Number of Deaths

[Machiavelli]

But......until we have a vaccine, there isn’t much more we can do than gradually head towards herd immunity....? There is no ‘cure’, only treatment of symptoms and intervention sufficient to allow the body to fight back - and, through lockdown, simply stopping it spreading.

The late timing of lockdown undoubtedly increased the spread enormously (dread to think how many cases Cheltenham led to) but any easing of lockdown will be a calculated punt on developing herd immunity vs not overwhelming the NHS with a spike in cases.

There is a good way of getting herd immunity: via a vaccine. The idea that we gradually head towards it without a vaccine (which is the other way of getting there) will result in 100,000s of deaths.

 

[Harry Wilson's tackle]

Please don't think me intentionally abrasive by the way. I rate you highly as a poster.

My understanding is that in at least a couple of the cases (Iceland) it is an antibodies test, not a "do you have it now" test, but again could be wrong.

Not at all!

OK, I just found a really interesting paper, big meta analysis of huge amounts of data. The suggestion is 16% of ordinary flu victims are asymptomtic, but the provenance of the data and variations between the published studies means that the real value could be as high as 28% . . . or as low as 4%.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/
The fraction of influenza virus infections that are asymptomatic: a systematic review and meta-analysis

So even with traditional flu there is some uncertainty, albeit most of the sources of error are likely to push the estimated value up rather than down. A bit like data on credible performance of magic - there are many examples, and the mean value is more than zero, yet we know the true value is zero (the work of James Randi).

Sorry if I come across as pompous and pedantic - it's because, unless I edit carefully, I am

 

[Mellotron]

Not at all!

OK, I just found a really interesting paper, big meta analysis of huge amounts of data. The suggestion is 16% of ordinary flu victims are asymptomtic, but the provenance of the data and variations between the published studies means that the real value could be as high as 28% . . . or as low as 4%.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4586318/
The fraction of influenza virus infections that are asymptomatic: a systematic review and meta-analysis

So even with traditional flu there is some uncertainty, albeit most of the sources of error are likely to push the estimated value up rather than down. A bit like data on credible performance of magic - there are many examples, and the mean value is more than zero, yet we know the true value is zero (the work of James Randi).

Sorry if I come across as pompous and pedantic - it's because, unless I edit carefully, I am

Not at all. You clearly understand a lot of this a lot better than me, particularly reading data.

 

[Blue Valkyrie]

Any ideas why the UK with 84,000 cases has had 10,600 fatalities yet Germany with 127,000 cases has had only 2,900 fatalities ?

What are we doing wrong and what are they doing right ?

Idea : Germany used to be East and West Germany.

Did East Germany use a Soviet TB vaccine that has made a difference ?

 

[Machiavelli]

I'm assuming Germany are only counting official tests done, much like us and other European countries - i.e. phone diagnoses won't count towards totals, I'd be very surprised if they did but willing to be put right if so.

Re: Asymptomatic - Have you seen the studies from Iceland (50% asymptomatic, based on mass testing of the entire population), https://edition.cnn.com/2020/04/01/europe/iceland-testing-coronavirus-intl/index.html

Italy - case study of a town of 3,000 called Vo all being tested, suggesting 75%-80% were asymptomatic

and (less reliable but) China (paper published that suggested 78% asymptomatic across Wuhan)?

That is exactly what they have done in a number of places - tested people showing no symptoms. My 50-80% is based on the higher and lower figures from these studies. Quite a bit of reference to these studies in the Coronavirus Good News Thread. Of course they do not prove anything in and of themselves yet...but they are starting to show more promising trends, and it absolutely should be factored into our thinking.

Another example that I'm still looking for the source article on, which was in the Good News Thread - in Italy they took blood from 60 people in the Bergamo area who were all completely asymptomatic and believed themselves to have not had the virus. The antibodies count showed that 40 of them had had COVID-19.

I remain sceptical of the asymptomatic rates that you're quoting. I might be wrong, but it's too early to say. That Vo study is tiny, and I had a look at that Iceland report, and it reads more like a sales pitch for a company. Only 5% of Iceland's population have been properly tested, according to that.
But, hopefully we can agree that all of this shows just how crucial the mass production of tests (antibody in this instance) will be before we know how to start easing/lifting this lockdown.

 

[Harry Wilson's tackle]

There is a good way of getting herd immunity: via a vaccine. The idea that we gradually head towards it without a vaccine (which is the other way of getting there) will result in 100,000s of deaths.

Unless the current declared number of COVID cases in the UK and Germany are both out by a factor of ten (due to this much hoped-for existence of symptomless infection in, well, 90% of those infected), to hit 65% 'exposure' by the herd immunity route, there will be between one and five million deaths (see my calculation elsewhere). With that sort of scenario, you'd expect to find most of the world in serious lock down until we have a vaccine.

Oh.

 

[vegster]

Idea : Germany used to be East and West Germany.

Did East Germany use a Soviet TB vaccine that has made a difference ?

Yet Russia have 18,000 cases yet 2,500 fatalities ? That's that theory up in smoke !

 

[Blue Valkyrie]

Yet Russia have 18,000 cases yet 2,500 fatalities ? That's that theory up in smoke !

Putin's figures are fantasy.

And according to worldometers the declared figure is 148 deaths anyway.

 

[Harry Wilson's tackle]

Idea : Germany used to be East and West Germany.

Did East Germany use a Soviet TB vaccine that has made a difference ?

No. Very different epitope. Different species (one is a bacterium, the other a virus). A bit like expecting eating fish to cause someone to become allergic to cats.

 

[vegster]

Putin's figures are fantasy.

And according to worldometers the declared figure is 148 deaths anyway.

Ignore my post, quoted the wrong column ! Russian vaccine still on !

 

[Blue Valkyrie]

No. Very different epitope. Different species (one is a bacterium, the other a virus). A bit like expecting eating fish to cause someone to become allergic to cats.

The article I read :


https://www.reuters.com/article/us-health-coronavirus-tbvaccine-explaine-idUSKBN21K372

There are many similar articles.

At least scientists are actively investigating.

 

[Green Cross Code Man]

The article I read :


https://www.reuters.com/article/us-health-coronavirus-tbvaccine-explaine-idUSKBN21K372

There are many similar articles.

At least scientists are actively investigating.

That's very interesting. However if it was used since the 1920's, it isn't proving very effective for older people is it?

 

[Blue Valkyrie]

That's very interesting. However if it was used since the 1920's, it isn't proving very effective for older people is it?

I don't have the answers, but they do propose that it depends on the "strain"? of TB vaccine used - there is not just one it seems.

 

[RossyG]

That's very interesting. However if it was used since the 1920's, it isn't proving very effective for older people is it?

It wasn’t widely used in the U.K. until 1953. That still means that most people from 30-80 would’ve had it, though.

 

[dazzer6666]

There is a good way of getting herd immunity: via a vaccine. The idea that we gradually head towards it without a vaccine (which is the other way of getting there) will result in 100,000s of deaths.

So.....lockdown for an unspecified period, but according to the majority of experts 12-18 months then ?

 

[beorhthelm]

re BCG/TB vaccine, there is evidence the vaccine give the immune system a general boost that may help with all virus and infection. its not supposed to taken as a possible therapy for coronavirus.

 

[RossyG]

So.....lockdown for an unspecified period, but according to the majority of experts 12-18 months then ?

There’s no guarantee that a vaccine will be around even then. It took almost 30 to get one for HIV and a decade for one for SARS.

It might be herd immunity or nothing.

 

[Harry Wilson's tackle]

I remain sceptical of the asymptomatic rates that you're quoting. I might be wrong, but it's too early to say. That Vo study is tiny, and I had a look at that Iceland report, and it reads more like a sales pitch for a company. Only 5% of Iceland's population have been properly tested, according to that.
But, hopefully we can agree that all of this shows just how crucial the mass production of tests (antibody in this instance) will be before we know how to start easing/lifting this lockdown.

Yes. These studies are not reliable.

I am slowly developing a postition on this that, like yours, is based on skepticism, and using the only global data set that is comprehensive (even if unreliable), the BBC daily set.

It is very hard for people not used to reading original research publications (as opposed to editorials, or even lay summaries, or newspaper articles) to be able to smell the coffee from the bullshit.

Here's the thing. In hospitals 99% of what goes on between doctor and patient is the delivery of healthcare. Collecting data for scientific purposes is not high on the agenda. It isn't done all that well in most cases even when it is top of the agenda (testing a new drug, for example).

So what we have here is a hospital agenda focused on making a clinical judgement about whether to hospitalize based on symptoms, not lab testing, and in which diagnostic lab testing is ordinarily done for selecting treatment not for collecting data for governments. This is the same in the UK, Germany, India, America, you name it. Furthermore, there is little point a doctor lab testing to prove COVA infection right now because it would not affect treatment (we have no 'treatment' for this virus, only a pathway to alleviate symptoms - so-called 'palliative intervention') and diagnosis would be of value to the patient only if it could prove the infection were something else, i.e., bacterial, that could be treated (with a specific antibiotic). So the data we get for modelling is always going to be 'poor'. But would it be too poor to assist in planning? No, I wouldn't say that.

My first hypothesis is that the poverty of the data has a systematic bias element so, when it comes to estimating how many people have had the virus so that the rate of death (that number divided by the number of bone fide COVA deaths) can be calculated is concerned, although it will always be a bit 'back of the envelope', we can expect a systematic (and therefore correctable) bias based on an assumption that the total number of proven cases is an underestimate, and the number of recorded COVA deaths is an underestimate. But by how much?

Some are modelling based on an assumption that the number of reported cases is only 10% of the total, due to a large number of asymptomatic folk who never become 'patients'. Based on data I have seen, and equivalent (but MUCH better) data for ordinary flu, I would expect this to be way out - and the real value closer to 80-90%. In other words the number of reported cases (in countries other than the big fat liars like Russia) are almost certainly no more than a 50% underestimation, and probably much less than that (i.e., pretty reasonable, on a log scale).

Meanwhile what about the number of reported COVA deaths? Might this be an overestimate or an underestimate? With death certification out of hospital based on clinical diagnosis (not 'testing'), and out of hospital data in the UK accurate only once every few days, I nevertheless doubt the true values are wrong by no more than 20% (plus or minus). Radio 5 just stated that out of hospital values are not part of the total mortality data. So there we are then, officiel death numbers in the UK will be systematically underestimated. By how much I don't know.

With total numbers of cases in the UK likely underestimated, but the true values no more than double the numbers reported, and with total COVA deaths greater than those reported, my feeling is that in the UK, Germany, France, Italy and Spain the number of deaths per cased (1 death in 8 cases) is about right for now.

But will it (the rate of death in those infected) get bigger (less 'lethal') in the next few weeks?. I'll have to think about that.

 

[Machiavelli]

So.....lockdown for an unspecified period, but according to the majority of experts 12-18 months then ?

I don't quite follow what you're asking in this post, but I'll go with what I think you're saying.
I suspect that we'll have a much more relaxed form of lockdown until the vaccine arrives. This might entail easing the lockdown in certain areas (eg Brighton & Hove has a relatively low total of confirmed cases, currently 213) and/or for certain sectors (eg construction returns to work) and/or demographic profiles (eg 70yo+ no, or strictly limited, human interaction).
All of this lifting will be facilitated by the production (or purchase) of mass tests (both antibody and to see if you have the virus), and it's this that is key to minimisation of economic disruption. Unfortunately, we're looking poor on this front. Better news is the app the government is considering which will constitute a drastic infringement of civil liberties (and will generate a lot of hostility), but will enable economic activity to pick up, and the lockdown to be eased, without mass transmission.

Edit: this is a much better summary of the options ahead:

https://www.theguardian.com/comment...-the-coronavirus-crisis-end-lockdown-pandemic

 

[WATFORD zero]

...