Urban75 Home About Offline BrixtonBuzz Contact

Coronavirus in the UK - news, lockdown and discussion

And this is my own attempt to graph 3 different sorts of figures for deaths per day.

The big gap between the ONS and others at the peak is the reason I couldnt take the others seriously.

Then later on, for a long time I assumed that the daily government figures (without 28 day limit) were running higher than the ONS ones because they were still catching up with some of the deaths they missed earlier, but I dont know if thats the actual and only reason why those number ended up higher than ONS ones after a certain point. The 28-day limit number has gotten the daily figure closer to the ONS figure, but for the wrong reasons.

Theres probably also differences between which nations are included in the figures, but I can overlook that at this stage o the pandemic since the number of daily deaths from Scotland etc is so low at the moment.

Screenshot 2020-09-07 at 13.03.18.png
 
And this is my own attempt to graph 3 different sorts of figures for deaths per day.

The big gap between the ONS and others at the peak is the reason I couldnt take the others seriously.

Then later on, for a long time I assumed that the daily government figures (without 28 day limit) were running higher than the ONS ones because they were still catching up with some of the deaths they missed earlier, but I dont know if thats the actual and only reason why those number ended up higher than ONS ones after a certain point. The 28-day limit number has gotten the daily figure closer to the ONS figure, but for the wrong reasons.

Theres probably also differences between which nations are included in the figures, but I can overlook that at this stage o the pandemic since the number of daily deaths from Scotland etc is so low at the moment.

View attachment 229498
Probably being a bit slow here...but I can't really see why the ONS data would go from significantly >than (no limit) Govt data to significantly < (no limit) Govt. data, unless they too have altered their methodology mid data set?
 
Probably being a bit slow here...but I can't really see why the ONS data would go from significantly >than (no limit) Govt data to significantly < (no limit) Govt. data, unless they too have altered their methodology mid data set?
The 'no limit' count includes anyone who has ever tested positive for C-19 who later dies. At first almost everyone who tested +ve and then died was probably a Covid death. As time passed and people recovered then just because they'd previously had Covid didn't mean that was the reason they died. Old people are still dying of old age. People continue to die for all the same reasons they died before Covid. That means a growing number of people on that measure were counted as a Covid death when they weren't. The ONS figure has continued to only include people with Covid on their death Certificate.
 
Probably being a bit slow here...but I can't really see why the ONS data would go from significantly >than (no limit) Govt data to significantly < (no limit) Govt. data, unless they too have altered their methodology mid data set?

The ONS has not altered its methodology, but the nature of their methodology means that changes elsewhere can affect their numbers.

By this I mean that ONS figures are based on Covid-19 being mentioned on the death certificate, and what ends up on the death certificate is influenced by the feelings of those responsible for filling in such paperwork at the time. I would expect this to vary somewhat over time as the impression of current pandemic phase changes over time and this is bound to influence opinion on whether a particular death was down to Covid-19.

In every epidemic and pandemic we know that there is almost never anything close to a perfect measure of deaths, they are almost always undercounted by some margin. This is why even the authorities themselves also pay attention to excess death statistics, where we can judge the total number of deaths at any time against the number who died on the same date in other years.

And so in this pandemic like any other, the rule of thumb is that the actual number of Covid-19 deaths is somewhere in between the number shown on death certificates, and the total excess deaths from all causes. And the number of excess deaths has been well over 60,000. I had it at more like 65,000 for the UK but some people will have a different number because they take away amounts for weeks where the total deaths was below the 5 year average.

It might also be worth pointing out that when number of deaths per day by any measure falls to quite low levels, all the flaws and noise in the methodology and data are going to have a more obvious influence. Excess deaths also become a harder measure to make use of during these periods. Because there is always some fluctuation in number of excess deaths compared to other years anyway (eg we had less deaths than average in the period before the pandemic got going, possibly due to last winters flu season being earlier than usual). And indeed after the massive amount of Covid-19 death dwindled, total deaths had a number of weeks where they fell back below the average. Some people, who had an interest in diminishing the deaths in this pandemic, eg by suggesting most of the deaths were just people dying a few months earlier than they otherwise would, were I think expecting to see a prolonged period in summer where the number of deaths in total was well below average, and they would use that to make their point. But it hasnt panned out like that so far, the fall below the average was too modest and in more recent weeks the ONS figures returned to being slightly above the norm, not below.
 
The ONS has not altered its methodology, but the nature of their methodology means that changes elsewhere can affect their numbers.

By this I mean that ONS figures are based on Covid-19 being mentioned on the death certificate, and what ends up on the death certificate is influenced by the feelings of those responsible for filling in such paperwork at the time. I would expect this to vary somewhat over time as the impression of current pandemic phase changes over time and this is bound to influence opinion on whether a particular death was down to Covid-19.

In every epidemic and pandemic we know that there is almost never anything close to a perfect measure of deaths, they are almost always undercounted by some margin. This is why even the authorities themselves also pay attention to excess death statistics, where we can judge the total number of deaths at any time against the number who died on the same date in other years.

And so in this pandemic like any other, the rule of thumb is that the actual number of Covid-19 deaths is somewhere in between the number shown on death certificates, and the total excess deaths from all causes. And the number of excess deaths has been well over 60,000. I had it at more like 65,000 for the UK but some people will have a different number because they take away amounts for weeks where the total deaths was below the 5 year average.

It might also be worth pointing out that when number of deaths per day by any measure falls to quite low levels, all the flaws and noise in the methodology and data are going to have a more obvious influence. Excess deaths also become a harder measure to make use of during these periods. Because there is always some fluctuation in number of excess deaths compared to other years anyway (eg we had less deaths than average in the period before the pandemic got going, possibly due to last winters flu season being earlier than usual). And indeed after the massive amount of Covid-19 death dwindled, total deaths had a number of weeks where they fell back below the average. Some people, who had an interest in diminishing the deaths in this pandemic, eg by suggesting most of the deaths were just people dying a few months earlier than they otherwise would, were I think expecting to see a prolonged period in summer where the number of deaths in total was well below average, and they would use that to make their point. But it hasnt panned out like that so far, the fall below the average was too modest and in more recent weeks the ONS figures returned to being slightly above the norm, not below.
Appreciate the time and effort required for such cogent responses; thanks.

I just feel that I want to be abreast of the various data sources/methods before we start to see the Autumn/Winter numbers rise.
 
The 'no limit' count includes anyone who has ever tested positive for C-19 who later dies. At first almost everyone who tested +ve and then died was probably a Covid death. As time passed and people recovered then just because they'd previously had Covid didn't mean that was the reason they died. Old people are still dying of old age. People continue to die for all the same reasons they died before Covid. That means a growing number of people on that measure were counted as a Covid death when they weren't. The ONS figure has continued to only include people with Covid on their death Certificate.
Clear; thanks.
 
So my office is reopening start of next month (just in time for next wave!) experimentally for a few people in senior management and they're trying to get a booking app running. The Management have said there will be no expectation on people to return - the fact is they can get so few people in that it's hardly worth them insisting. Kitchens and cafe will be shut, and presumably meeting rooms too, so not massively sure what the point is in coming in if you can't collaborate with people physically anyway.
 
Appreciate the time and effort required for such cogent responses; thanks.

I just feel that I want to be abreast of the various data sources/methods before we start to see the Autumn/Winter numbers rise.

No problem. Well one problem is that no matter how much I waffle on, I always feel like I've left some other aspects out.

In this case, there are often variations in methodology between the different UK nations, but beyond some basic differences I would probably need to be a fly on thousands of walls in order to grasp all the detail and reasons for variations and inaccuracies.

And I didnt talk much about the various different strengths and uses for the various different measures of deaths. Well it is already obvious which ones I find best to use for proper judgements about total deaths in the fulness of time. But ONS figures are especially laggy, so some of the daily figures, whilst only providing an incomplete picture, may be more useful in terms of providing more timely indicators that the situation has gotten worse. Its even possible that the 28-day limit will help any new uptick in deaths to stand out in a more obvious way, with less ambiguities about what could actually be behind the rise.

As ever when I start talking about death stats being laggy, thats partly down to how long it takes people to get sick and die from Covid-19, and so I end up advising everyone that we should be filling in the gaps between number of positive tests and number of deaths with all sorts of hospital numbers. But again these numbers and the underlying methodologies vary between nations, and unlike deaths there is far less info available about the generation of these statistics that we could use to inform ourselves of the differences and limitations. Plus, unfortunately, there were times during the first pandemic wave where we didnt have much hospital data, or the government stopped publishing certain numbers for long periods of time, driving me cray especially sine the press raraly noticed and complained. I dont know if that will ever happen again but I cannot rule it out, and I also cannot judge how well any of these indicators will show us the next emerging picture in a timely and accurate way.

One thing for sure is that I wont stop going on about all these different sorts of numbers, so however things evolve in the coming months I am likely to be posting an excessive quantity of graphs etc, with an emphasis on trying to zoom into emerging phenomenon as soon as they clearly show up in particular numbers.
 
Up to just under 3,000 cases a day and the weather's getting worse, kids back to school, universities starting, and many more people back to work.
Any bets how this is going to go...?

A subtle blend of stay absurd, a front door is better than any Hancock, the UK is thought to be 4 weeks behind reality, lets campaign for airport tests even though the rest of the testing system is already creaking under demand, lets do the same with parliament by calling for MPs to get tested every day so they can huddle around and give Johnson the crowd noise he requires to power his bluster, lets report on looming quarantines as though racing home before the deadline at great personal expense is a matter of life and death, pay attention to the private parties not the pubs, lets find some 'experts' who were wrong before to have another go at selling us their disgraceful agenda, its only a matter of time till PPE is back in the news, will the spectre of the daily pandemic press briefings return, and at what stage of a second wave will this happen and will last orders at the bar be called? And maybe some special guest appearances and a few surprises. Rumours on this front include a pandemic edition of the masked singer, featuring Baroness Didos dud dodo routine.
 
Maybe I'm just bitter that Jonathan Van-Tam wasnt given his own spin-off show after the main series ended.

It was May 30th that he warned 'don't tear the pants out of it', so by some measures we have done well to go about 100 days since then before reaching this moment of concern. Then again, that advice was considered obsolete in some tory circles because the emperors new pants were tear-proof by virtue of their illusory nature.
 
If there is ever a series 2 then I'd probably give Van-Tam star billing. Preferably he podium would also be seated atop a cage containing Cummings. And I wouldnt ask members of the cast who joined part way through series 1, like John Newton or Dido Harding, to return for series 2.
 
I thought that we could create a Gladiators/Battle Royale game show where Matt Hancock had to rescue a Johnson trapped in a cage from being given an electric shock whenever he banged his fists on the table or made one of his terrible metaphors. If either said the words 'world beating' they would be placed in a vat of slime like what you used to get in kids game shows, only they wouldn't be allowed to leave afterwards.
 
Last edited:
I thought that we could create a Gladiators/Battle Royale game show where Matt Hancock had to rescue a Johnson trapped in a cage from being given an electric shock whenever he banged his fists on the table or made one of his terrible metaphors.

Could mix it up a bit. Each week a randomly selected key-worker who put their life on the line in the pandemic gets to spin, spin, spin the wheel of justice. Possible fates for Johnson could include a bungee freezer jump where the weight of the freezer and length of the cord is calculated using an unsophisticated model from mediocre experts with the wrong data fed into it. Or a continuation of Johnsons inadvertently helpful public health demonstrations. First time round he showed us how easy it was to catch it at work by being reckless and behind the curve, and then helped everyone to learn that 7+ days of symptoms is the danger period where hospitalisation may be required. Next time he might fancy demonstrating the ability of the same person to be infected more than once, and whether the subsequent illness is milder or deadlier than the first.
 
Ah yes the BBC has something on that now:

The county borough of Caerphilly is to be placed under a local lockdown from 18:00 BST on Tuesday after a "rapid" rise in coronavirus cases.

People will not be able to leave or enter the borough without good reason, the Welsh Government said.

Face masks will be required for everyone over 11 in shops and people will not be able to meet indoors.

Seeing others within extended households will be banned, and no overnight stays will be allowed.

 
Not looking good across Wales. Partly, because our numbers have been so low that even a comparatively small spike looks huge. But the schools reopening seems to have really kicked things into a higher gear - I wouldn't be surprised if the Caerphilly outbreak is school-related.

 
But the schools reopening seems to have really kicked things into a higher gear - I wouldn't be surprised if the Caerphilly outbreak is school-related.

Sounds like they are finding cases with a variety of settings, which is partly what will fuel their concerns about general community transmission. Schools and nurseries already part of the mix there for sure:


Two Tesco workers have tested positive for coronavirus in Caerphilly.

More coronavirus cases have also been confirmed at a school, nursery and a pub in the area.
 
Caerphilly goes under lockdown

The county borough of Caerphilly is to be placed under a local lockdown from 18:00 BST on Tuesday after a "rapid" rise in coronavirus cases.

People will not be able to leave or enter the borough without good reason, the Welsh Government said.

Face masks will be required for everyone over 11 in shops and people will not be able to meet indoors.

Seeing others within extended households will be banned, and no overnight stays will be allowed.


There have been 133 new Covid-19 cases in Caerphilly county in the last seven days.

The Welsh Government said that at 55.4 cases per 100,000 population, it was the highest rate in Wales and one of the highest in the UK.

Wales' Health Minister Vaughan Gething said the virus was circulating in the community and it was "only a matter of time before we start to see more serious cases, which need hospital treatment".

The restrictions mean people cannot enter or leave Caerphilly county borough - which includes Caerphilly town, Ystrad Mynach, Blackwood, Newbridge and Risca - without a "reasonable excuse".

That includes work, if people are unable to work from home, or making a compassionate visit to a loved one, or to give care.

 
But the schools reopening seems to have really kicked things into a higher gear - I wouldn't be surprised if the Caerphilly outbreak is school-related.

I should also have said that its still a relatively early stage where there hasnt even been that much time for a lot of spread within educational settings to happen and be discovered yet. We are seeing cases with links to education, but in a lot of those the infections have happened elsewhere and the schools etc are mostly connected to the cases by virtue of who has to self-isolate as a result.

This isnt supposed to be a reassuring load of bullshit like the government etc have tried to make out in regard schools, its just a point about timing really, and all the other settings where the virus has had greater opportunities to spread in recent times than it had months back.

In other words, a lot of the alarming signals we have seen of late still represent the pre-schools-reopening phase, and other drivers of infection, so we might expect that we arent even seeing much of the impact from schools reopening yet, implying a much worse trajectory yet awaits. But schools etc are already starting to be responsible for an increase in self-isolation and demand for tests, and it wont be long before clusters of infection directly linked to schools and nurseries will have had enough time to occur more and be identified.
 
Back
Top Bottom