2. Professor Sir Diamond

PROFESSOR SIR IAN DIAMOND (sworn).

Questions From Lead Counsel to the Inquiry for Module 2B

Lady Hallett: Thank you for your continuing help, Professor.

The Witness: It’s a privilege to have the opportunity again.

Mr Poole: Could you please start, Professor, by giving us your full name.

Professor Sir Diamond: Yes, I am Sir Ian Diamond and I’m the National Statistician.

Lead 2B: Now, you are in fact Professor Sir Ian Diamond. I was proposing to call you Professor Diamond, is that –

Professor Sir Diamond: I am very happy for you to call me whatever you wish.

Lead 2B: Now, Professor Diamond, I think you know the drill from Module 2, but if you can keep your voice up so that we can hear you but also so that your evidence can be recorded. If I do ask you something you don’t understand, please ask me to rephrase it.

Now, Professor, you have been good enough to provide a detailed witness statement for this module, Module 2B, and we can see it there on screen. You signed that statement on 8 January of this year. Is that statement true to the best of your knowledge and belief?

Professor Sir Diamond: Yes, it is.

Lead 2B: You also provided two witness statements to Module 2 and you also gave evidence in Module 2 on 10 October, and the transcript of that evidence is obviously available on the Inquiry website. If I can start, though, with just a few questions about yourself. You are the chief executive of the UK Statistics Authority and since August 2019 you have been the National Statistician; is that right?

Professor Sir Diamond: That’s right.

Lead 2B: You are also head of the Government Statistical Service and Analysis Function, and in that capacity you provide, am I right, overall leadership for the Office of National Statistics and the statistics profession across the UK Government?

Professor Sir Diamond: Yes, I would just clarify I’m head of the Government Statistical Service and I’m also head of the Government Analysis Function. They are two separate – the analysis function includes the statisticians but also includes economists, operational researchers, social researchers, actuaries and geographers.

Lead 2B: I’m grateful.

If we can start, then, please, with some questions about generally data gathering in the UK but also Wales. Dealing first then with the UK Statistics Authority, what is the UK Statistics Authority?

Professor Sir Diamond: The UK Statistics Authority consists of the Office for Statistics Regulation, which is the regulator of statistics, but, more importantly for this bit, the Office for National Statistics, which is the operational arm of the UK Statistics Authority. The Office for National Statistics has a responsibility to provide official statistics across largely the economy and population and society and to produce those statistics in order to provide the evidence on which policy can be based.

Lead 2B: Can you please describe to us the way in which the UK Statistics Authority operates in relation to the UK Government but also the devolved administrations.

Professor Sir Diamond: Yes, I can. With regard to the UK Government, the UK Statistics Authority is independent of government, although we do have a line to the Cabinet Office, but we are entirely independent and have a board with an independent chair, Sir Robert Chote.

Statistics is a devolved responsibility to the devolved administrations, and – I, though, still have, you may call it a pastoral responsibility to the three chief statisticians of the devolved administration, and we meet regularly and we talk and we have a concordat between the ONS and each of the three devolved administrations where we agree to work together to provide statistics where appropriate which have comparability right across the United Kingdom.

Lead 2B: So in the case of Wales, the Chief Statistician for Wales during the relevant period was Glyn Jones, who was then succeeded by Stephanie Howarth, who we will be hearing from a bit later on.

During the pandemic, did you have much contact with the Chief Statistician for Wales?

Professor Sir Diamond: Yes, very much. I personally had contact in a number of ways, both formal, so that the Inter Administration Committee – I know that’s a mouthful – meets quarterly, and that includes everyone, to talk. In addition we had informal contact whenever that was necessary, and we met on regular bases. I would also say that my colleagues across the Office for National Statistics had very regular meetings with colleagues in the Welsh Government; indeed, around the large Covid Infection Survey they were meeting daily to discuss results and to discuss potential analyses.

I would also say that we try very hard to produce statistics for the Welsh Government in some areas, particularly mortality. In other areas we are ready to respond to requests.

Lead 2B: Now, Professor, I’m right in saying that you attended a SAGE meeting on 16 April 2020, and the importance of understanding the R number or the reproduction number and the community prevalence for the following two to three weeks was discussed at that meeting, and you understood that this information would be crucial information that would inform the government decision-making; that’s right, isn’t it?

Professor Sir Diamond: Yes.

Lead 2B: Following that SAGE meeting, the ONS was commissioned to deliver the Covid-19 Infection Survey that you’ve just referred to. Can you just tell us, what’s the importance of the Covid-19 Infection Survey?

Professor Sir Diamond: At that time, colleagues may remember that test and trace was, I think the best way to say, stretched, and indeed GP services were very stretched. Therefore we did not have an accurate measure of how much Covid was in the population, what the proportion of the population was who at any moment were positive.

And that’s – as a statistician, when I was asked what one would do, I’m afraid the knee-jerk reaction is to say “Let’s do a survey”. And at that time it was unclear, on 16 April, whether one could do a survey, a household survey, where one could do swabs and get that level, but I felt it was possible, and my colleagues rallied round, government said “Let us do this”, and so we started. We drew a sample, we recruited a large number of people. For the first three months I would have to say it was England-only, because this was a pilot, and we ran a household survey which enabled us to estimate the degree of positivity.

With regard to this module, it was successful and so we then started to do it for Wales, and started to produce data on a weekly basis for Wales. And the logistics of going to a household, taking the swabs, then getting them to the laboratory, getting the results – and it is not simply, I would say, a question of, if you like, ticking positive/not and then dividing by the number, there’s quite a lot of statistics that goes into estimating the prevalence, and we did all that and we made estimates twice a week.

Lead 2B: Just to put some dates on that, I think I’m right in saying that the Covid infection study started April 2020?

Professor Sir Diamond: Yes.

Lead 2B: But as you’ve just alluded to, I think field work didn’t commence in Wales until late June, I think –

Professor Sir Diamond: That’s right.

Lead 2B: – 29 June, and then it started producing infection data for Wales beginning of August.

Why was there that delay in respect of Wales?

Professor Sir Diamond: Well, it was I would say not a delay. As I indicated in my last response, we went into a pilot initially. It wasn’t clear whether people would respond, it wasn’t clear that we could get the logistics right, so it was right to do a pilot. And when it was clear that this was a successful pilot, that’s when other devolved administrations decided they wished to join and Wales was the first of those.

Lead 2B: And how did that survey ensure it was able to appreciate the specific types of data that the Welsh Government would need?

Professor Sir Diamond: Well, as I indicated, Welsh Government statisticians were at our daily meetings, they were also at weekly meetings that happened, and we aimed always to respond to requests. So if there was an analysis that Welsh Government statisticians wanted on that Covid Infection Survey, then either at the daily meetings or at the weekly meetings they could say “These are matters which are important to the Welsh Government we really need to get some information on them”.

Lead 2B: How was information from the Covid infection study conveyed to Welsh officials? Was it through Welsh statisticians attending meetings –

Professor Sir Diamond: No, no, no, formally through – because of the importance of pace here – I mean, at times, for example, as one moves forward, some of the Omicron variant doubling time was about two and a half days, so one couldn’t wait a long time before letting government know the results. And so what we agreed with the regulator was that we would let management information go to government, and that typically went on a Tuesday evening, although we published a couple of days later. Why did we publish a couple of days later? That’s because there was still quality assurance that needed to be done and work needed to be done to really make sure that everything was fine, and to get it ready for publication. But the broad data, a slide pack of a rather large number of slides went to government right across all four administrations, went to the Welsh Government every Tuesday evening, and they then had those data to work with immediately.

Lead 2B: Now, Professor, I want to ask you some questions about ONS statistics on fatalities. Now, I appreciate, as we’ve said earlier, you’ve given evidence in Module 2 and I don’t want to make you repeat everything that you’ve said, but there will be some people following your evidence today who won’t have seen your evidence in Module 2 or be familiar with it.

By way of a very brief overview of ONS data on fatalities during the pandemic, is it right that throughout the relevant period, the period we’re concerned with, there were two main published data sources available on deaths?

Professor Sir Diamond: Sorry, I thought you were going to tell me what there were.

Lead 2B: Professor –

Professor Sir Diamond: I’m quite happy –

Lead 2B: Professor, perhaps if I tell you, I tell you what they are –

Professor Sir Diamond: No, no, no –

Lead 2B: – and then you tell me if I’ve got them right.

Professor Sir Diamond: Yes.

Lead 2B: There’s the daily Department for Health and Social Care Covid-19 deaths data, that’s published 2 pm daily for the UK; is that –

Professor Sir Diamond: That’s right.

Lead 2B: And that data was drawn from NHS England, Public Health Wales, Health Protection Scotland –

Professor Sir Diamond: That’s right.

Lead 2B: – and Public Health Agency in Northern Ireland.

And then the second source, ONS weekly death registrations data for England and Wales, and that was what was released every Tuesday at 9.30 am, and that related to the week –

Professor Sir Diamond: Yes.

Lead 2B: – that ended 11 days prior; have I got that right?

Professor Sir Diamond: That’s right. Exactly so. And the distinction is that the ONS data cover all deaths, and with all places of death, whereas initially the DHSC data were for hospitals, first point.

The second point is that they are – the DHSC data would be on the day the hospital was – recorded, and so what you tended to find, and I think DHSC data are really good at very quick estimates, is that at weekends not so many are recorded so there was always, you know, a bit of a weekend gap and then …

The ONS data for there, our data come on date of registration and we receive them and we are able then to produce them with cause of death, because the death certificate, as you will be aware, has the opportunity to write a cause of death and an underlying cause of death, and that’s where we got much, almost all, of our information on Covid mortality, whether the physician registering the death recorded Covid either as the prime or underlying factor.

Lead 2B: So, Professor, which measure, the DHSC data or ONS data, would be more accurate or more helpful to understand what was happening in Wales during the –

Professor Sir Diamond: Well, I think they’re both – I think they’re both helpful. So during the pandemic you would be getting the DHSC data very quickly, on a daily basis. The ONS data, which I would argue have, if you like, more granularity, give more place of death and clearly more cause of death, that comes on a weekly basis, so it’s a little slower but at the same time gives you an enormous amount of information.

So I do think it is important to recognise that, you know, very often in statistics we’re able to provide quick information which we are clear about what the pluses and minuses are of, while a little later you have, if you like, a much more detailed and better information. If you’re happy to wait those 11 days, as most people were, then that’s what I would use.

Lead 2B: Understood.

Now, the Inquiry understands that on 31 March 2020, the ONS gave an exemption to provide the DHSC with provisional data on deaths registered weekly in England and Wales. I think the idea was that that would help ministers better understand the spread of Covid-19.

Was an equivalent exemption made for provisional data to be shared with the devolved administrations? Obviously specifically –

Professor Sir Diamond: Er –

Lead 2B: – the Welsh Government.

Professor Sir Diamond: Not clear and I would need to check on that.

Lead 2B: Now, before we address fatalities, can we just look at Welsh demographics as they were in 2020, and you outline in your witness statement a number of data point estimating Welsh population demographics in mid-2020. Just before we come to those figures, can you just explain how those estimates were developed.

Professor Sir Diamond: Sure. So in 2020 we were working from the 2011 census, using what we call a cohort component method to update year on year. That basically means we start with the 2011 census, we add on births, we take off deaths and we make an allowance for migration. I’d have to say that by 2020 you’re about as far away from the previous census that you get. We are still very proud of those estimates, but then in 2021 we did an unbelievably good census in Wales, and so there will be a distinction between 2020 and 2021.

Lead 2B: Now turning to the actual figures, and you’ve set these out at paragraph 11 of your witness statement for this module, and I don’t need you to turn it up, but the estimated usually resident population for Wales in mid-2020 was 3.17 million.

Professor Sir Diamond: Yes.

Lead 2B: What is the importance of the “usually resident” measure for somewhere like Wales?

Professor Sir Diamond: Well, put pretty simply, that doesn’t include houses, for example, that are second homes. So, you know, we don’t have people there. There may also be people who report that, you know, they are working or living somewhere else but maintain a house in Wales, which may be their first house. So the “usually resident” population is one that we use a lot.

Sometimes local authorities make a point of saying “Well, actually we want to know how many services to deliver”. So let us take a place like Cardiff, the number of people usually resident, shall we say, on a Sunday evening is rather different to the number of people that Cardiff has to provide services for on a Wednesday lunchtime. So the usually resident population is a clear distinction of those people who are there, not necessarily those people who will be there at different times during the day and week. And of course does not include people who are short-term visitors, eg tourists.

Lead 2B: Now, you say in your report that the median age of the population of Wales, 42.4 years, are you able to help us with how that compares to the UK as a whole?

Professor Sir Diamond: Yes, it’s a little older, and indeed Wales has a slightly higher proportion of people over 65, and I think that’s worth saying.

Lead 2B: In terms of demographic spread, how does that compare to the rest of the UK?

Professor Sir Diamond: Well, Wales – Wales is a very heterogeneous place, so what you have around South Wales and particularly around the old mining areas north of Cardiff is a very highly dense population and one with relatively high degrees of deprivation. On the other hand, as you go north from there or west, it becomes a very rural population. As such, you have a real mix. So that South Wales area looks not unlike, for example, the area going from Manchester across to Leeds and around, whereas the northern and western group much more like the Lake District.

Lead 2B: Professor, you have produced a number of figures also in your witness statement about Welsh economic activity. Where do those figures come from? Is that the census or the two censuses –

Professor Sir Diamond: Well, partly from the census, where we’re able to ask people: what do you do? (inaudible) proportion, but we also run a labour force survey, which is a very large survey which tells us about activity and inactivity, and those data we use as well.

Lead 2B: It’s right to say, isn’t it, that a greater percentage of usual residents aged 16 and over in Wales were economically inactive compared to those in England?

Professor Sir Diamond: That’s absolutely right, and again this is one of the reasons that I spent a little time a moment ago talking about the geography of Wales, a lot of that inactivity is in that area, that old industrial area to the north of Cardiff.

Lead 2B: Changing topic slightly, and talking about four nations co-operation now, if I may, at the very beginning of the pandemic, so January to early March, you say there was less immediate contact between the UK Government and the devolved administrations. Can you just explain a bit what you mean by less immediate contact.

Professor Sir Diamond: Well, I think initially, as I indicated earlier, we do have good relations, but statistics is a devolved responsibility, we meet as chief statisticians quarterly, and initially the Welsh Government was – Statistician was working on Welsh Government issues, we were much more focused into Whitehall, and it was only as things started to evolve that we said “Come on, we need to get together here and really work together”.

Lead 2B: Were there any specific challenges or delays in establishing collaboration with Wales and the Chief Statistician for Wales?

Professor Sir Diamond: No, no.

Lead 2B: I’d like to now move on to talk about mortality data, if I can.

If we can have, please, INQ000396876, and this is table 6. This – as it says at the top, “Death registrations involving Covid-19, March 2020-February 2022, UK, England, Wales, Scotland and Northern Ireland and region of England”.

In terms of what we can see in that table, Wales has the second highest age-standardised mortality rate of deaths involving Covid-19 144.6.

Lady Hallett: Sorry, just before you go on, Mr Poole, can we just say, “involving Covid”, does that mean Covid appears on the death certificate?

Professor Sir Diamond: Yes.

Lady Hallett: Thank you.

Mr Poole: Thank you, my Lady.

So that was greater than the UK average, which was 143.2. England slightly higher at 145. Was that difference between Wales and England would you say that’s statistically significant?

Professor Sir Diamond: No.

Lead 2B: On the other hand, was Wales’ age-standardised mortality rate significantly higher than the rates for Scotland and Northern Ireland?

Professor Sir Diamond: Yes.

Lead 2B: Now, it might help to break this information down by wave.

Could we, please, have table 13 from your Module 2 witness statement, which is INQ000271436, please. So the table straddles those pages.

What’s the source of this data, Professor?

Professor Sir Diamond: So what we are looking at here is what we call excess deaths, and what one is doing there is taking the difference between the deaths that occurred and a measure of expected deaths. Now, that’s the measure that you might expect to have had. And for this table what we were doing was taking the mean of the deaths in the previous – in that week, in the previous five years. So you take the difference between the two, and of course that could be either positive or negative. If it’s positive then you’ve got more deaths than you would have expected, and if it’s negative then fewer. And in this case, we are reporting for those particular periods much higher mortality than would have been expected.

Lead 2B: If we can just see a few things from this table, England have the highest percentage excess death registrations when looking at the whole time period, and also two of the three lockdown periods. So the periods March to June 2020 and January to May 2021. That’s right, isn’t it?

Professor Sir Diamond: Yes.

Lead 2B: This was true when looking at excess deaths using either numbers of death registrations or age-standardised mortality rates?

Professor Sir Diamond: That’s right.

Lead 2B: During the second lockdown, so August to December 2020, we can see Northern Ireland had the highest percentage above average when looking at numbers of death registrations but in fact Wales had the highest percentage when looking at age-standardised mortality rates; is that right?

Professor Sir Diamond: That’s right.

Lead 2B: Meaning that when one controls for the age composition of Northern Ireland and Wales, Wales fared the worst during the second wave than all of the other –

Professor Sir Diamond: During the second – yeah, no, you’re absolutely right. During the second wave mortality in Wales was the highest of the four administrations across the UK.

Lead 2B: If we can please look at another chart, it’s INQ000412042.

This shows weekly – excess weekly deaths in Wales. Can you perhaps describe what we can see in this chart, Professor?

Professor Sir Diamond: So what this chart does is not only give you the picture of excess deaths, but, by using different shades, and in this case the blue, the blue are those deaths which involve Covid. And so what you can see, I would suggest, quite clearly, are three things: firstly, in that first wave, there was a very strong peak of deaths in April 2020, and that that peak, the excess was largely driven by Covid.

The second thing I would say, and you’ve just referred to it, is a very high peak in the autumn and early part – autumn of 2020 and early part of 2021, and again that was largely driven by Covid.

The third point I would make is that following those two very, very big peaks, the numbers of deaths later on in 2021 and into 2022 do not have those peaks of excesses. While sadly there remain a number of Covid deaths marked in blue, a – due to a number of things, improved treatment, improved – the brilliance of vaccination and other things, the actual mortality due to Covid went down greatly while at the same time as some of the new variants, I’m thinking particularly of the Omicron variant came in, the actual proportion of the population who had Covid got very much bigger than it had ever been before.

So, if you go back to March/April 2020, the proportion of the population overall we would suggest with Covid was relatively small but if you got it, particularly if you were old or frail, then very, very sadly, the prospect of mortality was high. By the end of the period the probability of actually having Covid had increased, but due to all those factors, vaccination, better treatment, and indeed perhaps the evolution of the virus, the probability of mortality had reduced greatly.

Lead 2B: That’s very helpful, thank you, Professor.

Can we please have another chart on screen, it’s INQ000412042. I think it’s at page 2 of those slides.

Now, this is showing “Daily deaths with Covid-19 on the death certificate”, comparing Wales and then comparing it to the UK.

What does this comparison show us, Professor?

Professor Sir Diamond: Well, much, I would have to say, as what I’ve said previously. I would say also very clearly that we have put different Y axis scales there, so there’s not a complete comparison, so don’t think that the numbers in Wales were rather bigger than the numbers in England, but it’s making the point, I think, very, very clearly, firstly, that the trends largely mirrored across the UK and in Wales, but secondly you really can see that big peak towards the end of 2020, beginning of 2021, which is the point we’ve just been making about mortality in Wales sadly being rather bigger than anywhere else during that second wave.

Lead 2B: If we can move then to talk about deaths in Wales by age, and do so by reference to another chart.

It’s – I think it’s the same INQ but page 7 – yes, I’m grateful.

This shows “Age-specific death rates involving Covid-19 by [a] five-year age group, [in the period] March 2020 to June 2022 …”. Can you please just talk us through what we see here. What does this tell us about the age distribution of Covid-19 deaths in Wales?

Professor Sir Diamond: It tells us very, very simply that mortality in Wales was very largely restricted to the elderly.

Now, we know from other places that often there were comorbidities that may have played a role in mortality for younger people. We don’t see that very much. That’s not to say, clearly – there are small numbers of people aged 40-44 and 45-49 who, very sadly, would have died, but basically what this is showing is that mortality in Wales was restricted to the elderly.

Lead 2B: Now, I think you conducted your own work into excess deaths per age group and you’ve set this out helpfully in your witness statement and – to show the effect of Covid on different age groups compared to deaths in non-Covid years.

Am I right if I was to summarise the work you carried out as concluding that the highest excess mortality was observed in those aged 45 to 49 years old, with age-specific rates between this period March 2020 to June 2022 being recorded as 16.8% above the five-year average?

Professor Sir Diamond: I think that’s right but I would have to say, and it comes back to the point I’ve just made, and I’m not trying to make a statistical point but a – if you have a very small number, then it doesn’t take a very big number for that to be quite a big proportion. So yes – and, you know, you’re right, the good news is that people aged 45 to 49 do not have very high mortality, and so that small amount of mortality is a relatively high percentage. But it does not, as that graph that I’ve just talked to shows, become a high mortality compared to those older ages.

Lead 2B: I understand.

If we can move then away from age and focus on place – we can have the chart, thank you, it’s page 8 of the same document – and just talk us through again what we see there, in particular, if you could, the negative figures for hospital and hospice.

Professor Sir Diamond: Right, well, when we are looking here is at the excess deaths by where the death occurred, and I will come, if I may, to “Hospitals” and “Hospices”, but if I could just make a point about the positive ones first.

We saw a significant increase in deaths at “Home”. Now, some of that could be Covid, others of it could be, for example, cardiovascular disease or whatever, you know, where people had not gone into hospital. We also see a significant increase in care homes than we might have expected.

And I just want to say a few words again about “Other communal establishments” and “Elsewhere”, because the percentages are high, but, as the point I made earlier, they are high percentages of small numbers. And the “Other communal establishment” is a very big group which includes all kinds of things, including, for example, student halls of residence where there was no mortality or almost no mortality, but it does include sheltered housing, and that’s where we think much of that increase comes, you know, where, one, again – because people are looking after themselves, although they are very elderly, mortality there is relatively low. And the “Elsewhere”, which includes all kinds of things, does include those people who were pronounced dead, sadly, on arrival at hospital. And again, we would suggest that that could include a wide range of areas, including cardiovascular disease or whatever, but, you know, sadly, that’s there.

So let me then return to “Hospital”. The first thing to say is that, yes, the numbers are below zero but they’re tiny below zero and I might suggest, and can I just stress in what I’m about to say that I am speculating, I do not have firm evidence, but we do know that a lot of illnesses people didn’t go to hospital, for all kinds of reasons, because the hospitals were absolutely stretched doing wonderful things dealing with Covid patients, and so that could have led to just a small reduction despite the fact that we had looked at those peaks.

And very similarly with hospices, I would suggest that people were choosing perhaps to stay at home if they had, for example, sadly, terminal cancer, rather than going into hospices.

Lead 2B: Professor, in your statement – and we can take that chart down, thank you – in your statement you describe how there was only limited understanding of the way in which socioeconomic characteristics contributed to deaths in Wales unlike in England. Can you just start by explaining why that is.

Professor Sir Diamond: Very simply, in order to make good estimates, there needs to be enough data to make estimates from, and so we did make estimates as best as we could during the – during 2020, often linking mortality data back to the 2011 census. Now, that presents problems, because, if you think about it, many people who were in the 2011 census would have moved home, would have migrated internationally, so we had to use a number of quite complex statistical techniques to be able to make those estimates, and we were able, for example, to make some estimates at that period around ethnicity. We were able then to make longer – we then moved to using data that we were able to get in England from Public Health – from GP records, and we used those for England in the future.

What that showed was a high degree of association between mortality and ethnicity and mortality and deprivation, and we see that in Wales as well, where we’re able to look. And we did not do work following 2020 for Wales but we did offer to work with colleagues in Wales and offer the code to them, and the 2011 census data were made available in Wales to the SAIL Databank, and to Digital Health and Care Wales.

Lead 2B: I understand, because I think that links to a question I was going to ask. Stephanie Howarth suggests in her statement to the Inquiry that the SAIL Databank that you’ve just referred to has access to GP data obviously in Wales. Would that have enabled some parity between statistical publications for –

Professor Sir Diamond: Yeah.

Lead 2B: – Wales and England? Is the lack of this linked information in Wales, in your view, a significant gap in identifying and understanding socioeconomic or public health trends –

Professor Sir Diamond: I do think it’s as a mortality. The better you are able to link data, the more granular the information that you can get, and the better that information is.

Lead 2B: What have you been able to measure to gain an understanding of sociodemographic trends in Wales?

Professor Sir Diamond: Well, we were able to look at geography, and I think it is important to note that the areas with the highest levels of mortality, Rhondda Cynon Taf, Merthyr Tydfil, are those areas where one would find associations with deprivation and to something you mentioned earlier, inactivity, which could be due to ill health. So we’re able to say that.

We did find some early data around ethnicity, and we were also able to look at what is called the Index of Multiple Deprivation. Now, this is an index which is put together from a large number of variables, typically those in censuses, and which is able to go to a relatively small geography, and one is able to then – what one typically does is take the five quintiles, and, say – let’s go from the areas of most deprivation to the areas of least deprivation, and what we’re able to say is that there is a clear gradient, a clear gradient, between the most deprived and the least deprived areas in terms of mortality.

Lead 2B: Do you have any other breakdown of deaths by, say, religion, disability status or occupation –

Professor Sir Diamond: No.

Lead 2B: – group?

Professor Sir Diamond: I would say – just a point I would make is that my colleagues would say that they can see no real reason for some of the things in those areas that we found in England not translating across to Wales, but it would be for the Welsh Government to have done that work.

Lead 2B: I think, just to illustrate a point you’ve made, if we can have, please, INQ000396876, it’s figure 5, from your witness statement to this module. This shows proportion of excess deaths by local authorities in Wales. Does that largely accord with your findings about deprived quintiles –

Professor Sir Diamond: Yes.

Lead 2B: – that you’ve just referred to?

Professor Sir Diamond: I mean, as I say, these are local authorities, so if you look at somewhere – I mean, given it’s just down the road, Newport, there are some pretty poor areas in Newport, but there are also some less poor areas. So overall if you look at Newport it looks like it’s in the middle. On the other hand a place like Merthyr Tydfil is much more uniformly deprived.

And so I think it is important when you look at local authorities to understand that the heterogeneity in terms of deprivation across those local authorities so that – you know.

And it is also the case, I mean, there was no question at all, that places like Powys have very – relatively low areas, but let us not pretend that there is not deprivation in rural areas, it’s just it can be hidden compared with urban areas.

Lead 2B: Understood.

I want to move away from mortality data and talk a bit about infections data. And we touched upon this when we were talking about the Covid-19 Infection Survey, and remembering then that that only began to publish data for Wales in August 2020.

Can we, please, have INQ000412042 on display. Thank you.

Professor, can you just talk us through what this chart shows.

Professor Sir Diamond: What it shows is the percentage of the overall population of the four administrations who tested positive in any period. Note the four arrows to the left-hand side which indicate, as you’ve rightly pointed out, when each administration started to collect data. So what you clearly see is an increase in positivity in the autumn of 2020 – and I don’t know, sir, if we’re going to discuss the firebreak at any time?

Lead 2B: By all means. I mean, this accords with the firebreak of –

Professor Sir Diamond: Yes, sorry –

Lead 2B: – September 2020 –

Professor Sir Diamond: For what it’s worth I will just add it. I didn’t know if you wanted to address it later.

You can see for Wales that the percentage goes up and then goes down again and then goes up again, and that accords with the firebreak that the Welsh Government brought in.

Things then flatten off in the early summer of 2021. There is then an increase. And then in late 2021, when the Omicron epidemic came strongly in towards – that you actually see the highest percentages that we have seen.

And I would point out that, again, these are national data, and if you were to look at the age-specific numbers, which have higher confidence intervals around them, you do at times get above 10% of the population in some age groups at that time testing positive.

Of course, as I’ve indicated earlier, that has less of an impact on mortality thanks to the brilliance of vaccination and also better treatment, but once you’ve got very, very high proportions of people who are testing positive and therefore out of work, you are actually starting – or not able to work for that period, you are actually starting to have enormous impacts on the economy, and that is something I think that we need to remember in that latter period.

Lead 2B: Thank you, Professor.

And we can take that chart down, please.

The Covid-19 Infection Survey was, I think I’m right in saying, able to generate data about likelihood to become infected based on some sociodemographic characteristics; is that right?

Professor Sir Diamond: That’s right.

Lead 2B: But that data was UK-wide data rather than Wales-only data; is that correct?

Professor Sir Diamond: That’s right.

Lead 2B: Can we please have another chart displayed. It’s INQ000396876. It’s – thank you, yes. It’s that figure 10:

“Likelihood of testing positive for Covid-19 by core demographic characteristic, UK, 29 August to 11 September 2021.”

Again, Professor, perhaps you can just talk us through what this chart shows us.

Professor Sir Diamond: Well, what it shows, and it comes back to many of the points that I’ve been making, and that is that if you are in a larger household, the probability of testing positive increases. If you are in a more deprived area, the probability of – well, this graph shows the lower the deprivation, the lower the rate. The opposite of that, clearly, for higher deprivation. Also shows that the major – the big urban areas and, indeed, the slightly less big urban areas had higher probabilities. And it also shows that there is a variation but very little difference in terms of the probability of being positive between non-white populations as a whole, people of colour, and the white population.

Lead 2B: I was going to move on to ethnicity data. Are estimates able to be produced for mortality in Wales across ethnic groups?

Professor Sir Diamond: We have produced them for 2020. I’d have to say the numbers of people of colour, broadly defined, are in single figures, and so it would be very hard to make any strong assumptions and we have not done it post 2020. You’d need to talk to Stephanie Howarth about that.

We have shown for – in that period for England and Wales, that there were strong differences by ethnicity of mortality. And we would argue for a number of reasons that – we cannot hide from the fact that in our country people of colour are more likely to live in deprived areas, are more likely to be in multigenerational households, all the kind of things – and are also less likely to be in occupations which enable them to work from home, and all those reasons would have contributed to those factors.

Lead 2B: I was talking about mortality just then, but what about infectiousness in Wales, were estimates able to be produced for infectiousness across ethnic groups?

Professor Sir Diamond: No, well, we found that very difficult and we tried in many ways. The reason for that is while we’ve got quite a number of people in our sample, the proportions positive are often very small, and so actually trying to fit strong models becomes very, very difficult because you’ve got very, very small numbers of people who were positive.

So, you know, you’ve got – you know, once you start getting sort of very tiny numbers of people in different socioeconomic groups, for example, you can’t actually model, so we just produce individual data and we suggest that many of the things we find for the UK as a whole would hold in the four nations.

Lead 2B: So is – does that explain why that – it could be done, for example, for England, but it –

Professor Sir Diamond: Yes.

Lead 2B: – couldn’t be done for Wales?

Professor Sir Diamond: The sample size – I mean, these kinds of statistical models are driven by how much data you’ve got, and the sample size in England was extremely big, the sample size in Wales – and remember that it’s not just the sample size but the proportion testing positive means that you’re actually working with relatively small numbers, and that makes estimating some of the models almost impossible.

Lead 2B: Now, data gaps, Professor, were identified for those with protected characteristics in Wales by the Equality and Human Rights Commission’s 2018 paper “Is Wales Fairer?” Are you aware of that report and its finding on this question of data inequalities for –

Professor Sir Diamond: I’m aware it talks about disability.

Lead 2B: A point to note, I mean, moving to – that was a 2018 paper. Moving forward two years, during the pandemic there were no datasets – or no datasets that permitted any meaningful comparison were available for the impact of the pandemic on ethnicity, occupation, religion, disability status; that’s right isn’t it?

Professor Sir Diamond: I mean, disability, I would have to say, is a major data gap for our country. And we do have a question on the census which asks about limiting long-term illness, and that gives you some information but it doesn’t actually help, in terms of the granularity, as to whether you might have – be hard of hearing or whether you have musculoskeletal problems.

So we do, I believe, have a data gap around disability, and indeed we at the ONS have been trying to engender some conversations about that.

Lead 2B: What actions did the ONS take to support the Welsh Government assess the unequal impact of the pandemic?

Professor Sir Diamond: We offered support at any time. We offered at different times to share some of the code that we had developed. And, as I indicated, we enabled the census data to be in Wales and we have the Covid Infection Survey data which were held in our Secure Research Service, and Welsh statisticians were able to access those data very easily and very quickly.

Lead 2B: Now, a point that we touched on yesterday with Professor Ogbonna, would it assist if ethnicity data was recorded by coroners and registrars on death certificates?

Professor Sir Diamond: I’m not completely convinced about that. And the reason I am not convinced about it – let me start by saying I’m 100% convinced that we need to get mortality by ethnicity, but the point I would make is that when one gets to the death certificate, the person who most knows about their ethnicity is sadly no longer with us. Which is fine, you know, if it is a very close relative who is reporting, but it doesn’t necessarily need to be that.

So I’m personally not convinced about putting ever more data onto death certificates. What I would prefer to see is that we had a system, for example through the health service, which routinely collected good ethnic data and where we committed on a very regular basis to link those data together, and to be able to publish differentials in mortality by ethnicity.

So I’m – let me be clear, I’m 100% convinced about the need to produce more – you know, I’m not just – I would suggest ethnicity, I would also add disability or add other areas. But I would argue the best way to do it is through having those data available through, for example, the health service and then linking them in. Which we can do very easily and very quickly.

Lead 2B: Did you or your colleagues at the ONS collect any data from Wales on Long Covid, Professor?

Professor Sir Diamond: Yes, we did. And it is self-reported, and again one of the advantages, I would argue, of the Covid Infection Survey that we’ve talked about thus far during this morning’s conversation is that it was longitudinal in nature. What does that mean? That means that we go back to the same households over time. That enables us to follow up and to ask people: have you still got the symptoms? And what are the symptoms? And that enabled us to make overall estimates for Wales. And to answer perhaps your following question “Could you get down to socio-economic groups?”, no. But, we were able to identify levels of Long Covid self-reported, and that peaked at 4%. 4%. One in 25 people reporting that they had experienced Long Covid, and that’s self-report, ie “I’d had symptoms for at least 12 weeks”, and in some cases for a year.

Lead 2B: The Inquiry understands that information was requested from Number 10 about Long Covid in February 2021 and from that point the ONS provided updates at DHSC ministerial round tables on Long Covid. Has the Welsh Government made a comparable request?

Professor Sir Diamond: I’m not clear whether there was a request – there was certainly – I’m not aware of a request formally to me. But as I indicated right at the beginning, statisticians in the ONS were meeting on a very regular basis with statisticians in Wales, and those conversations regularly included discussions about analyses, which would have been asked for.

And indeed I think the other thing to say about the Long Covid is that we were able to make some small disaggregations and to say that Long Covid was more likely to occur amongst women and also amongst those in the more deprived areas.

Lead 2B: Professor, just returning to a topic we touched on earlier, about collaboration with ONS and Welsh Government, and we can see from your statement that UKSA and the ONS worked closely with scientific and expert groups within the UK Government, and at paragraph 153 of your witness statement you state that your engagement with scientific and expert groups in the devolved administrations was much more limited. Briefly, why do you think your engagement was more limited with those groups in the devolved administrations?

Professor Sir Diamond: Well, I mean, very simply, because we, if you like, were not able to just say “We’re coming”. And so, yes, we did go and present to the Technical Advisory Cell of the Welsh Government, and we did have conversation, but we are waiting, in many cases, to be asked, whereas with the UK Government you know, I was part of SAGE, I was – as indeed were colleagues from the Welsh Government. I would have to say that also particularly the chief data officer of Public Health Wales, a woman called Fliss Bennee, attended them of the same meetings that I attended, with, for example, the Joint Biosecurity Centre and all kinds of things. So the Welsh were at many of the meetings. But we only attended, for example, the Technical Advisory Cell when invited.

Lead 2B: Professor, just finally then, from me, we heard evidence yesterday from Helena Herklots CBE that deaths of people in care homes were not counted early in the pandemic. This might well be a question for Stephanie Howarth rather than you, but are you aware of why those deaths weren’t counted in Wales?

Professor Sir Diamond: No. We were able to do, as – I mean, we’ve looked at the data that I’ve shown you already, and we do get place of occurrence. What I can’t tell you is why they weren’t counted initially.

Mr Poole: Professor, thank you, they’re all the questions I have for you.

The Witness: Thank you.

Lady Hallett: Thank you, Mr Poole.

I think we have some pre-approved Rule 10s. I think it’s Ms Gowman.

Questions From Ms Gowman

Ms Gowman: Thank you, my Lady.

Professor, I ask questions on behalf of Covid-19 Bereaved Families for Justice Cymru.

My first question relates to the data available to SAGE to inform its advice, and indeed you’ve confirmed in your evidence that you were part of SAGE. Do you agree with the evidence of Professor Ailsa Henderson in Module 2 that, firstly, SAGE focused overwhelmingly on data from England, and, secondly, that sometimes data from England was described as UK data for the purpose of SAGE advice?

Professor Sir Diamond: I would have to say that I would agree that a lot of the data which were looked at were England-centric. I would also, though, say that Welsh, Scottish and Northern Irish colleagues were at every SAGE meeting, and on very many occasions I can recall Sir Patrick Vallance making an effort specifically to bring in those colleagues to ask … and as I said before, the Welsh Government did have a Technical Advisory Cell and we presented to that cell when invited.

Ms Gowman: But focusing specifically on SAGE, and you’ve accepted candidly that SAGE did seemingly focus on data from England, but you’ve rightly pointed out that there was attempts to draw in data from Wales, what was the breakdown between English-centric focus and the attempts made to bring in Welsh data? Where did it go –

Professor Sir Diamond: Well, I think – I mean, I think – I mean, just for absolute clarity, I didn’t say that it was totally English-centric –

Ms Gowman: No –

Professor Sir Diamond: – I did say –

Ms Gowman: Yes.

Professor Sir Diamond: And I would also say that much of the modelling that went on used the Covid Infection Survey that I’ve had the privilege to talk about this morning. And that’s UK-wide. So much of the data on which, for example, the mathematical models were used, was based on UK-wide data and therefore included Wales.

So I wouldn’t – I would not like, with respect, to put a percentage on things. I am conscious that Welsh colleagues, Scottish colleagues, Northern Irish colleagues, had every opportunity to input and that much of the modelling that went on would have used UK-wide data.

Ms Gowman: Thank you.

Moving on to my next question, and for context, the Welsh Government liaised with the UK Government to provide daily aggregate data to feed into the UK Government’s Covid-19 dashboard, and with that in mind, do you agree with concerns raised by some in Welsh Government, for example former Chief Statistician Glyn Jones, that “definitions of data items were not always clear at the outset” and that this “posed a risk of misinterpretation by assuming data across four nations comparable” when that was not always the case?

Professor Sir Diamond: I don’t think there was – actually I disagree with – Glyn Jones is someone I respect enormously. I think – look, initially there were different definitions. I have no doubt about that. And indeed I got a group together in, I recall, June, I think it was June, but in the summer of 2020, said, “Look, we need across our four administrations, we need a very clear definition”, and there were different definitions and I brought that together and I made it happen. Previous to that, you know, it may be that people misinterpreted, but, you know, typically the footnotes would have been – made it very clear what those data were.

I do think, and I’ve said this many, many times, very early on in the pandemic there were a lot of data moving around and I don’t think always that the visualisation of those data was absolutely brilliant, and we worked – we being ONS and many others – worked very hard to move from what I would call a data deluge into insight by moving to really ask questions.

I mean, this morning has been an absolutely fine example of that, where you can – you get some really good questions and you can say what the data says about them. And that I think was something that happened very quickly, but early – you know, in February/March, early April 2020, there were a lot of data around, which was one of the reasons we started the survey.

Ms Gowman: And something that we can learn moving forward –

Professor Sir Diamond: I think it’s a real lessons learned. I mean, I do try to address in my witness statement some lessons learned. I do think that whoever – I mean, I very much hope there isn’t another pandemic like this, but I do believe that the National Statistician should be right at the heart very, very early. I do believe that we need to make sure we are sharing data much more easily and much more quickly. And indeed I do believe we should be sharing those data now so that we are ready. And I do believe we need very much to be learning.

Perhaps, I mean, you and your colleague who talked before would not need to learn this, but I think sometimes it is the question that we need to make sure we are asking questions. If you just came to tell me something interesting about whatever, I might tell you something interesting but it might not be the answer you needed. So really focusing on questions and learning to focus on questions is something that we, as a nation, need to improve in our data literacy.

Ms Gowman: Thank you, Professor.

Very, very briefly, one final question on the Covid-19 latest insights tool that you’ve already mentioned. Did that tool incorporate data from Wales and analysis specific to Wales?

Professor Sir Diamond: Yes, it did, where – in those datasets which were UK-wide.

Ms Gowman: I’m grateful.

Professor Sir Diamond: So Covid Infection Survey, Opinions and Lifestyle Survey, designed UK-wide, able to produce data which would include Wales.

Ms Gowman: And more specific because it was answering questions, as you put it?

Professor Sir Diamond: 100%.

Ms Gowman: Thank you, Professor –

The Witness: Thank you very much.

Ms Gowman: Thank you, my Lady.

Lady Hallett: Thank you very much, Ms Gowman. I think there were suggested possible extra questions, but I have been told what they were, and one of them I think, about infection surveys in schools, could be asked of the next witness, because it’s a Welsh-specific question, and I think the other question that I have been told about I’m afraid is not for this witness who is not an expert in public health, so there are no further questions.

Thank you very much indeed, Professor, I’m very grateful. I’m not going to give you a guarantee I’m not going to ask you again to help, but –

The Witness: Were you to ask me again, let me be very clear, it is a privilege to be able to –

Lady Hallett: Very kind of you to say so. Thank you very much indeed.

The Witness: Thank you very much.

Thank you very much.

(The witness withdrew)

Lady Hallett: What time is it? 1.55.

(12.54 pm)

(The short adjournment)

(1.55 pm)

Lady Hallett: Right.

Ms Spector: My Lady, please can I call Stephanie Howarth.

Lady Hallett: Thank you.