Covid-19 Gambles

[denso13]

This one is a good read.

As is this https://www.nature.com/articles/d41586-020-02801-8.

 

[JohnW2]

Can you provide any citations for this, beyond the mannequin / droplet studies? (Flawed by the way since it is aerosols which count).

As you must be aware, there is a large amount of literature 'out there' with a whole spectrum of differing conclusions. However, that does not alter my intuitive feeling that they have to be 'better than nothing'. Also, in addition to whatever direct effect they have, they probably also serve a 'psychological' benefit, in reminding people that things are not 'normal', and that we need to remain cautious about our degree of contact with other people.

Throughout my professional life (many decades!) there has been a debate about the usefulness of face masks. However, I don't recall having ever heard a serious suggestion that we should abandon the wearing of masks (traditionally pretty crude, usually 'paper', ones) in operating theatres and when undertaking sterile procedures.

Kind Regards, John

 

[echoes]

This one is a good read, https://www.pnas.org/content/118/4/e2014564118

Thanks Denso. I've not read all of it yet, but I wonder which particular parts of this you find most compelling, and which studies referred to provide the best evidence?
I'm struggling to find conclusive stuff among the studies reported. No shortage of "suggest", "could" and "may".

 

[JohnW2]

The 'exponential' word is cropping up far more than the actual data permits. ... The rising part of these curves is never exponential. The falling part is.

I'm not quite sure what you are saying.

On a theoretical basis (assuming that each infected person infects an average of N others - the 'R' value), the rise (when R>1) should be exponential and the fall (when R<1) should be negative exponential.

However, and perhaps this is what you're saying, it gets more complicated than that since this is not really what we have actually been seeing. As can be seen in the curves I showed in post #6, other than at the start of the rise, the rise has been essentially linear, as has been the fall (until its late stages). The same (essentially linear increases) can be seen in the curve I showed for recent times in post #6.

Is that what you were referring to?

Kind Regards, John

 

[denso13]

Thanks Denso.

No problem. I tend to skim read and go to the conclusions which do seem pretty clear.

I do think mistakes have been made over this saga, but on the whole, I don't dispute the underlying thinking about this novel virus.

 

[JohnW2]

Thanks Denso. I've not read all of it yet, but I wonder which particular parts of this you find most compelling, and which studies referred to provide the best evidence? I'm struggling to find conclusive stuff among the studies reported. No shortage of "suggest", "could" and "may".

It is, as I said, a messy situation, with little convincing evidence in any direction, presumably at least partially because of the difficulty in studying the question - to get a truly decisive answer would probably require some 'unethical' studies!

However, in a situation such as we are discussing, it probably makes sense to 'err on the side of caution' so, quite apart from my gut feeling that "masks must be better than nothing", perhaps I should ask you whether you are aware of any convincing evidence that wearing masks can do harm?

Kind Regards, John

 

[echoes]

You are surely not suggesting that the rapid falls in cases which followed soon after our implementation of lockdowns (in March, November and early January) were 'co-coincidences', or even the result of seasonal factors, are you?

Yes I am. You'll see exactly the same curve in Sweden, which did not lockdown. It's just what happens.
Since you've done a good deal of analysis, you will have noticed that the infection rate in March 2020 had already peaked before lockdown.

 

[Swwils]

They will, but obviously more frequently in populations/countries with a high prevalence of infection. As I said, variants can only arise during replication in infected people, so, the more infected people there are .... !!
I'm no virologist, so I can't comment on that but, on the basis of what I do know, I don't see. mechanistically, why that necessarily has to be be the case
That would be true if the variant itself was relatively 'benign' (as well as having 'vaccine escape') but, as above, I don't know how confident we can be that such will necessarily be the case. Them main 'variants of concern' we have so far seen all seem to exhibit some degree of 'vaccine escape', but have tended to be more, rather than less, infectious.
That is, of course, one of the main arguments being presented for 'now' - I just hope that it proves to be a prudent approach - although, as you will realise, I have my uncertainties about that.

Kind Regards, John

Varients of concern tend to appear in immunocompromised people*, but this is not set in stone and is not a very well understood area - essentially your immune system is very good at stopping this happening but it still happens all the time. One positive is that we have extremely good surveillance of this in the UK and is one essential reason we are able to trace infection lineages.

Imagine the virus has some structures/mechanisms that allow it to enter cells. Covid19 essentially has structures A, B and C available to it.

By far in large we are concerned about it using structure/mechanism B to gain entry, as it is very prolific and common in humans, especially in tissues we care about. Structure B is in this example a traditional yale style key and the target cells have these yale locks that covid19 can access with its key.

The vaccine is specifically targeted to the handle and shaft of the key, rather than the teeth - so if the virus varient changes the teeth of the key (the most likely part that would change), the vaccine still works as the handle and shaft are the same.

If the virus has its handle changed, then it is very likely that it's no longer really the structure/mechanism B and hence the Yale locks are no longer at risk.

 

[denso13]

It is, as I said, a messy situation, with little convincing evidence in any direction

Well this https://www.pnas.org/content/118/4/e2014564118 and https://www.nature.com/articles/d41586-020-02801-8 suggest it isn't.

 

[Swwils]

Here's the whole sequence of covid19.

7.4kb.

1,273 aminos make up the target spike protein, about 130 of those are of interest in the varients.

The bar is high, and the vast majority of mutations and changes are not beneficial in anyway.

 

[echoes]

I'm not quite sure what you are saying.

On a theoretical basis (assuming that each infected person infects an average of N others - the 'R' value), the rise (when R>1) should be exponential and the fall (when R<1) should be negative exponential.

No need for theoretical bases and R0 when we have real data, and you'll find that it's impossible to fit an exponential to the rising parts of the curves.
This means that R0 is not a reliable predictor, because it does not remain constant.

 

[Swwils]

No need for theoretical bases and R0 when we have real data, and you'll find that it's impossible to fit an exponential to the rising parts of the curves.
This means that R0 is not a reliable predictor, because it does not remain constant.

R0 is not intended to be used in this way, for instance the R value in a village in the cotswolds will almost always be different than say a population on an oil rig.

 

[JohnW2]

Yes I am. You'll see exactly the same curve in Sweden, which did not lockdown. It's just what happens.

Everyone cites Sweden, and I don't think anyone has much of a clear understanding of what happened there, and why (do you?) - there was a rapid, apparently 'spontaneous' fall about 7-8 months after an equally rapid rise. Can you cite any country other than Sweden where there was a marked fall in the absence of appreciable 'control measures?

Since you've done a good deal of analysis, you will have noticed that the infection rate in March 2020 had already peaked before lockdown.

Well, for a start, back then we had no idea as to how many cases there were, since we were essentially only testing (hence counting) those ill enough to be in hospital, hence vastly underestimating the true number of cases. However, in terms of the numbers we do have, I really don't understand your suggestion that infections peaked prior to lockdown (vertical rewd line indicates lockdown.

Kind Regards, John

 

[JohnW2]

No need for theoretical bases and R0 when we have real data ...

R0 is, in itself, of no relevance, since that is the 'reproduction (transmission) rate' in a totally free-ly mixing population (no 'control measures') which is totally susceptible )i.e. no-one with any immunity). What matters is the prevailing R, which is influenced by both the control measures and the degree of population immunity.

... and you'll find that it's impossible to fit an exponential to the rising parts of the curves.

I've already agreed with that, as far as the UK is concerned. The early parts of the rises and the late parts of the falls have been essentially exponential, but the best fit to the (important) bit in-between has been essentially linear (both up and down). I don't think anyone yet fully understands this, but it has presented on of the headaches for modellers.

However, that's been far from the case in all countries - for example, in China's initial outbreak, both rise and fall were essentially exponential.

This means that R0 is not a reliable predictor, because it does not remain constant.

Quite, as above, it's 'R' (which varies) which matters, not R0. However, that does not really help us to understand the shape of the curves we've seen in the UK, since, for example, the near-linear parts of the rising curves imply a progressively (but unexplained) reduction in R. .

Kind Regards, John

 

[echoes]

I really don't understand your suggestion that infections peaked prior to lockdown (vertical rewd line indicates lockdown.

Interesting. Worldometer data that I was plotting at the time showed a clear peak in the rate of increase on or around 18 March.