Tweet

Kit Yates

29 Jan, 10 tweets, 5 min read

@chrischirp

Something strange is going on with the positivity data and I'm not sure what to make of it.
Might be of important for anyone who's following the data.
Perhaps you have a better idea about what's going on than me
@chrischirp @jburnmurdoch @COVID19actuary @RP131 @fascinatorfun

Case numbers appear to be coming down quickly (almost linearly over the last few weeks - UK).
Nothing too unusual in that, although it's always good to check the positivity rates in case this is caused by a decrease in testing.
Turns out testing hasn't decreased much recently.

Looking at English regional positivity rates on the COVID dashboard:
coronavirus.data.gov.uk
Paints a positive picture of decline in almost all regions (with the exception of the North West).

The first big discrepancy comes from looking at the positivity data from PHE (assets.publishing.service.gov.uk/government/upl…) we see a steeper decline in all of the regions in comparison to the dashboard data.

This is perhaps even clearer when we look across the whole of England and compare the positivity rate data on the same graph.
PHE clearly shows positivity coming down faster.

Adding the ONS infection survey into the mix paints a different picture still.
Although these two "positivity rates" are not directly comparable, you would expect them to do something broadly similar all other things being equal...
ons.gov.uk/peoplepopulati…

But ONS has shown relatively little change in positivity over the last week and even since the end of December the fall hasn't been huge (even in the raw data - forget the models).

REACT results released yesterday are now also starting to show a slightly steeper decline: spiral.imperial.ac.uk/bitstream/1004…
But none of these sources of data seem to paint a coherent picture.

Understanding how cases are changing is clearly important for assessing the impact of this lockdown and will inform how we start to release restrictions.
I don't know what to make of it all. So the question I am asking is, what the hell is going on?
Any insights much appreciated.

Thanks to Bob Hawkins for preparing the dashboard and PHE positivity comparisons.

• • •

Missing some Tweet in this thread? You can try to force a refresh

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

More from @Kit_Yates_Maths

Kit Yates

@Kit_Yates_Maths

29 Jan

@sky_news

Short clips from my @sky_news interview earlier.
Firstly, on R (0.7-1.1). This wide range masks a more mixed picture regionally, but even at a regional level there are only a few regions whose upper-end estimates are above 1.
R reflects what was happening a couple of weeks ago.

Death rates are still high with over 1000 people a day dying (28 days from a positive test - by date of death).
But daily death rates are now coming down reflecting the earlier falls in cases.
Over 1/4 of the UK’s covid deaths in this pandemic have occurred in the last 4 weeks.

Deaths will start to come down slowly, reflecting falls in hospital admissions/occupancy, but we are coming down from and extremely high level, so we can expect 10s of thousands of deaths to be reported over the coming months.

Read 4 tweets

Kit Yates

@Kit_Yates_Maths

25 Jan

A quick good news update on the state of play with vaccine delivery.
After a dip at the start of last week, delivery has continued to rise to almost half million doses a day. No dip on Saturday this week at least.
It's almost exclusively first doses being delivered now. 1/5

This steady performance is reflected in the cumulative number of doses delivered - up at almost 7 million as of Saturday.
2/5

Broken down by first and second dose coverage we can see that almost 10% of the UK population had received their first dose by Saturday.
3/5

Read 5 tweets

Kit Yates

@Kit_Yates_Maths

22 Jan

@IndependentSage

Finally lets finish with some good news with slides from the @IndependentSage briefing.
The delivery of the vaccine is ramping up.
You can see clearly where second doses (orange) were starting to be given before the strategy was changed and now its almost exclusively first doses.

Broken down by nation you can clearly see the rate (just first doses now) is increasing significantly. We're not up to the target rates we need yet, but if this continue then it won't be too long.
This seems positive for our ability to vaccinate the most vulnerable by Feb.

This is reflected in the rates of vaccination per hundred thousand people in each of the four home nations. It's increasing quickly in each country.

Read 4 tweets

Kit Yates

@Kit_Yates_Maths

22 Jan

@IndependentSage

Is this lockdown 3.0 as tough as lockdown 1?

Here are a few pieces of data from the @IndependentSage briefing which suggest that despite tackling a much more transmissible virus, lockdown is less strict, which might explain why we are only just keeping on top of cases.

Firstly, schools.
Respiratory infection outbreaks have increased since the new year.
Outbreaks in nurseries are as high as they were before Christmas.
Its similar story for SEN schools.
ONS data shows infection rate increasing in primary ...

... and early years children.
Despite schools nominally being closed, attendance rates are 4 times higher in primary and secondary schools than they were in lockdown 1.
21% of children in primary school vs 4%
5% of children in secondary school vs 1%
Schools are, in fact, open.

Read 7 tweets

Kit Yates

@Kit_Yates_Maths

22 Jan

@IndependentSage

A thread on the current death data and how we can interpret this in comparison to the first wave, with slides from the @IndependentSage briefing.

We saw headline-making, record numbers of reported daily deaths (within 28 days of a positive covid test) this week - 1820 on Weds.

There are issues due to the weekly fluctuations. Monday saw just (just!!!) 600 deaths reported.

The reporting artefacts mean its more consistent to look at deaths by date of death.

However, "reported-by" is the most up-to-date and the metric we've been using from the start.

Looking by date of death gives a much smoother picture with fewer fluctuations, which means is is more comparable day to day.
We are averaging over 1000 deaths a day at the moment (again 28 days since a COVID test). The average never went over 1000 a day in the first wave....

Read 8 tweets

Kit Yates

@Kit_Yates_Maths

22 Jan

@IndependentSage

Case/positivity data from the @IndependentSage briefing today.

Firstly, the raw UK cases figures show steep declines, but are a bit misleading.

Positivity data paints a slightly different picture of shallower decline (as does the ONS data overlaid).

Cases per hundred thousand of population broken down by the home nations also makes it look like cases are declining rapidly in each nation.

Northern Ireland (black) looks to be coming down particularly fast.

A slightly different picture emerges when you look at positivity data (i.e. positive cases divided by tests).

NI still declining quickly, still down but England and Scotland less steeply and Wales appears to be plateauing.

Worth keeping an eye on.

Read 10 tweets

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Share this page!

Kit Yates

Try unrolling a thread yourself!

More from @Kit_Yates_Maths

Kit Yates

Kit Yates

Kit Yates

Kit Yates

Kit Yates

Kit Yates

Did Thread Reader help you today?

Like this author's thread?