3. Mr Gavin Freeguard

MR GAVIN FREEGUARD (affirmed).

Questions From Counsel to the Inquiry

Mr O’Connor: Could you give us your full name, please.

Mr Gavin Freeguard: Gavin James Freeguard.

Counsel Inquiry: Mr Freeguard, you have kindly prepared a detailed report for the Inquiry on the subject of data processing and the use of data in government during the Covid pandemic. We have the first page of the report on screen. You can see it there. And we can see that just above your name and the date, you’ve included a statement explaining your understanding of your duties as an expert, and you’ve indicated there that the detail within the report is true to the best of your knowledge and belief. Is that right?

Mr Gavin Freeguard: That’s correct.

Counsel Inquiry: Mr Freeguard, you are currently a freelance data consultant, I think?

Mr Gavin Freeguard: Yes, that’s right, I tend to work on data policy, research and advocacy, and also focus on digital government as well.

Counsel Inquiry: Thank you.

You’ve set out in your report, we don’t need to look at it, a range of roles that you undertake at the moment, including that you are an associate at the Institute for Government, and also a special adviser at the Open Data Institute?

Mr Gavin Freeguard: Yes.

Counsel Inquiry: You’ve also indicated that prior to undertaking freelance work, you were in fact an employed researcher at the IFG holding the post, latterly, of programme director for data transparency and digital government?

Mr Gavin Freeguard: Yes, that’s right.

Counsel Inquiry: Mr Freeguard, as I’ve said, the focus of this report, which I should perhaps have said is a lengthy one, it’s 70 or so pages long, is to do with data matters and how they affected the taking of core political and administrative decisions during the Covid pandemic, and you are aware, because these were all part of your instructions, that the list of issues for Module 2 of this Inquiry relate to matters such as identifying the key sources of data available to decision-makers; yes?

Mr Gavin Freeguard: Yes.

Counsel Inquiry: Questioning whether decision-makers had access or adequate access to reliable data, statistics and so on; asking to what extent the data collection during the pandemic addressed the impact of Covid-19 on vulnerable and at-risk groups; and finally, addressing the question of the adequacy of sharing of data within government structures during the same period, and all those matters are addressed in your report, are they not?

Mr Gavin Freeguard: They are, yes.

Counsel Inquiry: It’s right at the outset to make it clear, as you do in your report, that your expertise in this field is very much focused on matters relating to Westminster government and institutions. So, as we know, we were addressing devolution matters yesterday, that covers matters relating to England and the UK, but you’ve explained in your report that your expertise doesn’t really extend to similar matters within the devolved nations, Wales, Scotland and Northern Ireland?

Mr Gavin Freeguard: That’s right, I occasionally touch upon that in my report and think it’s an extremely important topic, but my work has been focused on Westminster and Whitehall.

Counsel Inquiry: So you have fairly made the point that you can’t help us in an expert way with those matters.

Just finally by way of introduction, it’s important to make it clear, and for the sake of those who go to your report, which we will of course adduce in evidence in full, that you were at the outset instructed to prepare this report on the basis of your own understanding of these matters and materials that were publicly accessible.

You weren’t provided with the very considerable amount of evidence that we have now obtained, the reason for that being you were instructed before we had all the evidence.

You have latterly seen some of the evidence that the Inquiry has obtained, and we’ll look at some of those documents together, but any reader of your report should not assume that the judgements you express in the report are made on the basis of a review of all of our material. That would not be right, would it?

Mr Gavin Freeguard: That’s correct.

Counsel Inquiry: Thank you.

I want to start then, Mr Freeguard, by some fairly broad summaries of the early stages of your report. You make the point at the outset of your report, and perhaps it’s a theme, perhaps the most important theme running through it, that data is a vital tool for policy making, particularly in times of crisis?

Mr Gavin Freeguard: Yes, that’s right. I think for policy making, in terms of bringing evidence to that, in terms of evidence of what works when it comes to policies that are being implemented, but also to the operational side of government, and I think we see that through the pandemic as well, particularly when it comes to healthcare data.

Counsel Inquiry: Yes, thank you.

When you answer my questions, if you could try and speak slowly, that will assist all of us, including the stenographer.

You explain in your report that in recent decades, the last decade or two of the data revolution, the potential desire for government accessing, sharing, manipulating these very large datasets has grown alongside the availability of datasets of that nature?

Mr Gavin Freeguard: Yes, that’s right, and my report begins by looking at, I think, around 30 different reports that we’ve seen since the mid-90s about how government might improve, how it makes the most of some of those opportunities, and also that we’ve tended not to follow up on quite a lot of those or looked in detail at which recommendations have succeeded and which ones have not, and the reasons behind that as well.

Counsel Inquiry: Let’s perhaps just – I’m certainly not going to take you to all of those reports, in fact possibly only one of them, Mr Freeguard, but if we go to page 8 of your report, I think it’s fairly obvious, even just glancing at the pages – so we see your listing those reports. The earliest I think is in 1996.

Then if we can just scroll forward quickly through the pages, we can see all of these different reports – thank you, just if we can keep going – it’s a long list, Mr Freeguard, of reports, in the large part commissioned by central government, asking themselves questions as to how they can use data more efficiently.

You’ve touched on this point, but the question is: is there almost a failing in itself by noticing quite how many reports there are within a relatively short period?

Mr Gavin Freeguard: I think so, yes. Again, it’s not been for lack of trying, it’s not been for lack of interesting ideas about what might need to be changed and what needs to be done in order to change things, but it does feel like we often repeat recommendations.

There’s some very strikingly similar language even from the late ’90s that we could find in a report published, you know, in the last couple of years. Government just doesn’t quite seem to grasp some of the opportunities sometimes.

And I think we’re also quite poor at learning from the successes and failures of those different recommendations.

Counsel Inquiry: You in your report have tried to crystallise some of the key issues that come up, as you say, time and again in these reports: routine failures in the data field and, if you like, barriers – I think that’s a word you use – to data sharing within government.

I want to ask you about them just briefly in turn.

First of all, you talk about technical barriers, and by that I think you mean interoperability problems, the fact that government perhaps can’t speak to it – or different parts of government can’t speak to itself very well, or at least transfer data between areas of government.

Is that right?

Mr Gavin Freeguard: Yes, that’s right. So that often stems from a lack of data standards, so people in different parts of government might be recording similar data but doing it in very different ways, which makes it much more difficult, as you say, for interoperability, to compare and collate those various different things. Sometimes the data simply isn’t collected, which is definitely a technical problem. But we also have issues of legacy technology, where some of the computer systems and some of the data systems that have been in use in government have been in use for many decades, it becomes more and more difficult to extract data from those or you need much older expertise, and then trying to reform those systems and bring in newer technology becomes more challenging.

Counsel Inquiry: So those are the technical barriers. You refer to legal barriers. In a few sentences what do you mean by that?

Mr Gavin Freeguard: That tends to be where the law does not provide or is not clear that it provides the sort of locus for different parts of government being able to share or access data. I think that’s one area, and you will see this in the sort of chronology of reports, where some of those problems have been tackled over the last decade or so. Things like the Digital Economy Act and some of the clarity brought by the General Data Protection Regulation have sort of solved some of those problems, but there are still some challenges with there being a lot of different pieces of legislation that can allow data sharing, lack of clarity about how to make the most of some of those routes as well.

Counsel Inquiry: Yes.

Mr Gavin Freeguard: And I think just a final sort of legal issue is I think because different parts of government are different legal entities, that can sometimes make sharing across those different organisations more challenging.

Counsel Inquiry: Thank you.

That does, I think, lead into the third of those barriers that you describe, which is cultural and organisational barriers, and at least one of the points that you make in that regard is that, going back to the legal barriers, even if in fact as a matter of legality it is possible to share, let’s say, a dataset from one part of government to another, you might find that the general sort of concern about legality of sharing means that there is a sort of inertia there and that datasets that might be shared end up in fact not being shared, and that’s one of the points you make under cultural and organisational barriers of that nature?

Mr Gavin Freeguard: That’s right, and in a sense that comes from a very good place. I think that’s people within government organisations not wanting to get something wrong, they’re quite fearful that they might accidently share something that they shouldn’t be, and it speaks to a need for greater guidance and more support for people that are trying to do that.

I think some of the other cultural and organisational barriers that it’s worth picking out are that, you know, some of this is not about data, it’s about the structure of government and the fact that siloed government departments don’t tend to work together as effectively as they might, and sometimes the incentives aren’t there for different departments to share or make their data available.

So, for example, one department might be the one that has to invest in a particular dataset, the benefits might accrue to a different department, that will not necessarily be reflected in sort of funding or anything else.

Lady Hallett: Mr O’Connor used the word “inertia”, I wonder if “caution” might be a better word.

I mean, I remember when we first had to start dealing with data management for the judiciary, and some extremely able lawyers found some of the provisions quite complicated, and therefore that may have led to a degree of caution.

Mr Gavin Freeguard: Yes, I think that’s right, and certainly traditionally that complexity, especially when there were a lot of different legal gateways, led to some of those challenges.

I think we’re also in a relatively early stage, still, of some of that legislation bedding in, which means some of the legal precedents are still being developed and perhaps the specialism isn’t quite there in all parts of government yet.

Mr O’Connor: Then, Mr Freeguard, the final generic barrier that you describe is political. You refer to other priorities within the political space. Can you expand on that?

Mr Gavin Freeguard: Yes, that’s right. I think over the last sort of seven years or so obviously this has been quite a lot of turnover at the top of British Government. I think we have had 12 different ministers for the Cabinet Office since 2015, that being the organisation most responsible for the use of data within government. Things like Brexit and then, of course, the pandemic itself have absorbed an awful lot of civil service and political time. And another example of that is the lack of a chief data officer for government. That’s something that’s been promised at least since 2017 and there has not been a permanent appointment. And again, that sort of role could bring some political focus and strategic leadership on some of the issues that we’ll be talking about, and that has been lacking.

Counsel Inquiry: Thank you. I’m going to come back to the point about the chief data officer in a moment, but just finishing then this point, you’ve referred to those four really very generic barriers. Those are points which emerged to a greater or lesser extent in many or all of those reports at which we looked, all flashing on the screen, and those who want to can look at in more detail by looking at your report itself. We are obviously going to come on and talk about the experience during the pandemic, but is your view that those four issues, as it were, also had their part to play during the pandemic?

Mr Gavin Freeguard: Yes, that’s definitely right. I think as well it’s worth saying that the pandemic helped overcome some of those barriers. The fact that there was such urgency, there was such a clear purpose and such a collective purpose from different parts of government meant that perhaps some of that cross-government working happened where perhaps we haven’t seen it in the past. Obviously there was a strong political imperative to get things done as well.

So we definitely saw some of them being overcome in a way that perhaps we haven’t during business as usual periods in government.

Counsel Inquiry: Thank you very much.

Mr Freeguard, I did mention that I might ask you about one of these many reports and in fact as it happens it’s one of those which is shown on the screen at the moment, it’s the national data strategy, so I wonder if we could just highlight that or enlarge that subparagraph in the middle there. Yes, all right. We can go – thank you, there it is.

So, Mr Freeguard, that, as we can see on the screen, was a report, a strategy, that was published in fact during the pandemic, although you make the point that it had been commissioned some two years earlier, and I think it had been – responsibility for the review had been passed between government departments during the time it was being considered.

But tell us a little about that strategy.

Mr Gavin Freeguard: So that came from, as you say, a sort of machinery of government change and I think a renewed focus on what better use of data could do for government. And it was designed deliberately as a framework strategy, that is setting some overall missions for the use of data in government, better use of data being based on particular pillars – I’ll come back to one of those in a moment – and then there being lots of different activities, specific activities, to be conducted by different parts of government underneath that.

One of the pillars that I mentioned is – probably of most interest to the Inquiry – about data availability. That is a term that the strategy uses for data sharing, accessing data, finding data, more generally. Again, as a framework strategy, it was very much there to be an overarching thing that other parts of government might be able to use the principles from as well.

I was at a roundtable last year, a roundtable discussion organised by Civil Service World, which suggested that lots of departments were still looking to some of the principles in the strategy to try to improve their own use of data, although it does feel like over the last year or so it’s received less prominence. It’s now the responsibility of the Department for Science, Innovation and Technology, and when that was created this was not one of the initiatives that was sort of highlighted as moving from the Department of Digital, Culture, Media and Sport to DSIT.

Counsel Inquiry: I don’t think it’s clear from your list, but has this strategy yet been replaced by yet another review or machinery of government exercise?

Mr Gavin Freeguard: It’s still there, there are still some teams within DSIT working on particular missions within it, I’ve spoken to some of them quite recently, and I think, again, the fact that it’s a framework strategy, with other parts of work mentioned within it, means that it’s still living in some way.

So, for instance, Cabinet Office, the Central Digital and Data Office specifically, is responsible for a roadmap for government data and digital, and that’s one of the many parts of the strategy which continues in some way.

Lady Hallett: So we have roadmaps, we have strategies; any action?

Mr Gavin Freeguard: There is some action, there is definitely action. For instance, Cabinet Office recently have been doing work around data maturity, that’s helping different parts of government to understand how prepared they are for doing more work with data.

Mr O’Connor: Let me ask you about a different aspect of action arising from this strategy and that’s the point you made earlier about data officers.

First of all, I think I’m right – well, first of all, it’s right, isn’t it, that this national data strategy does encourage the appointment of data officers first of all within government departments and then a chief data officer to oversee data within the government generally?

Mr Gavin Freeguard: That’s right, there is a lot in there generally about upskilling civil servants to work with data and, as you say, that very specific chief data officer for government pledge.

Counsel Inquiry: I’ll come back and ask you about the chief data officer, but it’s right, as we understand it, that before 2022, so last year, the Treasury, which is, clearly, one of the departments that had a large role to play during the pandemic, didn’t have a chief data officer at all. I mean, it was only last year that a chief data officer for the Treasury was appointed, no doubt as a result of this strategy.

Are you able to help us with whether that, as it were, was indicative of a lack of control of data within the Treasury before the appointment of a chief data officer within that department?

Mr Gavin Freeguard: I think it certainly brings more focus to the use of data within the department, and means that there is somebody who is responsible for thinking about it, which may not quite have been the case before.

I think something we find sometimes across government is there may be people playing some of the roles associated with a chief data officer. They may have different names, they may not be quite the right seniority to be able to do anything with it, and that sometimes does make the picture more difficult to understand.

Counsel Inquiry: Then what about the chief data officer for the government as a whole? That post still hasn’t been filled, as we understand it, despite the recommendation dating back to 2020?

Mr Gavin Freeguard: That’s right, there have been a few interim holders of that role, there have been a few attempts, I think, since 2020 to fill it permanently and, again, since earlier than that to try to fill it as well. But as far as I’m aware, I think there is a process that has been ongoing fairly recently, but I don’t think an appointment has yet been announced.

Counsel Inquiry: I’m going to move on, Mr Freeguard, thank you for that.

You have mentioned briefly in that summary you gave us some of the statutory underpinning of the data world, if I can put it that way. We are all perhaps familiar, to a greater or lesser extent, with the Data Protection Act, GDPR, now the UK GDPR. You also refer in your report to the privacy and electronic communications regulations in that context as part of the basic underpinning.

I’m not going to go to the detail of any of those provisions, but one of the provisions I do want to ask you about is another one you’ve mentioned, which is the Digital Economy Act 2017. That’s at page 13 of the report.

If we can expand paragraph 14, the provisions of this Act are, I think, one of the areas that you mentioned a moment ago that perhaps hasn’t fully bedded down, or it’s not fully understood exactly what functions it serves and, if you like, the permissions that it grants.

Tell us, if you will, what innovation this Act represented.

Mr Gavin Freeguard: So the Digital Economy Act of 2017 covered issues beyond those of interest to the Inquiry, so things around broadband and digital connectivity as well. But when it came to data sharing within government, it did sort of provide powers for different parts of government to be able to share data with one another, and in fact I think very much a response to some of those many reports that we’ve already touched on about the lack of legal clarity and around powers for data sharing.

So it introduced powers for sharing data through the purposes of digital governments and public service delivery. It does not cover sort of health and social care, adult social care provision, so it does cover other purposes but not those ones. And there is, under certain parts of the Act, a register that is published by Cabinet Office, so we have some insight into some of the data that’s being shared across government.

Counsel Inquiry: Thank you.

So perhaps two important points: one is the general point you make that under this Act is a freeing up of the ability within – for – well, of sharing of data within government, but important to note, particularly of course given our focus on the pandemic, that that general liberalisation of sharing of data did not apply to health or social care data; is that right?

Mr Gavin Freeguard: Yes.

Counsel Inquiry: That sharing of data of that particularly sensitive nature can be done but there is a special procedure which you refer to in your report, I’m not going to take you to the provisions, but it’s the Health Service Control of Patient Information Regulations which permit a Secretary of State to, as it were, grant specific and very focused permission for the sharing of, for example, healthcare data. In fact this morning we heard Sir Ian Diamond talking about what are known as COPI regulations, and I will ask you some short supplementary questions about those in a few moments.

My Lady, I wonder if that would be a convenient point for a break?

Lady Hallett: Certainly, perfect timing, Mr O’Connor. I shall return at 3.15.

(3.00 pm)

(A short break)

(3.15 pm)

Lady Hallett: Mr O’Connor.

Mr O’Connor: My Lady.

Mr Freeguard, I want to move and ask you some more sort of focused questions about issues that arose, difficulties that arose in relation to using and sharing data at the outset of the pandemic, and as you’ve already suggested it may be that we see represented there some of those generic barriers that we discussed before the break.

The first of these is simply a lack of data available on which to make decisions, and for these purposes perhaps if we can look at your report, page 41, paragraph 79.

We can see, starting four lines down, you refer to the well known decision, on 12 March 2020, to stop community testing, the test and trace initiative, and the very simple consequence of that, from a data point of view, was that the tap was turned off. And so you refer there to a modeller making the comment, “if you’re only seeing the tip of the iceberg - hospitalisations, deaths - you don’t know how broad that iceberg is”. The understanding of the spread of the virus in the community was simply not something that was available as a matter of data at that stage?

Mr Gavin Freeguard: That’s right, and I’m not sure I can put it any better than that quote from the modeller.

Obviously in order to properly monitor the outbreak and understand how to tackle it, that sort of data is what you need decision-makers to have, and the decision to stop testing meant that case data was much more difficult to come by.

Counsel Inquiry: As we heard in Mr Keith’s opening, only a couple of weeks or so later the discrepancy between cases, known cases of Covid, which had been the subject of tests, and what retrospective estimates of what was in fact the position, had grown, a bit like the iceberg, so that there were some 6,000 or so positive test results, but the estimate was that it was something close to half a million people were infected by that stage. The data simply wasn’t available to inform the position.

Mr Gavin Freeguard: That’s right.

Counsel Inquiry: So that’s perhaps one example of a challenge faced.

Another slightly different challenge, but which again goes back to those barriers we were discussing, was a reluctance within government, or an inability perhaps, to share data that was available. That’s where we come back to the Digital Economy Act that we were discussing, which on the one hand had freed up the sharing of most data within government, but not healthcare data.

Thank you. It’s paragraph 48 of your report, which refers to the first Control of Patient Information notice being issued or COPI notice as we heard Sir Ian Diamond refer to them as.

This was a notice issued by the Secretary of State under these regulations which, in effect, put healthcare data into the same category as other forms of data under the Digital Economy Act, although for an express purpose and a limited period of time, and I think you refer to the fact that these notices had to be renewed, I don’t know if it’s every three months, every six months, something of that nature?

Mr Gavin Freeguard: That’s right. I think it’s worth underlining that the General Data Protection Regulation and the Data Protection Act meant that there was the flexibility there for data sharing, including in health, to happen, but from things that I’ve heard and included in my reports, the COPI notices were extremely important in sending a signal across the whole system that it was extremely important to be able to share particular types of health data, and even if – I also note that some people were still reluctant, it still took a little bit of time to overcome their caution and to give them that legal reassurance, but it did act, in the words of some people that we interviewed for an Institute for Government project, as a badge of honour and a security blanket, it made them feel much safer in being able to share a lot of that health patient data.

Counsel Inquiry: Thank you.

A further safety blanket, if that’s what we want to call it, was provided by the Information Commissioner, at around the same time. This is something that you refer to at paragraph 28 of your report on page 17.

We haven’t mentioned the Information Commissioner so far, but again I think most of us in this room understand that that is someone who is tasked with enforcing these various data regimes that we’re discussing.

Is it right that at the outset of the pandemic the Information Commissioner issued a notice encouraging people to share data?

Mr Gavin Freeguard: Yes, that’s right. There was a short statement in March 2020 where I think the Commissioner said she would be pragmatic in how she applied the existing principles to try to get people to share data that needed to be shared. That was something that was welcomed by the National Data Guardian, a sort of watchdog for the use of health data, who listens quite a lot to patients.

And then as paragraph 28 shows, there was a sort of longer document, though still a short one, in April 2020 that again acknowledged that there were severe pressures facing people in government and elsewhere during the pandemic, and also that data did need to be shared for health purposes, particularly during the early days of the pandemic.

Counsel Inquiry: And again, from your assessment on the ground, did this have an effect?

Mr Gavin Freeguard: It seems to have – again as you said, it’s another security blanket that meant that people knew that there was some protection there, there was guidance there, there was support there for what they needed to do.

Counsel Inquiry: Thank you.

I’m going to move on again to a slightly different context, which takes us back to the sort of systemic or technical barriers within government. Even if the data exists, even if there is permission to share it sometimes the infrastructure simply isn’t there to allow that sharing to take place.

For these purposes, yes, it’s page 40 and paragraph 75 of your report where I wanted to pick up, and you refer there, we’ve covered data that didn’t exist, but it’s the formatting, it’s the process issues, and you refer to data flows not having been thought about or tested as part of pandemic preparedness exercises. I’m going to come back to Operation Cygnus in a moment.

If we can go on to paragraph 76, you refer to:

“Several … Downing Street figures [having] spoken about the lack of developed data pipelines and processes at the start of the pandemic. Data collection and collation was instead ad hoc, with officials emailing and being emailed Excel files, Word documents … making phone calls … “

Writing it down, pulling it together. All very inefficient.

Does that seem to have been the process that was taking place in, for example, Downing Street and the Cabinet Office in the early stages of the pandemic?

Mr Gavin Freeguard: Certainly at the start, as you can see, Simon Case, the Cabinet Secretary, among those speaking afterwards about the fact that that’s how it was working. There weren’t those sort of reproducible pipelines, APIs, another way of being able to share data more quickly and more easily. Instead, a very fragmented process of emails, Word documents, Excel spreadsheets, phone calls, people reading off scraps of paper onto whiteboards in Downing Street, rather than having the flows there ready to go. I think in his statement Tom Shinner expresses his surprise that so much effort had to be put into bringing that data in, rather than just being able to enact protocols that already existed.

Counsel Inquiry: You might have to educate us a little bit, Mr Freeguard. I mean, to some of us, the idea of emailing an Excel spreadsheet might feel like a sort of pretty sophisticated technical thing to do, but you’re describing it here as something that was not sophisticated enough.

Mr Gavin Freeguard: When it’s coming from hundreds of different health organisations at a rate of knots, it’s much more difficult to keep hold of all of that, having to copy and paste it. We know as well, for instance, Public Health England at one point was compiling its line lists – that is a spreadsheet with a line for each patient – from individual emails. A lot of those things should be much more automatic. There should be processes in place where that data is being brought together much more easily, sort of automatically, rather than a lot of effort having to be spent collating – especially in a moment of extreme pressure – from all of those different sources.

Counsel Inquiry: You’ve mentioned APIs. We perhaps don’t need to get into the detail, but essentially this is a piece of software which manipulates automatically, it’s a feed; another term you use is a data pipeline. It bypasses the need for any particular person to look at spreadsheets or documents and amalgamate them; it does it all automatically?

Mr Gavin Freeguard: That’s right, application programming interfaces which, as you suggest, allow sort of different computers to talk to one another without the need for all of that copying and pasting.

Counsel Inquiry: There was a reference in paragraph 75 to the fact that these lessons perhaps should already have been learnt, and you refer to Operation Cygnus, which is something that the Chair and the Inquiry is familiar with from Module 1 of this Inquiry, but it may be that this particular aspect to it is worth emphasising.

If we can go to paragraph 80 of your report, please, on page 41, and picking up four or five lines down, you say:

“… there are some factors [beside the general ones you refer to at the beginning of the paragraph] more specific to the pandemic. Previous pandemic preparedness exercises touched on data, but not in much practical detail; where they made recommendations, they appear not to have been taken up. One of the best known drills [Exercise Cygnus in 2016, the flu simulation] recommended establishing a cross-government working group to clarify the ‘process and timelines for providing and best presenting data on which responders will make strategic decisions’ since participants ‘were unclear about how epidemiological information would be produced and disseminated’.”

In other words, it seems, addressing just exactly this point: how is this sort of data to be received and used centrally in the context of a disaster or a pandemic?

But you go on to make the point that Cygnus was set in week 7 of an epidemic. Why is that important?

Mr Gavin Freeguard: I would hope that you’d be able to get those data flows up and running, or would need those data flows up and running much earlier in the process, and I think it shows that although there were – you know, there was high-level commitment to better data, the thing that was missing was actually thinking about those practicalities: what data did you need to get to which people, and how were you actually going to do it?

Counsel Inquiry: May it have been that Operation Cygnus just rather assumed that the data flows would be coming in, but they didn’t think about in fact whether that would be the case in the event of a pandemic?

Mr Gavin Freeguard: Quite possibly. And again, there are other exercises which also thought about what data might be needed, what data might be there. In some cases, it was perhaps not the data that was expected when it came to Covid. A lot of the exercises had been based about – around different types of pandemic. So there may have been some surprises about what data was available when it came to healthcare in particular. But, again, there seems to be that lack of granularity and practicality in thinking about how those flows would actually work.

Counsel Inquiry: We may be able to get a fairly practical insight into this by looking at a document known as a CRIP, which was one of the collations of data that was in play in the early stages of the pandemic. So if we could have up on screen, please, INQ000056187. Thank you.

Are you familiar with what a CRIP is, Mr Freeguard?

Mr Gavin Freeguard: Yes.

Counsel Inquiry: What does it stand for?

Mr Gavin Freeguard: Commonly Recognised Information Picture.

Counsel Inquiry: That wasn’t a very good question, was it, because the answer was there in front of you?

Mr Gavin Freeguard: Right on the screen.

Counsel Inquiry: But it’s a term of art, is it not, in sort of contingency planning, and CRIPS are intended to be a sort of regular daily bringing together of the critical information that decision-makers need in order to make the decisions that fall for them?

Mr Gavin Freeguard: Yes, and I think an attempt to create a single source of truth so that all of those decision-makers are at least starting in the same place.

Counsel Inquiry: This, we note, is a CRIP dated Wednesday 18 March. As we will – as we’ve already heard, but we will certainly come to hear in more detail, that was a critical moment in the early stages of the pandemic. We will hear evidence about the SAGE meeting on Friday the 13th, a few days before, where the true scale of the decisions facing the country perhaps became apparent; crisis meetings over the weekend, we’re in the last few days before the decision to implement the first national lockdown were made early the next week, so it’s really a sort of crisis moment.

We can also see, of course, it’s CRIP number 28, so this CRIP system has been in play for a little while by that stage.

Before we look at the detail, Mr Freeguard, an important point to bear in mind is that this document is, if you like, a static document. Can you explain what we mean by that?

Mr Gavin Freeguard: So a snapshot of data at a particular point that were being circulated and printed out, as opposed to what we saw later in the pandemic, which was something more interactive. So, for instance, the Covid-19 dashboard and the various internal equivalents, which could be updated more regularly.

Counsel Inquiry: We’ll certainly come to look at that. So it’s a static document, simply it stands for itself, but it’s also the case, isn’t it, that someone actually needs to populate the various tables and charts or bullet points within the document? It doesn’t happen automatic, someone’s got to type in the numbers and the detail.

Mr Gavin Freeguard: That’s right, with something like a dashboard the pipelines would be running to bring the data through, whereas, as you say, somebody would have to sit down and paste things in for this.

Counsel Inquiry: It’s a relatively short document, six, seven, eight pages long, something of that nature. If we just look at the next page, please, we see a situation update. I just want to draw your attention to, I think it’s the fifth bullet point down at the top. There is an estimated population infected figure there of 5,000 to 10,000. Of course we know, looking back, that that was a gross underestimate, but that isn’t really the point I want to make.

If we bear that figure in mind, 5,000 to 10,000, if we can just go on to the next page, we see in the third box down exactly the same variable, “Estimated population infected”, but instead of 5,000 to 10,000, it’s 30,000 to 40,000.

So it would seem that whoever was typing in the numbers into the box, or perhaps it was more than one person, perhaps from a different source, we’ve ended up with a short document with two very different figures for the same variable. One might have thought one of, if not the most important variable in the document: how many people do you think in this country have got Covid at the moment? You look at one page, it says 5,000 to 10,000; the next page, well, it could be 30,000 to 40,000, this being the document that the Prime Minister is looking at in order to make his decisions.

It’s quite a striking example, isn’t it, of the sorts of problems we were discussing?

Mr Gavin Freeguard: It is, and this is exactly what those hundreds of individual Excel spreadsheets or emails or Microsoft Word documents gets you to, it’s that people copying and pasting, there not being the right quality assurance around the data that’s coming in.

Counsel Inquiry: We imagine someone sitting sort of besieged by emails, he’s printed some of them off, he’s looked at some of them on their phone, and mistakes happen?

Mr Gavin Freeguard: Quite.

Counsel Inquiry: If we could go on to page 5 of the document, please, again one might have thought a rather important page for the key decision-making that was going on, trying to understand what the health and social care situation is in the country.

It’s quite striking that, of the columns which are actually completed, in other words data is available, one of them is for the percentage of NHS 111 calls answered within 60 seconds. Not, one might have thought, the most important data on which to decide whether to, for example, lock down the country or not.

Moving along, we also see full data for the number of urgent operations that have been cancelled. But between those two columns, the area on the sheet for the number of ICU beds occupied, the percentage of ICU beds occupied by Covid-19 patients – and bearing in mind that one of the priorities, if not the priority, that had been identified by the government at that point was protecting the NHS from collapse because of being overwhelmed by Covid patients – no data seems to have been provided?

Mr Gavin Freeguard: Indeed, and I think some of the data that you would need to work out the percentages there would actually just be things like the overall number of beds in the system. Even data like that, which should have been easier to come by, was also missing at the start.

Lady Hallett: Can I just go back? Mr O’Connor, did you say these documents were prepared for the Prime Minister and Cabinet?

Mr O’Connor: Yes. Well, we’ll hear evidence about it, but these were –

Lady Hallett: Okay. And this was number 28.

Mr O’Connor: Yes.

Lady Hallett: So are we assuming that that data was missing in numbers 1 to 27 as well?

Mr O’Connor: I think, to be fair to Mr Freeguard, he may not know the answer – or certainly doesn’t know the answer to that question, and in fact I don’t know whether we have a full set of these CRIPS in evidence, but it’s certainly something we can look at.

Lady Hallett: Extraordinary.

Mr O’Connor: It is, though, is it not, Mr Freeguard, of a piece with the evidence we were looking at earlier about frantic activity in Number 10, emails, jotting things down on pieces of paper? And I’m going to take you to another document now, which is an email from Dominic Cummings to two of the Number 10 staff expressing a degree of frustration about these types of matters.

This is INQ000174715, please. If we can pick up about halfway down where it says “Tom/Marc”.

So this is an email on 25 March, so a week or so later, just after the lockdown has been announced. He is talking about the dashboard, we’ll come back to that, but it’s the same sort of data or types of data that we are discussing. He says:

“The NHS numbers remain a mess. I look at the daily COBRA reports, I scribble down numbers people say 815 or 915. I look back and none of them are consistent.”

I’m not going to read it all out. If we cast our eyes down, we see he is trying to make sense of the reports.

Then if we can go to the next page, please, he is talking about beds forecast to be free, so similar type of data that we noticed wasn’t available. Then he says:

“No deaths – number reporting seems a shitshow, different times every day. WTF use is a number for 9am that’s then updated retrospectively …

“There should be a number calculated in a sensible way and done at the same time every day and is intuitively sensible. This shouldn’t be too much to ask of NHS.”

Then, perhaps picking up on the same theme as the observation by my Lady a few minutes ago:

“I’m truly amazed that even after weeks this most basic stuff is so hard for us to get our hands on…

“Am I being unreasonable?”

If we can go straight to the response he gets, that’s INQ000048313, page 31, please. You’re ahead of me, thank you. No, you’re not. It’s page 31.

So we see in fact that at the top half of the page is the last bit of that email, and then Tom Shinner replies, bottom half of the page:

“You are not being unreasonable. It is a mess; it should be possible to fix all this. You’ll see some of it tomorrow.”

He refers to:

“… deliberate obfuscation … in the room and the boss letting MH [perhaps Matt Hancock] [and others] get away with throwing him a random graph with subtly different numbers none of us have seen.”

And so on.

Thank you.

So both in the CRIP and perhaps in that exchange of emails we see, at least in the outset, some very serious data availability, data handling problems in Number 10?

Mr Gavin Freeguard: And as I think the Dominic Cummings email shows as well, some of those issues that come from perhaps the same thing being recorded in different ways in different places and not being able to bring it together so easily, and the time and effort that it takes to come to that single version of the truth, when you should be – you should be having time to discuss the response to all of it, you’re still arguing over the numbers to begin with.

Counsel Inquiry: Now, we saw that one of the recipients of that email from Dominic Cummings was Tom Shinner, and he was the person or at least one of the people responsible for developing the dashboard that you’ve referred to, and that was a product that was designed to do something very similar to the CRIP in terms of being, in your phrase, a single source of truth, but it was a very different type of product, was it not?

Can you give us, just in a few sentences – we’ll have a look at it in a moment, but tell us how it was different.

Mr Gavin Freeguard: Certainly over time a lot of the data pipelines became much more automated, so we avoid some of the problems that we’ve already discussed. It’s much more visual as well, rather than just numbers in a table we can see it brought to life in actually very clear charts and graphs as well, and I think some of the – some of the key issues around the caveats behind any data, the sources of the data, when it’s updated, that becomes much more standardised and regular and clearly marked as well.

Counsel Inquiry: Just to emphasise the point about automation, and we’ll come to it, but instead of it being a product of people receiving emails, spreadsheets and manually populating tables, perhaps as the name “dashboard” suggests, this was an automated process where the information simply fed into the system and the graphs and so on were dynamically generated?

Mr Gavin Freeguard: Certainly much more automation.

Counsel Inquiry: Let’s have a look, please, it’s INQ000 – yes.

Now, this was a – first of all, it’s an internal product for the use – for use of the Prime Minister and other members of the government. Is that right?

Mr Gavin Freeguard: Yes.

Counsel Inquiry: We see it’s got “COBR” at the top. The reason I mention it there was a different – I think later there was a public-facing dashboard which provided similar, albeit less detailed, information to the world at large, but this is the internal government version of the product; is that right?

Mr Gavin Freeguard: Yes, that’s right.

Counsel Inquiry: It was – we’ve described it as being dynamic, so it was updated perhaps all the time but certainly it would be possible to look at a different dashboard for every day; is that right?

Mr Gavin Freeguard: Yes, and across various different measures as well.

Counsel Inquiry: Yes, and we can see that the date here in small type is was almost the prototype and it was before it was actually being used. So it was in fact, if you remember the email between Dominic Cummings and Tom Shinner was I think 25 March, so this is around the time that the dashboard is being set up and this is, as it were, a prototype. But as you say, we see immediately it’s a very different sort of product, much more accessible. Can we go on and look, please, at pages 11, 12 and 13 in series. So there, as it happens, we see the way in which the dashboard presents the information which was missing from that CRIP: the number of ICU beds, how many are being used, therefore how many are available. And similar – if we can just scroll forward to pages 12 and 13, we see similar data relating to those matters.

Lady Hallett: How did they make the transition from the obviously inadequate CRIP to the dashboard and the automated – “populating” I think is the word, isn’t it? Did they have the software available, or did they buy it as a matter of emergency? How did suddenly we make this transition?

Mr Gavin Freeguard: My understanding is that they brought some more expertise into the centre, so people with sort of data science skills, I think there were a few private March. This, in fact, I think was the very first, it companies who may have brought software in as well that was able to, I think to begin with, patch together some of those various different sources that were coming in, and then over time develop those pipelines that were much more automated to bring the data through.

Mr O’Connor: I’m sure we’ll hear more evidence about this, Mr Freeguard, but it’s fair to say that none of the things that happened in setting up the dashboard could only happen once the pandemic started; these were all things that could have been done in advance in readiness for a pandemic?

Mr Gavin Freeguard: That’s correct.

Counsel Inquiry: Just before we leave this, it’s right, isn’t it, that once this dashboard in fact became operational, within a week or two of this date, over time it became much more sophisticated? I think when one prints off this document it’s 20 or so pages long, certainly it doubled, trebled in size, as the product became more sophisticated, more data was added to it during the course of the pandemic; is that your understanding?

Mr Gavin Freeguard: Yes, that’s right and, as you suggested earlier as well, there became a public-facing version on a website which was much more interactive, very well sourced, lots of notes about how it was being put together, and allowed any member of the public to dive into the data in more detail.

Counsel Inquiry: Yes. Thank you.

Just following on from this, I want to ask you about just two of the other government initiatives during the pandemic to improve the use and sharing of data. There are a number that you mention in your report. I just want to ask you about two.

First of all, what’s known as No 10DS, so if we can turn to page 19 of your report, paragraph 35(c). So the DS stands for data science team. Can you tell us a little about this initiative?

Mr Gavin Freeguard: Yes, so this was an attempt to bring in people with sort of greater data science expertise, and a sort of variety of skills around that to improve the use of data and analysis within Downing Street, so to make sure that the Prime Minister and the Cabinet were getting the best analysis they could, thinking about more innovative ways of using data to develop an even clearer picture, and I think as well ensuring that Downing Street had the data capability to hold other parts of government to account as well.

Counsel Inquiry: And I think it’s right to say that at least one of the people who was responsible for this data science team was a man called Ben Warner; is that right?

Mr Gavin Freeguard: Yes, that’s right. There were other initiatives alongside this such as JDAC, the Joint Data and Analysis – sort of – Centre, which again fulfilled some similar roles.

Counsel Inquiry: Yes, I think it was one of the other addressees of that Dominic Cummings was Marc Warner, who was Ben’s brother, and they were both working alongside Dominic Cummings in Number 10 at around this time, and Ben Warner will be giving evidence to the Inquiry in a couple of weeks’ time.

So that’s the No 10DS.

On the next page, page 36 – sorry, no, in fact on page 59, paragraph 36 of your report, if I’ve got that right. It may be that I haven’t. But in any event, yes, the establishment of the Joint Biosecurity Centre. Now, we heard something about this yesterday in the context of its jointness, the fact that it drew in expertise from not only England but the devolved nations as well. From a data angle, what was the significance of the Joint Biosecurity Centre?

Mr Gavin Freeguard: So it was designed to be a sort of single analytical function on behalf of the sort of public health side of things, to bring together various different health data flows, and be able to advise the Prime Minister and Cabinet and others based on that information.

Counsel Inquiry: Thank you.

I want to move on, Mr Freeguard, and just touch on all – or sort of survey the different types of dataset that were available to the government during the pandemic, and this is something that you explain at some length in your report. In fact, several of these areas have been covered with Sir Ian Diamond this morning, which means we can take them very shortly.

If one classifies the various datasets – and as I’ve said, there were very many of them as the pandemic progressed, were there not?

Mr Gavin Freeguard: Yes, and of very different types as well.

Counsel Inquiry: Different types, some with sort of interesting sounding names like SIREN and Vivaldi, others with less poetic descriptions. But, in any event, one might describe as testing and healthcare datasets, starting most obviously with, as you said, line data from hospitals, details of admission numbers, deaths, availability of ICU beds and so on. But also then, as the pandemic progressed, various prevalence studies, and those Sir Ian did discuss, including of course the ONS Coronavirus Infection Survey which the ONS ran itself, so we don’t need to take time on that.

One of the other datasets relating to prevalence that you refer to in your report was waste water datasets. Tell us something about that.

Mr Gavin Freeguard: Yes, so again a sort of innovative use of data from the JBC and later the UK Health Security Agency where various sewage centres across the country, the water there was tested to understand if there were remnants of Covid, because obviously people infected with Covid would shed sort of samples in various different ways, waste water being one of them, and that was used to indicate whether there might be a sort of rise in cases to be expected.

Counsel Inquiry: Did that enable some sort of – presumably some sort of regional variation to be identified, depending on where the testing was taking place?

Mr Gavin Freeguard: Yes, that’s right. I think it covered around three-quarters of England, I think I’m right in saying.

Counsel Inquiry: The other area of datasets that I wanted to pursue with you, which again we didn’t get into with Sir Ian, was what you describe as non-health data. So, leaving aside all of the studies about whether people had Covid or not and vaccines and deaths, these were a really rather different source of data that were nonetheless useful, were they not?

Mr Gavin Freeguard: Yes, that’s right.

Counsel Inquiry: Another term which is I think used to describe at least some of this type of data is mobility data?

Mr Gavin Freeguard: Yes, so one of the data sources that was used, for example, by some of the groups modelling the reproduction number and other statistics, was Google mobility data. So this is where people with sort of Google on their phones have not disabled location on there and that’s able to give an indication of sort of where people are moving.

Counsel Inquiry: So just pausing there, nothing to do with searching for things on the internet, but Google also has this slightly less well known function of knowing where you are?

Mr Gavin Freeguard: That’s right, the sort of thing that enables people to see where they are on Google Maps and so on.

Counsel Inquiry: So, just go on, how was that used?

Mr Gavin Freeguard: So that and various other big studies as well, such as CoMix, CORSAIR, I think the Scottish Contact Survey as well, and also polling, were all used as indications of to what extent people were mixing, how many social contacts were they seeing, and that could be used as part of modelling the spread of the disease.

Counsel Inquiry: You also referred to telecommunications operators providing data of this nature. Was that also simply relating to mobility or was that something else?

Mr Gavin Freeguard: Yes, I think that tended to be the same purpose.

Counsel Inquiry: It may be that we hear more from the modellers about this, but as well as being a factor that could feed into the R number, it’s right, isn’t it, that this data also provided an insight into the simple question of whether NPIs were working or not?

Mr Gavin Freeguard: Yes, that’s right.

Counsel Inquiry: So how much less movement around the country is there this week than there was last week?

Mr Gavin Freeguard: Exactly.

Counsel Inquiry: It’s also something you mention in your report that educational establishments, schools, universities, fed in this type of data?

Mr Gavin Freeguard: That’s right. So, again, some of the R modelling groups and others made use of school attendance data that the Department for Education got up and running, I think asking schools to fill in forms at the start and then I think also automating that process as far as they could, again as an indicator to see how people were moving around but also how many people might be absent with Covid.

Counsel Inquiry: Yes. Thank you.

I’m going to move now to a slightly separate subject, which is one of those subjects you were asked to address in your report, which is that of data sharing between the central UK Government and the regions, the regions and local authorities in England.

For this purpose, if we could go to page 44 of your report, please, it’s paragraph 85. Thank you. Paragraph 85.

Now, you say, Mr Freeguard, that the UK Government was slow to share data with local government during the pandemic, with the result that many parts of local government in England did not have the data they needed at a sufficiently granular – it might mean level – to understand and combat the spread of Covid in their areas, even when some of those areas were entering local lockdowns, although you then go on to say it was eventually shared, and there are some examples of innovation and so on.

First of all, the first sentence of that paragraph, you refer to something that didn’t happen as well as it might have done, the sharing down to local and regional levels.

I want to just try and draw together some of the points you make in the rest of that page, and it seems first of all that one of the problems that you’ve identified is that local authorities were regarded as, your words, an afterthought in government planning and that central government’s institutions lacked an understanding or a sufficient understanding of local government to facilitate sharing of this type?

Mr Gavin Freeguard: Yes, that’s right. I think “afterthought” is a quote from somebody we spoke to –

Counsel Inquiry: Yes.

Mr Gavin Freeguard: – in local government, and I think they felt they hadn’t been included in any of the planning processes, and again we’ve seen the effects of not properly planning and thinking about this.

As you say, there’s also that sense they had that central government didn’t necessarily understand what they might need data for, how they might need to use it, and the benefits that it might have for them.

I think it’s worth saying this is something that also happened before the pandemic. I’ve spoken to a lot of people in local government who are required to submit various types of data up to central government and often don’t feel like they’re able to make the most of it, it takes a very long time for it to come back to them, or it’s very much a one-way relationship and they’re not supplied with information that could be valuable to them in running their local areas at the best of times.

Counsel Inquiry: You also make the point that it isn’t a straightforward thing to do for something called central government to share data with local authorities and regional authorities, precisely because those lower level, lower tier authorities are very various, they have different capabilities, different infrastructures, different maturity when it comes to dealing with data, and so it’s actually quite a difficult process, it’s not just a question of flicking a switch?

Mr Gavin Freeguard: That’s right, and again it will be quite different for different datasets, data in different areas. As you say, there are different parts of local government with sort of different levels of data maturity and being able to use that, but I think even those that are very experienced in this sort of thing felt they weren’t getting what they needed. And again, you know, this is a sort of longer standing problem than just the pandemic.

I think there is also a lot of overlapping jurisdictions. I think one of the reports I refer to in the annex, somebody said that at one point during the pandemic they were asked to share data with a local resilience forum, to which their reply was, “We’d be very happy to, if we knew what one of those things was”.

I think there are, as you suggest, lots of different types of organisation and, again, understanding which of those needed what at what particular moments was quite difficult.

Counsel Inquiry: Then lastly, of these themes about problematic sharing – actually if we could just look, please, further down the page at 86(c) – you’ve referred to some of these problems being long-standing, Mr Freeguard, and emanations of difficulties that had existed before, but this point is specific to the pandemic, is it not, that these problems were, as it were, exacerbated by the government setting up, for example, the Test and Trace scheme independently of local public health authorities?

Mr Gavin Freeguard: That’s right. I mean, despite – local government again would say that the funding cuts that they’ve experienced over the last decade or so has been a real problem in a lot of this. They still have a lot of expertise, including when it comes to sort of tracing the spread of disease. Things like Test and Trace being set up outside those existing systems and not building on that expertise created further problems.

Counsel Inquiry: It’s going to be the subject of another module so we won’t get into it, but from – there may be lots of good reasons to set up Test and Trace in the way that it was, but what you’re saying here is that from a data point of view it was in effect creating a new silo?

Mr Gavin Freeguard: That’s right.

Counsel Inquiry: Those are, I think, the key themes, negative themes, early problems; but it’s right to say, as you indicated, or as recorded in that first paragraph, that things did get better during the pandemic and these problems were overcome, to a greater or lesser extent?

Mr Gavin Freeguard: There seemed to be improvements. Again, it would probably vary with different datasets in different parts of the country, but I think some things were probably learnt from the earlier days, as new data came online it may have been dealt with differently, and there were efforts from central government, particularly what is now the Department for Levelling Up, to sort of try and support others across the country.

Counsel Inquiry: I’ve just got two more areas I want to address with you, Mr Freeguard, and the first of those two is the question of data on race inequality disparities during the pandemic.

Perhaps we can turn, for those purposes, to page 46 of the report, paragraph 89. You say here, quite bluntly, that:

“The system for capturing data around race, ethnicity and other characteristics was no more unified than any other data systems.”

So perhaps another way of saying it shared a lot of the problems we’ve been discussing for the last hour or so?

Mr Gavin Freeguard: That’s right, whether it was not collected as part of certain datasets or, again, whether there were inconsistencies in how it was collected; and I think there is also a definitional issue here which is that often different ethnic minorities can be brought together under the same sort of data headings when actually the experience might be quite different for different groups.

Counsel Inquiry: Yes. We have heard already a certain amount of evidence, of course, about the disproportionate impact of Covid on certain ethnic groups. We heard some evidence last week from a FEHMO witness who emphasised the, as it were, almost immediate understanding amongst certainly healthcare workers of the disproportionate impact that black and ethnic minority healthcare workers were experiencing, and we saw correspondence from one of their organisations to Chris Whitty and other senior members of the sort of healthcare establishment, amongst other things, pointing to the lack of data, the urgent need for analysis and understanding of what exactly was going on.

You refer in your report at paragraph 90 to the PHE, the two PHE reports that then followed quite swiftly during 2020 addressing these matters: one, the disparities in the risk and outcomes of Covid-19, and then later the report entitled “Understanding the impact of Covid-19 on [black and minority ethnic] communities”.

So that was the process, but is it right, though, that in terms of data both of those reports essentially concluded that there wasn’t enough data and that, albeit that no doubt they addressed other steps that could be taken, in terms of the data what was really needed was more?

Mr Gavin Freeguard: Yes, that’s right, and it’s again something that I’ve heard from members of modelling groups as well, that they would have liked to have been able to have that detail to understand how the disease was affecting different communities. The PHE reports that you mentioned again talk about possible steps such as standardising definitions across different government datasets, thinking about making it mandatory on death certificates to record ethnicity, and various other steps around the data as well.

Counsel Inquiry: In terms of progression during the pandemic, we can see in this paragraph that it’s been highlighted, just the last sentence or so, you do refer to several of the early Covid-19 studies allowing for analysis of particular characteristics, and then you refer to the ONS – the CIS, as Professor Diamond called it – the Covid Infection Survey having some fields for occupation, ethnicity and deprivation.

And then actually if we can look further on at paragraph 91, please, you’re referring here to the technical report produced by Sir Chris Whitty and Patrick Vallance but they note – I’m picking up on the last few words of this page, and we’ll go on to the next – they note in the technical report that by the second wave, so later in 2020, the weekly Covid-19 surveillance report incorporated more disparities data.

And then dropping down a line or two, from May 2021, so over a year after the start of the pandemic, the PHE were publishing what was known as the CHIME data, so Covid health disparities monitoring for England, which included reporting on disparities, although you say that it, like several other datasets, didn’t have access to underlying conditions.

So the picture seems to be, Mr Freeguard, but tell me if I’m right, that this problem of inadequate data even once the pandemic was well under way, even once the disparate effect was known and people were trying to do something about it, even as late as 2021 it seems there were still problems with these datasets?

Mr Gavin Freeguard: Yes, that seems to be the case. And I was struck, reading the technical reports, there’s an entire chapter on disparities, and it begins by saying that infectious disease outbreaks tend to exacerbate existing inequalities, and I think that does also again take us back to the questions about preparation and whether perhaps some of this could have been foreseen in terms of the demographic data might be needed.

Counsel Inquiry: You fairly, though, identify in your report that there are particular difficulties, perhaps, sensitivities around this type of data. So, for example, at paragraph 95 you refer – and again this is a reference back to that part of the technical report that you were describing – the technical report itself stating that the data is sensitive and it was being asked of communities with relatively low trust in government organisations and understandable concerns about privacy and the use of their data.

Mr Gavin Freeguard: Yes, and I think it’s fair to say as well that the Public Health England publications and others recognise that and speak quite a lot about the importance of engaging the community.

Counsel Inquiry: At paragraph 96 you expand on a point you referred to earlier, which is calls by the Equality Hub that a key lesson from the pandemic is that we shouldn’t be treating ethnic minorities as a homogeneous group and that data needs to be more sophisticated.

Mr Gavin Freeguard: Yes, that’s right and, as you suggest, the Equality Hub, the Race Disparity Unit and others have started to do quite a bit of work on what that might actually look like in practice.

Counsel Inquiry: Then, finally for these purposes, paragraph 98. You go back to the technical report, and this does also bring us back to some of the evidence that Sir Ian Diamond was giving, first noting that ethnicity recording on NHS systems is of poor quality, and then referring to reports from the Race Disparity Unit, their final report on addressing Covid-19 health inequalities. It says:

“… key lessons include[d] improving the quality of health ethnicity data so patterns and trends can be spotted more quickly, and continuing to monitor the impacts of Covid-19 by ethnic group.”

And their recommendations – if we just briefly go over the page – again familiar points, points – at least some of them – raised by Sir Ian Diamond. We see (b), for example:

“The ONS is to collaborate with others to consider how linking Census and health data could be improved …”

So that’s the innovative – either innovative or experimental, depending on who you listen to – technique that he was describing and a way of providing more robust ethnicity data and combining different – the census data with healthcare data to that end; is that right?

Mr Gavin Freeguard: That’s right, yes.

Counsel Inquiry: Then also we see, dropping down two points, the issue – again to which you’ve referred – of collecting ethnicity on death certificates. I’m not sure you were in the room, but that fact was something Sir Ian Diamond did not support; he was much keener on trying to obtain the data from NHS records.

Lady Hallett: You keep saying “Sir Paul”, I think it’s Sir Ian.

Mr O’Connor: I’m sorry.

Lady Hallett: It’s all right. The stenographer has been reporting “Sir Ian”.

Mr O’Connor: She might correct it.

That’s what I wanted to ask you on ethnicity data, and that just leads me to the last point, Mr Freeguard.

You have – at various points in your report you describe the pandemic, despite the challenges that we have been discussing, as the high water or a high water mark for the use of data, the sharing of data. Just, perhaps it’s obvious, but what do you mean by that?

Mr Gavin Freeguard: So again that’s a quote from the foreword – the ministerial foreword to the National Data Strategy. At that point the minister was Oliver Dowden, he is now the Deputy Prime Minister, and he says – he describes it as a high water mark because so many people in government, who perhaps hadn’t understood data or hadn’t taken it seriously, were suddenly very aware of the benefits that using it properly might have in combatting the pandemic, whether that was in monitoring the spread or mitigating the effects.

Counsel Inquiry: And so, perhaps precisely because they were responding to these challenges in a rush – and I think you describe it as mending the plumbing or something?

Mr Gavin Freeguard: Fixing the plumbing, yes.

Counsel Inquiry: Fixing the plumbing. But the imperative to make data work meant that it was used more broadly and understood better, perhaps, than it had been before?

Mr Gavin Freeguard: That’s right. Again, there was a very clear purpose, there was a real urgency to using data and improving some of the systems behind it to fight the pandemic.

Counsel Inquiry: The reason I ask you this is because of a second observation that you make at points in your report, which is that the risk now is, if you like, that the tide is going in, and that some of the advantages and the forward steps that were taken during the pandemic are actually being lost?

Mr Gavin Freeguard: That’s right. Several public servants that I’ve spoken to have expressed their concern that without that urgency, without that problem that needed to be solved instantly, political attention might wander, some of the barriers that were temporarily overcome may return to business as usual, and I’m struck as well by – I note one piece of evidence from Professor Keeling where he looks at monkeypox experience –

Counsel Inquiry: Yes, I was going to ask you about that.

Let’s go, if we can – just the last document – to INQ000056476. As you say, Professor Keeling is an epidemiologist at the University of Warwick who is in fact coming to give evidence later this week. What we’re looking at is a response that he provided at an early stage of our own processes, a questionnaire response, but he does make an interesting observation which we will ask him about when he comes, but let’s ask you about it now.

So if we can go to page 39, please, and perhaps zoom in on the whole of that paragraph numbered 2, so from the – yes. He makes the point – which no doubt we will hear from him and his fellow modellers in due course – that models are only as good as the data that feeds into them and modern models are often data-hungry.

There is then a section of this paragraph which I can summarise as him explaining that the data flow that they received from the NHS during the pandemic was not exactly as they would have wished, they were given some data but not all.

Then importantly for our purposes, there is the passage starting:

“I had hoped that many of these difficulties …”

He says – perhaps I should have said, we don’t need to go back to it, but this questionnaire response was in fact dated September 2022, so a little over a year ago.

He says:

“I had hoped that many of these difficulties would have been resolved for the Monkeypox outbreak [that was going on at the time he wrote the questionnaire response], but if anything, the data access issues are worse. Admittedly SAGE and SPI-M are not directly involved in Monkeypox modelling, but the academic community has still been asked for its help. With Monkeypox, the UK data is only available to UKHSA affiliated staff with a UKHSA laptop, and is again siloed so that the entirety of the datasets are not available to all users.”

Does that perhaps provide a concrete example of the concern that you were expressing that that sort of spirit of the pandemic making data work not only hasn’t been maintained but is perhaps regressing?

Mr Gavin Freeguard: It does, yes. I think it comes back to a point we discussed earlier as well, I think it may be due to caution, clearly, UKHSA wants to ensure that the data is not shared inappropriately, but we can see in this example it means that academics that could bring a huge amount of insight to it appear not to be able to access it.

Counsel Inquiry: Without asking you to solve the entire problems of data in government, what would be your immediate suggestions as to how we can stop these advantages gained over the pandemic being lost?

Mr Gavin Freeguard: I think the meta answer – and it goes back to the very first set of questions – is learning from not just the pandemic but looking at all of those different recommendations that have been made over the last few decades, understanding why those that have stuck have stuck, and why those that keep being made over and over again have not actually been successfully operationalised.

I think – we were talking about the chief data officer earlier; I think having those sort of leadership positions which are able to keep a focus on these sorts of issues and have the power and the levers to make a change, having those positions would be hugely helpful.

We’ve seen already development in the civil service of sort of greater skills in all civil servants – not just those working in the data profession, so senior leaders as well – giving them a sense of data literacy. You don’t need to know exactly how to write these algorithms and understand how to build these pipelines, but you need to have sense of what is being done to do it, and what can work and what can’t.

And I think just ensuring that, you know, these questions about data are built into the design of policies and operational systems from the start to ensure that, you know, what we might need at some point will be available whenever it’s built.

Mr O’Connor: Thank you very much, Mr Freeguard, that’s been very helpful.

My Lady, those are all my questions, and I don’t believe there are any Rule 10 questions.

Questions From the Chair

Lady Hallett: And I listen to the voices in the wilderness, Mr Freeguard. I hope you don’t feel as if you’re like that, because certainly what you’ve said has been extremely interesting, and I’m very sorry to hear that you think that we’re not making the same kind of progress or we’re going back; the opposite, in fact, we’re deteriorating again. I think that would be extraordinary.

Mr Gavin Freeguard: I think it’s worth saying there are definitely lots of initiatives going on at the moment which I think give me more cause for hope, and there are – there is now a generation who takes this really seriously because of what’s happened, but hopefully the work of the Inquiry will stop it sliding any further.

Lady Hallett: We’ll do our very best, but just one last final question, therefore, from me: if a pandemic hit tomorrow – and I have been told in another module that it’s not a question of if, but when – would we revert to the CRIP situation or would we have in place the ability to have the dashboard? In other words, do we have those, that capability still that could be put in place immediately? Because the CRIP thing seemed to me to be a complete waste of everyone’s time.

Mr Gavin Freeguard: I would very much hope the latter. I think, you know, the dashboard team and others who were supporting them did an excellent job, and I think in a sense that has become one of the success stories of the government response. If we can’t learn from that, then I think we’re in real trouble.

Lady Hallett: Thank you very much indeed.

The Witness: Thank you.

(The witness withdrew)

Lady Hallett: Thank you. 10 o’clock tomorrow, please.

(4.17 pm)

(The hearing adjourned until 10 am on Wednesday, 11 October 2023)