IOM Covid removing restrictions

[TheTeapot]

People in this thread showing the same ability to properly interpret statistics as Josem.

[thesultanofsheight]

2 minutes ago, Roger Mexico said:

It's amazing how many people only come to this conclusion once the figures that they have just been quoting have been disproved (which incidentally was Mark Twain's point)

I said he once said it, which he did, I didn’t say it was attributed to him. He attributed it wrongly to Disraeli, but he still said it as a quote in a book he published. 

[TheTeapot]

2 minutes ago, Gladys said:

What happened on 4 September?

Without going on to the ons site to check, I'd guess a data backlog from the late August bank holiday.

[Roger Mexico]

1 minute ago, Gladys said:

What happened on 4 September?

August Bank Holiday (these are week ending dates of course).  These are based on date of registration not death so any bank holidays throw the number of deaths registered off for the week because the offices were open for one day less.   You can see this shown in Figure 1 of the ONS report.

[TheTeapot]

Oh look!

[Roger Mexico]

2 hours ago, wrighty said:

Antibody studies here suggest only 16% of close household contacts of covid cases got it. And many of them were probably before we knew so much about it. It’s not that transmissable. 

But (simplifying madly) even that would mean in a family of four, if one arrived from across with the virus they would have a roughly 50% chance of infecting one of the others.  If you're trying to keep the virus completely out of the community, those aren't great odds.

That's before you consider the problems with antibody testing.  I've not seen the results of the larger study yet, but the pilot study showed that only about 75% who had tested positive with PCR for Covid[1] showed positive for antibodies.

The number of close contacts (presumably mostly family/household members) in the study was only fairly small (152) of these 13 had been PCR tested of whom 10 had been positive, though of course that 13 would have been tested for a reason on top of being a contact.  Of the remainder about 10% were antibody-positive.

 

[1]  They actually managed to retest about two-thirds of all the surviving cases on the Island, which I don't think there was enough bragging about.  Testing was on average 69 days from PCR test and more might have been positive earlier, but of course the same would also apply to close contacts.

[thesultanofsheight]

45 minutes ago, Roger Mexico said:

But (simplifying madly) even that would mean in a family of four, if one arrived from across with the virus they would have a roughly 50% chance of infecting one of the others.  If you're trying to keep the virus completely out of the community, those aren't great odds.

That doesn’t make sense. 

[Roger Mexico]

5 hours ago, Cambon said:

The virus is not as transmissible as they say. They are trying to scare people. Virtually all transmission is by contact.

No it isn't I'm afraid.  COVID-19 is thought to spread mainly through the respiratory route (both droplet and aerosol), after an infected person coughs, sneezes, sings, talks or breathes. Of course the nearer you are to an infected person and longer you are near to them, the more likely you are catch it that way.  And it is also possible to catch it from contact, but the main route is through droplets/aerosols.

[thesultanofsheight]

7 minutes ago, Roger Mexico said:

COVID-19 is thought to spread mainly through the respiratory route (both droplet and aerosol), after an infected person coughs, sneezes, sings, talks or breathes.

Thought .. 

[Gladys]

1 hour ago, Roger Mexico said:

August Bank Holiday (these are week ending dates of course).  These are based on date of registration not death so any bank holidays throw the number of deaths registered off for the week because the offices were open for one day less.   You can see this shown in Figure 1 of the ONS report.

Is the system updated in real time or weekly, because I wouldn't expect a one day closure (20%) would result in such an anomaly. I am not disputing the graphs, just that jumps out and unless you have a qualification in stats, it is not easily understood. (BTW, I did study stats as part of a business related qualification, but that was many, many years ago!)

[hampsterkahn]

It is a travesty that in a world so desperately in need of virologists, epidemiologists, clinicians, statisticians  and  even half- good clairvoyants they are all  to be found here, just wasting their time arguing and insulting each other  on Manx Forums.

[Utah 01]

11 minutes ago, hampsterkahn said:

so desperately in need of virologists, epidemiologists, clinicians, statisticians  and  even half- good clairvoyants

There's bleedin' tens of thousands of them.  They've all crawled out of the academia woodwork since March.  It was beyond imagination that there were so many of them closeted away in 'public health' departments and universities all over the planet.  What did they do all day apart from peer-review each others' clairvoyance abilities?

[Roger Mexico]

1 hour ago, Gladys said:

Is the system updated in real time or weekly, because I wouldn't expect a one day closure (20%) would result in such an anomaly. I am not disputing the graphs, just that jumps out and unless you have a qualification in stats, it is not easily understood. (BTW, I did study stats as part of a business related qualification, but that was many, many years ago!)

It's actually that much of an anomaly - it's the difference with the average of the five preceding years after all, not the absolute value, which probably fell by less than 20%.  The two surrounding weeks will have risen correspondingly.

I suspect it's exaggerated this year because the August Bank Holiday may have fallen in a different numbered week from what it normally does and the ONS will be comparing a four day Week 37 (say) with the average of five five day ones.

[Roger Mexico]

1 hour ago, thesultanofsheight said:

2 hours ago, Roger Mexico said:

But (simplifying madly) even that would mean in a family of four, if one arrived from across with the virus they would have a roughly 50% chance of infecting one of the others.  If you're trying to keep the virus completely out of the community, those aren't great odds.

That doesn’t make sense. 

I was probably over-simplifying far to much but the underlying idea is that if someone who lives with an infected person has a 16% chance of catching it, then if there are three people living with them the second and third also have a 16% chance each, so (simplifying) the chance of someone getting it is 3 x 16% = 48% ie about half.

Actually it's a bit less, because the probability of at least one of the three getting it is 100% - (84% x 84% x 84%) = 41%, but the exact figures don't really matter, the basic point is the commonsense one that the more people there are to be infected, the more likely one of them will be. 

[Gladys]

1 hour ago, Roger Mexico said:

It's actually that much of an anomaly - it's the difference with the average of the five preceding years after all, not the absolute value, which probably fell by less than 20%.  The two surrounding weeks will have risen correspondingly.

I suspect it's exaggerated this year because the August Bank Holiday may have fallen in a different numbered week from what it normally does and the ONS will be comparing a four day Week 37 (say) with the average of five five day ones.

It's actually a lot of an anomaly, even if it is a comparison of an actual 4 day week with an average of 5 days you would be able to explain only a 20% or thereabouts deviation, surely?