Corona Virus main thread

[BigFish 1,986]

4 hours ago, Barbe bleu said:

Gimmick  is a harsh word.  I see an app as one tool among many. It is my no means the only tool we need but, used correctly, I cant see a problem.

Maybe you are correct that gimmick is a bit harsh, @Barbe bleu. Thing is Hancock is now pushing this as the next big thing following the testing debacle. What is needed is a coherent strategy, not a collection of random tools. Within in a strategy it is a tool that I could agree with you, that it could serve a pupose, but Hancock doesn't have a strategy.

 

[Barbe bleu 814]

51 minutes ago, Surfer said:

 

https://www.euromomo.eu/graphs-and-maps/#excess-mortality

 

You really are going to have to give an explanation of these graphs.

[ricardo 7,347]

[Surfer 1,547]

7 hours ago, Barbe bleu said:

You really are going to have to give an explanation of these graphs.

Details in this link.. https://www.euromomo.eu/graphs-and-maps/#excess-mortality. The Z value (vertical axis) is explained there. 

However, the concept of "excess mortality" is quite simple. We know from historical records what the total mortality from all causes is  on a week by week basis throughout the year. So even without breaking it down into mortality by disease / accident etc - which gets into all sorts of issues like did you test that person for Covid 19 or not - we can understand if there is a greater mortality this week / month than there was in prior years this week / month. With Covid 19 running riot that will cause additional deaths because of the disease itself, and because of delayed treatments of other diseases, but it will also cause fewer deaths because of reduced road travel for example. So it is a true measure of the effect of the disease. The horizontal axis is weeks ... i.e which week this year.

With respect to the charts - if the curves are above the dashed line that says there are more deaths, if below there are fewer deaths. Britain right now has the highest peak value above normal of any European country, and this peak value is being sustained, in other countries it may have trended close to that same peak, but has since declined, in the case of Holland it is now below the normal.

The measure of how bad you have performed is the area under the curve. Our area is the biggest, QED we are performing worse. 

Ricardo's data suggests Italy and Spain performed worse than UK, but either way it is close, and as their excess deaths are now close to zero and ours are clearly not, we will undoubtedly overtake them even if we are behind as his referenced charts suggests. 

[Surfer 1,547]

8 hours ago, Bill said:

" Well we do have a source for excess mortality - and it does not make good reading for anyone in the UK "

That measure is still only a rough indication - given that it would need to be broken down by comparing work/road related deaths.

How many deaths are those who would have died due to 'illness' but Covid sped that process ?

With respect, a key advantage of "excess mortality" as a measurement is that it does not need a causal breakdown. We can just say that "during the pandemic we saw X more deaths" It is a measure of the overall impact of the disease across the whole community.

 

[Bill 1,788]

That excess is however not weighed against  a normal average. If that average is, say 1500 deaths per week then 2000 would suggest that there is an extra 500 (Covid) deaths.

However, at the moment there is likely to be a below average f.death figure due to fewer accidents...work/road so until all deaths are quantified by type, then it cannot be known what the Covid death number is, other than a suggestion that it is greater than the extra deaths over the average deaths

This probably could be better written as algebra ....................................... or probably just better written 😉

 

[Herman 9,715]

Just under 300 people that came into the country pre lockdown were quarantined. 

[Indy 3,283]

3 hours ago, Bill said:

That excess is however not weighed against  a normal average. If that average is, say 1500 deaths per week then 2000 would suggest that there is an extra 500 (Covid) deaths.

However, at the moment there is likely to be a below average f.death figure due to fewer accidents...work/road so until all deaths are quantified by type, then it cannot be known what the Covid death number is, other than a suggestion that it is greater than the extra deaths over the average deaths

This probably could be better written as algebra ....................................... or probably just better written 😉

 

Add to this how many in this extra above death would have died in 2020? As the year goes on and Covid is under some control, will the death rate fall under the average rate as Covid took out those who were already destined to pass away this year? could it be when 2020 is measured the average for this year might not be so far from 2019 average death rate.

Agree with bill on this one.

Edited May 5, 2020 by Indy

[Jim Smith 2,314]

Its being reported that French hospitals have re tested samples from patients diagnosed with pneumonia in December and January and have apparently found a sample from 27th December which is positive for Covid 19. 43 year old Franch/Algerian man who was diagnosed with pneumonia. Now recovered and had not travelled to anywhere prior to getting ill except for Algeria in September.

Interesting if true.

Maybe wishful thinking but a lot of people were struck down by a bad cough/virus in December that they couldn't shake off for 2-3 weeks. 

[Van wink 2,994]

All cause mortality data is what we will need to look back on to really try to understand how this disease has behaved and has affected us. Disease specific data is not capturing the full picture, it would appear from the excess deaths that many people are dying at home or in social settings from causes associated with CV 19 but unrecorded as such, but this isn’t the whole story.The value of analysis of the data will be to help get a fuller understanding of how the epidemic progressed in the U.K. so that we can be properly prepared for something similar in the future, also comparisons with other health systems to learn from best practice. 
The number of excess deaths ( outside of hospital settings ) plus the study published in Germany yesterday and the recent findings of earlier infection dates in France  makes me think that community infection rate is higher than many believe. I have also read papers suggesting herd immunity may be achieved at well below 60% infection so possibly a faint glimmer of hope for the future.

Edited May 5, 2020 by Van wink

[Aggy 752]

3 hours ago, Bill said:

That excess is however not weighed against  a normal average. If that average is, say 1500 deaths per week then 2000 would suggest that there is an extra 500 (Covid) deaths.

However, at the moment there is likely to be a below average f.death figure due to fewer accidents...work/road so until all deaths are quantified by type, then it cannot be known what the Covid death number is, other than a suggestion that it is greater than the extra deaths over the average deaths

This probably could be better written as algebra ....................................... or probably just better written 😉

 

Agreed on this , although there are things which work both ways.

Whilst the reduced road and work related deaths might suggest covid is more deadly than the excess deaths suggest, you could look at increased deaths relating to mental health, people not going to hospital to get treated as they’re scared of covid etc to balance it back out.

Even before lockdown through January to mid- March we were recording excess deaths this year for things completely unrelated to “covid, flu or pneumonia”. So we were actually seeing increased excess deaths all year anyway (above the five year average), so it’s probably a safe bet that not all of the excess deaths now are just down to coronavirus.

When covid came along, until April 10th, there were still fewer deaths each week from “flu, coronavirus and pneumonia” than there normally are from just “flu and pneumonia”. So until April 10th coronavirus was not a cause of any excess death statistically  - yet we were still recording more total deaths on a weekly basis over the first 3.5 months of the year than the five year average.

Since April 10, coronavirus has caused deaths from “flu, corona or pneumonia” to become higher than the five year average number of deaths from “flu or pneumonia”.  However, we are still (or when I last checked the figures a week or so ago) seeing excess deaths from other things too.

On that basis, I don’t agree with Surfer’s statement that excess deaths is a measure of overall impact of the disease across the community. We were already seeing excess deaths from things completely unrelated to covid all year long, and that won’t have changed.

So the excess deaths figure is impacted by coronavirus, yes, but as it is also heavily impacted by other things, it simply isn’t a measure of the impact of the disease only.

[Barbe bleu 814]

6 hours ago, Surfer said:

Details in this link.. https://www.euromomo.eu/graphs-and-maps/#excess-mortality. The Z value (vertical axis) is explained there. 

However, the concept of "excess mortality" is quite simple. We know from historical records what the total mortality from all causes is  on a week by week basis throughout the year. So even without breaking it down into mortality by disease / accident etc - which gets into all sorts of issues like did you test that person for Covid 19 or not - we can understand if there is a greater mortality this week / month than there was in prior years this week / month. With Covid 19 running riot that will cause additional deaths because of the disease itself, and because of delayed treatments of other diseases, but it will also cause fewer deaths because of reduced road travel for example. So it is a true measure of the effect of the disease. The horizontal axis is weeks ... i.e which week this year.

With respect to the charts - if the curves are above the dashed line that says there are more deaths, if below there are fewer deaths. Britain right now has the highest peak value above normal of any European country, and this peak value is being sustained, in other countries it may have trended close to that same peak, but has since declined, in the case of Holland it is now below the normal.

The measure of how bad you have performed is the area under the curve. Our area is the biggest, QED we are performing worse. 

Ricardo's data suggests Italy and Spain performed worse than UK, but either way it is close, and as their excess deaths are now close to zero and ours are clearly not, we will undoubtedly overtake them even if we are behind as his referenced charts suggests. 

I think there is a reason you didn't try to define 'z score'  and its because the website helpfully gives this completely impenetrable definition

"In general, Z-score = (x-mean of the population)/Standard deviation of the population, which could be approximated in our context by S-score = (number of deaths - baseline) / Standard deviation of the residuals (variation of the number of deaths around the baseline) on the part of the series used to fit the model, used as the standard unit."

 

honestly, I cannot understand these graphs.  Yea they look bad but as I don't understand the y axis I really cant get too excited about them!

 

[Yellow Fever 3,747]

BB - As stated the units for the z score are standard deviations in effect from the 'mean'.

Hence the values given are hugely statistically significant (as is obvious from any graph in any form of excess deaths) 3 sd would be significant, 5 for a proof of a theorem in physics etc above chance.

[T 190]

Yes you are comparing the average number of deaths with the average variance in deaths  So if your average number of deaths is close to your average variation in deaths then it come be just normal variations in deaths but the higher your average deaths are compared to the average variance is then the greater amount of excess deaths compared to the average number of deaths  So if you average is close to the average variance it could just be natural fluctuations but if you have a high score is means it is highly probably you have a lot of excess deaths  

 

I take that if you have more deaths now you will lower the rate going forward  But you take that to the extreme if we all die this year then the average number of deaths next year will be zero so it doesn’t  mean it is a good thing! Boris die before your time is actually s good way of putting it  It is why I like the average age of death over time as a good measure  

 

[keelansgrandad 6,679]

Just to let you know in case you were fretting not knowing, Gordon Ramsay is managing to drive the 50K each way to check on how his new restaurant is coming along.

[T 190]

Infections and R continuing to fall following initial lifting of restrictions so further gradual lifting of restrictions will be  implemented in Germany which should be good news for everyone on Lifting restrictions everywhere  

RKI saying death rate per 100000 is 8 in Germany, 21 in US, 43 in U.K. and 54 in Spain

[BigFish 1,986]

In the early phases, if we had managed to ramp testing capacity quicker, it would have been beneficial.

For all sorts of reasons that did not happen. And I think it’s clear you need lots of testing for this

This looks like damning criticism of the governments decision to halt testing earlier in the pandemic. Perhaps Hancock would like to explain at todays Downing Street briefing what those reasons were.

[Bill 1,788]

The bottom line is that 30.000 have died and that figure is probably higher.

However, the important part is that whatever the number is, it is the number while we have been in lockdown. If the death rate continues to fall as relaxation increase then it points to the thought that we have it under control, to a degree - and it will allow a respite so as to prepare ourselves better for what may come this autumn.

Getting caught up in the minutiae of this means that the bigger picture is being lost. It is that which needs all our focus, maybe not on here, but certainly outside.

Put bluntly has it killed all who are to be killed, or the greater  number, or will it increase again once we begin to interact more closely Because they are too many variables to suggest the numbers give an accurate indication of spread and/or deaths.

[Bill 1,788]

4 minutes ago, BigFish said:

In the early phases, if we had managed to ramp testing capacity quicker, it would have been beneficial.

For all sorts of reasons that did not happen. And I think it’s clear you need lots of testing for this

This looks like damning criticism of the governments decision to halt testing earlier in the pandemic. Perhaps Hancock would like to explain at todays Downing Street briefing what those reasons were.

https://www.independent.co.uk/news/uk/politics/coronavirus-lockdown-uk-patrick-vallance-news-update-a9499196.html

[Herman 9,715]

18 hours ago, TeemuVanBasten said:

Looks like we did a pretty good job of achieving a flat curve, wasn't avoiding a sharp peak and to keep ICU wards under full capacity the point of lockdown? We only got to 80 percent capacity nationally I believe (not including nightingale). We certainly didn't get the point where we had to decide who to treat and who not to treat like Italy, so I object to the idea that we've performed worse than all other EU countries. 

We didn't have to decide not to treat over 60's like Italy, and we didn't get care workers abandoning residents like Spain, we haven't had lockdown protests like Germany. We've done ok as a nation, I'm proud. 

 

That aged well. 

[The Real Buh 3,423]

Th weird competitive nature of this, along with the political point scoring, is a timely reminder that the obsession of political activism is the realm of low intellect. It’s genuinely made a lot of people mentally ill which is such a shame. I look at the figures in our society on both sides of the political divide and can only think “what a waste of life”

[Barbe bleu 814]

3 hours ago, Yellow Fever said:

BB - As stated the units for the z score are standard deviations in effect from the 'mean'.

Hence the values given are hugely statistically significant (as is obvious from any graph in any form of excess deaths) 3 sd would be significant, 5 for a proof of a theorem in physics etc above chance.

So is it correct to say that if the excess deaths occur at a time of high natural death (in flu season for instance) the graph will show a lower peak than if the same (or less) deaths occurred at a time of lower ' natural' mortality?

If we are comparing nations based on these graphs we must be careful then that we account for natural variation?

[sgncfc 1,221]

21 minutes ago, Bill said:

The bottom line is that 30.000 have died and that figure is probably higher.

However, the important part is that whatever the number is, it is the number while we have been in lockdown. If the death rate continues to fall as relaxation increase then it points to the thought that we have it under control, to a degree - and it will allow a respite so as to prepare ourselves better for what may come this autumn.

Getting caught up in the minutiae of this means that the bigger picture is being lost. It is that which needs all our focus, maybe not on here, but certainly outside.

Put bluntly has it killed all who are to be killed, or the greater  number, or will it increase again once we begin to interact more closely Because they are too many variables to suggest the numbers give an accurate indication of spread and/or deaths.

There is quite a lot of information in the scientific or AI domain which is not yet in the public domain. For instance, the first actual cases of Covid-19 so far known occurred in June 2019, not December 2019; the medical profession has developed a way of caring for the most serious ill in hospital which has reduced deaths - not a treatment, but (for instance) more regular chest draining, higher doses of Vit D, laying on front, not putting people on ventilators etc.

Unless we were to autopsy everyone who died since last June we will never know how many have died because of or with the virus, but the best useable data is excess deaths - this does show that the UK has performed poorly against other European countries overall, but the actual reasons why are not yet proven.

There will be an inquiry eventually which will not be favourable, I suspect. For the moment we need to look forward - the immediate problem we have is that numbers of new cases are not coming down quickly enough. The increased numbers of testing is contributing to that, but the main people being tested are those in a target or key worker group or with symptoms so as the risk to those with health conditions or obesity is so high this is a real problem.

The game changers are whether those who have had it have immunity, and indeed how many have had it. The Roche antibody test is likely to be the one everyone will use and that will be available by the end of this month. Our data by the end of June will be much more informative and therefore the chance of any second or third wave being as devastating in terms of the death rate is small.

[baldy09 156]

I getting so fed up with it all - Im starting to think I just go out and risk it all,  back to work etc, that's how bad I'm feeling .I know its wrong but how much more can one take

[Barbe bleu 814]

37 minutes ago, Herman said:

That aged well. 

I think the point that he was making was that the NHS was not overwhelmed to the point that it was rationing treatment.  I'm not sure that anyone disputes this.

Whether or not more lives could have been saved is a different point.  The best answer is probably 'yes' but it will be a long time yet before we know which nation  dealt with it best.

Edited May 5, 2020 by Barbe bleu

[Bill 1,788]

12 minutes ago, Barbe bleu said:

I think the point that he was making was that the NHS was not overwhelmed to the point that it was rationing treatment.  I'm not sure that anyone disputes this.

Whether or not more lives could have been saved is a different point.  The best answer is probably 'yes' but it will be a long time yet before we know which nation  dealt with it best.

whereas in the real world it was

operations were cancelled by the hundreds

anecdotal evidence from those medical staff working spoke of how life and death decisions were being made based on equipment shortages

and elderly patients were sent back to care homes due to lack of beds

it is a shame that you don't have anything better to do than constantly peddle lies on here......or is this reaction what you crave ?

[Surfer 1,547]

58 minutes ago, Barbe bleu said:

So is it correct to say that if the excess deaths occur at a time of high natural death (in flu season for instance) the graph will show a lower peak than if the same (or less) deaths occurred at a time of lower ' natural' mortality?

If we are comparing nations based on these graphs we must be careful then that we account for natural variation?

Excess deaths is simply the total number of deaths in a time period compared to the average number of deaths in prior years for the same time period. So the answer is no. 

Edited May 5, 2020 by Surfer

[ricardo 7,347]

[BigFish 1,986]

38 minutes ago, ricardo said:

Meanwhile the number of deaths in care homes has doubled in a week to 5,890.

Yet the PM calls this a success

Go figure

[Yellow Fever 3,747]

2 hours ago, Barbe bleu said:

So is it correct to say that if the excess deaths occur at a time of high natural death (in flu season for instance) the graph will show a lower peak than if the same (or less) deaths occurred at a time of lower ' natural' mortality?

If we are comparing nations based on these graphs we must be careful then that we account for natural variation?

I think you'll find they account for seasonal variations - it's already in the 'mean'.

 

Edited May 5, 2020 by Yellow Fever
Just saw Surfers reply - the point is the 'mean' is not an annual mean but a time/date dependent function (the x axis) ergo normal seasonal variations are already included. Of course excess deaths up/down may be due to other factors as well.