Scrutable

MartinDurkin wrote: Tue Apr 14, 2020 9:33 am This from David Spiegelhalter looks worrying, particularly "Covid not mentioned" figures. Data only to 3rd April.
https://twitter.com/d_spiegel/status/12 ... 2692096003
EVjXaKrXkAAd-pa.jpg

Even that pisses me off. Why can't he be clear whether official figures are in this chart or not?

Martin Y wrote: Tue Apr 14, 2020 9:43 am

MartinDurkin wrote: Tue Apr 14, 2020 9:33 am This from David Spiegelhalter looks worrying, particularly "Covid not mentioned" figures. Data only to 3rd April.
https://twitter.com/d_spiegel/status/12 ... 2692096003
EVjXaKrXkAAd-pa.jpg

It's very revealing. Either April stared with a sudden spate of 2,500 remarkable coincidences per week or the real Covid-19 death rate was about 6,000 per week.

Short BBC story here. Sorry if covered already/elsewhere; just looking in quickly with a cup of tea.

tenchboy wrote: Tue Apr 14, 2020 10:23 am

Martin Y wrote: Tue Apr 14, 2020 9:43 am

MartinDurkin wrote: Tue Apr 14, 2020 9:33 am This from David Spiegelhalter looks worrying, particularly "Covid not mentioned" figures. Data only to 3rd April.
https://twitter.com/d_spiegel/status/12 ... 2692096003
EVjXaKrXkAAd-pa.jpg

It's very revealing. Either April stared with a sudden spate of 2,500 remarkable coincidences per week or the real Covid-19 death rate was about 6,000 per week.

Short BBC story here. Sorry if covered already/elsewhere; just looking in quickly with a cup of tea.

Short Guardian story here.

Some of the excess non covid deaths are likely down to people not seeking or being unable to access healthcare. We know A+E attendance is way down, some of this will be minor injuries and illnesses, but some is likely serious conditions and people are avoiding hospital to the point of endangering themselves.

lpm wrote: Tue Apr 14, 2020 10:22 am

MartinDurkin wrote: Tue Apr 14, 2020 9:33 am This from David Spiegelhalter looks worrying, particularly "Covid not mentioned" figures. Data only to 3rd April.
https://twitter.com/d_spiegel/status/12 ... 2692096003

Even that pisses me off. Why can't he be clear whether official figures are in this chart or not?

Well the link in his tweet goes here.
https://www.ons.gov.uk/peoplepopulation ... ek-1-to-14
What other "official" figures are you referring to?

The reason we need detailed death statistics is because:

- the govt halted testing, except for limited number of testing in hospitals
- so we don't know how fast it is spreading
- which we need to know in order to decide on lockdown measures
- but we can use deaths as a proxy for covid cases
- because all deaths are recorded somewhere
- but the govt only gives deaths in hospitals
- so we need to add non-hospital covid deaths
- which we can get from (a) death certificates mentioning covid
- and (b) modelling of excess deaths over the average
- but (a) has the problem that signers of death certificate might not know
- and (b) would include non-covid deaths caused by the lockdown
- so there needs to be very clear analysis of the numbers
- with people stating exactly what they are looking at, how the analysis is done, what the timeframes are, and their assumptions
- because without this we are in the dark about the current trajectory of infections

lpm wrote: Tue Apr 14, 2020 11:57 am The reason we need detailed death statistics is because:

- the govt halted testing, except for limited number of testing in hospitals and close relatives of senior Tories

FTFY

lpm wrote: Tue Apr 14, 2020 11:57 am The reason we need detailed death statistics is because:

- the govt halted testing, except for limited number of testing in hospitals
- so we don't know how fast it is spreading
- which we need to know in order to decide on lockdown measures
- but we can use deaths as a proxy for covid cases
- because all deaths are recorded somewhere
- but the govt only gives deaths in hospitals
- so we need to add non-hospital covid deaths
- which we can get from (a) death certificates mentioning covid
- and (b) modelling of excess deaths over the average
- but (a) has the problem that signers of death certificate might not know
- and (b) would include non-covid deaths caused by the lockdown
- so there needs to be very clear analysis of the numbers
- with people stating exactly what they are looking at, how the analysis is done, what the timeframes are, and their assumptions
- because without this we are in the dark about the current trajectory of infections

Are there any countries that are measuring in this detail? Can we learn from other countries and extrapolate accordingly?

Just saw this on the BBC live feed:

New plans for recording UK care home deaths
The UK has been releasing daily coronavirus death tolls, but the figures have been hospital deaths only - they have not included deaths at people's houses or in care homes.

New figures from the Office of National Statistics published on Tuesday provide more information on deaths in care homes, but they are about two weeks out of date. That's because they rely on death certificates, which take a while to process.

Health officials are now trying to find ways to provide more up-to-date information from care homes.

From later this week, deaths that are suspected or confirmed to involve Covid-19 will be recorded daily, England's care regulator has said.

The Care Quality Commission (CQC) says these new figures should provide a "clearer picture of deaths" and highlight worst-hit regions.

The CQC says it will work with the ONS and Public Health England.

It is currently unclear whether these figures will be released publicly.[\quote]

headshot wrote: Tue Apr 14, 2020 12:00 pm

lpm wrote: Tue Apr 14, 2020 11:57 am The reason we need detailed death statistics is because:

- the govt halted testing, except for limited number of testing in hospitals
- so we don't know how fast it is spreading
- which we need to know in order to decide on lockdown measures
- but we can use deaths as a proxy for covid cases
- because all deaths are recorded somewhere
- but the govt only gives deaths in hospitals
- so we need to add non-hospital covid deaths
- which we can get from (a) death certificates mentioning covid
- and (b) modelling of excess deaths over the average
- but (a) has the problem that signers of death certificate might not know
- and (b) would include non-covid deaths caused by the lockdown
- so there needs to be very clear analysis of the numbers
- with people stating exactly what they are looking at, how the analysis is done, what the timeframes are, and their assumptions
- because without this we are in the dark about the current trajectory of infections

Are there any countries that are measuring in this detail? Can we learn from other countries and extrapolate accordingly?

Every aspect of countries' collection of death records will have variations over what is and isn't recorded. There'll need to be a huge benchmarking effort to make sure like is being compared with like. Simple differences might include:
- whether the death is byf or with Covid
- whether it was tested or is suspected
- was it extrapolated from a sample, or from exhaustive records
- was it confirmed by a coroner or doctor, or merely by relatives
- etc

The good news is that there are fewer patients in intensive care here every day, and recoveries are outstripping deaths.

The meh news is that the field hospital in the Fiera probably wasn't necessary (still better to have it and not need it).

The bad news is the indeterminate numbers of people left to die in care homes. The police are investigating. I've heard things like "not wearing masks because it scares the residents".

Cases were down today but so were tests.

So the government claimed, re transmission in hospitals that "good practice was being followed in terms of separating patients with coronavirus from those without. "

IT CAN BE PASSED ON ASYMPTOMATICALLY YOU f.cking tw.ts. IF YOU DON'T TEST EVERYBODY, THEN ASSUME LOTS OF NON-COVID PATIENTS HAVE COVID

Couple this with that stupid prick Matt Hancock's claim that the NHS were overusing PPE - as far as I can tell he means using it for non-covid patients (see above re: their actual covid status) - and it's going to rip through hospitals like a wildfire, killing doctors, nurses, porters, cleaners and patients.

lpm wrote: Tue Apr 14, 2020 11:57 am The reason we need detailed death statistics is because:<snip>

Yeah. The underlying problem is that you can't manage the situation at all well unless you know how and where it's spreading. You good get that data through an extensive testing regime, but (a) there aren't enough tests to do that and (b) they may not be reliable enough anyway.

So for gross numbers, we are using deaths, which ought to be reliable but evidently aren't. We've got the official daily numbers, but they have multiple problems: only hospital deaths, only where there was a positive Covid test. We've got the death certificate numbers, but they are even worse: far too slow to come out, cause of death is often vague and not based on an examination of the patient, etc. So both of these are underestimating the number of Covid deaths and confusing the picture about incidental deaths.

What we need is an ad-hoc website, where for every death certificate written, a form is filled in giving:
- cause of death as per certificate
- probability that death was directly due to Covid infection (so if the circumstances are suggestive but there hasn't been a test and the doctor hasn't been able to do a proper examination, a likelihood can be given)
- probability that the death was indirectly due to the situation, e.g. lack of medical care due to reprioritisation, effects of lockdown such as unexpected suicides/domestic violence, etc.

By probabilising the numbers we will get a much more accurate total, a better understanding of the various contributory causes to the increased number of deaths, and much faster results than the certificate stats.

meh, they gave up on a £1bn NHS IT project years ago. People want live streamed probabilistic data but in reality we have forms in triplicate, faxed to an underling who then manually types them into a database (probably one of those huge black screen green ink dos jobs)

EACLucifer wrote: Tue Apr 14, 2020 6:35 pm So the government claimed, re transmission in hospitals that "good practice was being followed in terms of separating patients with coronavirus from those without. "

IT CAN BE PASSED ON ASYMPTOMATICALLY YOU f.cking tw.ts. IF YOU DON'T TEST EVERYBODY, THEN ASSUME LOTS OF NON-COVID PATIENTS HAVE COVID

Couple this with that stupid prick Matt Hancock's claim that the NHS were overusing PPE - as far as I can tell he means using it for non-covid patients (see above re: their actual covid status) - and it's going to rip through hospitals like a wildfire, killing doctors, nurses, porters, cleaners and patients.

And as you are well aware, if one is parsimonious with disposable PPE, it stops being a protection and starts being a source of contamination.

Trump has announced he's pulling US funding from the WHO, basically because he's decided they're to blame for any problems America now has because obviously nothing is ever his fault.
https://www.bbc.co.uk/news/world-us-canada-52289056

What a prick.

EACLucifer wrote: Tue Apr 14, 2020 6:35 pm Couple this with that stupid prick Matt Hancock's claim that the NHS were overusing PPE - as far as I can tell he means using it for non-covid patients

Certainly, locally, staff are overusing PPE and wanting to wear full aerosol-generating procedure kit rather than the standard Covid-patient PPE. Speaking to them it's either because they are unaware of the differences or because they don't trust the advice.

Very, very dubious about this (pre-print) paper that links BCG vaccination with low mortality rates of COVID-19. Chances of correlation not equalling causation are quite high without a definitive mechanism. Plus, I'd like to see where the UK is on those charts...

The commander of the Islamic Revolution Guard Corps has unveiled a handheld device that he said could identify people infected with the coronavirus within 100m (330ft), Tasnim news agency reports.

"The basis of this device is to create a magnetic field based on a bipolar virus inside the device, so its antenna can focus on any place within a 100m diameter that is infected by the virus, and identify the infected place in five seconds," Maj-Gen Hossein Salami was quoted as saying.

He added that its accuracy was "80%”, but provided no evidence.

Experts say the only reliable way to test for Covid-19 currently is by taking a swab of the nose or throat, which is sent off to a lab to look for signs of the virus’s genetic material.

From BBC Live News page timed at 11.38

With a pic

This looks like a 'magnetic' version of that bomb detector with the empty box and two dowsing rods that the british bloke sold and demonstrated all over the world and got caught for.

This is/will be a disaster if they give people the all clear with this

tenchboy wrote: Wed Apr 15, 2020 10:49 am

The commander of the Islamic Revolution Guard Corps has unveiled a handheld device that he said could identify people infected with the coronavirus within 100m (330ft), Tasnim news agency reports.

"The basis of this device is to create a magnetic field based on a bipolar virus inside the device, so its antenna can focus on any place within a 100m diameter that is infected by the virus, and identify the infected place in five seconds," Maj-Gen Hossein Salami was quoted as saying.

He added that its accuracy was "80%”, but provided no evidence.

Experts say the only reliable way to test for Covid-19 currently is by taking a swab of the nose or throat, which is sent off to a lab to look for signs of the virus’s genetic material.

From BBC Live News page timed at 11.38

With a pic

This looks like a 'magnetic' version of that bomb detector with the empty box and two dowsing rods that the british bloke sold and demonstrated all over the world and got caught for.

This is/will be a disaster if they give people the all clear with this

How can the BBC even report that with only a token comment for "balance" at the end from "experts".

tenchboy wrote: Wed Apr 15, 2020 10:49 am

The commander of the Islamic Revolution Guard Corps has unveiled a handheld device that he said could identify people infected with the coronavirus within 100m (330ft), Tasnim news agency reports.

"The basis of this device is to create a magnetic field based on a bipolar virus inside the device, so its antenna can focus on any place within a 100m diameter that is infected by the virus, and identify the infected place in five seconds," Maj-Gen Hossein Salami was quoted as saying.

He added that its accuracy was "80%”, but provided no evidence.

Experts say the only reliable way to test for Covid-19 currently is by taking a swab of the nose or throat, which is sent off to a lab to look for signs of the virus’s genetic material.

From BBC Live News page timed at 11.38

With a pic

This looks like a 'magnetic' version of that bomb detector with the empty box and two dowsing rods that the british bloke sold and demonstrated all over the world and got caught for.

This is/will be a disaster if they give people the all clear with this

Oh, sh.t. It's like a sick prank. How do you suppose, all these years later, these pieces of magical junk can still be around and the leader of the Revolutionary Guards can be bamboozled into showing them off to the press? Heads may roll, in a very literal way. Maybe that's even the least-bad outcome.

shpalman wrote: Wed Apr 15, 2020 11:10 am

tenchboy wrote: Wed Apr 15, 2020 10:49 am

The commander of the Islamic Revolution Guard Corps has unveiled a handheld device that he said could identify people infected with the coronavirus within 100m (330ft), Tasnim news agency reports.

"The basis of this device is to create a magnetic field based on a bipolar virus inside the device, so its antenna can focus on any place within a 100m diameter that is infected by the virus, and identify the infected place in five seconds," Maj-Gen Hossein Salami was quoted as saying.

He added that its accuracy was "80%”, but provided no evidence.

Experts say the only reliable way to test for Covid-19 currently is by taking a swab of the nose or throat, which is sent off to a lab to look for signs of the virus’s genetic material.

From BBC Live News page timed at 11.38

With a pic

This looks like a 'magnetic' version of that bomb detector with the empty box and two dowsing rods that the british bloke sold and demonstrated all over the world and got caught for.

This is/will be a disaster if they give people the all clear with this

How can the BBC even report that with only a token comment for "balance" at the end from "experts".

Because, like most journalists, they are ignorant about basic science and thus haven't a clue as to how to report it so they fall back on the "he said, she said" narrative.

lpm wrote: Fri Apr 10, 2020 6:35 pm

The researchers were ordered to stop testing because the swabs were not collected for coronavirus, it was an ethical violation. Also the labs were not certified for that kind of testing.

That winds me up. Useful info chucked during a global emergency.

Certified makes sense - if they ain't certified, then it's not necessarily useful info. It's like the marseilles chloroquine study or that fraud who used to always return positive tests for MRSA for journalists.
But the ethical approval bit is f.cking stupid. If you're testing without consent for something like huntingdon's chorea, there you now know someone will develop a fatal degenerative illness in the future, then that's unethical (how can you tell them if they don't wanna know? How can you not tell them? you have added stress and a large amount onto life/health insurance costs), but for COVID-19, an illness that they either had or hadn't had and the only info from a personal level that will come out is either 'you're in the same boat as before - keep on social distancing' or 'good news, you had it so now you're immune', then there's nothing truly unethical in that - it's only box-ticking brainlessness to call it unethical

tom p wrote: Wed Apr 15, 2020 3:56 pm

lpm wrote: Fri Apr 10, 2020 6:35 pm

The researchers were ordered to stop testing because the swabs were not collected for coronavirus, it was an ethical violation. Also the labs were not certified for that kind of testing.

That winds me up. Useful info chucked during a global emergency.

Certified makes sense - if they ain't certified, then it's not necessarily useful info. It's like the marseilles chloroquine study or that fraud who used to always return positive tests for MRSA for journalists.
But the ethical approval bit is f.cking stupid. If you're testing without consent for something like huntingdon's chorea, there you now know someone will develop a fatal degenerative illness in the future, then that's unethical (how can you tell them if they don't wanna know? How can you not tell them? you have added stress and a large amount onto life/health insurance costs), but for COVID-19, an illness that they either had or hadn't had and the only info from a personal level that will come out is either 'you're in the same boat as before - keep on social distancing' or 'good news, you had it so now you're immune', then there's nothing truly unethical in that - it's only box-ticking brainlessness to call it unethical

What's particularly odd about using the consent angle to prevent looking at the samples for SARS-CoV-2 is that the Seattle Flu Study was presented to potential participants as an investigation into "flu" - not named influenza viruses.They asked for participants with "flu-like symptoms" such as "fever, cough, muscle or body aches, headaches, fatigue" and told participants that they were interested in "why certain flu strains are stronger and more dangerous than others".

They didn't refer specifically to "influenza" in the publicity materials - and they didn't name specific influenza viruses.

https://web.archive.org/web/20190527105 ... lu.org/faq

It could be argued, that testing to see whether someone's "flu-like symptoms" are caused by a coronavirus rather than influenza virus is really quite consistent with what participants were agreeing to be involved with.

If POTUS is happy to call COVID-19 a strain of "flu" surely it's not unreasonable to think that participants would think it's the sort of thing that might be tested for in a "flu" study.

FlammableFlower wrote: Wed Apr 15, 2020 10:36 am Very, very dubious about this (pre-print) paper that links BCG vaccination with low mortality rates of COVID-19. Chances of correlation not equalling causation are quite high without a definitive mechanism. Plus, I'd like to see where the UK is on those charts...

They say "BCG vaccination has been shown to produce broad protection against viral infections" and reference this review: https://www.ncbi.nlm.nih.gov/pubmed/31055165 [see quote below]

(They seem less confident there's a specific effect when they say this: "we also found that BCG vaccination was correlated with a reduction in the number of COVID-19 reported infections in a country suggesting that BCG might confer some protection specifically against COVID-19")

Non-specific effects of BCG vaccine on viral infections wrote:Numerous epidemiological, clinical and immunological studies demonstrate that BCG vaccination impacts the immune response to subsequent infections, resulting in reduced morbidity and mortality. Important lines of evidence indicating that BCG protects against viral pathogens comes from experimental studies in mice showing that BCG offers protection against various DNA and RNA viruses, including herpes and influenza viruses. Recently, the effect of BCG on an experimental viral infection in humans has been demonstrated. These effects are thought to be mediated via the induction of innate immune memory and heterologous lymphocyte activation, resulting in enhanced cytokine production, macrophage activity, T-cell responses and antibody titres.