agent69
Member of DD Central
Posts: 5,587
Likes: 4,182
|
Post by agent69 on Jul 12, 2020 9:50:14 GMT
There may have been a change in reporting that may be quietly reported on the PHE site over the coming days (it wouldn't be the first time). The missing piece of the puzzle at the current time is the backlog at coroners. I wonder if a batch of deaths were released as a result of the coroners getting back to work? The person who appears to have the best handle on this is statistician Jamie Jenkins, see his twitter page. He currently estimates 69,500 excess deaths with only about 0.5% being premature deaths: The problem with covid-19 is that it has spawned an almost infinite number of experts, all looking for their 15 minutes of fame. On what basis have you decided that My Jenkins has the best handle on this? I'm curious why his estimate of premature deaths has doubled in 4 days, and why he is only counting people over 85?
When it comes to looking at numbers I still recall the mantra I was taught at school - there are lies, dam lies and statistics.
|
|
|
Post by dan1 on Jul 12, 2020 10:03:00 GMT
There may have been a change in reporting that may be quietly reported on the PHE site over the coming days (it wouldn't be the first time). The missing piece of the puzzle at the current time is the backlog at coroners. I wonder if a batch of deaths were released as a result of the coroners getting back to work? The person who appears to have the best handle on this is statistician Jamie Jenkins, see his twitter page. He currently estimates 69,500 excess deaths with only about 0.5% being premature deaths: The problem with covid-19 is that it has spawned an almost infinite number of experts, all looking for their 15 minutes of fame. On what basis have you decided that My Jenkins has the best handle on this? I'm curious why his estimate of premature deaths has doubled in 4 days, and why he is only counting people over 85?
When it comes to looking at numbers I still recall the mantra I was taught at school - there are lies, dam lies and statistics.
I pity you agent69. Keep safe.
|
|
|
Post by Deleted on Jul 15, 2020 8:42:50 GMT
Chatting to my German friends they say they were lucky to be invaded by a group of horny young male Americans, French, Canadians and Brits in around 1944 who spread sexual diseases throughout the country just as the first antibiotics came along, leading to the development of some serious contagion tracking systems. Switching the sexually transmitted tracing system to contact tracing was pretty easy. V pleased to read that, by omission, the Aussies and Kiwis were hygienic and "clean" then. 😄 Sorry yes, of course the Kiwis and the Aussies had their own share of syphilis as did the Czechs and Russians.
See, I didn't mention sheep once there, oops.
|
|
r00lish67
Member of DD Central
Posts: 2,691
Likes: 4,048
|
Post by r00lish67 on Jul 16, 2020 10:59:32 GMT
Apologies if this has been covered before / is obvious.
I was idly curious, looking at Worldometers just now, as to whether new deaths per day vs new cases per day for each country was converging at all across different countries*
So I took out the table into a spreadsheet and divided new deaths by new cases and sorted by that %..
and, to my genuine surprise (I don't follow this closely) it looks like the UK is at the very top of the table for yesterday and the day before, and probably all of last week.
New deaths / New cases for UK = 15.8% yesterday, 34.7% the day before.
Only Mexico was close to us yesterday (11.8%), whilst the day before no-one was anywhere close (France 2nd = 14%).
So my questions to the learned people here are:
1) Given the imperfection/lag etc of the data, is this at all meaningful?
2) If it is meaningful, why is this the case? Is the UK just not testing enough people? Are we as a nation more fat/older than everyone else? Or is it something else?
*in answer to my original curiosity, the answer seems to be 'not really'. Perhaps the data is just too messy and the testing approaches too mixed to be useful. The average was about 4.5%, but the rates per country were anywhere from 0.4% to 12%, or more typically 4-8%).
|
|
cb25
Posts: 3,521
Likes: 2,665
|
Post by cb25 on Jul 16, 2020 11:14:04 GMT
r00lish67 I don't see it as a meaningful figure, but that might be because I'm not one of the learned. One could easily get the deaths/cases ratio down by (say) infecting 1000 people in a day. Ratio would drop with any deaths to come days later.
|
|
|
Post by bracknellboy on Jul 16, 2020 11:41:40 GMT
Added deaths/cases for 11/07/2020: 148/820 Not good, big jump in both deaths & cases over figures from 7 days ago Not wise to read too much into a single days figures.
However, I do feel the urge to emigrate to Scotland getting ever stronger. there is only one problem with that.....
|
|
|
Post by dan1 on Jul 16, 2020 11:47:43 GMT
Apologies if this has been covered before / is obvious. I was idly curious, looking at Worldometers just now, as to whether new deaths per day vs new cases per day for each country was converging at all across different countries* So I took out the table into a spreadsheet and divided new deaths by new cases and sorted by that %.. and, to my genuine surprise (I don't follow this closely) it looks like the UK is at the very top of the table for yesterday and the day before, and probably all of last week. New deaths / New cases for UK = 15.8% yesterday, 34.7% the day before. Only Mexico was close to us yesterday (11.8%), whilst the day before no-one was anywhere close (France 2nd = 14%). So my questions to the learned people here are: 1) Given the imperfection/lag etc of the data, is this at all meaningful?
2) If it is meaningful, why is this the case? Is the UK just not testing enough people? Are we as a nation more fat/older than everyone else? Or is it something else? *in answer to my original curiosity, the answer seems to be 'not really'. Perhaps the data is just too messy and the testing approaches too mixed to be useful. The average was about 4.5%, but the rates per country were anywhere from 0.4% to 12%, or more typically 4-8%). No, at least not in my opinion. There may been something to glean by comparing the data on countries that are in the same stage of their outbreak but comparing a country with infections on the rise to those where it's dwindling is probably grossly misleading. But even this should be taken with a pinch of salt when comparing large and diverse nations (just think USA where lumped together are states like NY/NJ where infections are reducing (I think) with FL/TX/AR where they're on the rise). It's largely a product of the lag, as you mention, between infection and death. Perhaps you could look at cases 3 weeks prior to deaths now (best to look at moving 7-day average I guess) but even that will depend on reporting delays (many countries take a great deal longer to report deaths than we do). And as the number of deaths reduces to a lower level then you get issues such as those infected many weeks/months prior dying contributing disproportionately to the statistics as compared to when there were greater infections/deaths. One issue specific to the UK may be that coroners inquests have restarted so we may be seeing increased numbers of "reported" deaths when in fact the date of death was several weeks/months prior (coroners only just getting back to work I guess). Again, when the number of infections/deaths are low these inquest deaths will contribute disproportionately. Of course, we should be looking at excess deaths not just those where a positive test has been performed. I've no real feel for how we compare to other countries (except Belguim where suspicions are reported in the death stats).
|
|
cb25
Posts: 3,521
Likes: 2,665
|
Post by cb25 on Jul 16, 2020 15:31:02 GMT
Added deaths/cases for 16/07/2020: 66/642 7-day average deaths has dipped below 75 (73.9) for the first time since 26 March
|
|
|
Post by bernythedolt on Jul 16, 2020 23:12:38 GMT
Apologies if this has been covered before / is obvious. I was idly curious, looking at Worldometers just now, as to whether new deaths per day vs new cases per day for each country was converging at all across different countries* So I took out the table into a spreadsheet and divided new deaths by new cases and sorted by that %.. and, to my genuine surprise (I don't follow this closely) it looks like the UK is at the very top of the table for yesterday and the day before, and probably all of last week. New deaths / New cases for UK = 15.8% yesterday, 34.7% the day before. Only Mexico was close to us yesterday (11.8%), whilst the day before no-one was anywhere close (France 2nd = 14%). So my questions to the learned people here are: 1) Given the imperfection/lag etc of the data, is this at all meaningful? 2) If it is meaningful, why is this the case? Is the UK just not testing enough people? Are we as a nation more fat/older than everyone else? Or is it something else? *in answer to my original curiosity, the answer seems to be 'not really'. Perhaps the data is just too messy and the testing approaches too mixed to be useful. The average was about 4.5%, but the rates per country were anywhere from 0.4% to 12%, or more typically 4-8%). In effect I believe what you're observing is a snapshot of the Case Fatality Rate. The CFR is defined as number of deaths recorded/number of cases recorded. In the UK, we have recorded a total of 45,119 deaths from 292,552 cases, giving a CFR since day one of 15.4%. I posted some while ago that the UK's CFR together with France, Belgium, Spain, Italy, Hungary... were clustered together around 14%-15% and were considerably higher than certain other European states like Germany, Portugal, Poland, Austria, Ireland, Denmark, Romania, Czechia, Slovakia, Slovenia... which are clustered around 5% or lower. There are noticeably few nations that sit between these two figures. Most were... and still are today... either ~15% or <~5%. (This led me to ponder at the time whether there were two distinct strains of the virus operating in Europe). You may find this resource interesting. Select the Case-Fatality Ratio tab beneath the map. Then click on the white dots to see the 'all-time' CFR for UK = 15.4%, England = 16.1%, France = 14%. Figures remarkably close to those you observed in your snapshot of current data.
|
|
cb25
Posts: 3,521
Likes: 2,665
|
Post by cb25 on Jul 17, 2020 17:23:35 GMT
No daily deaths figures published today for 16 July 2020 due to "The Secretary of State has today, 17 July, asked PHE to urgently review their estimation of daily death statistics. Currently the daily deaths measure counts all people who have tested positive for coronavirus and since died, with no cut-off between time of testing and date of death. There have been claims that the lack of cut-off may distort the current daily deaths number. We are therefore pausing the publication of the daily figure while this is resolved." ( link)
|
|
starfished
Member of DD Central
Posts: 296
Likes: 216
|
Post by starfished on Jul 17, 2020 17:44:39 GMT
No daily deaths figures published today for 16 July 2020 due to "The Secretary of State has today, 17 July, asked PHE to urgently review their estimation of daily death statistics. Currently the daily deaths measure counts all people who have tested positive for coronavirus and since died, with no cut-off between time of testing and date of death. There have been claims that the lack of cut-off may distort the current daily deaths number. We are therefore pausing the publication of the daily figure while this is resolved." ( link) It will be interesting to see how this play out and if they stop publishing entirely which is what I suspect the government wants. Identifying a covid death was always a bit of challenge especially given its higher impact on those that had co-morbidities. As had been mentioned before, excess deaths over the 2020 to 2022 and economic growth over 2020 to 2022 probably are the cleanest tools available to judge covid response, impact and resilience by country but it does mean us all being a bit more patient...
|
|
agent69
Member of DD Central
Posts: 5,587
Likes: 4,182
|
Post by agent69 on Jul 17, 2020 19:24:19 GMT
No daily deaths figures published today for 16 July 2020 due to "The Secretary of State has today, 17 July, asked PHE to urgently review their estimation of daily death statistics. Currently the daily deaths measure counts all people who have tested positive for coronavirus and since died, with no cut-off between time of testing and date of death. There have been claims that the lack of cut-off may distort the current daily deaths number. We are therefore pausing the publication of the daily figure while this is resolved." ( link) There was an interesting article on Sky news earlier today regarding how PHE calculates Covid related deaths.
Apparently in Scotland, Wales and NI if you recovered from Covid more than 28 days ago and subsequently die it doesn't go in the stats. But in England if you recovered from Covid 2 months ago and die in a car crash tomorrow it gets counted as Covid related.
|
|
benaj
Member of DD Central
Posts: 4,855
Likes: 1,591
|
Post by benaj on Jul 17, 2020 19:29:54 GMT
|
|
cb25
Posts: 3,521
Likes: 2,665
|
Post by cb25 on Jul 17, 2020 19:47:09 GMT
No daily deaths figures published today for 16 July 2020 due to "The Secretary of State has today, 17 July, asked PHE to urgently review their estimation of daily death statistics. Currently the daily deaths measure counts all people who have tested positive for coronavirus and since died, with no cut-off between time of testing and date of death. There have been claims that the lack of cut-off may distort the current daily deaths number. We are therefore pausing the publication of the daily figure while this is resolved." ( link) There was an interesting article on Sky news earlier today regarding how PHE calculates Covid related deaths.
Apparently in Scotland, Wales and NI if you recovered from Covid more than 28 days ago and subsequently die it doesn't go in the stats. But in England if you recovered from Covid 2 months ago and die in a car crash tomorrow it gets counted as Covid related. Guardian article states "A Department of Health and Social Care source said: “You could have been tested positive in February, have no symptoms, then be hit by a bus in July and you’d be recorded as a Covid death.”"
|
|
Greenwood2
Member of DD Central
Posts: 4,241
Likes: 2,686
|
Post by Greenwood2 on Jul 17, 2020 19:54:16 GMT
There was an interesting article on Sky news earlier today regarding how PHE calculates Covid related deaths.
Apparently in Scotland, Wales and NI if you recovered from Covid more than 28 days ago and subsequently die it doesn't go in the stats. But in England if you recovered from Covid 2 months ago and die in a car crash tomorrow it gets counted as Covid related. Guardian article states "A Department of Health and Social Care source said: “You could have been tested positive in February, have no symptoms, then be hit by a bus in July and you’d be recorded as a Covid death.”" Isn't this good news, Covid deaths may be lower than recorded, so far I thought everyone seemed to be saying the government are underestimating infections/deaths.
|
|