it will be, the really useful thing for us to model is to help them decide, because there won't be enough vaccine immediately to inoculate everybody. But the chief medical officers and Ministers and various others will need to know what the priority level is, and our modelling can help assess the risk of people being super-spreaders or people being most vulnerable and, therefore, those who should receive the vaccine in order to protect society as a whole.

Thank you very much. I just wanted to add that, obviously, with the Pfizer vaccine, one of the important things to note is the temperature that it needs to be stored at. And I'm sure that we'll be doing some—. Well, we've started some initial modelling to look at stock levels and how they could be transported, and the impact that that would have on how available the vaccine could be, and then how, essentially, it could be provided to patients. So, there is a vaccine group that has been established, chaired by Public Health Wales colleagues, and they've asked us to start looking into some of those sorts of aspects so that they can start to bring that together. 

Okay. We'll break there for now, and we'll break for the two minutes' silence. I'll ask our sound operator to co-ordinate things. Then, we'll come back after the two minutes and just close the public meetings. So, with your indulgence—

—a allaf i alw ar bwy bynnag sydd yn rheoli'r sain i'n rheoli ni nawr am y ddau funud o dawelwch? Diolch yn fawr.

—I call whoever is controlling the sound to direct us for the two minutes' silence. Thank you.

11:00

Cynhaliwyd dau funud o dawelwch. 

A two minutes' silence was held.

Felly, dyna ddiwedd y sesiwn. Diolch yn fawr i chi i gyd. Dyna ddiwedd y ddau funud o dawelwch, unwaith eto, am flwyddyn arall. A dwi hefyd yn cymryd mai dyna ddiwedd y cwestiynau. Oeddet ti'n hapus yn fanna, Rhun? Dyna ti.

Felly, gallaf i gloi'r sesiwn yma drwy ddiolch yn fawr iawn i bob un ohonoch chi am eich cyfraniad y bore yma—mae wedi bod yn sesiwn hynod ddiddorol. Diolch yn fawr iawn i chi. Yn ôl ein harfer, mi fyddwch chi'n derbyn trawsgrifiad o'r trafodaethau, er mwyn i chi allu gwirio eu bod nhw'n ffeithiol gywir. Ond gyda chymaint â hynna o ragymadrodd, diolch yn fawr iawn i chi i gyd. A dyna ddiwedd yr eitem yna, ac rydych chi'n gallu mynd yn awr. Diolch yn fawr.

So, that concludes the session. Thank you very much to all of you. That brings the two minutes' silence to a close. Thank you very much. I also take it that that concludes the questioning. Are you content, Rhun? I see that you are.

So, may I conclude this session by thanking all of you very much for your contributions? It's been an exceptionally interesting session. Thank you. You will, as is customary, receive a transcript of the discussion, for you to check for factual accuracy. But with those few words, thank you very much to all of you. And that brings that session to an end—you can go. Thank you very much.

3. Cynnig o dan Reol Sefydlog 17.42(ix) i benderfynu gwahardd y cyhoedd o weddill y cyfarfod
3. Motion under Standing Order 17.42(ix) to resolve to exclude the public from the remainder of this meeting

Cynnig:

bod y pwyllgor yn penderfynu gwahardd y cyhoedd o weddill y cyfarfod yn unol â Rheol Sefydlog 17.42(ix).

Motion:

that the committee resolves to exclude the public from the remainder of the meeting in accordance with Standing Order 17.42(ix).

Cynigiwyd y cynnig.

Motion moved.

I'm cyd-Aelodau, rydyn ni wedi cyrraedd eitem 3, a chynnig o dan Reol Sefydlog 17.42(ix) i benderfynu gwahardd y cyhoedd o weddill y cyfarfod yma. Ydy pawb yn gytûn? Dwi'n gweld bod pawb yn gytûn. Diolch yn fawr. Ac felly, dyna ddiwedd y cyfarfod cyhoeddus.

To my fellow Members, we've reached item 3, and a motion under Standing Order 17.42(ix) to resolve to exclude the public from the remainder of this meeting. Is everyone content? I see that everyone is indeed content. Thank you very much. And that brings the public meeting to a close.

Derbyniwyd y cynnig.

Daeth rhan gyhoeddus y cyfarfod i ben am 11:03.

Motion agreed.

The public part of the meeting ended at 11:03.