Coronavirus - General Chat - No statistics

[tjh290633]

You missed one -

4. Lockdowns do work - but only if enough people abide by the rules for long enough for the restrictions to be effective...

Cheers,

Itsallaguess

Which might imply that over-long lockdown is self-defeating.

TJH

[johnhemming]

I understand that Wales, despite its 2 week lockdown, found that cases continued to rise.

This leads to a number of conclusions.

1. Lockdown does not work, something else does.

2. 2 weeks is not long enough, compare the effect of England's 4 weeks.

3. Localised restrictions may be more effective than widespread restrictions.

TJH

There is, of course, an alternative conclusion. That is that the measurement of "cases" is not reliable.

If you look at hospital admissions and deaths in Wales:
https://coronavirus.data.gov.uk/details ... Name=Wales
https://coronavirus.data.gov.uk/details ... Name=Wales

What you see is
a) The figures are a bit of a mess (which is a known issue as they produce the figures in a different way to England)
b) That if anything admissions peaked in early November and deaths just after that.

I don't know of any other Welsh statistics, but what I do know is that prevalence testing (which is what most of the "cases" figures are) is pretty useless for lot of reasons.

We know separately from a study in the US Marines that restrictions can reduce the amount of virus transferred, but not stop it.

[XFool]

My reference to other countries (usually Sweden!) is less to say: "We should be doing what they are doing" or "We should not be doing what they are doing", than to point out how some of these alternative explanations or recommendations for UK policy don't always seem to stand too much scrutiny - even before we attempt to allow for differences between nations and societies.

That's precisely why you need to do it AFTER allowing for the differences!

And how am I supposed to do that? Realistically, how is anyone here in a position to properly do that?

[dealtn]

My reference to other countries (usually Sweden!) is less to say: "We should be doing what they are doing" or "We should not be doing what they are doing", than to point out how some of these alternative explanations or recommendations for UK policy don't always seem to stand too much scrutiny - even before we attempt to allow for differences between nations and societies.

That's precisely why you need to do it AFTER allowing for the differences!

And how am I supposed to do that? Realistically, how is anyone here in a position to properly do that?

How are we supposed to "attempt to allow for differences" instead of "don't seem to stand too much scrutiny"?

Well I guess I would at least try to understand how some factors that aren't common between samples might have an influence rather than simply ignoring them. I would imagine it isn't easy, and even those much more expert than anyone here would have difficulty.

But this is what was said

"So, if my fellow posters are going to refer to other countries / populations and cite them as examples of what the UK should do / have done, it would be really helpful to recognise and point out any variables that might need to be considered along with the difference in the raw data. Otherwise, it seems to me, that such comparisons are just making unfounded assertions, rather than actually being interested in the problem of what might be the best response to this pandemic".

Which you promptly replied to with

I agree with this.

So if country A had no hospitals, yet country B did, we might draw some conclusions, and hopefully those more expert than ourselves would be able to "test" a hypothesis on why this variable matters, by how much, and what should be done about it. (And we might find a link we could share that demonstrated this.)

Similarly if A was older than B, more densely populated than B, was an Island and B wasn't, had a different ethnicity mix etc.

So "realistically" it is hard. But that is still preferable to ignoring those differences.

[XFool]

But this is what was said

"So, if my fellow posters are going to refer to other countries / populations and cite them as examples of what the UK should do / have done, it would be really helpful to recognise and point out any variables that might need to be considered along with the difference in the raw data. Otherwise, it seems to me, that such comparisons are just making unfounded assertions, rather than actually being interested in the problem of what might be the best response to this pandemic".

Which you promptly replied to with

I agree with this.

Meaning, on my behalf, there are going to be quite a few differences between different societies and therefore complex and difficult to account for them. I might, like vrdriver, be able to point them out, but that is a very long way from realistically accounting for them. I wouldn't even try and, as I am not one of those putting forward mathematical models, I don't really need to.

So if country A had no hospitals, yet country B did, we might draw some conclusions, and hopefully those more expert than ourselves would be able to "test" a hypothesis on why this variable matters, by how much, and what should be done about it. (And we might find a link we could share that demonstrated this.)

Similarly if A was older than B, more densely populated than B, was an Island and B wasn't, had a different ethnicity mix etc.

Which, one might assume, various specialists are, and will be, doing. Personally, I generally prefer to leave these matters to those who are the appropriate and relevant experts. Though that seems to be where some seem to have a problem.

[swill453]

I'm not sure this site has been well-publicised, but in case any fellow Scots want to know about local Coronavirus cases and aren't aware of it:

https://public.tableau.com/profile/phs. ... 0/Overview

It has the number of cases detected in a "neighbourhood" in the last 7 days. A "neighbourhood" is a subdivision of a local authority, and each has approximately 3000 to 6000 population.

It's a bit clunky to use, as the raw data isn't available at that level for reasons of privacy. So you have to look at it on a map.

Each time you click it can take a second or two to react so it pays to do it slowly.

Click on "Cases by neighbourhood".

In the left dropdown box, select your local authority.

It is now shown in a map section subdivided into neighbourhoods. The colours relate to the number of cases per 100,000, but are not particularly useful at this granularity. Instead, hover the mouse over the one you're interested in and an information box will pop up showing the actual number of cases detected in the last 7 days. This will be either 0, 1-4, or the actual number (1-4 is used for privacy reasons to prevent individual identification).

There's a link on the page to find out which neighbourhood area a postcode is in, since the names are sometimes obscure and there are no features marked on the map. Alternatively, I found having Google Maps open in a separate browser window was useful.

Scott.

[Johnspenceuk]

I'm not sure this site has been well-publicised, but in case any fellow Scots want to know about local Coronavirus cases and aren't aware of it:

https://public.tableau.com/profile/phs. ... 0/Overview

It has the number of cases detected in a "neighbourhood" in the last 7 days. A "neighbourhood" is a subdivision of a local authority, and each has approximately 3000 to 6000 population.

It's a bit clunky to use, as the raw data isn't available at that level for reasons of privacy. So you have to look at it on a map.

Each time you click it can take a second or two to react so it pays to do it slowly.

Click on "Cases by neighbourhood".

In the left dropdown box, select your local authority.

It is now shown in a map section subdivided into neighbourhoods. The colours relate to the number of cases per 100,000, but are not particularly useful at this granularity. Instead, hover the mouse over the one you're interested in and an information box will pop up showing the actual number of cases detected in the last 7 days. This will be either 0, 1-4, or the actual number (1-4 is used for privacy reasons to prevent individual identification).

There's a link on the page to find out which neighbourhood area a postcode is in, since the names are sometimes obscure and there are no features marked on the map. Alternatively, I found having Google Maps open in a separate browser window was useful.

Scott.

Hi that seems similar to the one I started using a few weeks ago but everything is on the main page. https://www.travellingtabby.com/scotlan ... s-tracker/ it also has a UK one as well but I haven't tried it yet.

John

[swill453]

Hi that seems similar to the one I started using a few weeks ago but everything is on the main page. https://www.travellingtabby.com/scotlan ... s-tracker/ it also has a UK one as well but I haven't tried it yet.

I've been using travellingtabby and I do like it, but unless I've missed it I can't see anything there that has granularity lower than the Council area (population 183,000 in my case)

The site I mentioned has it down to basically half my provincial town (population 3426).

Scott.

[Johnspenceuk]

Hi that seems similar to the one I started using a few weeks ago but everything is on the main page. https://www.travellingtabby.com/scotlan ... s-tracker/ it also has a UK one as well but I haven't tried it yet.

I've been using travellingtabby and I do like it, but unless I've missed it I can't see anything there that has granularity lower than the Council area (population 183,000 in my case)

The site I mentioned has it down to basically half my provincial town (population 3426).

Scott.

I have been playing about with it. I see your point takes it down to my local council ward. Thanks

John

[Itsallaguess]

I personally said earlier this year that I thought as with other Coronaviruses this one would be seasonal.

Hence an increase in infections a seasonal (and in the UK second) wave (even though it is more of a Gopertz curve than a sine wave) is something I predicted.

What I did not predict was when it would peak. We now know that in the UK was over a month ago.

John,

Could you please give a YES / NO answer to the following question -

1. Do you think that the Tier 3 restrictions in place for the most prevalent areas prior to the start of the November lockdown will have helped to reduce community-transmission of the virus in those areas?

Cheers,

Itsallaguess

[johnhemming]

Could you please give a YES / NO answer to the following question -

1. Do you think that the Tier 3 restrictions in place for the most prevalent areas prior to the start of the November lockdown will have helped to reduce community-transmission of the virus in those areas?

Yes I could.

[jfgw]

Could you please give a YES / NO answer to the following question -

1. Do you think that the Tier 3 restrictions in place for the most prevalent areas prior to the start of the November lockdown will have helped to reduce community-transmission of the virus in those areas?

My answer: YES.

Where on the scale from negligible to highly significant the effect of the restrictions occurs is a bit more difficult to determine.


Julian F. G. W.

[Itsallaguess]

Could you please give a YES / NO answer to the following question -

1. Do you think that the Tier 3 restrictions in place for the most prevalent areas prior to the start of the November lockdown will have helped to reduce community-transmission of the virus in those areas?

My answer: YES.

Where on the scale from negligible to highly significant the effect of the restrictions occurs is a bit more difficult to determine.

Thanks Julian - my interest in this primarily stems from the fact that when people wish to look at the drop-off in England second-wave infections, and then work back and suggest that the November lockdown started too late to 'cause' that drop-off, such people are ignoring the fact that there were already Tier 3 protocols in place in the hardest-hit areas earlier than the November lockdown date itself...

If there were absolutely no protocols in place, and then we saw the second wave starting, and then we locked down, then such a work-back would be valid, but where pre-November-lockdown protocols were already in place, and where some of those protocols included a Tier 3 level that was lockdown in all but name, then I think it's incorrect to use the start-date of the November lockdown to justify such an argument.

Cheers,

Itsallaguess

[zico]

So, if my fellow posters are going to refer to other countries / populations and cite them as examples of what the UK should do / have done, it would be really helpful to recognise and point out any variables that might need to be considered along with the difference in the raw data. Otherwise, it seems to me, that such comparisons are just making unfounded assertions, rather than actually being interested in the problem of what might be the best response to this pandemic.

VRD

As a statistician (actually operational researcher - but nobody knows what that is!) the issue here is about comparing the differences in results to the differences in populations. For example, school league table performance can be very well explained by non-teaching factors, such as catchment area, parents' professions, average salaries etc etc. But in the case of the pandemic, we have a lot of very similar countries getting widely different results, and the biggest difference is in their policy responses, not their national characteristics.
For example, New Zealand - very successful results, 15% of population aged 65+. UK - very poor results, 18% of population aged 65+.
Now, we could say, let's adjust the New Zealand results to allow for their lower percentage of people aged 65+. Let's assume something very unlikely, that if New Zealand had 18% (as in the UK) instead of 15%, there would be twice as many deaths.
That would change the New Zealand death toll from 25 to 50 (individuals, not thousands) compared with UK's 75,000 deaths.

We suffered a lot from "exceptionalism" at the start of the pandemic. China's experience would be much worse than ours, because lots of Chinese men smoke. Then, when Italy suffered, people said Italians like to hug and have their grannies live with them. But then our results were worse than Italy.

Africa is the one area where demographics do seem to make a big difference. Africa has a very young demographic, with just 3.6% of people aged 65+ and their Covid deaths are massively lower than those experienced in "1st world" countries. (Africa's average life expectancy is just 62 years old). But most African countries also reacted more quickly than the UK to the pandemic. It would be very difficult to compare our response with that of Africa's, but comparatively easy to compare ourselves with European countries with similar age profiles, health provisions, and income levels.

[johnhemming]

If there were absolutely no protocols in place, and then we saw the second wave starting, and then we locked down, then such a work-back would be valid, but where pre-November-lockdown protocols were already in place, and where some of those protocols included a Tier 3 level that was lockdown in all but name, then I think it's incorrect to use the start-date of the November lockdown to justify such an argument.

The argument, however, is that the lockdown was an unnecessary step simply because for whatever reason the number of infectionshad already peaked.

I actually don't mind accepting that restrictions have an effect on infection rates even if you cannot actually stop the virus, but only reduce the rate of infection.

I have said this on a number of occasions. Hence "Yes" it does have an effect.

The question, of course, is how big an effect.

New Zealand and Australia have used border controls. The UK introduced them too late to have an effect. (the virus was spreading in the UK in December 2019).

[Nimrod103]

As a statistician (actually operational researcher - but nobody knows what that is!) the issue here is about comparing the differences in results to the differences in populations. For example, school league table performance can be very well explained by non-teaching factors, such as catchment area, parents' professions, average salaries etc etc. But in the case of the pandemic, we have a lot of very similar countries getting widely different results, and the biggest difference is in their policy responses, not their national characteristics.

As a statistician, why do you dismiss the very apparent differences between countries, their population density and distribution, and their propensity to get close together or stay far apart, and all other different habits? I can think of no two similar countries. Even within the UK, Yorkshire people are known to be more tactile.

For example, New Zealand - very successful results, 15% of population aged 65+. UK - very poor results, 18% of population aged 65+.
Now, we could say, let's adjust the New Zealand results to allow for their lower percentage of people aged 65+. Let's assume something very unlikely, that if New Zealand had 18% (as in the UK) instead of 15%, there would be twice as many deaths.
That would change the New Zealand death toll from 25 to 50 (individuals, not thousands) compared with UK's 75,000 deaths.

New Zealand is an isolated island 3 hours flying time from its nearest neighbour, which is also relatively isolated. It has relatively little in and out traffic, especially not duing the summer. I guess most New Zealanders holiday locally. By contrast most of the infection entry to the UK came from returning skiers from the Alps and sunseekers from Spain. In their hundreds of thousands.

We suffered a lot from "exceptionalism" at the start of the pandemic. China's experience would be much worse than ours, because lots of Chinese men smoke. Then, when Italy suffered, people said Italians like to hug and have their grannies live with them. But then our results were worse than Italy.

Britain's problem is that they pay little heed to the requirements of isolation, and obeying the law not to go out and socialize. Brits mostly think they know better, and rules are for other people. That is their exceptionalism.

Africa is the one area where demographics do seem to make a big difference. Africa has a very young demographic, with just 3.6% of people aged 65+ and their Covid deaths are massively lower than those experienced in "1st world" countries. (Africa's average life expectancy is just 62 years old). But most African countries also reacted more quickly than the UK to the pandemic. It would be very difficult to compare our response with that of Africa's, but comparatively easy to compare ourselves with European countries with similar age profiles, health provisions, and income levels.

I am aware of many other reasons why Africa has not been badly affected so far (though I have heard on John Campbell's videos that Kenya is not in a good way). Other reasons are that genetically it has been suggested that genes derived from our Neanderthal past make us vulnerable to infection. Africans have no such genes. Another suggestion is that parasite infections (very common in Africans) offer protection against immune system over-reaction.
Simplistic statisticss don't seem too credible.

[servodude]

If there were absolutely no protocols in place, and then we saw the second wave starting, and then we locked down, then such a work-back would be valid, but where pre-November-lockdown protocols were already in place, and where some of those protocols included a Tier 3 level that was lockdown in all but name, then I think it's incorrect to use the start-date of the November lockdown to justify such an argument.

The argument, however, is that the lockdown was an unnecessary step simply because for whatever reason the number of infectionshad already peaked.

I actually don't mind accepting that restrictions have an effect on infection rates even if you cannot actually stop the virus, but only reduce the rate of infection.

I have said this on a number of occasions. Hence "Yes" it does have an effect.

The question, of course, is how big an effect.

New Zealand and Australia have used border controls. The UK introduced them too late to have an effect. (the virus was spreading in the UK in December 2019).

Not exclusively
- when the virus was spreading in the wider population they could very well have decided "sod this they're all old and unproductive"
- they didn't

do you really believe that that "have an effect" stuff? or are you just doing that "using words in a special way thing"?

The effects of each and every intervention compound
- surely if you're on a website such as this that can't be beyond your comprehension?

-sd

[johnhemming]

do you really believe that that "have an effect" stuff? or are you just doing that "using words in a special way thing"?

The effects of each and every intervention compounds
- surely if you're on a website such as this that can't be beyond your comprehension?

It is not necessarily the case that every intervention compounds, nor necessarily the case that every intervention has an effect. Some interventions may even be counter productive.

In the end I rely on the evidence. The reliable evidence are things like the hospital admissions, deaths and people going to primary care.

In this seasonal wave we initially see an exponential growth then a tailing off are it plateaus and we are probably now in a gradual decline.

The effect of increasing immunity levels is easy to mathematically model and it results in a curve of infections which is much like that which we see. Other interventions will have some effect, but we should be able to see that in the stats.

We can actually see the impact of the March lockdown by comparing the English death statistics to the Swedish statistics. The curves essentially match until around Mid April when the English figures decline more rapidly.

We can also see the effect of greater infection in London by comparing the London chart for admissions to other NHS regions.

Now in the end I will trust the maths more than I will trust your assertions.

[GoSeigen]

Could you please give a YES / NO answer to the following question -

1. Do you think that the Tier 3 restrictions in place for the most prevalent areas prior to the start of the November lockdown will have helped to reduce community-transmission of the virus in those areas?

My answer: YES.

Where on the scale from negligible to highly significant the effect of the restrictions occurs is a bit more difficult to determine.

Thanks Julian - my interest in this primarily stems from the fact that when people wish to look at the drop-off in England second-wave infections, and then work back and suggest that the November lockdown started too late to 'cause' that drop-off, such people are ignoring the fact that there were already Tier 3 protocols in place in the hardest-hit areas earlier than the November lockdown date itself...

If there were absolutely no protocols in place, and then we saw the second wave starting, and then we locked down, then such a work-back would be valid, but where pre-November-lockdown protocols were already in place, and where some of those protocols included a Tier 3 level that was lockdown in all but name, then I think it's incorrect to use the start-date of the November lockdown to justify such an argument.

There are other people who view the concept of a second wave as faulty. It has all been one wave of infection working itself out.

Restrictions on movement and other measures like masks and social distancing reduce the rate of spread but I strongly disagree that the second lockdown was necessary or had any appreciable effect limiting deaths in recent weeks. I predicted simply looking at numbers of deaths both in the UK and comparable countries that the death rate would top out about 400pd/2800pw. The rate has gone slightly higher but is consistent with the view that this is the end of the virus and there is no need at all to fear a resurgence of deaths from the virus. The second lockdown will have negligible effect on deaths IMO because victims on the whole will be vulnerable people who succumb despite getting excellent care, and that whether they contract the illness now or a month or two later would really not make much difference to either their care or the outcome.

The benefits of lockdown are negligible compared to the outrage of liberty caused to people who have nothing to fear from the virus.

GS

[johnhemming]

There are other people who view the concept of a second wave as faulty. It has all been one wave of infection working itself out.

My view is that the first phase/wave (in the UK) was a standard viral infection wave for a virus in its normal progression according to a Gompertz curve. Then later in the year the virus is moves into a seasonally virulent phase and you have a seasonal phase/wave with its associated Gompertz curve of infections.

I would expect to see another seasonal wave next September/October/November in the UK, but not before that and not as big as this one.