Generic Virus Thread

[Stevo985]

9 minutes ago, bannedfromHandV said:

Well not really, you like to present your view as absolute fact and then ridicule people who oppose it with basic analogies, as though they just don’t get it but wait, Stevo’s here with an analogy about sun cream and sunburn.....ah yeah now I get it, thanks so much for reducing it to a level my small brain can comprehend.

 

Well I’m sorry if that’s what you inferred. Wasn’t my intention. 

[bannedfromHandV]

19 minutes ago, Stevo985 said:

Well I’m sorry if that’s what you inferred. Wasn’t my intention. 

You know what, I don't even know what point I was trying to make and I'm probably wrong anyway, usually am.

Apols for going in on it.

 

[Troglodyte]

37 minutes ago, Stevo985 said:

Well I’m sorry if that’s what you inferred. Wasn’t my intention. 

 

16 minutes ago, bannedfromHandV said:

You know what, I don't even know what point I was trying to make and I'm probably wrong anyway, usually am.

Apols for going in on it.

[blandy]

1 hour ago, KentVillan said:

Correlation is not causation.

In the instance of lockdown, it's as close to it as makes no odds. We know for a fact that if there is no contact between people then there is no further spreading. We know from the evidence of lockdown in various nations that for each it has reduced the spread. At some point, multiple correlations become sufficient evidence that causation (in terms of lockdown causing reduction) is no longer speculative. Now you may or may not feel that the point has been reached, which is an individual thing. But even allowing for all the other factors, it's pretty much held as a truth amongst the epidemiologists and scientists and medics that lockdown works in reducing spread and consequently potential cases. There are arguments, valid ones about the impact of lockdown on other aspects - mental health, the economy and all the rest, but not about it's effect on the single issue of reducing the rate of infection.

[KentVillan]

No, unfortunately not. I’d need a couple of books on probability and statistics to explain why not, but I’ll try.

The problem is that each unit you’re looking at (a country) did not behave independently, and was not subject to its own independent experience of Covid-19 (since it was a global pandemic). They then implemented *loads* of different measures within short timeframes, often at the same time.

You can’t then isolate all the measures and all the countries and their contextual factors from each other, so this breaks all the normal principles of statistical inference.

If lockdown is the main explanation, then why did Japan largely avoid a severe outbreak? Korea had a different approach as well. I don’t throw these out as conclusive examples that lockdown doesn’t work, but they illustrate how hard it is to disentangle everything.

It’s obviously true that the further you stand from someone the harder it is to catch the virus from them. But that’s different from saying that nationwide, top-down lockdown policies are the key predictor. Hand washing guidelines seem to have had a huge impact in the weeks before lockdown.

What you’re seeing all over the media is a handful of studies (two I think) by British scientists who have huge professional incentives to say “told you so” or “wasn’t my fault”.

Just to prove that this is not a settled question, here is a paper by a respected statistician (a pre-print, so not evidence or anything conclusive, but just to illustrate that the question is still open): https://arxiv.org/abs/2005.02090

Quote

A Bayesian inverse problem approach applied to UK data on COVID-19 deaths and the published disease duration distribution suggests that infections were in decline before UK lockdown, and that infections in Sweden started to decline only a short time later.

Correlation and confirmation bias are incredibly hard to overcome, and sometimes a pattern looks remarkably convincing, before it is eventually debunked as a causal relationship.

A good primer on this is “Fooled By Randomness” by Nassim Taleb (before he became a Twitter troll). A really good read.

Edited June 12, 2020 by KentVillan

[snowychap]

3 hours ago, blandy said:

Exactly  - I was trying (badly, obviously) to highlight that they were looking at one thing - ICU patients and saying 8% from one set of stats....which is fine. It gives a particular ball park using a particular method for a particular population group.

But if you want to know about how that relates generally to Covi,  to care homes and "would have died anyway" or even the wider population and "whda" even if you start with a similar 8% (which would likely be pessimistic, even for care home residents) you're still missing a whole bunch of data to help refine that figure to something more representative or accurate. 

Probably still  not written that as clearly as I'd like.

Hmm. I think I'm getting what you're on about here - similar to the valid points that @KentVillan made.

I still don't get the '8% being people who missed treatment' bit from earlier but then again I picked up a broccoli spear instead of my fork when eating lunch earlier (that's a genuine admission) so I might just be having one of those days.

Edited June 12, 2020 by snowychap

[snowychap]

2 hours ago, KentVillan said:

Correlation is not causation.

You are @leviramsey and I claim $5. I also agree.

[blandy]

1 hour ago, KentVillan said:

The problem is that each unit you’re looking at (a country) did not behave independently, and was not subject to its own independent experience of Covid-19 (since it was a global pandemic). You can’t separate them all from each other, so this breaks all the normal principles of statistical inference.

If lockdown is the main explanation, then why did Japan largely avoid a severe outbreak? Korea had a different approach as well.

It’s obviously true that the further you stand from someone the harder it is to catch the virus from them. But that’s different from saying that nationwide, top-down lockdown policies are the key predictor. Hand washing guidelines seem to have had a huge impact in the weeks before lockdown.

Each Country did behave independently, though there was guidance from WHO, there was no across the board imposition of measures, or owt else. Each Country had its own experiences, as you yourself say - mentioning Japan and S.Korea etc.

I've not said lockdown was the main explanation have I? I didn't mean to  - Only that we know where it was introduced it was effective in reducing transmission rates. Youre right that other steps ALSO helped. But it is fact that no contact at all stops it dead. Put everyone in a hermetically sealed bubble and it stops. Lockdown, in all its various guises across nations is a much less strict version of that. Looking at the UK, it was only when lockdown was introduced that there started to be a clearly visible change to the infection rate. The other steps were in palce before lockdown, and continued during.

Japan and other nations - we also know to an extent why their different measures succeeded (so far). Acting early when the numbers are smaller means lockdown and widspread isolation isn't necessary - the known infected can be found, isolated, treated, prevented from spreading it. The asymptomatic can't, but as soon as they infect someone and they get sick, the S. Korean track and trace kicks in. Japan, like New Zealand and Australia introduced border enforcement, it caught it early, and you're right that things like not shaking hands, high standards of hospitals and medical stuff, a population that adhered to hand washing and sanitising guidelines, that wears masks in Cities (because of pollution) didn't need to lockdown.

My point is that once/if the other measures available prove ineffective due to non-observance or other reasons, lockdown works. There's nowhere it hasn't worked. It's extreme, but I repeat what I said, it's clear that it reduces the infection rate by reducing the contact through which the virus spread (contact including proximity). It's not about being a predictor. If I was saying that, you'd be right, but that's not my proposition. It works as a mitigation. It's not a predictor.

[snowychap]

1 hour ago, KentVillan said:

It’s obviously true that the further you stand from someone the harder it is to catch the virus from them. But that’s different from saying that nationwide, top-down lockdown policies are the key predictor. Hand washing guidelines seem to have had a huge impact in the weeks before lockdown.

I think one of the difficulties of the lockdown v no lockdown argument is what each thing actually means - that's before you get to the quality of the data and how you analyse it.

Not only are there varying degrees of 'lockdown' but there is also a wide range to 'no lockdown'.

[blandy]

18 minutes ago, snowychap said:

I still don't get the '8% being people who missed treatment' bit from earlier

Last go then I'll leave it.

Forget the 8%, just consider that there is some number, some percentage, of the population who statisitically die in a period of time. Now if you introduce a new fatal virus AND you take lockdown measures two things (well many more than two, but I'm trying to be simplistic) will happen - the virus will kill additional people. Also the lockdown will kill additional people suicide, lack of hospital places for heart disease and cancer etc). the UK totla here is something like 61,000 unexpected deaths (over the normal rate). Of those 61,000 statistically some "whda". But the actual individuals, individual souls, and the measuring of the individuals compiled up to top level figures, it hasn't caught (unsurprisingly) the reasons each perished - many through virus, sure. Others with virus (known), others with virus (suspected) others with virus, but not known (or recorded), other of various ailments and accidents. It maybe (extreme case, v. unlikely) that all the really sick people have somehow hung on to life, but a lot of people who didn't catch virus have died as a result of lockdown measures. It may be that (opposite extreme, also v. unlikely) all the really sick people whda have been "reaped" (horrible term) by virus.

So I was trying to sort of put across that there's a whole lot of stuff that we don't know, or can't see in and from the stats put out. As Kent said, the article was interesting, but it is necessarily limited in what it could address or use to make its conclusion, and there may (my words) be a lot more to come out which will change the picture or let us update what we understand.

[KentVillan]

34 minutes ago, blandy said:

Each Country did behave independently, though there was guidance from WHO, there was no across the board imposition of measures, or owt else. Each Country had its own experiences, as you yourself say - mentioning Japan and S.Korea etc.

I meant statistical independence not political independence.

https://en.m.wikipedia.org/wiki/Independence_(probability_theory)

Quote

Two events are independent, statistically independent, or stochastically independent^[1] if the occurrence of one does not affect the probability of occurrence of the other (equivalently, does not affect the odds). Similarly, two random variables are independent if the realization of one does not affect the probability distribution of the other.

Clearly every country was influenced, to some extent, by what other countries did.

Some measures were very widely adopted (e.g. encouraging elderly to avoid social gatherings, hand washing) and others varied more (e.g. widespread tracing, face masks).

The paper I linked to suggests that infections peaked before lockdown. The “clear” effect of lockdown on transmission could just be a coincidence - especially when you consider that deaths are a lagging indicator of infections (I think it takes about two weeks on average from infection to death?).

[blandy]

4 minutes ago, KentVillan said:

The “clear” effect of lockdown on transmission could just be a coincidence

If you consider only statistics, then that is valid. Once you also consider other information, it becomes less and less credible.

its like, I dunno, saying there is a correlation between the numbers of cars and trucks and buses on a city road and the level of air pollution on that road. Could just be a co-incidence. But when you consider that the diesels and petrols Each emit pollution, you start to move towards it being less likely to be coincidence. Then you look at roads in France and Italy and the USA and see the same effect. Every time there’s a traffic increase, there’s a pollution increase. You get variations around the average age of vehicles, the mix of heavy diesels to petrol cars, sure. And it’s true that the wind or weather on a particular day may also have an impact, but at some point you move from  (changing analogy) “there’s a correlation between smoking and shortened life expectancy” to “smoking is bad for your health”.

[snowychap]

37 minutes ago, blandy said:

Last go then I'll leave it.

Forget the 8%, just consider that there is some number, some percentage, of the population who statisitically die in a period of time. Now if you introduce a new fatal virus AND you take lockdown measures two things (well many more than two, but I'm trying to be simplistic) will happen - the virus will kill additional people. Also the lockdown will kill additional people suicide, lack of hospital places for heart disease and cancer etc). the UK totla here is something like 61,000 unexpected deaths (over the normal rate). Of those 61,000 statistically some "whda". But the actual individuals, individual souls, and the measuring of the individuals compiled up to top level figures, it hasn't caught (unsurprisingly) the reasons each perished - many through virus, sure. Others with virus (known), others with virus (suspected) others with virus, but not known (or recorded), other of various ailments and accidents. It maybe (extreme case, v. unlikely) that all the really sick people have somehow hung on to life, but a lot of people who didn't catch virus have died as a result of lockdown measures. It may be that (opposite extreme, also v. unlikely) all the really sick people whda have been "reaped" (horrible term) by virus.

So I was trying to sort of put across that there's a whole lot of stuff that we don't know, or can't see in and from the stats put out. As Kent said, the article was interesting, but it is necessarily limited in what it could address or use to make its conclusion, and there may (my words) be a lot more to come out which will change the picture or let us update what we understand.

I'm not sure I disagree with the underlying thought: in essence, the more we know, the more we can tell.

My issue was about the point you'd made about the 8% and that this suggested some wrong inferences from the analysis the authors of that piece had provided.

I'm less certain about the apparently equitable treatment between deaths caused by the virus and deaths indirectly caused by measures taken. That's really going to require much more data at a better level to determine things.

[Brumerican]

38 minutes ago, KentVillan said:

The “clear” effect of lockdown on transmission could just be a coincidence - especially when you consider that deaths are a lagging indicator 

Top Gear Christmas Special 2002 ?

[KentVillan]

3 hours ago, blandy said:

If you consider only statistics, then that is valid. Once you also consider other information, it becomes less and less credible.

its like, I dunno, saying there is a correlation between the numbers of cars and trucks and buses on a city road and the level of air pollution on that road. Could just be a co-incidence. But when you consider that the diesels and petrols Each emit pollution, you start to move towards it being less likely to be coincidence. Then you look at roads in France and Italy and the USA and see the same effect. Every time there’s a traffic increase, there’s a pollution increase. You get variations around the average age of vehicles, the mix of heavy diesels to petrol cars, sure. And it’s true that the wind or weather on a particular day may also have an impact, but at some point you move from  (changing analogy) “there’s a correlation between smoking and shortened life expectancy” to “smoking is bad for your health”.

I think we're slightly arguing at cross purposes here, and maybe interpreting words in different ways.

I'm not saying there's no evidence that keeping your distance from people, minimising social contact, etc. reduces the spread of the virus.

I'm saying there's still no clear evidence that the nationwide policy of govt imposed lockdown was what caused the UK's Covid-19 epidemic curve to flatten out. It's a subtle difference, but it's important.

It's completely possible (and many people are still arguing this) that the virus started to burn itself out before lockdown measures were introduced - e.g. because "super spreaders" were the first people to die or become immune, because a critical mass of immune people was reached, or because minor changes to public behaviour were already starting to take effect.

On top of that, we have to question the long-term benefits of lockdown, given that this is a disease that is likely to become endemic.

There is still a school of thought that all you achieve by social distancing and self-isolating is delaying the inevitable. Let's say no vaccine is ever discovered (or at least, no vaccine is discovered for a good couple of decades), and that our treatment of the disease only improves marginally. At some point you are likely to be exposed to the disease, and sadly, your fate is in the hands of the gods. Was it worth delaying that moment of reckoning for 6 months or 12 months or whatever?

Remember the original rationale behind lockdown was not stopping people from becoming infected, but stopping too many people from becoming infected at once.

We now seem to have shifted to a position where lockdown actually saves lives. Nobody was really arguing that in the early days - except in the sense that it might buy you time to develop a treatment or vaccine. It now seems clear that there will be no silver bullet treatment or vaccine (Chris Whitty has said as much), but likely a slow trickle of incremental improvements over several years. Does that justify persisting with lockdown? Maybe, but it seems like the case is getting less and less clearcut by the day.

Edited June 12, 2020 by KentVillan

[snowychap]

12 minutes ago, KentVillan said:

We now seem to have shifted to a position where lockdown actually saves lives. Nobody was really arguing that in the early days - except in the sense that it might buy you time to develop a treatment or vaccine.

People were arguing that in the early days and still are.

It wasn't just about flattening the curve.

[HanoiVillan]

1 hour ago, KentVillan said:

There is still a school of thought that all you achieve by social distancing and self-isolating is delaying the inevitable. Let's say no vaccine is ever discovered (or at least, no vaccine is discovered for a good couple of decades), and that our treatment of the disease only improves marginally. At some point you are likely to be exposed to the disease, and sadly, your fate is in the hands of the gods. Was it worth delaying that moment of reckoning for 6 months or 12 months or whatever?

Remember the original rationale behind lockdown was not stopping people from becoming infected, but stopping too many people from becoming infected at once.

We now seem to have shifted to a position where lockdown actually saves lives. Nobody was really arguing that in the early days - except in the sense that it might buy you time to develop a treatment or vaccine. It now seems clear that there will be no silver bullet treatment or vaccine (Chris Whitty has said as much)

These seem like fairly big assumptions here? You go from hypothesising a vaccine not being discovered for an extremely long period of time to 'it now seems clear that there will be no  . . . vaccine', which is not my understanding of people's assessment of the prospects for a vaccine?

And I don't really believe that 'our treatment of the disease improves only marginally' is something I would be banking on either. A huge improvement in treatment of the disease has already been made, in rolling patients onto their stomachs rather than putting them on ventilators. It's certainly not impossible that further changes in medical practice could have a big effect on mortality rates even in the absence of medicinal developments.

[KentVillan]

1 hour ago, HanoiVillan said:

These seem like fairly big assumptions here? You go from hypothesising a vaccine not being discovered for an extremely long period of time to 'it now seems clear that there will be no  . . . vaccine', which is not my understanding of people's assessment of the prospects for a vaccine?

And I don't really believe that 'our treatment of the disease improves only marginally' is something I would be banking on either. A huge improvement in treatment of the disease has already been made, in rolling patients onto their stomachs rather than putting them on ventilators. It's certainly not impossible that further changes in medical practice could have a big effect on mortality rates even in the absence of medicinal developments.

Why did you remove "silver bullet" from the quote? It completely changes the sense of the sentence you quote. Vaccines vary from those that offer partial protection (e.g. seasonal flu vaccines), to those that completely wipe out a disease (e.g. smallpox).

I don't think it's remotely unorthodox to suggest that we may never discover a vaccine to this virus, since that is the baseline assumption for most viruses.

As for proning, what makes you think it has been a "huge improvement in treatment" and not just a marginal gain?

Edited June 13, 2020 by KentVillan

[TheAuthority]

6 hours ago, KentVillan said:

I think we're slightly arguing at cross purposes here, and maybe interpreting words in different ways.

 

3 hours ago, KentVillan said:

As for proning, what makes you think it has been a "huge improvement in treatment" and not just a marginal gain?

[blandy]

8 hours ago, KentVillan said:

Remember the original rationale behind lockdown was not stopping people from becoming infected, but stopping too many people from becoming infected at once.

Absolutely. Still is.