Coronavirus - General Chat - No statistics

[johnhemming]

That's not my experience . Have you had a look at the footfall in M&S or Waitrose? And none of us golden oldies have any choice whether or not we want to attend a theatre or a concert hall.

If you follow the recent work of Carl Heneghan (https://www.dailymail.co.uk/news/articl ... -says.html or https://www.spectator.co.uk/article/cor ... table-why-)

He is from the University of Oxford Centre for Evidence Based Medicine, and explains things a lot better than I can.

I'm afraid I give little credence to newspaper articles - scientific work requires open scrutiny by experts in the field - i.e. published in a reputable journal - or at least in pre-published form, awaiting scrutiny.

CEBM publish their work here:
https://www.cebm.net/oxford-covid-19-evidence-service/

[bungeejumper]

More on retail footfall, from the agency with probably the best surveying mechanisms: https://fashionunited.uk/news/retail/uk ... 0082450469

National footfall for the week to 22nd August was down by around 30% year on year, but central London was down by 61.2%, and regional cities reporting a 49.8% drop.

Incidentally, ignore the disastrously misleading headline "UK footfall increases fourfold" - it means that the rate of increase in late August (4.1%) was five times that of the previous week (0.8%). Sheesh, what does it take to instil a bit of mathematical sense into idiot copywriters?

BJ

[UncleEbenezer]

... do more of the same!

Measures in place in Bolton et al have seen a sharp rise in covid cases. So they're to have more of the same measures.

[scotia]

John - I'll try again - are there any refutations of this major study (Ward et al)

Please give a link to it.

Sorry - I thought the Ward et al report was voraciously pounced upon by all who were interested in Covid-19. So I thought the reference was well known - and I certainly posted it in a previous response, and I was sure you had already commented on it. Look back to August 17th for the reference, and your comments on it. But to save you the trouble here it is again:-
https://www.imperial.ac.uk/media/imperial-college/institute-of-global-health-innovation/Ward-et-al-120820.pdf

[Nimrod103]

That's not my experience . Have you had a look at the footfall in M&S or Waitrose? And none of us golden oldies have any choice whether or not we want to attend a theatre or a concert hall.

If you follow the recent work of Carl Heneghan (https://www.dailymail.co.uk/news/articl ... -says.html or https://www.spectator.co.uk/article/cor ... table-why-)

He is from the University of Oxford Centre for Evidence Based Medicine, and explains things a lot better than I can.

I'm afraid I give little credence to newspaper articles - scientific work requires open scrutiny by experts in the field - i.e. published in a reputable journal - or at least in pre-published form, awaiting scrutiny. I well remember of the idiocy of cold fusion - issued as a press release.
However I freely agree that the deaths per Covid-19 infection is falling - probably due to the improvements in treatments, and in the most affected sections of the populations (oldies and the infirm) taking significantly more care in social interactions.

Personally, I would not dismiss Carl Heneghan, Professor and Head of the University of Oxford Centre for Evidence Based Medicine, as any old newspaper hack. He has put his reputation on the line in published articles, rather than hiding his predictions and analysis in academic papers (many of which recently on Covid, seem to be based on few samples, and get released without peer review). He says there are not many real improvements in treatment, and anyway hospitalizations are still on a declining trend, not only deaths. Many oldies may be keeping a low profile, but there are enough around, in shops and in pubs, to be having an effect on the death rate now, if it the infection rate was really rising.

[johnhemming]

Sorry - I thought the Ward et al report was voraciously pounced upon by all who were interested in Covid-19. So I thought the reference was well known - and I certainly posted it in a previous response, and I was sure you had already commented on it. Look back to August 17th for the reference, and your comments on it. But to save you the trouble here it is again:-
https://www.imperial.ac.uk/media/imperial-college/institute-of-global-health-innovation/Ward-et-al-120820.pdf

I wanted to be certain as to which one you were referring to. I did tell you previously that this particular report did accept that it had potential for error.

In estimating the IFR, we may have underestimated the number
of infected individuals (leading to higher estimates of IFR), as a result of weakened or absent
antibody response in some people, and waning antibody over time.

So what they say in the report is that they "may have underestimated the number of infected individuals" because of fading or weakened antibody response or absent antibody response.

Hence we have to ask the question as to
a) Do antibodies fade
b) Do some people defeat the disease without antibodies

We know the answer to both of these questions is yes. Hence the figures in the report are wrong.

However the conclusion of the report is:

The pandemic of SARS-CoV-2 infection in England disproportionately affected ethnic minority groups
and health and care home workers. The higher risk of infection in these groups may explain, at least
in part, their increased risk of hospitalisation and mortality from COVID-19.

Hence the underestimation of the number of infected individuals does not affect the conclusion of the report although it does mean that the other figures in the report should not be taken as significant.

[scotia]

And I'll add John - did you ever use real data - i.e. Excess deaths which are significantly larger than Covid-19-recorded deaths. If I remember correctly I think you only ever used Covid-19 recorded deaths in hospitals. Not an impressive data set when it is estimated that around 40% of the UK registered Covid-19 related deaths were in Care Homes.

The data is "real data". I use this particular data because it is consistently measured. It is important, for example, to look at the date that someone dies rather than the date upon which the death is reported.

It is the same as the ONS data on covid infections. They have a subset which is randomly selected. Only that data has the chance of giving anything close to a reliable result. There are, however, issues anyway with PCR tests.

It may be consistently measured - but what is measured? It misses out all recorded deaths at home, and in Care Homes , and the very substantial excess deaths that are certainly present but are not recorded as Covid-related. And from that selected sample you make assumptions on time between infection and death. Do you not think that there may be substantially different times in Care Homes than in Hospitals?

[johnhemming]

It may be consistently measured - but what is measured?

Actually from that I and various other people conclude as to when the peak rate of infection in the community occurred and hence from that what the effect of the lockdown was (and whether it was necessary).

[scotia]

Personally, I would not dismiss Carl Heneghan, Professor and Head of the University of Oxford Centre for Evidence Based Medicine, as any old newspaper hack.

I don't - but newspaper reports I take with a pinch of salt.
I appreciate that his group are issuing numerous online reports - with the "not peer reviewed" caution.

[scotia]

It may be consistently measured - but what is measured?

Actually from that I and various other people conclude as to when the peak rate of infection in the community occurred and hence from that what the effect of the lockdown was (and whether it was necessary).

But how do you know that from a selective subset of deaths? Are you suggesting that all other excess deaths outside of hospitals (where there is no intensive care) have an identical time scale of infection to death?
Let me be clear - the only accurate measure we have of Covid-19 deaths is the excess mortality statistics. And that may relate to the shape of the infection curve -adjusted for the time between infection and death. But selecting that time from only Hospital deaths does not seem a sensible policy.

[JamesMuenchen]

Personally, I would not dismiss Carl Heneghan, Professor and Head of the University of Oxford Centre for Evidence Based Medicine, as any old newspaper hack.

I don't - but newspaper reports I take with a pinch of salt.
I appreciate that his group are issuing numerous online reports - with the "not peer reviewed" caution.

https://www.forbes.com/sites/geoffreyka ... er-review/
Earlier this year the editors of Britain’s two top medical journals spoke out forcefully about the peer review problem.
Speaking at the Royal Society in London in April, Richard Smith, a former editor of the BMJ, who edited the journal for more than a decade, made the case that pre-publication review was too slow, expensive, and burdensome to reviewers, and “lacking in evidence that it actually works in its chief goal of spotting errors.”

He cited an experiment conducted during his time at the BMJ, in which a short paper containing eight deliberate errors was sent to 300 reviewers, as evidence of how badly the system can fail to detect errors.

“No one found more than five, the median was two, and 20 percent didn’t spot any,”

[johnhemming]

But selecting that time from only Hospital deaths does not seem a sensible policy.

We obviously disagree about this. I think the hospital deaths are a representative and consistent sample of people infected by Covid-19 at least for the purpose of tracking the rates of infection.

If you have a different dataset that is prepared in an equally granular manner and similarly reliable then put that forward as an alternative.

What we have as tools for tracking infection rates are
a) Hospital admissions
b) Hospital deaths
c) The GP infection surveillance

All of these point to the same conclusion.

That's good enough for me.

[scotia]

has anybody on this board been attempting to extract Covid-19 data from the government source using the instructions contained in an on-line manual
https://coronavirus.data.gov.uk/developers-guide.
I find this manual opaque, and its suggestion that you need to develop an api to get at the data seems crazy.
It also is subject to changes which can trash an application - but there are no change lists in the updated guide to warn a developer
However I wanted to get at the data, and I set forth to carry it out, using curl on a Windows PC.
After a bit of a struggle, I got it working, then they changed "newDeathsByPublishDate" to "newDeaths28DaysByPublishDate" - and the application crashed. So I worked out from error messages what had happened, and got it back into action.
However a few days ago it ceased working - no error messages - and it simply returned zero data.
Is anyone out there using a similar app? Have you run into similar difficulties? Do you have a soluion?
I have emailed the address on the developers guide - but have had no response.

[sg31]

I've found over the years that people believe what they want to believe and read articles/ watch programmes that support that belief. They find reason to discredit things that go against their belief. Trying to persuade them that they are wrong is impossible or at least very difficult.

I've been reading research papers since this all started, they are generally way above my competence level but I have friends in America who explain the implications of much of the information.

As Scotia has said most general newspaper reports are pretty useless, unless they are employed by one of the specialist science or medical journals they just don't understand what they are reading. If they work for one of the broadsheets they are generalists by and large, expected to cover everything thrown at them. When it gets to sharp end of medical research they don't have a prayer of understanding the complex issues . When you get to the real specialist stuff even another researcher in another medical field will struggle.

As a layman it seems to me self evident that Covid is a seriously nasty disease, it has killed tens of thousands of people around the world, even more have complications from the illness that will be with them for the rest of their lives.

Just about every Government in the world has been forced into bringing in severe restrictions and lockdowns to reduce the infection rate and manage the pressure on hospital beds. These politiciand are fallible individuals but they did have expert advice. They generally followed the science and took the recommended action. I'm prepared to assume they knew more than most of us did and give them the benefit of the doubt. They were faced with a serious threat and dealt with it. The fact that our pro business Conservative Government introduced a lockdown that damaged business interests is suggestive that they had liitle alternative

It might not be what we desire but there are flare ups around the world that need dealing with. That isn't peoples imagination it is happening.

If you don't want to believe Covid is a threat that's fine, I'm not going to try to persuade you to change your mind.

I'm lucky I live in a rural location and I'm retired, there's no need for me to interact with others unless I consider it safe. I do appreciate that others have no such luck.

[johnhemming]

I've found over the years that people believe what they want to believe and read articles/ watch programmes that support that belief. They find reason to discredit things that go against their belief. Trying to persuade them that they are wrong is impossible or at least very difficult.

I think you are right about this as a general principle. I find online debates helpful to enable me to challenge my own views and ensure that I get facts right. I also tend to read material sourced from people who I don't agree with.

The issue of confirmation bias is a key issue in investment.

However, I do have quite a bit of experience of the public sector and systems of government and am perhaps not as confident as you as to the ability of most politicians in selecting which science to follow.

[scotia]

He cited an experiment conducted during his time at the BMJ, in which a short paper containing eight deliberate errors was sent to 300 reviewers, as evidence of how badly the system can fail to detect errors.

“No one found more than five, the median was two, and 20 percent didn’t spot any,”

Yes - regrettably true. Going back a long time, I found that peer reviewing a paper required a lot of work - especially since I didn't have the capability of online searches. And it became increasingly obvious that others were not taking a similar amount of care.
However - on the statistics you quoted above - I reckon if I had discovered 5 errors, I wouldn't have bothered looking for more - I would simply have reported it as not worthy of publication.
But yes - I still think a flawed system of reviewing is better than none.

[Nimrod103]

As Scotia has said most general newspaper reports are pretty useless,

Please, on the subject of Carl Heneghan, we are not talking about newspaper 'reports'. These are articles (the Spectator one at least) under his own name, and presumably carry his full authority and approval. If he is wrong, it is his neck on the line.

[scotia]

As Scotia has said most general newspaper reports are pretty useless,

Please, on the subject of Carl Heneghan, we are not talking about newspaper 'reports'. These are articles (the Spectator one at least) under his own name, and presumably carry his full authority and approval. If he is wrong, it is his neck on the line.

I should possibly bear part of the blame. I was referring to two newspaper references - one from the Daily Mail, the other from the Spectator. The headline in the Daily Mail would not, I hope, have been written by Carl Heneghan - so all trust in such an article evaporated very quickly. The Spectator article is considerably more scholarly - but my personal past experience of national newspapers and TV does not always support the claim that all that is reported is exactly as the author intended. So I'll stick to scientific journals - and am willing to read internal reports on authors' sites - although I hope they will subject their work to appropriate scrutiny.

[scotia]

Interesting newly published set of experimental results from Iceland in the New England Journal of Medicine
https://www.nejm.org/doi/full/10.1056/NEJMoa2026116?query=recirc_curatedRelated_article
and the accompanying editorial is
https://www.nejm.org/doi/full/10.1056/NEJMe2028079?query=recirc_curatedRelated_article

[dealtn]

and accounts of trips to the supermart are laden with horror stories about non-distancing and non-masking.

I think we were brought up on different versions of horror stories!