50 Comments

"...a much more reasonable question is why the Prime Minister (and the COVID response as a whole) was geared for so long to the idea of five-day doubling when it’s likely that the figure was closer to three. That to me is the killer difference." Indeed. It's worth remembering what John Edmunds said in his notorious March 13 interview on Channel 4 News when he was up against Tomas Pueyo.

Pueyo: It really depends on what’s happening at every moment, right now this thing is going really really fast in the UK, right? We have 800 cases, it’s growing at 33 percent every day, that means in three days you get 1,900, so it’s more than doubling in the next three days, and so what’s happening is early on when this thing is really catching up cases explode and they grow at say 2x, like they double every two days, and this is the situation in which the UK is today, it is the situation in which Italy was a week ago, it is the situation in which Spain is right now

Presenter: OK, all right, OK. John why are you shaking your head at those numbers?

Edmunds: It’s true if you just look crudely at the numbers that the number of cases are doubling about every two and a half days, but that’s because they’re doing more contact tracing, the actual underlying rate of doubling is more like about every five days.

We now know from the inquiry that Edmunds said in a e-mail *the previous day* that "the data are crap".

It would be interesting to hear what Edmunds makes of all this now but of course he wasn't asked about any of this by the toadying KC.

Expand full comment
author

The SPI-M-O statement from February 3rd had doubling time in Hubei as 3-5 days https://assets.publishing.service.gov.uk/media/5eccf270d3bf7f4601e5765e/15-spi-m-o-consensus-statement-03022020.pdf I think it's a good example of where China not being more transparent with the data didn't help, but I think that knowing that was the range there should have at least anchored the thinking in terms of worst-case scenarios.

Expand full comment

It was bizarre that the Pueyo exchange wasn't mentioned. Pueyo's take... which was based on a numerate overview of a very deep trawl of all the evidence out there... turned out to be remarkably accurate and prescient. His alarmed and proactive response also turned out to be entirely proportionate. Yet he was widely reviled as a non-expert trying to cause alarm. Edmunds' quibbles over every stat allowed him to miss the big picture, and lack imagination about what might be done about it, and this was the root of the problem in government as a whole. Also the refusal to show any alarm was a general issue, and lead to what amounted to deceiving both the public and their elected representative about what was coming, long after it was all but certain. Are there no ethical/practical issues raised by lying to an educated population about an impending disaster? Honestly telling them in a timely manner that a good chunk of them would die soon, if not of coronavirus then of another problem that couldn't be treated in collapsed health service, might have been another way of avoiding lockdown.

Expand full comment

I’m struck by how obfuscatory the language of that little meeting excerpt is compared to Cummings’s email. “There were differing views about the likelihood of..” You mean, you couldn’t agree? Just say that. I have often thought there’s a sort of special language - Bureaucratese - that functions as a gatekeeper to all this stuff, a Teflon for responsibility.

Compared to which Cummings just tells people “this, that, NOW.” It may be my many years in journalism but I find the latter more useful.

Expand full comment
author

Yes, I think it's maybe a consequence of scientists who are used to writing for each other in an understated way suddenly seeing their words getting read by the general public. Not sure you'd get away with underselling on the scale of "We wish to suggest a structure for the salt of deoxyribose nucleic acid (D.N.A.). This structure has novel features which are of considerable biological interest" these days though!

Expand full comment
Nov 2, 2023Liked by Oliver Johnson

I think it goes deeper. I contribute to a NICE study group. The medics are accustomed to communicating in risks and uncertainties. For example they know that a medical test is a compromise of specificity and sensitivity, so a test is never fully accurate. A similar point can be made regarding statistics - the data are always crap aka imperfect, it's how you understand and interpret them that matters. Politicians and the press prefer certainty and consequences; if a test isn't positive it must be negative (wrong); a risk of 0.1% is negligible (wrong, eg if you are a frequent flyer); you're not much use as an expert if you can't give me a definitive answer.

I really hope the enquiry digs fully into whether the skills and competencies of the key decision-makers were a match for the task, without necessarily seeking to attribute blame, else we won't learn....

Expand full comment

I have two -possibly contradictory - positions on one aspect of this; one is that, contra ISAGE, this wasn't about scientific decision making, it was about public policy formation based on a number of considerations, one of which was the emergent science adapting to real-world experience of the disease. The second is that the daily news conferences featured questions by lobby correspondents, not correspondents with a good knowledge of public health issues, and I remember the panic on Matt Hancock's face when a question came from New Scientist. Consequently, I see the 'science' being both over- and under- stated.

On the whole, I thought Helen Macnamara's evidence today was a lot more interesting than that of Mr Cummings yesterday.

Expand full comment
author

I haven't really caught up with what she said today (been in meetings a lot of the time) - though I hear there was some interesting stuff on PPE for example, so will try to l look into it later.

Expand full comment

I think DC's superpower is being able to drive a certain class of British person absolutely mad. He did it with the AV referendum and with the Taxpayers' Alliance and again with Brexit, and then throughout his time at No 10.

A while back but post-Brexit referendum I was at an open day for the King's College maths school in London that DC was instrumental in setting up, and at some point when we were traipsing around the building I conversationally mentioned DC's involvement to my son. The parents around me - who looked like good, well-intentioned Guardian readers to a man and woman - literally ^jumped^ and turned around and stared at me. I mean... yes? This is an example of DC analysing a need and putting himself in the right position to do something about it? Whether you agree with him or not have you not bloody noticed that this is what he does? That school gets the best A level results in the country. (From a highly unusual pupil base and with a lot of extra resources, admittedly.)

The most annoying thing is that these people - with whom I tend to agree, politically - never seem to learn that while they're pretending to be horrified by his use of the 'C' word (which I can guarantee most of them use themselves, probably about Cummings) he's actually DOING THINGS. It's his famous OODA loop thing innit. Bloody wish he was on my side.

Expand full comment
author

That's interesting - and I think there's probably a similar "political" reaction to Katherine Birbalsingh at Michaela. In fact, as a maths person, one of the lasting legacies of Cummings' time in Government was meant to be extra funding for maths research (and easier ways of accessing it), though I think the devil's turned out to be in the detail.

Expand full comment

Yeah, very similar to Birbalsingh. And Gove tbh, interesting that they all got something right about education. I wouldn’t want to go on a caravanning holiday with any of them, but I’m a bit suspicious of fellow lefties who wouldn’t want to go down the pub with them for an evening. (Or to a rave club, in Gove’s case.) How can you not be a little bit interested in how they’ve achieved the things they have, even if you don’t approve?

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

Cummings has now endorsed your post:

"this from @BristOliver (Director of Statistics Institute at Bristol) is almost the only thing I've seen that understands yesterday & the critical points - the point isn't my bad language, obviously SW1 much prefers to focus on that than Establishment fiascoes that I was describing with my bad language - & the MSM coverage almost totally misses almost every important point - 100% on brand... much more interested in swearing/psychodrama than WHY the core of the state melted down, WHY these problems were so hard to fix, and WHY they remained to then fuck up Ukraine, and now new crises... the dysfunction I describe remains, the Cabinet Office remains a dumpster fire today on counter-terrorism and much else... "

Expand full comment
author

Yes, I saw .. it's been a bit mad here today as a result!

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

Perhaps a public enquiry isn't the right context, but I would like to see much more investigation into which NPIs work(ed), and if they didn't, why not. I'd also like to see more comparisons with other countries -why, for example, has Sweden ended up with relatively lower excess deaths over the last 3 and a half years than almost any other European country ? Did other countries' health systems respond in similar ways to the pandemic as the NHS ? How do the rates of hospital-acquired Covid vary between hospitals and between countries, and why ? Have other countries seen similar post-pandemic problems (big increases in waiting lists, more mental health issues, etc. ) ?

Expand full comment

Exactly. These are the types of questions that should be being asked at the inquiry, not assuming we did everything correctly just not soon or hard enough.

Expand full comment
author

Yes, it's frustrating that we still don't know a lot of this stuff. I think inter-country comparisons are still tricky (because of different reporting standards etc), but we could have easily randomised some schools for different return dates, mask policies etc, rather than just guessing still.

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

imho it seems that the whole inquiry system fails. Obviously I'm biased but an investigation led by legal and lawyers is always going to lead to this sort of politically dramatic process rather than something akin to an air crash investigation

Expand full comment
author

I don't know the mechanics of air crash investigations, but I have heard them mentioned as a possible alternative model. There's some interesting stuff in the Chris Hadfield "Astronaut's Guide to Life on Earth" book about how astronauts have to report and learn from problems, which we could probably learn from too.

Expand full comment

thanks for the recommendation I will read that.

Expand full comment

RE: the doubling times discussion. Definitely some systematic failures. The most interesting dicussion, which is from those who were there at the time, is probably the "In Context" section of this paper.

https://royalsocietypublishing.org/doi/full/10.1098/rstb.2020.0264#d1e1618

Expand full comment
author

Gosh, that's interesting, I hadn't seen that. I remember hearing the "5 day" figure more or less live at the time, and thinking it didn't seem right .. https://twitter.com/BristOliver/status/1239597688167051266 But I appreciate it was a fast-moving situation

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

" a much more reasonable question is why the Prime Minister (and the COVID response as a whole) was geared for so long to the idea of five-day doubling when it’s likely that the figure was closer to three"

I very much agree! There is some interesting discussion about that in Christina Pagel's and Kit Yates' BMJ article on the role of modelling:

https://www.bmj.com/content/378/bmj-2022-070615

"The initial lack of data sharing could have contributed to mistakes made early on in the pandemic. Groups with access to poorer quality data did not feel able to challenge the conclusions of groups with access to better quality data, leading to poor modelling outcomes. In March 2020, for example, SPI-M overestimated the doubling time of the UK epidemic. Although some of the modelling groups were generating more accurate values, their estimates did not find their way to SAGE. Instead, an overestimated doubling time of 5-7 days appears in the SAGE minutes of 18 March.."

I'd add that the Very Simple Model (probably available in a @BristOliver log-scale tweet back then way before I followed him?) did suggest a doubling time of <=3 days back then, and it should have needed very strong evidence from more complicated models to discard this simple data.

Expand full comment
author

Yes, such a tweet does exist - more or less in real time as the press conference was going on! https://twitter.com/BristOliver/status/1239597688167051266

(I think there's a reasonable question about what the "true" growth rate was, given that testing was indeed going up at the time as well, but I don't think something in the vicinity of 3 days should have been off the table when thinking about plausible worst case scenarios)

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

One of the interesting parts of this is the memories people have of that time, and the certainty of their views. I remember a Radio 4 programme about the emerging pandemic in February, annoyingly I can't find it at the moment, where independent epidemiologists were interviewed and still trying to downplay it as a future risk.

Expand full comment
author

I think we've all blotted a lot of it from our minds, particularly in terms of how fast it all moved. I was on a residential course with people from different parts of the university on 4th and 5th March, and while it was clear that a few people were in a place where they were starting to think seriously about it all, for a lot of us it all felt pretty remote at that stage still.

Expand full comment

We were due to be travelling to the French Alps for a skiing vacation starting 12th March 2020. For the two weeks before my wife and I had been following the situation in Italy and the emerging cases in France and the U.K. We’re both scientists, I’m an immunologist, and my wife a cell biologist and medical science writer. We did a very simple crude estimate to inform our decision making, we looked at the reported case trajectories in Northern Italy and projected on to that the early reports from France and the U.K. We concluded that France would have a problem coinciding with our vacation and that the U.K. would follow within one to two weeks. We took the decision to cancel our vacation and because there were no French or British official Government warnings at the time we lost our money. Our decision was vindicated when the French imposed restrictions on the 14th March which would effectively have confined us to our apartment from the start of our vacation!

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

Great post. Leaving aside the fact that the main purpose - or should I say effect - of public inquiries is the transfer of vast sums of public money from taxpayers to lawyers via public funding of the core participants, the focus on trivia and swearing is very depressing.

You can see the cultural desperation from Hugo Keith to do that innocuous lead the witness down the path to the trvial 'gotcha' moment. Don't they want to actually learn?

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

The cross examination of Cummings covered in a somewhat condemnatory fashion the style of what was said & somewhat less on the content. The focus on style rather than content is very typically English. The was reality that Cummings & Co were facing a pandemic that could kill 100s of thousands with a PM unfit to be a PM, a private office ditto and cabinet office ditto. The impression gained was that this mattered somewhat less than the fact that Cummings used the work pigfuck. If you are wondering why UK Inc is going down the plug 'ole then prima facie evidence was there for all to see in the focus of the cross examination.

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

Really excellent blog - thanks for writing it.

This line in particular landed with me: "Instead of treating COVID like an episode of Love Island: Aftersun, this could have been a chance to learn from the kinds of systems failures that Cummings loves to discuss".

The headlines and stories you see are all about playground name calling in WhatsApp groups.

The real story is an exposure of systematic failures in government, and nothing has changed since they were exposed.

Expand full comment
author

Thanks, glad you enjoyed it!

Expand full comment
Nov 1, 2023Liked by Oliver Johnson

To a large extent, I agree with your frustration. I think some of the framing might be the way in which the Inquiry is broken into sections, with the current one being about how the Government dealt with the crisis (Core UK Decision-making and Political Governance (Module 2) – Public Hearings). Hopefully, more detail relevant to systems failures will come later (although I do think the political stuff is very important too since the impact of governance/government in a crisis can be critical.)

Expand full comment
author

That's true, I haven't looked too far into what the other modules are (apart from the fact that I saw that schools aren't due until Module 5 or something), so maybe some of this stuff will be covered then like you say.

Expand full comment
Nov 2, 2023Liked by Oliver Johnson

When everyone has always known that just about anything typed into a computer is as durable as it is possible to be and is likely to be made public, if anyone shows a flicker of interest in it, what makes people with responsibility type such dim things into email, WhatsApp etc. and with such haste, meanness or poor taste?

Expand full comment

I agree that the Covid inquiry is spending far too much time on the early days of the pandemic when we were all pretty much in the dark. It really needs to address how "many of the unforced errors in our COVID response came later in autumn and winter". The winter wave saw a larger death toll than the spring, but these issues have hardly been discussed "for lack of time". This is particularly frustrating for me because I want to understand why and how, after spending 2020 being extremely careful, as January 2021 started I was struck down and spent eight weeks in induced coma and six months in hospital overall. I am still recovering! Something must have gone very wrong but we haven't heard hide nor hair about it yet!

Expand full comment
author

That sounds pretty grim, hope that the recovery is working out for you now. Alpha was a pretty terrible time, and particularly ironic that it hit just when the vaccine was becoming available - the variant arriving a month or two later could have made a huge difference to the outcomes I think.

Expand full comment

Yes, the Pfizer-BioNTech vaccine was available from early December 2020 but the roll out was painfully slow. The first people to get it were over 80 or working in the health service with infected patients. I would like to know more about how and why these decisions were being made - especially now we are told that Johnson wanted the old to be allowed to die and younger people to be protected.

Expand full comment