Re: Coronavirus - Numbers and Statistics

[johnhemming]

The closest putative relation to it is probably Sweden, yet like the UK, Covid-19 ripped through their nursing homes.

An easy first order approximation is deaths per million population.

Please give me Sweden, UK and Italy.

[XFool]

Still don't believe there was a "milder summer form" (virus or illness?).

I accept that I cannot persuade you.

However, it is clear that the viral load affects the outcome. The viral load is lower in the summer (according to the figures).

QED

But if you disagree I don't mind.

I'm never sure whether I disagree with you, or with the way you put things!

[johnhemming]

I'm never sure whether I disagree with you, or with the way you put things!

I don't mind.

[Newroad]

Hi JohnHemming.

If you wish to set up straw men (to knock down) you can do your own leg-work.

There are many comparitive statistical approaches which might be adopted, yet unless they are comparing apples to apples, they are probably of marginal value.

I'm more than happy for you to produce any statistical data you wish, indeed, that is the genesis of this thread. My objections occur when questionable assertions/conclusions are drawn from the statistics (or not from the statistics, as the case may be).

Regards, Newroad

[johnhemming]

I'm more than happy for you to produce any statistical data you wish, indeed, that is the genesis of this thread. My objections occur when questionable assertions/conclusions are drawn from the statistics (or not from the statistics, as the case may be).

You seem to be new to this:

https://johnhemming.blogspot.com/2020/1 ... -data.html

[Newroad]

Not at all, JohnHemming.

I saw them in the original post - they ostensibly appear like good work, but are not my main area of interest. Macro assertions/conclusions relating to the Covid-19 response(s) are.

Regards, Newroad

[jfgw]

https://johnhemming.blogspot.com/2020/12/nhs-trust-covid-analysis-by-trust-data.html

The populations covered by the different trusts vary a lot. I haven't found a table but Googling them individually brings up a value for most (not all) trusts.

If I find time, I will see if I can do some sort of per capita analysis, even if it is on a small sample.

Some trusts are specialised and, therefore, will be atypical such as the Liverpool Heart and Chest Hospital NHS Foundation Trust which covers a population of 2.8 million, the highest I have found so far.


Julian F. G. W.

[servodude]

Italy went too soft, too late

At this point I lose any enthusiasm for debate.

In the Northern Hemisphere there were lockdowns in March/April. The more effective they were the more people died in the autumn winter period.

What you're really comparing are stricter "the world will end if we don't" lockdowns with "OK we need to look like we're doing something" lockdowns

The less effort put in to control the spread of the virus - the more the virus will spread

And being a pedant.... at 25°04′N 121°31′E you'll find Taiwan
- last time I looked that was in the Northern Hemisphere?

- sd

[redsturgeon]

Italy went too soft, too late

At this point I lose any enthusiasm for debate.

In the Northern Hemisphere there were lockdowns in March/April. The more effective they were the more people died in the autumn winter period.

What you're really comparing are stricter "the world will end if we don't" lockdowns with "OK we need to look like we're doing something" lockdowns

The less effort put in to control the spread of the virus - the more the virus will spread

And being a pedant.... at 25°04′N 121°31′E you'll find Taiwan
- last time I looked that was in the Northern Hemisphere?

- sd

As are South Korea, Japan, Vietnam. With death rates of 17, 26 and less than one per million respectively vs our 1000 plus. Why have we not learned lessons from their behaviours.

John

[Newroad]

Correct on both counts, RedSturgeon and Servodude.

Even Singapore is about 1' north of the equator. You need to go down to Jakarta to get about 6' South of the equator. It could sensibly be argued that the tropical countries aren't a reasonable comparator for Europe, but Wuhan, Taiwan, i.e. East and North East Asia, rather than South East Asia, are not unreasonable.

And yes, the degree and timing of hard lockdown (if one occurred at all) is very relevant. Take Australia for example - Western Australia had a hard border* with the other states for most of the year, with Queensland and South Australia not far off. If you made a scaled down version of that, looking at approximate size and population density, you would have Scotland with a hard border (i.e. WA) with Wales and Northern Ireland having a near hard border (SA and Qu.) From what I understand, WA hasn't had local transmission of Covid-19 for over 250 days.

Regards, Newroad

* it was challenged by Queenslander Clive Palmer in the Federal Court, if you wish to look it up - the WA Government position was substantively upheld

[johnhemming]

There are going to be other factors (such as prior Coronavirus infection) which will influence these things. However, we do have a way of comparing different small areas of England such as Lewisham
https://coronavirus.data.gov.uk/details ... HS%20Trust

Where the hospital admissions peaked on 7th April and hence arguably are influenced by the lockdown and Guys
https://coronavirus.data.gov.uk/details ... on%20Trust

Where hospital admissions peaked on 29th March and were not influenced by the lockdown. Lewisham still has a problem, Guys does not.

If I can get figures for the local populations served by each trust I can do some calculations as to what one would expect to occur.

However, I do think looking at different areas in London is a good way of avoiding problems from other factors influencing the effect.

[johnhemming]

Kings is another that peaked before the effect of the lockdown came in
https://coronavirus.data.gov.uk/details ... on%20Trust

[redsturgeon]

Kings is another that peaked before the effect of the lockdown came in
https://coronavirus.data.gov.uk/details ... on%20Trust

There are going to be so many variables at individual hospital levels that will affect admission numbers that I am not confident that too much can be read into the figures at that level of granularity.

For instance:

- How busy is admissions at the precise moment that a patient is brought in, is there even parking space for the ambulance at the nearest hospital or is it better to divert to a less busy place.

- Have any local processes been put in place where one trust gets certain cases before others

- Are the covid/ICU beds at or near capacity at one hospital vs another. eg, our local hospital in Hampshire was receiving cases from Kent.

- Has an in hospital transmission incident meant that no new cases can be accepted temporarily.

Unless you know details of what is happening at a trust level, the bare admission numbers may not tell the story that you think.

Speaking with a good friend who is an ICU consultant at our local hospital has informed my view on this.

John

[johnhemming]

There are going to be so many variables at individual hospital levels that will affect admission numbers that I am not confident that too much can be read into the figures at that level of granularity.

In the absence of any detailed figures explaining what adjustments should be made to the figures for any other reasons we need to rely on the figures.

If SAGE and PHE were half competent they would not just analyse these figures and identify any necessary corrections, but manage a database with where people are infected and when and use that to drive decision-making. It is not unreasonable to assume in the absence of other information that people were infected a fortnight before admission.

However, I recognise that there are people who don't like the conclusion that these figures point it. I am, therefore, not surprised when people raise unquantified, unevidenced criticisms. That does not make the figures invalid, however.

[redsturgeon]

There are going to be so many variables at individual hospital levels that will affect admission numbers that I am not confident that too much can be read into the figures at that level of granularity.

In the absence of any detailed figures explaining what adjustments should be made to the figures for any other reasons we need to rely on the figures.

If SAGE and PHE were half competent they would not just analyse these figures and identify any necessary corrections, but manage a database with where people are infected and when and use that to drive decision-making. It is not unreasonable to assume in the absence of other information that people were infected a fortnight before admission.

However, I recognise that there are people who don't like the conclusion that these figures point it. I am, therefore, not surprised when people raise unquantified, unevidenced criticisms. That does not make the figures invalid, however.

What conclusions are those? The figures are not invalid, as in they exist but as I pointed out they may not show what you wish them to show.

Looking at Guys covid admissions figures you then seem to be leaping to herd immunity in London...or have I got that wrong?

Perhaps when trying to make statements about London it would be a good idea to look at London Hospital admissions as a whole, this would take out much of the noise.

John

[johnhemming]

Looking at Guys covid admissions figures you then seem to be leaping to herd immunity in London...or have I got that wrong?

Perhaps when trying to make statements about London it would be a good idea to look at London Hospital admissions as a whole, this would take out much of the noise.

Remembering that the Herd Immunity Threshold goes up in the season for coronaviruses (because the virus is more virulent) what the Guys, Imperial and Kings figures (and there are probably others, but I have not tried to find them all) show is is for the local area they cover the amount of infection in March was so great that it came beyond the Herd Immunity Threshold for the seasonal wave - in those areas. I accept that this is not the case in other areas such as Lewisham.

What is important about them is that if we know the population served by the trust we can make an estimate as to the number of hospital admissions that will occur to hit the herd immunity threshold. That would enable predicting what is going to happen in other trusts.

You may notice that hospital admissions trends seem to preceed the governments PCR prevalence cases numbers. Hence there is a lot of useful information that can be obtained from this. Even more useful information would be available if the figures were analysed properly to exclude things like infection in hospital and to track the location of infection.

[redsturgeon]

...more useful information would be available if the figures were analysed properly to exclude things like infection in hospital and to track the location of infection.

I agree with this

John

[johnhemming]

...more useful information would be available if the figures were analysed properly to exclude things like infection in hospital and to track the location of infection.

I agree with this

Had I been in parliament I would have concentrated my effort on ensuring that the data was accurate. It is only when you have accurate data that you can work out what to do.

I have mentioned this to journalists, but they did not see the importance of this as an issue.

[Newroad]

Hi JohnHemming.

The issue is that you draw questionable/assumptions conclusions and pejoratively misuse terms, then somewhat amusingly accuse others of doing the same.

In the original link, you come up with this one without any meaningful attempt to substantiate: "Some trusts serve areas where there is a high level of immunity and hence few people being infected"
You then refer to a "milder summer form" of the virus, but when questioned, admit something along the lines that behaviours probably reduced viral loads upon infection which may thereafter influence the course of the infection - I don't have an issue with the revised logic per se
You then appear to assume that herd immunity can be achieved within a "season" for coronaviruses in general and more pertinently here, implicitly for Covid-19
And finally, I suspect you have misused another term in context, "virulence", unless you have some actual evidence you can point to that the Covid-19 pathogen follows a fundamentally different path in causing damage to the human host during "the season for coronaviruses" (whatever that is - I'm guessing you broadly mean winter) all things being equal

It is my supposition that you have the agenda/view and are trying to make the stats fit it, rather than most of the other posters thus far.

Regards, Newroad

[johnhemming]

The issue is that you draw questionable/assumptions conclusions and pejoratively misuse terms, then somewhat amusingly accuse others of doing the same.

We have been through discussions about the meaning of words on this site before. I take the view that words should be used with the dictionary meaning and my use of the word "virulence" fits with that, but if you wish to continue that argument please do me the favour of reading the previous discussion about that before. I probably won't respond, however.

Similarly it appears that the reason Coronaviruses are seasonal (in that they become more virulent in certain seasons) is linked possibly to relative humidity and in particular people switching the heating on. Again we have discussed this before.

Finally I suggest you use the "quote" facility in the forum. That where it is set up to do so results in people getting an email when they are quoted. That is more likely to elicit a response.