John Burn-Murdoch Profile picture
Stories, stats & scatterplots for @FinancialTimes | Mainly Covid for now | Visiting senior fellow @LSEdataScience | john.burn-murdoch@ft.com | #dataviz
BrakTalk AlanLB 🦋 ⚖ 🦋 Profile picture Chris🌟 😷 #TakeBackBritain #RejoinEU #FBPE💙 Profile picture DLlewellyn Profile picture Kevin Coates Profile picture 104 added to My Authors
30 Nov
NEW: we know cases rising fast in South Africa, but what about severe outcomes?

I spent yesterday pulling together hospitalisations for Gauteng province, so we can compare the fledging Omicron wave to those that preceded it.

So far admissions following ~same path as past waves.
Quick notes (1/2):
• For cases to be rising faster while admissions are on same pace may hint at a lower proportion of severe disease
• But this may also simply be the result of more infection-acquired immunity and vax
• Or of cases so far being predominantly younger people
And 2/2:
• Share of patients in ICU currently much lower than same stage of Delta wave, but may change if cases spread from being mainly young. Or may not, which would suggest T and B cells kicking in 🤞
• Please follow people on the ground incl @Tuliodna @tomtom_m @rjlessells
Read 6 tweets
25 Nov
Five quick tweets on the new variant B.1.1.529

Caveat first: data here is *very* preliminary, so everything could change. Nonetheless, better safe than sorry.

1) Based on the data we have, this variant is out-competing others *far* faster than Beta and even Delta did 🚩🚩
2) This is coinciding with a wider rise in cases in South Africa.

Again, currently we’re talking about small numbers (both of B.1.1.529 and of cases in SA overall), but there’s a clear upward trend. This may be a blip, but this is how waves start.
3) Aside from its rapid rise in South Africa, the other key reason for concern here is that the specific mutations that B.1.1.529 has are — based on what we’ve seen with other variants — known to be associated with greater transmissibility & immune evasion
Read 9 tweets
24 Nov
NEW: detailed thread on Europe’s winter wave and the contrast vs UK

What’s happening? Why the difference? Can boosters help?

First, the wave itself: cases, hospitalisations & deaths surging in Europe, several western countries shooting past UK 📈

Story: ft.com/content/974487…
The situation is even clearer when plotted on a log scale:

UK is broadly a flat line, with European countries cutting up steeply through it. France, Italy & Spain all on course to pass UK for cases. Germany now above UK for daily deaths and Netherlands set to follow.
So why these exponential surges across Europe but not in the UK?

There’s increasingly little difference in social mixing behaviour between the countries, and where we do see differences e.g in mask-wearing, they’re generally more virus-friendly in the UK 🤔
Read 24 tweets
18 Nov
The good thing about this story is that any time I see someone sharing it as evidence that wearing masks cuts Covid incidence by 53%, I know they haven’t read the paper it’s based on and aren’t fussed about the quality of evidence as long as it produces the right number 🙃
What do I mean by this?

Let’s start by saying that yes, we have good evidence that masks reduce Covid incidence 😷🦠📉😀

BUT it points to nowhere near a 53% reduction. As today’s @bmj_latest states, the best evidence — randomised controlled trials — point to more like a 10% cut
Indeed, here is the BMJ’s take on today’s paper:
• Most of the studies included in the meta-analysis from the Guardian story are poor quality and subject to major biases
• We need more & better research [like the mask-wearing RCTs which have consistently found a smaller effect]
Read 6 tweets
15 Nov
NEW: @UKHSA study finds Pfizer booster is extremely effective against symptomatic infection, both compared to the unvaccinated and to those with 2 doses ft.com/content/8330da…

Whether first 2 doses were AZ or Pfizer, a Pfizer booster sends vaccine efficacy up to 93-94% 💪 Image
Study was on people aged 50+, comparing those boosted ~5+ months after dose 2, to those @ 5+ months unboosted.

AZ efficacy was 61% after dose 2, waning to 44% @ 5 months.

Pfizer was 82% after dose 2, waning to 63% @ 5 months.

2 wks after Pfizer booster, both groups -> 93-94%!
Best way to think about booster impact is not to look at going from 44 to 93 with AZ, i.e roughly doubling, but invert the numbers and go from (100-44) to (100-93), i.e from relative risk vs unvaxxed of 56% to just 7%

That’s an 87% increase in protection *relative to two doses!* Image
Read 8 tweets
11 Nov
NEW: England has recorded 18 successive days of week-on-week declines in cases, its longest sequence of declines since February, suggesting its autumn/winter wave may have peaked ft.com/content/e11add…
Crucially, hospital admissions, patient numbers and deaths are now also trending downwards, as the fall in case numbers has shifted from being youth-driven into all age groups.

These acute indicators look to be topping out at 10-20% of last winter’s peak levels.
A key factor here has been England’s booster rollout.

Antibody levels in the oldest groups (vaccinated the earliest) had been slowly eroding as the months passed, but in the last 5 weeks they have shot back up as third doses have gone into arms 💉💉💉💪💪💪
Read 18 tweets
5 Nov
NEW: Covid cases, hospitalisations & deaths on the rise again across Europe, with rates of all three metrics surpassing the UK in many countries

Starting in the west: Belgium, Netherlands & Germany in particular experiencing sharp increases in not only cases but ICU & deaths too
And the picture worsens as we move further east.

In central Europe, cases were a fraction of UK levels over summer, but have now rocketed past, with ICU occupancy and deaths also climbing fast.

Vaccine coverage is generally lower here than in western Europe.
In eastern Europe, the situation is dire.

Romania, Bulgaria & Latvia all set new records for daily deaths in recent weeks, and deaths are 10x current UK levels.

In much of western Europe it can feel like the pandemic is an echo of its past self. Try telling that to the east
Read 8 tweets
3 Nov
🚨all of ft.com is free to read today, so if you like the glimpses you see on Twitter but want to go deeper, now's your chance!

Beyond Covid, read our best work on climate, politics, inequality & culture. Apparently there's even stuff on economics & finance! 🚨
Throughout the day I'll be sharing links to some of the pieces I've enjoyed most this year, starting with this brilliantly clear column by @TimHarford demonstrating how a carbon tax would be a game-changer in the battle to curb emissions on.ft.com/3EqFsZ1
Another great piece to read 🔓#FTfreetoday:

Off the back of Squid Game and Parasite, @crsdavies took a deep dive into the huge success story that is South Korea’s entertainment export industry, and the decades of strategising that led to this point ft.com/content/7c3c53…
Read 5 tweets
2 Nov
NEW: as of the end of September, Russia has recorded 753,583 excess deaths over the course of the pandemic.

This is almost 4x its official reported Covid death toll, and is second only to the US in absolute terms among countries with reliable data ft.com/content/f1a270…
• Russia’s vax rates remain stubbornly low, only 33% fully-vaxxed amid widespread mistrust
• Society largely carrying on as normal despite 2k daily excess & 1k daily reported deaths

Full details in the story from @polinaivanovva, me & @mroliverbarnes ft.com/content/f1a270…
We’ve also updated our excess deaths data for all countries: ft.com/content/a2901c…
Read 6 tweets
1 Nov
One of the most startling Covid charts I’ve made in a long time, on vaccine inequality:

Rich countries have given out more booster shots in the last 3 months, than poor countries have given out total doses all year.

@donatopmancini’s story: ft.com/content/35a3d4…
In rich countries, just 18% of people aged 12+ are yet to have any dose, and 5% have now had a third dose.

In poor countries, 94% are yet to have a single dose.
The result?

Stark inequality in outcomes.

In countries with ~70% fully vaccinated, about one in 250 cases results in death.

At 50-60% fully vaxxed, around one-in-75 results in death.

At 10-15% fully vaxxed, it’s one-in-25.

And in parts of Africa, as high as one-in-15.
Read 5 tweets
17 Oct
NEW: there’s been a lot of chatter about why cases, hospitalisations and deaths are much higher in the UK than elsewhere in Western Europe.

I think a lot of the commentary has been overly simplistic, politicised and at-times flat-out wrong.

Let’s see if we can do better:
Yesterday we published a story comparing the situation in the UK vs a selection of Western Europe peers: ft.com/content/345825…

Here are the top-line stats:
• Cases among older people are 7x higher in UK
• Hospital admissions are 6x higher
• Deaths are 3x higher

Not great! Image
Why so much higher in UK?

Here’s a look at what those countries are doing differently to reduce transmission:
• % of people never wearing masks has rocketed in UK but stayed very low elsewhere
• % of people attending large gatherings in UK is surging way ahead of elsewhere Image
Read 25 tweets
14 Oct
Folks, if you see anyone quoting Alex Berenson / Joe Rogan on how case rates in England are higher among vaxxed than unvaxxed, that data is incorrect, as explained here.

The table they shared is real, but it's wrong.
Also, FFS this is why UKHSA's persistence in publishing knowingly misleading data is such an utter disaster.

You now have several million people confident in their belief that vaccines don't work, and waving around the official UK documents that prove it.
And for those asking "at least they didn’t use the specific @UKHSA report, right? ...right?"

They did
Read 5 tweets
11 Oct
NEW: people obsess over vaccine uptake stats, eagerly comparing one country to others to see which has jabbed the highest share of its population, but what if I told you many — perhaps most — of those stats are wrong?

Time for a thread on bad Covid data and how it can cost lives
Let’s start with the most obvious sign of the problem:

In several European countries, the share of elderly people who have been vaccinated exceeds 100%.

To state the obvious, this is not possible.

But it’s not just a funny quirk, it has big implications.
In reality these countries won’t have reached 100%, so there are tens of thousands of unvaxxed elderly that are invisible, and no-one is going to visit someone to make sure they’re jabbed if they think everyone is jabbed.
Read 25 tweets
2 Oct
Really interesting story from @mjruehl in Singapore (82% of population fully vaxxed), where the transition from Zero Covid to living with the virus is not going smoothly to say the least. on.ft.com/3maIJnW

Lots of lessons here for Australia, New Zealand and others
As @mjruehl writes, Singapore's problem stems from the interplay of two factors:
• A population that has become extremely nervous and worried about *any* Covid
• A policy of mass testing asymptomatic people, many of whom rush to hospital after a positive, even if symptom-free
The local reaction has been, as ever, polarised. Healthcare professionals have criticised the mass testing policy [given the context in which it's being rolled out], while 25% of the population want a return to lockdown, and there are petitions to reintroduce quarantine hotels.
Read 7 tweets
15 Sep
NEW: lots of news recently on waning immunity against infection, but a study has now landed from Public Health England on how vaccines are faring against *severe disease & death*

This chart summarises key findings, but the paper is a real goldmine, so let’s dig into more detail:
First up, fresh data on protection against symptomatic infection. Key results:
• We knew protection started out lower among older groups. Now we know they also see the most waning
• Waning much more muted (if happening at all) among under-65s
• Moderna > Pfizer > AstraZeneca
Sticking with symptomatic infection, they also looked at a hot topic: the impact of the interval between first and second dose

Very short intervals (3 weeks, e.g in US) produce lower protection than longer intervals (e.g UK & Canada), though note overlapping confidence intervals
Read 23 tweets
23 Aug
NEW: in the last couple of weeks there have a *lot* of new studies out assessing vaccine efficacy, many of which have touched on the question of waning immunity.

Unsurprisingly, these have prompted a *lot* of questions.

Time for a thread to summarise what we do and don’t know:
Let’s start with last week’s Oxford paper, the most significant study to date on waning immunity to Covid.

The researchers found signs that vaccine efficacy against symptomatic infection erodes over time, and that waning may occur faster in some vaccines than others.
Unsurprisingly, this prompted a lot of questions.

Could it just be that the first people to be vaccinated were older, and perhaps more vulnerable to waning?

Nope. The study controlled for age.
Read 33 tweets
30 Jul
NEW: lots of attention on ONS Infection Survey today, but some confusion over how it should (and should not) be used to asses whether England’s fall in cases is "real"

Quick thread:

Most attention has gone on ONS “% of people testing positive” metric showing a continued rise
But "testing positive" is a lagging indicator of cases. It estimates how many *have* Covid today, not how many are *catching it* today.

Fortunately, ONS has re-introduced its incidence data (blue line), which is a much better yardstick for cases, though always 2 weeks old 😩.
So how to resolve issue of one lagging indicator, and one that’s 2 wks old?

Look to Scotland, where cases peaked 2 wks before England, so ONS indicators have had time to catch up

Turns out ONS incidence fell at exactly same time as cases 🙂. ONS positivity likewise, just lagged
Read 10 tweets
27 Jul
There’s a wild story about the women’s gymnastics at the Sydney Olympics in 2000, which I think is very relevant to what we’re hearing about Simone Biles, and the wider point of how the top level of elite sport is just as much mental as it is physical.
In the women’s all-round final in 2000, the organisers set the vault at the wrong height. Two inches too low. This was a pretty huge deal.

For competitors who have done thousands, maybe tens of thousands of vaults at a specific height, a two inch difference is night and day.
In the first round, 17 of 36 finalists fumbled the vault

One landed on her back. Clear gold-medal favourite, Russia’s Svetlana Khorkina (comfortably won qualifying) landed on her knees.

Total chaos, and nobody knew why. Athletes second-guessing themselves.
Read 13 tweets
27 Jul
Correcting an important misconception (this is my chart, but misleading commentary):
•There were thousands more cases among young men than women after ⚽️ matches, showing impact of Euros on transmission
•But not due to attending matches. It was indoor gatherings to watch games
Of course, that still means the transmission bump was driven by the football, but match attendance is only a small part of the cause. The bulk is mixing in pubs, bars, homes etc, plus some from crowded transport to and from those indoor gatherings (and matches).
Thoughts on implications:
• I would guess these watch parties happen at a much larger scale for England games at major tournaments than they do for typical club games, but we've not had pubs fully open during the season yet so that will be worth monitoring.
Read 5 tweets
25 Jul
I feel like I've seen this before somewhere 🧐🤔
Lol that they couldn't even be bothered to change a single one of the numbers.
Read 4 tweets
24 Jul
NEW: people worry when they hear "40% of hospitalisations are fully vaxxed", but this chart shows that's actually good news.

The more people you vaccinate, the higher their share of hospitalisations, but the *total* number in hospital is a fraction of what it would otherwise be
If fewer people are fully vaccinated, a smaller share of hospitalisations will be fully-vaxxed too, but this is not a good thing:

Overall there will be a lot more people in hospital because far more of the population is unprotected.
In other words: if you want to know whether the vaccination program is working, don't focus on whether the fully vaxxed make up 40% or 12% of hospitalisations.

Focus on whether the hospitalisation rate is 270 per million or 684 per million.
Read 10 tweets