THREAD on cases, hospital admissions and why so many scientists & NHS leaders are worried.
Case study of London - and what is behind the alarm!
1/10
The key bit is that it takes about 10-14 days from infection to needing hospital. And if you have symptoms, you'll probably test positive 4-7 days into infection.
So there's roughly a week from testing positive to becoming a hospital admission. 2/10
So - cases in London have risen *very steeply* - but *mainly* in the last week. And only in the most recent week has Omicron been dominant.
But cases to 19 Dec (incomplete!) are already more than double previous week. 3/10
If we look at Covid hospital admissions to London up to 19 Dec we see a more modest rise - but that's largely because they do *not* reflect last week's high numbers - it's too soon! 4/10
If we line up admissions under cases a week earlier we can see they've been very consistently about 2% of confirmed cases. From now on, most admissions will be Omicron.
*If* it's not any milder, then that 2% ratio should keep holding. 5/10
Now - in fact most cases in week to 19 Dec have been younger adults. So let's reduce hospitalisations a bit to 1.5% (note, might still be too high). But even that still leads to a large jump *next week*. 6/10
And there'll be much more intergenerational mixing over Christmas. What happens to cases then? Esp with Omicron now dominant?
And admissions the first week of January? Will they double again? more? 7/10
If weekly admissions double just *twice* more from w/e 19 Dec (1.5K), we are back at last January peak in London (6K). We're likely to get a chunk to 1st double next week.
What is scaring people is that with Christmas mixing there is no real reason to think they won't go⬆️ 8/10
Omicron would have to be a *lot* milder to change this. Older adults more likely to need hospital have been boosted for a while - their boosters won't change hospitalisations over next few weeks.
Cases in London already going up in older adults. 9/10
So we've seen cases sky rocket in London this week. We know that they are likely to spread to older adults more over Christmas. Admissions will follow.
Hospitals already struggling with staff who have covid & rest are exhausted.
NHS is sounding the alarm for good reason. 10/10
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Short THREAD on log scales and why I tend not to use them.
I've been criticised at times for not plotting case numbers etc on a log scale. I want to explain why.
TLDR: essentially it's because the burden of covid ill health and on NHS is people not log people. 1/8
When you've got exponential growth (e.g. in new covid cases or hospital admissions, it will be a straight line if you put the vertical axis on a "log scale" (normally equally spaced powers of 10).
Charts show same exponential growth on a normal scale and a log scale. 2/8
Log scales are really useful for e.g:
a) seeing if growth is exponential (straight line?)
b) looking for acceleration or slow down in growth
c) comparing growth between countries at different stages of epidemic 3/8
THREAD: on omicron, UK cases, London & what to do next...
Last post-briefing tweet thread of the year! 1/18
First Omicron... as of 11 December, Omicron was most common in Scotland and England but starting its growth in Wales & N Ireland. With its growth speed, shares of cases will now be much higher in all regions. 2/18
In England, UKHSA "S gene dropout" data shows it was 40% of cases by 13 DEcember. It will be dominant in England by now.
WHO first designated it a variant of concern & named it Omicron 3 weeks ago today. Crazy. 3/18
SHORT THREAD: Some basic musings about hospital admissions and exponential growth.
88K cases reported today - let's assume just under half are Omicron - so ~40K. Delta hospital admissions have been running at about 1.7% of reported cases in last month. 1/5
Let's assume Omicron was half as likely to cause hospital admission as Delta. So 0.8% of reported cases end up in hospital. From 40K cases today that would result 340 admissions.
just FOUR doublings takes you to 5,440 admissions - far higher than last January peak. 2/5
The 40K cases today represent people infected about a week ago. Omicron has been doubling every 2 days. Even if that's slowed this week to 3 days, that's likely 2 more doubling baked into admissions. That means that FROM NOW we there'd be two more doublings exceeding jan peak 3/5
Four charts about why boosters can't do everything on their own.
1. There is a massive difference in booster uptake between the most and least deprived areas. And also a massive difference in those entirely unvaccinated.
What is plan to reach them? 1/4
2. There are big regional variations. London has the highest unvaccinated and least boosted population by far. It is also where epicentre of Omicron currently sits. There are far too many vulnerable people left in London. 2/4
3. There are big age differences that will take time to smooth out - even with our current booster acceleration.
In London (below), far too many teens remain unvaccinated and far too few over 50s boosted. 3/4
Quick thread on some omicron thoughts with @BBCNews today.
1. At population level, Omicron's sheer growth advantage is likely to outweigh any reduction in severity from existing immunity, at least in short term. More covid patient to NHS -> less NHS for everything else.
2. We can do more and we should do more - but how much more is the really hard question.
3. But *when* govt might do something is much harder to answer because Christmas is round the corner. If this were any other time, I suspect we'd there would be more public health measures in place already.