Coronavirus Sars-CoV-2/Covid-19 Megathread

[Sheilbh]

First case in France traced back to December 27th. Maybe there's more to this nasty flu going around in January/February after all.

The really striking thing is he hadn't been to China and had no obvious link to China - the strongest link from what I've seen is that his wife (who was not infected) worked at a sushi stall with Chinese members of staff - which suggests there was already a level of community transmission in France in December.

I also find the WHO's comment that "this is not surprising" a bit rich, given that they were only notified by China of this virus on 31 December

[DGuller]

Mutation has been detected that causes the virus to spread more quickly. It also directly affects the mechanism which vaccines in development are aiming to disrupt, so this likely delays that effort. 

Long and interesting article at the LA times.

https://www.latimes.com/california/story/2020-05-05/mutant-coronavirus-has-emerged-more-contagious-than-original

Well, that sucks.  This might show that China's numbers are not cooked after all, they just had a more containable virus to deal with.

[Josquius]

See there's a bit social media campaign for people to not use the NHSX trace-and-track app. The main line seems to be that the UK is going it alone and that the app is actually a secret Dominic Cummings' data harvesting project. I swear to God Brexit broke the minds of our middle class - we're going to be in lockdown forever at this rate

It's quite something the number of people who, I imagine, use Google Maps every week who are swearing they'll never touch this "dodgy, data-harvesting app". Also it's bullshit that the UK is going it alone - there is a divide and there are issues on both sides - between centralised and decentralised models. The UK and France are taking a centralised (as was Germany until about a week ago), other countries are lining up behind the Apple and Google decentralised approach. Needless to say I am dubious about the great privacy enhancing model being developed by *checks notes* Google

Edit: Actually I just checked Twitter and Cambridge Analytica (and for some reason Bobby Sands) is trending. Brexit broke us

I haven't seen any campaign. But I trust google more than I trust the tories. And as you say they already know where I'm going and who else is there.
I'm torn on whether I'll download it or not. Its certainly really dodgy that they don't just use the off the shelf solution. Might put it on my work phone only.

[Sheilbh]

There is no off the shelf solution. The Apple-Google project is still very much in development, the NHSX App (and I believe the similar one the French are launching - StopCovid) is something that is ready for testing.

The other big reason is I say there are privacy and security issues with both approaches. The French/UK argument for a centralised model is that model also allows for public health authorities to see where outbreaks are developing, and that the mechanism is overall more secure and protective of people's information. The decentralised argument is that its structure makes it more secure and protective, in part precisely because it doesn't allow public health authorities to see outbreaks.

And where do you think the Tories got people's information from if it wasn't things like the entire adtech industry that Google's developed over the past 10 years

[crazy canuck]

I don't think that's accurate. Our first case was in January.

When was Italy?

And Washington shut the state down fairly quickly. Their numbers have been flat for a couple of weeks now. New York City, however, got a later start on getting the virus and shutting things down.

We need to stop treating the US as a monolith. Every state has handled things differently, and it shows. That should be accounted for.

To illustrate your point further look at the difference between Ohio and Michigan

https://www.nytimes.com/2020/05/05/opinion/coronavirus-ohio-amy-acton.html?action=click&module=Opinion&pgtype=Homepage

Also, 200k new cases and 3000 deaths per day by June due to states opening up too early.

https://www.nytimes.com/2020/05/05/us/coronavirus-updates.html?action=click&module=Spotlight&pgtype=Homepage#link-26b9256b

[Josquius]

And where do you think the Tories got people's information from if it wasn't things like the entire adtech industry that Google's developed over the past 10 years

IIRC it was some facebook quiz that was the primary source of their harvest?

There is no off the shelf solution. The Apple-Google project is still very much in development, the NHSX App (and I believe the similar one the French are launching - StopCovid) is something that is ready for testing.

However its in development as a standard.
This seems the best way to go. What of for instance border communities having to use two different apps? One would be more sensible.

The other big reason is I say there are privacy and security issues with both approaches. The French/UK argument for a centralised model is that model also allows for public health authorities to see where outbreaks are developing, and that the mechanism is overall more secure and protective of people's information. The decentralised argument is that its structure makes it more secure and protective, in part precisely because it doesn't allow public health authorities to see outbreaks.

I'm pretty sure I recall reading the google/apple one though mostly working locally will also share basic information with the government?
TBH its not something I've paid overly much attention to.

[jimmy olsen]

I don't think that's accurate. Our first case was in January.

When was Italy?

And Washington shut the state down fairly quickly. Their numbers have been flat for a couple of weeks now. New York City, however, got a later start on getting the virus and shutting things down.

We need to stop treating the US as a monolith. Every state has handled things differently, and it shows. That should be accounted for.

To illustrate your point further look at the difference between Ohio and Michigan

https://www.nytimes.com/2020/05/05/opinion/coronavirus-ohio-amy-acton.html?action=click&module=Opinion&pgtype=Homepage

Also, 200k new cases and 3000 deaths per day by June due to states opening up too early.

https://www.nytimes.com/2020/05/05/us/coronavirus-updates.html?action=click&module=Spotlight&pgtype=Homepage#link-26b9256b

If there are 200k cases a day we're going to be seeing 12k dead a day, not 3k.

[Syt]

https://twitter.com/BBCNews/status/1256205212248870913

5 Royal Mail postboxes have been painted blue in honor of the NHS.

[Legbiter]

The number of cases and deaths per country have so little correlation with severity of quarantine measures (and even population size) that I doubt they're of much use at all.

It's maddening yeah. Sweden should have lost over 40.000 people today according to the first Imperial College model. They've lost around 2.700 and peaked 2 weeks ago.

[Tamas]

The number of cases and deaths per country have so little correlation with severity of quarantine measures (and even population size) that I doubt they're of much use at all.

It's maddening yeah. Sweden should have lost over 40.000 people today according to the first Imperial College model. They've lost around 2.700 and peaked 2 weeks ago.

Do we know their excess death statistics?

[Sheilbh]

The number of cases and deaths per country have so little correlation with severity of quarantine measures (and even population size) that I doubt they're of much use at all.

It's maddening yeah. Sweden should have lost over 40.000 people today according to the first Imperial College model. They've lost around 2.700 and peaked 2 weeks ago.

Yeah - still higher than their neighbours but we don't know all-cause excess mortality and we don't know over the life of this pandemic.

Similarly Japan - limited testing, no lockdown, elderly population and very few deaths. I think it's really tough to draw really obvious conclusions at this point until we look at the end of it - for example AR has flagged that density may be an issue which intuitively makes sense, but another factor is probably underlying comorbidities and age profile. I don't know how we start to untangle all of these for different countries.

[Sheilbh]

The number of cases and deaths per country have so little correlation with severity of quarantine measures (and even population size) that I doubt they're of much use at all.

It's maddening yeah. Sweden should have lost over 40.000 people today according to the first Imperial College model. They've lost around 2.700 and peaked 2 weeks ago.

Do we know their excess death statistics?

This is the latest I've seen (being updated with new UK numbers today):

[Legbiter]

The number of cases and deaths per country have so little correlation with severity of quarantine measures (and even population size) that I doubt they're of much use at all.

It's maddening yeah. Sweden should have lost over 40.000 people today according to the first Imperial College model. They've lost around 2.700 and peaked 2 weeks ago.

Do we know their excess death statistics?

Yeah, it's considerably above normal but they're not dumping them into mass graves or anything and they hardly had to use those military field hospitals they set up in the beginning.

Here the overall death rate is considerably lower than usual because our oldies were successfully shielded (so far) and there's been a big drop in respiratory infection rates. Also since traffic volume fell by 40% over the last couple of months we don't have the usual background traffic fatalities as well.

[Sheilbh]

Yeah on all cause there are odd effects - huge decline in accidents everywhere, probably an impact on lower pollution. In the UK we pushed people out of intensive care units and into care homes which is probably partially why the non-covid deaths have increased (some of those people being pushed out weren't tested and this was probably a vector for covid getting into care homes), there's a big decline in non-covid deaths in hospitals and a big increase in care homes from the average. I find it interesting that in the UK flu/pneumonia deaths aren't increasing which seems to indicate that covid is being identified (even when not tested) rather than seeing other respiratory conditions kick up. Plus, in the UK, the cancelling of routine out-patient care and people not going into A&E when they should.

I mean it'll vary everywhere and be difficult to unpack. But the biggest example of that I've seen is South Africa where accidental and violent deaths are down so significantly that it basically balances out the covid deaths and they're not seeing any change to their average mortality.

Edit: Incidentally looking at government slides it seems that covid diagnoses in hospital (including people diagnosed pre and post admission) was below 1,000 yesterday for the first time since March

Edit: One other quick thought on the Imperial model is I don't know how that's been applied to Sweden. It's not a statistical model, it's based on the data around how contagious the disease and how that leads to hsopitalisations etc and a set of assumptions about behaviour. So, for example the UK is actually more compliant with the various suppression actions than the model assumed. In Sweden's case they are not going through the unmitigated ICL model because they are socially distancing quite a lot so just looking at the CityMapper mobility index, Stockholm is about 30% of normal I imagine in certain settings that's even more pronounced. From my understanding the better comparison wouldn't be Sweden unmitigated but probably one of the flatten the curve scenarios which assume a significant degree of actions to reduce the R but not as many as a suppression strategy.

[Legbiter]

Seems like the UK playbook for getting out of lockdown is shaping up.

Employers will be told to close canteens and other communal areas, with staff encouraged to eat packed lunches on their own.

Shift patterns will be staggered to limit the number of people at work at any one time and reduce rush-hour pressure on public transport. Access to communal facilities like photocopiers will be limited, areas of the floor will be taped off to keep people two metres apart and people could be barred from sitting opposite one another.

Hot desking will be banned and workers will not even be allowed to lend each other a pen for fear of spreading the virus.

As a result of the stringent measures, millions of staff will be told to keep working from home indefinitely 'if at all possible'.

And firms have been told that 'vulnerable' people, including the over-70s, pregnant women and the severely obese, should work from home – even if this means finding them a new role.

The government is believed to have warned industry groups that the restrictions could stay in place for six to 12 months - dashing hopes of a swift return to normal life.

https://www.dailymail.co.uk/news/article-8286345/UKs-work-blueprint-Hot-desking-banned-no-sharing-pens-canteens-closed.html