Coronavirus: Follow virus advice or 'tougher measures' likely, says PM

[sasguru]

We could start with the last sentence here. Surely you can work out why the mortality rates as reported for covid-19 is much higher than the real figure? Have a think about it, and let me know if you need any help

I've already factored in the unknown denominator you thick twunt. Those are conservative figures. I've come to the conclusion that you're not the sharpest tool in the box.

[Whorty]

I've already factored in the unknown denominator you thick twunt. Those are conservative figures.

Show us your workings and assumptions then, rather than just posting soundbites and getting all hissy-fitty. Is this what you're like at work when you come up with some analysis and people want to know how you got there?

I've come to the conclusion that you're not the sharpest tool in the box.

Seems you're a bit slow on the uptake in many ways .... even Shauny beat you to this conclusion, years ago

[sasguru]

Show us your workings and assumptions then, rather than just posting soundbites and getting all hissy-fitty. Is this what you're like at work when you come up with some analysis and people want to know how you got there?

The 1% is a figure that most epidemiologists have arrived at as a reasonable figure, when all the dust has settled.
The WHO figure of 3.4% is too high.
In part based on analysis of Diamond Princess where the denom was known.
I've read several academic papers on this, I suggest you do the same.

[sasguru]

Seems you're a bit slow on the uptake in many ways .... even Shauny beat you to this conclusion, years ago

Yes I suppose your Remainer tendencies which blinded me were just random noise

[jamesbrown]

The 1% is a figure that most epidemiologists have arrived at as a reasonable figure, when all the dust has settled.
The WHO figure of 3.4% is too high.
In part based on analysis of Diamond Princess where the denom was known.
I've read several academic papers on this, I suggest you do the same.

Good article here on fatality rates and conditionalities (and uncertainties and unknowns):

https://www.cebm.net/global-covid-19...atality-rates/

[sasguru]

Good article here on fatality rates and conditionalities (and uncertainties and unknowns):

https://www.cebm.net/global-covid-19...atality-rates/

Hmmm their methodology seems complete arbitrary. IFR of 0.2% based on :

"Therefore, to estimate the IFR, we used the estimate from Germany’s current data 22nd March (93 deaths 23129) cases); CFR 0.40% (95% CI, 0.33% to 0.49%) and halved this for the IFR of 0.20% (95% CI, 0.17% to 0.25%) based on the assumption that half the cases go undetected by testing and none of this group dies"

Let alone the arbitrariness of simply halving CFR, there are severe problems of bias using just German death data (which has one of the lowest CFRs to date) - for one its health system probably has the most capacity of any system in the world (largest per capita no of ventilators and ICU beds)

[jamesbrown]

Hmmm their methodology seems complete arbitrary. IFR of 0.2% based on :

"Therefore, to estimate the IFR, we used the estimate from Germany’s current data 22nd March (93 deaths 23129) cases); CFR 0.40% (95% CI, 0.33% to 0.49%) and halved this for the IFR of 0.20% (95% CI, 0.17% to 0.25%) based on the assumption that half the cases go undetected by testing and none of this group dies"

Let alone the arbitrariness of simply halving, there are severe problems of bias using just German death data - for one its health system probably has the most capacity of any system in the world (largest per capita no of ventilators and ICU beds)

My understanding is that the IFR is, by definition, harder to calculate than the CFR precisely because it requires these assumptions about "true" infections.

Where have you seen better or less arbitrary assumptions as the basis for modeling? Genuine question, because this is not my area, but I take an interest from a statistical modeling POV.

[sasguru]

My understanding is that the IFR is, by definition, harder to calculate than the CFR precisely because it requires these assumptions about "true" infections.

Where have you seen better or less arbitrary assumptions as the basis for modeling? Genuine question, because this is not my area, but I take an interest from a statistical modeling POV.

Not my area either and for obvious reasons of speed not many peer reviewed papers out yet. Obviously everything is an informed guess but I think 0.5%-1% is very reasonable.

Some indications (albeit from an obviously biased sample: rich 50-somethings healthy enough to go on a cruise, but we have the denom for this unusually, so its a kind of controlled experiment):

Cruise ship outbreak helps pin down how deadly the new coronavirus is | Science News.

But the 1% is quoted by Whitty and I think the Imperial College team have his ear.

Also the S. Korea example is informative, but they had very sophisticated techniques to stop spread, this has CFRS:

https://www.ijidonline.com/article/S...150-8/fulltext

And most obviously, this which Ive posted elsewhere:

https://www.imperial.ac.uk/media/imp...16-03-2020.pdf

[d000hg]

I am surprised the police cannot disperse crowds.

One imagines all the people went out expecting it to be quiet. The thing is you CAN have a heck of a lot of people outside maintaining 2m distance (over 1000 can fit in a football pitch!) but you inevitably get pinch points as they enter and exit.
A looped walk could be very busy if you could stop people bunching in the car-parks.

Best is just find your nearest footpath, not some national beauty spot.

[jamesbrown]

Not my area either and for obvious reasons of speed not many peer reviewed papers out yet. Obviously everything is an informed guess but I think 0.5%-1% is very reasonable.

Some indications (albeit from an obviously biased sample: rich 50-somethings healthy enough to go on a cruise, but we have the denom for this unusually, so its a kind of controlled experiment):

Cruise ship outbreak helps pin down how deadly the new coronavirus is | Science News.

But the 1% is quoted by Whitty and I think the Imperial College team have his ear.

Right. The discrepancy between CFR and IFR should reduce as the testing becomes more expansive, but 0.2% does seem low. Either way, it's the impact of comorbidities (conditional rate) that stand out to me, albeit even harder to estimate, and I don't know if they're particularly different for other SARS-like CVs.