We’re starting this thread a couple of days early, because here in the United Kingdom a variety of things have changed regarding the reporting of Covid-19 “statistics”.
First of all the daily reports from the Department of Health and Social Care: now include “all deaths where a positive test for COVID-19 has been confirmed”, not just those “in hospital”. As the explanation on the DHSC web site puts it:
From 29 April 2020, DHSC are publishing as their daily announced figures on deaths from COVID-19 for the UK a new series that uses improved data for England produced by Public Health England (PHE). These figures provide a count of all deaths where a positive test for COVID-19 has been confirmed, wherever that death has taken place. Figures for Scotland, Wales and Northern Ireland have already begun to include deaths outside hospitals, so this change will ensure that the UK-wide series has a shared and common definitional coverage. This updated statement explains what the new data are and how they differ from both the data series previously published by DHSC and the figures produced by the ONS.
Separately, to improve the timely availability of data on deaths in care homes involving COVID-19, the ONS and the Care Quality Commission (CQC) agreed to publish from 28 April 2020 provisional counts of deaths in care homes, based on statutory notifications by care home providers to CQC. A separate explanatory statement about these new data has been published jointly by the ONS and CQC.
In addition to that change there is now a new section of the UK Government web site devoted to “National COVID-19 surveillance reports“. The most recent report at the moment:
Summarises the information from the surveillance systems which are used to monitor the Coronavirus Disease 2019 (COVID-19) pandemic in England.
The report is based on week 16 (data between 13 April and 19 April 2020) and where available daily data up to 22 April 2020.
COVID-19 is the disease name and SARS-CoV-2 is the virus name.
The report includes sections devoted to UK “Community surveillance”, “Primary care surveillance”, “Secondary care surveillance”, “Virological surveillance” and “Mortality surveillance”. In the latter section it points out that:
In week 16 2020 in England, statistically significant excess mortality by week of death above the upper 2 z-score threshold was seen overall, by age group in the 15-64 and 65+ year olds and sub nationally (all ages) in all regions (North East, North West, Yorkshire & Humber, East & West Midlands, East of England, London and South East & West regions) after correcting GRO disaggregate data for reporting delay with the standardised EuroMOMO algorithm (Figure 18). This data is provisional due to the time delay in registration; numbers may vary from week to week.
The recent “spike” appears to be statistically significant! However despite that, this report and the rest of the virtual paperwork emanating from the UK Government still fail to address the question we posed last week:
How will the UK Government build on the amazing community solidarity we have seen in the past few weeks and reassure the citizens of the nation that lifting the lockdown won’t result in further out-of-control outbreaks?
Whilst we continue to wait for an answer, here are the new “positive test” daily death numbers from the DHSC:
As of 9am 29 April, there have been 818,539 tests, with 52,429 tests on 28 April.
632,794 people have been tested of which 165,221 tested positive.
As of 5pm on 28 April, of those tested positive for coronavirus across all settings, 26,097 have sadly died. pic.twitter.com/Qw1GB5s3Wc
— Department of Health and Social Care (@DHSCgovuk) April 29, 2020
They do of course beg the question of how many “untested” deaths due to Covid-19 are occurring. Regular readers will be aware that we have been following the modelled future death projections of James Annan. Here’s his first update using the DHSC’s new numbers:
Twitter threading isn’t the best..here is what I get with the new time series of all reported deaths (still a huge undercount, but includes deaths outside hospitals) pic.twitter.com/1xWX4A8v1O
— James Annan 🇪🇺 (@jamesannan) April 30, 2020
The Office for National Statistics have released their latest weekly “death certificate” data. Here are the “main points” from their report:
- The provisional number of deaths registered in England and Wales in the week ending 24 April 2020 (Week 17) was 21,997, a decrease of 354 deaths registered compared with the previous week (Week 16); this is the first decrease in the number of deaths since the week ending 20 March 2020 (Week 12) but is 11,539 more than the five-year average for Week 17.
- Of the deaths registered in Week 17, 8,237 mentioned “novel coronavirus (COVID-19)”, which is 37.4% of all deaths; this is a decrease of 521 deaths compared with Week 16 (39.2% of all deaths).
- The number of deaths in care homes (from all causes) for Week 17 was 7,911, which is 595 higher than Week 16. The number of deaths in hospitals for Week 17 was 8,243, which is 1,191 lower than Week 16.
- In London, over half (50.5%) of deaths registered in Week 17 involved COVID-19; the North West and North East also had a high proportion of COVID-19 deaths, accounting for 38.8% and 38.0%, respectively, of deaths registered in these regions.
- In Wales, there were 413 deaths registered in Week 17 involving COVID-19, accounting for 36.7% of all deaths registered in Wales.
- Of deaths involving COVID-19 registered up to Week 17, 71.8% (19,643 deaths) occurred in hospital with the remainder mainly occurring in care homes (5,890 deaths), private homes (1,306 deaths) and hospices (301 deaths).
Here are those numbers in graphic detail:
The latest weekly Covid-19 update has been issued by the Office for National Statistics. Here’s the summary on Twitter:
33,337 / 21,647 = 1.54
The latest weekly update from the ONS has been published:
Our weekly deaths data for England show of all deaths occurring up to 8 May (registered up to 16 May), 37,154 involved #COVID19
— Office for National Statistics (ONS) (@ONS) May 19, 2020
For the same period
▪️ @DHSCgovuk reported 28,250 COVID-19 deaths
▪️ @NHSEngland reported 23,528 hospital deaths
➡️ https://t.co/XbMyXqwGkg pic.twitter.com/EyUzZUo3iX
Plus some additional detail:
Deaths in care homes made up 36.0% of all deaths in Week 17, 35.7% in Week 18 and 33.6% in Week 19 (Figure 7). Between Week 18 and Week 19, the number of deaths in care homes decreased by 33.7% to 4,248. However, the proportion of care home deaths that involved COVID-19 continued to increase, and 39.2% of all deaths in care homes involved COVID-19 in Week 19.
Watch this space!
35 replies on “Covid-19 in the UK in May 2020”
Thursday’s figures have been released:
With the lockdown continuing for the moment, here’s the current projection:
Let’s start with the basics shall we? Here are the UK Government’s published daily tested/deceased numbers:
and here is James Annan’s latest Covid-19 “forecast”:
As we discuss in a companion article the DHSC’s “122,347 tests on 30 April” is somewhat disingenuous.
As my old mate Andrew Neil put it:
Perhaps it won’t come as complete shock to discover that according to today’s DHSC update fewer tests were completed/posted than on the previous day?
Here too is James Annan’s latest “forecast”:
Let’s be charitable shall we, and assume that “the weekend effect” explains the apparent precipitous drop in the number of Covid-19 tests “administered” across the UK yesterday?
Here too is the latest Blue Skies Research update, now with “weekend effect” smoothing incorporated:
Another day with less than the Government’s “100,000 tests per day” target performed/posted:
The 288 daily deceased figure is the lowest for a long time, but in all probability subject to the “weekend effect”. Plus the usual forecast from James Annan:
The “weekend effect” has worn off. According to the Covid-19 Data Dashboard the “Daily number of COVID-19 associated UK deaths” is
693, but the DHSC Twitter feed hasn’t been updated with the usual summary information yet.
Here’s the new Blue Skies forecast:
P.S.
Sadly the number of tests performed/”in the post” is still less than Matt Hancock’s “100,000 per day” target.
Wednesday’s numbers are up on the Data Dashboard:
Plus somewhat later in the evening the DHSC daily tweet
which reveals another day of testing way below Matt Hancock’s alleged “target” for the end of last month. Plus the usual forecast:
Thursday’s official numbers:
Still < 100k tests. Plus the unofficial forecast, now with added bug fix!
The number of daily tests is creeping upwards, but sadly so is the number of daily deaths:
Plus the forecast from Blue Skies Research:
Allegedly people have been dancing in our streets in celebration of the 75th anniversary of Victory in Europe Day:
Meanwhile here are the latest round of the once United Kingdom’s daily deceased statistics:
And the latest forecast, necessarily not yet accounting for the unfortunate side effects of the recent Great British danse macabre:
BoJo has broadcast to the nation. “Stay alert!” Precisely what for remains to be seen. Acrobatic aerial porcines perhaps?
Meanwhile the “weekend effect” is with us once again:
Plus the perennial forecast, currently with no “Stay alert, don’t stay home!” amendments:
What with all the excitement today I almost forgot to mention. The number of tests “carried out” is officially still less than the 100,000 per day target for the end of April. Allegedlt that will have risen to 200,000/day by the end of May. Yet more aerial porcines on the way?
The numbers are out earlier than usual today. From the DHSC:
Note that allegedly the number of daily tests is finally back above the 100k target. Meanwhile from James Annan:
The “weekend effect” has evaporated once again:
“A big jump” as James Annan puts it:
The numbers are emerging from the DHSC earlier and earlier! Here are today’s:
Plus James Annan’s verdict:
Especially for Laurence Reed of BBC Radio Cornwall, here are the latest UK Covid-19 daily deceased numbers:
Plus James Annan’s forecast:
Daily deaths have dropped below 400 again. Intriguingly the number of tests performed is nearly twice the number of people tested:
Plus the latest forecast:
James’ cryptic comment refers to this news earlier in the day:
Daily tests are up slightly, but people tested is still woefully short of 100,000. Daily tests are up a lot also:
Plus the forecast:
The “weekend effect” is with us once again:
Plus the forecast based on “smoothed” data:
The “weekend effect” still in full swing:
Plus the forecast:
What will tomorrow’s numbers reveal?
Unfortunately it comes as no surprise that with the weekend now well and truly over daily deaths have leaped upwards once again:
The “painfully slow decline”:
Number of tests has jumped, but after yesterday’s brief hiatus the number of people tested has declined:
Very strange. Plus today’s analysis:
Another day of non zero deaths attributed to Covid-19:
James Annan is contemplating how best to modify his model to accommodate the “eased” UK lockdown:
Daily deaths similar to yesterday:
At least the number of people tested has increased, but then so has the number of positive tests.
James Annan has added a “breakpoint” to his model on May 11th, anticipating a potential increase in deaths following the “easing” of the UK’s lockdown:
The “bank holiday weekend effect” kicking in early, or significantly fewer deaths than last Friday?:
More data is required to forecast the next phase of the Covid-19 epidemic in the UK:
The “weekend effect” is definitely effective today. What’s more the gremlins that have stolen the “people tested” numbers have still not been vanquished:
Plus the suitably smoothed forecast:
Another weekend affected report. Another absence of people tested figures. Also an absence of daily tests over 100,000. Dare one mention the 200,000 target for the end of May?
Not enough “post breakpoint” data yet on the Blue Skies Research model for it to produce a useful “prediction”:
The latest weekly “death certificate” data has been released by the ONS:
Still no “people tested” numbers today, plus a probable “bank holiday effect”?
Plus the “forecast”:
The “long weekend” effect has well and truly worn off, and still no “people tested” numbers:
Meanwhile the plotted point on the “forecast” graph has leapt from on the lower bound to above the upper bound in one day:
The official daily deceased numbers seem to have settled down after the long weekend, albeit in the continuing absence of “people tested” figures:
And as James Annan puts it:
Rishi Sunak has announced the daily death number already, but the DHSC web site and Twitter feed have yet to be updated. Hence by way of a refreshing change, here is the forecast:
At 9:20 PM the DHSC Twitter feed still hasn’t updated, so here’s a screenshot instead:
Still no “people tested” numbers, and with the end of the month rapidly approaching still nowhere near 200,000 tests per day.
It took a long time to arrive again today. Here are the DHSC numbers:
Still no “people tested” numbers, and still nowhere near the magic 200,000 daily tests. At least the number of daily deaths has reduced. Plus the forecast:
Far fewer than 200,000 tests again today:
Plus the forecast, accompanied by a health warning:
The numbers are in for the last day of May, and very strange they look too:
39,045 – 38489 = 556. Not 111! The explanatory notes explain that:
Hence this revised forecast: