Covid-19: My good, bad, and to-be-determined takes
Taking stock of 18 months of opinions & predictions about the pandemic
It’s been more than 18 months since SARS-nCoV-2 started making the news - at first as yet another story from afar to make memes about, and later on as, well, you know. All that.
I’ve followed developments quite closely (and at times too closely), and tried to form and revise opinions on the relevant matters - lockdowns, masks, comms, economic response, vaccines, and so on. So I thought I’d take stock of how I’ve fared in assessing the situation as it evolved - the times I got it wrong, the times I got it right, and the times that, well, it’s still too soon to tell. I also want to try to probe on why I got it right or wrong; for me this is an exercise in evaluating thought processes, not just outcomes (but y’know, outcomes are decent indicators too sometimes…).
Where I think I got it wrong
Jan 2020: Liberal gut reaction to lockdowns and calls for travel suspensions
On 23 January, Wuhan entered a state of general quarantine. A month later, certain parts of Northern Italy went into lockdown - a temporary suspension of civil liberties most of us are familiar with by now, but at the time was completely unheard of for a Western liberal democracy. I remember finding both lockdowns (but especially the Italian one) quite concerning.
The pinnacle of liberal society is individual freedom - I should be free to do what I want, because I know what’s best for me (as long as I don’t infringe of the rights of others, because that interferes with their freedoms). The citizens of a healthy liberal society should thus be extremely skeptical of any measure that attempts to curtain individual freedoms, but also recognise that there exist scenarios where such measures are justified.
When I saw what was happening in Wuhan, I thought I was seeing an extreme overreach by the CCP. When something similar took place in Northern Italy, I thought “ah, shit, the far-right government is using this virus as an excuse to erode civil liberties and test the waters for whatever it’s planning to do next”. Similarly when many were calling for travel suspensions from China, and later Italy, I saw that as the political Right (because that’s who was calling for those measures at the time) attempting to legitimise xenophobic measures under the guise of public health.
Was it wrong to have those gut reactions? I don’t think so - vigilance is good and important, and if anything we need to be more vigilant of attempts to erode civil liberties and democratic institutions. I think the mistake was in having a gut reaction and stopping there, instead of deciding to examine the situation more critically and in greater depth.
Ultimately, it’s a tall ask - there’s so much going on in the world, and between politically-driven media agendas, misinformation, and not always knowing who to trust, we’ll never be able to always be up to date on everything, let alone have critically-examined positions on it all. Nonetheless, I think it’s something we should all aspire towards.
Almost no one ever talks about the great covid realignment, completely forgotten to time pic.twitter.com/xUBjUv0gDm— Ed West (@edwest) October 20, 2020
Feb 2020: Underestimating exponential growth
Until 8 March, I wasn’t particularly worried about Covid-19. A mixture of “eh cases are low” and “it’s not the flu, but it’s not not the flu either”. My feed was full of posts like the below, which was cause for concern, but not grave concern - we were all looking mostly at mortality, not transmissibility:
Then I read this thread on Twitter which simply talked through US case numbers and how they were projected to grow under a relatively static rate of exponential growth. At the time, the thought of millions (in the US alone) of cases, hospitalisations, and deaths was completely alien.
I think most people aren’t aware of the risk of systemic healthcare failure due to #COVID19 because they simply haven’t run the numbers yet. Let’s talk math. 1/n— Liz Specht (@LizSpecht) March 7, 2020
I accepted it as a real possibility while at the same time wishing there was some undeclared or unconsidered assumption Dr Specht was making that was causing her numbers to be so large. In retrospect, they added up pretty well, at least until the end of April milestone:
We’re looking at about 1M US cases by the end of April, 2M by ~May 5, 4M by ~May 11, and so on. Exponentials are hard to grasp, but this is how they go. 4/n— Liz Specht (@LizSpecht) March 7, 2020
Reported cases in the USA on:
30 April: 1,081,105 (vs 1M predicted)
May 5: 1,191,749 (vs 2M predicted)
May 11: 1,358,384 (vs 4M predicted)
Dr Specht’s estimate was assuming no lockdown - but of course by then the USA was in a state of lockdown, at least in many places.
Feb-Apr 2020: Thinking that the public stocking up on/wearing masks was likely to be net negative
This came from thinking Covid-19 was not airborne, and that instead transmission was mostly rooted in stuff like touching surfaces and then touching our face. Under that light, there were a few concerns:
- People would be more likely to touch their mask, and later their face. Prof. David Heymann (WHO advisor, expert epidemiologist who played a leading role in the SARS epidemic response) talked about this in this Chatham House interview, and I took it as gospel
- Reason #1 was reinforced by the worry of masks running out in the short term, and even healthcare worker mask supplies being stolen or supplies redirected
I don’t think the above two points were the ‘mistakes’ per se, though. The real mistake I made in retrospect was not really thinking about the risk of Covid-19 being in fact airborne, and why adopting the precautionary principle was likely to be prudent.
May 2020: Opting for cloth over surgical masks
I can’t remember exactly why I did this. I think it was because cloth masks were reusable? And maybe because I’d told myself a cloth mask with a disposable filter is as good if not better than a surgical mask, but in practice I rarely made use of my disposable filters. Well, that was wrong. Surgical masks are more effective on average, so I now never wear a cloth mask on its own (they add effectiveness in a double-mask setup, though)
Jan 2021: Being skeptical of the UK MHRA approving the Pfizer vaccine before the EMA
When the news that the MHRA had approved the Pfizer mRNA vaccine ahead of the European Medicines Agency coming out with a decision, my immediate reaction was skepticism and mistrust - not enough to think the vaccine wasn’t safe, but enough to think “in the hypothetical scenario where I could get vaccinated today, or wait a few weeks for the EMA to give the go-ahead, I’d wait a few weeks”. I was worried that the UK government, with its really poor track record on lockdowns, messaging, and mask mandates, might have pressured the MHRA. The real error there was trusting that the EU was consistently acting more rationally - as was soon revealed with the EU’s vaccination fiascos in the first few months of 2021.
(PS: good thread on why the EU got it wrong here)
Jun-Jul 2021: Overestimating the ability of the current vaccines to induce herd immunity and limit the risk of mutations
aka underestimating Delta.
It’s looking increasingly likely that even double-vaxxed people are highly likely to contract the Delta variant. This doesn’t mean the vaccine isn’t effective -its primary purpose has always been to limit hospitalisations and deaths- but it does mean that vaccinating your population doesn’t reduce transmissibility (at least not by much). I don’t regret acting on that assumption and e.g. getting a walk-in jab ~1.5 months earlier than I would’ve if I’d waited for appointments to open, and calling for others to do the same, but the (collectivist) logic for doing so is weaker than I’d thought (the ‘selfish’ logic is still the same, more or less).
Where I think I got it right
Late Feb 2020: Coming to grips with the reality of exponential growth (after reading the thread mentioned above), ditching public transport, and working from home 1-2 weeks before the UK finally entered its first lockdown [couldn’t find any tweets, and don’t have the patience to dig through fb comments… but here’s one from late March]
Early March 2020: Calling for in-person university debating tournaments to be cancelled.
Mid March 2020: Calling for a lockdown in the UK, and pointing out the impossibility of a ‘data-driven’ Covid response strategy (you can’t lock down when cases begin rising, because of the ~2 week reporting lag) [ example tweet for the former, couldn’t find something for the latter]
Late March 2020: Arguing against the ‘panic buying’ hypothesis in explaining supermarket shortages, and attributing the phenomenon to our just-in-time supply chains instead.
Sometime in March 2020: Realising that the tradeoff isn’t between “saving the economy” and “saving lives” (which initially I thought was the case) [and also how that’s also true with investing now to prevent a climate disaster]
September/October 2020: Backing a 2nd lockdown to coincide with the school holidays (don’t think I have any public proof). Instead, the UK government waited untill November, a decision which is projected to have cost an additional ~20,000 lives. But at least we got Christmas, right? Oh oops, no not that either.
Throughout the pandemic: Arguing against individuals concerned about their Covid risk blindly deferring to government restrictions to understand what’s safe or not safe to do (see my Jan blog post)
Q1 2021: Anything to do with AstraZeneca and how governments and regulators managed to convince people to avoid it (and how countries like the UK aren’t giving it to under-30s unless if you explicitly go to a walk-in centre and ask for it). Any serious side effect is still so many orders of magnitude less likely to happen than the risks of getting Covid-19 and not having vaccine-induced immunity (also less of a risk than women being on the pill, which I guess our totally-not-patriarchal society has deemed to be a much more ‘acceptable’ risk, right?)
Where I hope I got it right, but think the verdict isn’t out yet
- Jan 2021: Agree with the UK MHRA decision to prioritise 1st jabs by increasing the window between doses
- I think the data we had at the time already supported the maths behind doing this, at least if your #1 metric is preventing deaths - this also points to a recurring theme in this note: running the numbers, even if in crude back-of-the-envelope calculations, can be hugely informative
- Obviously I’m not gonna pretend I knew AstraZeneca would turn out to be marginally more effective after a longer dosage window
- What we still don’t know for sure is that this strategy has overall been more effective.
- For example, did it lead to more infections among the single-jabbed who let their guard down? (We might be able to answer this based on data so far, but I haven’t looked into it)
- More importantly, did it create the right conditions for the emergence of vaccine-resistant variants?
- Jun 2021: Thinking the UK should make it easier for people to get their 2nd jabs early. While it made sense to spread doses among as many people as possible at the start, the UK has more recently seen a slowdown in vaccination rates. If you have spare doses (especially doses that will be going off by the end of the day), why not make it really easy for anyone to get their 2nd jab between weeks 3 and 8? Instead (only the most keen) had to rely on a subreddit, and deal with the fact that the NHS would crack down on vaccination centres that let people get their 2nd jabs early.
- Throughout(?) 2021: Being pro-vaccine/covid passports (ideally proof of vaccination, as a compromise proof of vaccination OR very recent negative test)
- Vaccine passports are legitimate: If you can get jabbed (i.e. no medical complications or lack of availability of doses) and choose not to, then you are opting for being an agent of third-party harms. Getting vaccinated is not just about individual protection (like wearing a seatbelt), it is also about preventing yourself from being in a position of causing harm to others (like drunk driving). Your freedom ends where mine starts - there is no such thing as absolute freedom.
- I think vaccine passports can also work: A lot of unvaccinated people are on the fence, so to speak. If there are real tradeoffs they will be asked to make (no indoor dining, no travel, possibly losing their job), I think most will go for the thing they’d rather not do, but ultimately value less than the costs of not doing so.
Where I hope I’ve got it wrong
- I think the UK entered Stage 4 (no mask mandate, re-opening of nightclubs) too soon, for several reasons:
- Still too many projected deaths given vaccine uptake / overall immunity levels (i.e. factoring natural immunity from surviving a Covid-19 infection)
- Elevated risk of vaccine-resistant strains emerging. The current situation seems to be a perfect environment for vaccine-resistant strains to emerge, because so many (but not enough) people have been vaccinated, and so the selective pressure is for strains that can pierce through the current protection afforded by the mRNA and AZN vaccines.
- The rhetoric of ‘Freedom Day’ triggering a massive drop in covid-safe behaviours (mask wearing, avoiding indoors settings, avoiding large crowds etc.), both before the 19th of July and afterwards. Government headlines have strong signalling effects, both in a “people will take advantage of the fact that it’s now legal” sense, as well as the “people will think that since it’s legal, it’s safe to do”
- On top of the above, I also think the UK entered Stage 4 too unprepared. Why are nightclubs only getting a vaccine passport requirement in October? That’s two months of a majority-unvaccinated population (under-30s) being in what’s one of the best environments for vaccine transmission, but it also means that if a lot of under-30s will only get their jabs when they need their vax passport to clear them for indoors venues, they’ll be clashing with
Looking at the pandemic as a sort of meta-assessment of my ability to make sense of a complex situation rich in uncertainty and a largely unfamiliar domain, I’m reasonably happy with how I’ve fared. There were times, especially early on, when I disregarded the facts because the uncomfortable conclusions they were pointing to, but I was still quicker than most in revising those positions as the evidence drew more clear.
If I were to point to the three most important drivers of getting it right more often than not, and revising positions quickly, it would be: (1) Finding many credible experts that I could go to for updates and critical analyses, but also (2) following people who weren’t subject-matter experts but who have a strong generalist ability to come to grips with new domains, critically examine them, and think holistically (so like… people who are like me/like what I aspire to be, but better at it?). And (3) checking the numbers (relevant for e.g. understanding exponential growth, grocery shortages, rising/declining cases, increasing the window between doses)
Here’s a few people I follow that consistently help me be more up-to-date on the facts, and also help me think about those facts much better (i.e. they do most of the thinking for me)
- Prof Christina Pagel (professor of operational research at UCL; usually my first line of depressing-yet-vital news about new strains or how restriction lifting is going)
- John Burn-Murdoch (data journalist at the FT - everything he writes is so easy to understand, thanks to consistently excellent data visualization and storytelling)
- Richard Horton (editor-in-chief of The Lancet)
- Andy Slavitt (ran Obamacare under Obama, temp Sr Advisor for COVID Response under Biden)
Non-expert generalists/strong thinkers:
- Robert Wiblin (Director of Research at 80,000 Hours)
- Max Roser (development economist and founder of Our World in Data)
- Eliezer Yudkowski (research fellow at MIRI)
- Scott Alexander (the until-recently anonymous author of Slate Star Codex)
- Zeynep Tufekci (UNC & Harvard sociologist)
One last thought
The problem with facts is that they indeed do not care about your feelings - and there were times when I spent too much of my time and energy going through the facts, much to the detriment of my mental health. In retrospect, I shouldn’t have been checking case numbers, Twitter, or government press briefings as often as I was (which is why I toned it down after… May? June? something like that). It was a bit like Brexit in that sense: At some point, staying up to date is just depressing and frustrating, and learning about the latest government fuckup offers very little marginal benefit (since you cannot act on that new info anyway).