Worrying times

[JustinHorton]

People may find this useful if technical in places, click on boxes for some other charts (also, may take some time to load)

[Matt Mackenzie]

The "original" SARS had a death rate of 10% or more, no?

[Daniel Gormally]

We'll only know what the true death rate is when they instigate mass testing, of millions or more.

until then anything is speculation, although most of the evidence we have suggests it's nothing like 10 percent.

I think that's the problem with the virus- people don't really know where they stand. there is a lot of fear of the unknown.

[Paul McKeown]

Yes, but this is not disease is not the "original" SARS. The death rate of this disease is still subject to research by the qualified, but despite wild assertions here and no doubt elsewhere, it is not 10%. I do not wish to minimise the risks, but lurid scaremongering is not useful either. There are various estimates based on different data and different statistical methods by different authors in peer reviewed publications, all of which are easy to pursue in these times of enforced leisure. The consensus seems to be that when all the data is in, the death rate will be about 1% of those infected, with initial estimates of 2.5% or 3.5% probably reflecting lack of testing. A death rate of 1% is still an enormous toll, of course, and a fair justification for governments wreaking havoc upon their economies. This disease can be viewed as more dangerous than SARS, though, as although its death rate is lower than SARS, it appears to be much more transmissible.

[Ian Thompson]

The consensus seems to be that when all the data is in, the death rate will be about 1% of those infected, with figures of 2.5% or 3.5% probably reflecting lack of testing. A death rate of 1% is still an enormous toll, of course

The point has also been made that many of the victims are in poor health and wouldn't have lived much longer even if there had been no COVID-19. Perhaps the more relevant figure is how many people will die who wouldn't otherwise have done.

I believe figures are calculated for seasonal flu deaths that do allow for this.

[Christopher Kreuzer]

The technical terms, I think, are: Case Fatality Rate (CFR) and Infection Fatality Rate (IFR). Not sure what the term is for the adjusted mortality rates.

On the impact of lockdown across Europe, this is an interesting overview:

https://www.bbc.co.uk/news/world-europe-52025553

[NickFaulks]

The consensus seems to be that when all the data is in, the death rate will be about 1% of those infected

I shall be very surprised if it is as high as 0.1%, if properly calculated. By that I mean that the numerator should be, as Ian says above, the number of deaths of people who would not have died anyway and the denominator should be the number of people who have been infected, even if they showed unconvincing symptoms or none at all.

In Italy they are now testing post mortem. That may well be a useful thing to do, adding to the stock of knowledge. but the death rate gleaned from that group can never be anything but 100%.

[Kevin Thurlow]

Newspaper reports are always suspect, but at least this one has links to original papers (well done). If they're right, we should rush out and get estrogen.

https://www.theguardian.com/world/2020/ ... us-but-why

[JustinHorton]

The consensus seems to be that when all the data is in, the death rate will be about 1% of those infected

I shall be very surprised if it is as high as 0.1%, if properly calculated. By that I mean that the numerator should be, as Ian says above, the number of deaths of people who would not have died anyway and the denominator should be the number of people who have been infected, even if they showed unconvincing symptoms or none at all.

Well presumably the correct criteria in these categories are the same voices epidemiologists normally use?

[NickFaulks]

Well presumably the correct criteria in these categories are the same voices epidemiologists normally use?

I should not have written "if properly calculated". I meant "if calculated in an ideal way". Epidemiologists, like others trying to model social science, do the best they can with imperfect data.

For the numerator, they would use whatever it says on the death certificate. In different societies there may in cases of doubt be pressures either to stress a particular cause or to downplay it. For the denominator, they would use reported cases because that's all they've got.

I did see a piece of modelling which tried to estimate how many people might already be immune - unfortunately the conclusion was something like "between 10,000 and 10,000,000", which presumably shows how difficult it is.

[Daniel Gormally]

https://english.alarabiya.net/en/featur ... o-symptoms

[NickFaulks]

That's a good find!

[Daniel Gormally]

every blind squirrel finds a nut....

[David Robertson]

https://english.alarabiya.net/en/featur ... o-symptoms

This is complete and utter bo**ocks.

This is what the Government of Iceland says

A website of no known (and doubtful?) provenance nevertheless adds this:

Not all the results from Iceland’s tests have come through yet, but the ones that have, show that half of all cases are asymptomatic (at the time of testing). These results are indicated by a testing survey carried on an entire Italian (sic!!) town of Vo (population 3,300), where the results showed that more than 50% of all cases are asymptomatic.The whole population of the village was tested, and 3% of the residents tested positive. Then, after a two-week lockdown, the population was tested again. The transmission had been reduced by 90% and the results were confirmed: the majority of cases seem to be asymptomatic.

f**k me!

[NickFaulks]

Hard luck Danny, you don't get the approval of The Master. Still looks interesting to me.