Coronavirus outbreak

[roubaixtuesday]

It's obvious that more social interaction = more viral transmission.

So trivially, "Eat out to help Out" subsidised increased transmission of the virus.

Equally, quantification of this is difficult at best, and probably impossible, as the Warwick study attempt notwithstanding there are huge uncertainties.

It was, however, part of a pattern of behaviour of attempting to play off economic benefits with viral disbenefits that was widely challenged at the time and culminated in the catastrophic decision to keep shops restaurants and schools open until the U turn one day into the new school term.

 

[Ajax Bay]

many of the facts fall on deaf ears

I think encouraging the public via a £10 bribe to get them to eat-in to help-out was a fool hardy idea... but that's what happens when Boris follows the science. Any thoughts on the other other points I raised regarding Boris' decisions?

Would you like to curate a few 'facts' for us, so we can be clear whose ears may or may not have been 'deaf'? The complete disconnect between cases (not) rising during the whole period of 'Eat out to help out' has been pointed out upthread. Edit: Yet I agree with @rt about the affect on behaviour modes into the autumn when the tiered system failed to work as expected, in significant part because of the gradual dominance of the B.1.1.7 variant with its 40% greater transmissibility.
Hope you're not going to let the Duke of Edinburgh's demise affect your opportunity for a warm beer in your local pub garden this evening. Our Wednesday cycle ride-to-the-pub drink out to socialise will be enjoying the evening sunshine on, er, Wednesday.

 

[roubaixtuesday]

The complete disconnect between cases (not) rising during the whole period of 'Eat out to help out' has been pointed out upthread.

Cases rose throughout last summer, from the moment restrictions were reduced to their minimum.

People seem to insist the opposite frequently, I've no idea why.

 

[mjr]

Hope you're not going to let the Duke of Edinburgh's demise affect your opportunity for a warm beer in your local pub garden this evening. Our Wednesday cycle ride-to-the-pub drink out to socialise will be enjoying the evening sunshine on, er, Wednesday.

My most local pub isn't reopening, as it turns out. I expect we'll discover which others have collapsed over the next few weeks.

The next nearest is open but it's one of few (it's still low holiday season here, it seems) and only has a small garden so I am skipping it today because I expect it to be full and queuing. The one after that opens Friday but requires booking (so is double useless for tomorrow's cycle ride to the pub) and the next is not reopening until the May date (I guess they feel their garden isn't enough to cover costs). In another direction, I've heard nothing from two about reopening, then the third reopened at 5pm but is another one that requires table booking.

I doubt whether all these clunky booking systems that kill opportunistic pub stops will be enough to keep them all afloat, but let's wait and see.

 

[mjr]

I've seen it suggested that the biggest things that change by postponing locking down (as Boris has twice and Macron has, too, among others) is that you get more deaths and more economic damage. If a country (or even large enough region) uses short, sharp lockdowns then it may need more lockdowns and ultimately spend the same total time in lockdown because, basically, you have to spend the same amount of time lowering the rate as letting it rise, but total deaths is the area under the curve and so a series of low waves means fewer total deaths than a big hump.

I'm stuck at work trying to fix an (American, I think) server so I don't have time to test that on a spreadsheet but it sounds plausible. A quick search found a paper showing correlation of slow lockdowns and high deaths, but not something modelling different lockdown strategies in an easy-to-see graphical format suitable for a discussion like this. Has anyone else seen one?

And just to remind, WHO says lockdowns should be short and sharp but aren't the core of a good covid control, which should be personal hygiene, testing, contact tracing and quarantining of infected people. I've not seen the latest data on hygiene compliance and it looks like the UK is finally getting on top of testing, but the other two bits are still fails, aren't they?

Edit: OK, I just looked at the yougov hygiene surveys: extra hand-cleaning is down to 51%, avoiding crowds down to 60%, mask-wearing down to 71%, home-working down to 20% and avoiding handling public objects down to 40%. All of those are lowest or second-lowest scores since the first wave, except for home-working which blipped lower during Gove's go-to-work-to-save-Pret effort. Do people want to stress-test the vaccines or something?

 

[Ajax Bay]

Effective contact tracing and directed but unenforced domestic/home quarantining is dependent on population discipline, motivation and belief in their value (and individual economic circumstances). In our cosmopolitan and individualist society (urbanites: I'm looking at you) this has not been a UK strength. Nor has the UK's failure to protect care home residents, where so many have died before their time.
On hygiene: is there now consensus that fomite transmission is far less a factor than aerosol? I'm still spraying an wiping the supermarket trolley handle and hand sanitiser before and after shops, but that's about it.

 

[mjr]

Effective contact tracing and directed but unenforced domestic/home quarantining is dependent on population discipline, motivation and belief in their value (and individual economic circumstances). In our cosmopolitan and individualist society (urbanites: I'm looking at you) this has not been a UK strength.

I would have put those brackets the other way round, with "individual economic circumstances" as the primary reason many UK citizens aren't self-isolating and "population discipline, motivation and belief" as a parenthetical note, based on this commentary from the BMJ Blog https://blogs.bmj.com/bmj/2021/04/0...n-more-important-than-we-previously-realised/

Government should pay low-paid workers enough to isolate, in the interests of public health. The introduction of the £10,000 fine just deterred them from taking tests. The current situation is sick, bordering on class warfare.

On hygiene: is there now consensus that fomite transmission is far less a factor than aerosol? I'm still spraying an wiping the supermarket trolley handle and hand sanitiser before and after shops, but that's about it.

I don't know but while that might make the hand-washing drop less worrying, there's still those reductions in mask use, avoiding crowded places and teleworking which matter either way.

 

[Wobblers]

And of course that would be completely correct............well, if it wasn't totally wrong.

View attachment 583483

The data is not saying what you think it is.

At the beginning of August, the infection rate was very low - less than a thousand new infections per day. Deaths were in single digits per day. In order to see any effect on the ONS graph, the death raate would need to be at 100-200 times greater. At the time, and well into the autumn, the infection rate was doubling every week. That implies seven doubling periods (2^7 = 128). It would take 7 weeks for that to happen. Furthermore, it typically takes 4 weeks from infection until death, on average. Therefore, if Eat Out to Help Out had an effect on the fatality rates, that would be expected to appear 11 weeks after it commenced.

Eat Out to Help Out started on 3rd August. From the above discussion, you'd expect a significant spike in deaths to occur 11 weeks later in mid October. That is precisely what we see. A significant jump in fatalities is seen in the third week of October. The correlation is remarkably good. Were I to uncover similar such data in the course of my research, I'd be very pleased indeed. Come to think of it, I got surfactants into laundry products based on evidence that was less impressive than that!

In hindsight, Eat Out to Help Out was ill advised. It is unlikely to have helped, and looking closely at the data strongly suggests that it was, at the least, a significant contributor to the large increase in cases we experienced last autumn.

 

[PK99]

The data is not saying what you think it is.

At the beginning of August, the infection rate was very low - less than a thousand new infections per day. Deaths were in single digits per day. In order to see any effect on the ONS graph, the death raate would need to be at 100-200 times greater. At the time, and well into the autumn, the infection rate was doubling every week. That implies seven doubling periods (2^7 = 128). It would take 7 weeks for that to happen. Furthermore, it typically takes 4 weeks from infection until death, on average. Therefore, if Eat Out to Help Out had an effect on the fatality rates, that would be expected to appear 11 weeks after it commenced.

Eat Out to Help Out started on 3rd August. From the above discussion, you'd expect a significant spike in deaths to occur 11 weeks later in mid October. That is precisely what we see. A significant jump in fatalities is seen in the third week of October. The correlation is remarkably good. Were I to uncover similar such data in the course of my research, I'd be very pleased indeed. Come to think of it, I got surfactants into laundry products based on evidence that was less impressive than that!

In hindsight, Eat Out to Help Out was ill advised. It is unlikely to have helped, and looking closely at the data strongly suggests that it was, at the least, a significant contributor to the large increase in cases we experienced last autumn.

Interesting correlation - but as we know, correrlation and causation are not the same thing.

This international comparison over the time frame you consider shows a similar pattern across a number of neighbouring EU countries. The UK data on deaths over the period does not stand out as unusual.

Does that not imply a separate common factor?

 

[roubaixtuesday]

Does that not imply a separate common factor?

You can't separate our the various factors that were driving transmission.

But on what was happening in other nearby countries:

France was quarantined from the UK due to case levels, whilst Eat out to help out was in place.

So we knew that other countries with broadly similar restrictions in place were struggling to control cases. Yet we not only failed to tighten our own restrictions, we actively subsidised actions we knew were likely to increase infections.

And this wasn't a one-off. Next up, startlingly even at the time, the give launched its campaign to "stop skiving, save Pret, Catch Covid", aimed at getting city centres "buzzing" again(!!).

This is not hindsight:


https://www.cyclechat.net/threads/coronavirus-outbreak.256913/page-830#post-6115321

 

[Pale Rider]

The data is not saying what you think it is.

At the beginning of August, the infection rate was very low - less than a thousand new infections per day. Deaths were in single digits per day. In order to see any effect on the ONS graph, the death raate would need to be at 100-200 times greater. At the time, and well into the autumn, the infection rate was doubling every week. That implies seven doubling periods (2^7 = 128). It would take 7 weeks for that to happen. Furthermore, it typically takes 4 weeks from infection until death, on average. Therefore, if Eat Out to Help Out had an effect on the fatality rates, that would be expected to appear 11 weeks after it commenced.

Eat Out to Help Out started on 3rd August. From the above discussion, you'd expect a significant spike in deaths to occur 11 weeks later in mid October. That is precisely what we see. A significant jump in fatalities is seen in the third week of October. The correlation is remarkably good. Were I to uncover similar such data in the course of my research, I'd be very pleased indeed. Come to think of it, I got surfactants into laundry products based on evidence that was less impressive than that!

In hindsight, Eat Out to Help Out was ill advised. It is unlikely to have helped, and looking closely at the data strongly suggests that it was, at the least, a significant contributor to the large increase in cases we experienced last autumn.

Muddled reasoning.

The eat outers were not lab rats in a controlled environment.

In the subsequent weeks after their allegedly fatal meals a great deal was happening in terms of differing restrictions, different travel patterns, different levels of compliance, and new strains - to mention a just a few.

Nor do we have any idea how many people ate out to help out.

Judging their fates in isolation is impossible.

The timespan from infection to death also seems to vary wildly, so if some eat outers subsequently died, which is almost certainly true, we can have no idea how and when they caught Covid.

Oh, and the one eat out meal I had - an annual family celebration which fell by chance in the period - was appropriately socially distanced, so it's unlikely any of us caught Covid from it.

 

[roubaixtuesday]

Nor do we have any idea how many people ate out to help out.

160 million

£849 million had been claimed through the scheme by 30 September, providing discounts for over 160 million meals in August

https://www.gov.uk/government/stati...ics/eat-out-to-help-out-statistics-commentary

 

[roubaixtuesday]

it's unlikely any of us caught Covid from it

This is true of any gathering when cases are low, even if not distanced.

The mathematics of exponentials means judging risk by individual events results in a fatal underestimate of the overall societal risk. This should be one of the key lessons for future pandemics IMO.

 

[vickster]

160 million

£849 million had been claimed through the scheme by 30 September, providing discounts for over 160 million meals in August

https://www.gov.uk/government/stati...ics/eat-out-to-help-out-statistics-commentary

That’s how many meals, not individuals. I for one had several meals out, again very well socially distanced and impeccably managed by all staff

 

[roubaixtuesday]

That’s how many meals, not individuals

Well, yes, obviously as it exceeds the population 3x over...