Mr. Tea

Shub-Niggurath, Please
Staff member
45 cycles of PCR testing means any RNA traces are replicated 2⁴⁵ times to get a result. Seems legit.
So every one of these diagnoses is false, you're clever enough to have spotted it and it hasn't occurred to a single person who does this stuff for a living? Seems legit.
 

mixed_biscuits

_________________________
Oxford CEBM: 'We deduce that a reported “case” is most probably simply the result of a positive PCR test. The new guidance is meaningless unless it provides a clear threshold for the limits of detection.'
 

Mr. Tea

Shub-Niggurath, Please
Staff member

"
Given the variation in the data, it’s now much harder to be confident, by simply looking at the graph above, that the proportion of positive tests is increasing. Carl Heneghan, a professor at the University of Oxford’s Centre for Evidence-Based Medicine, collating cases by the date of sample rather than the date of reporting, has even suggested that the proportion of positive cases might be decreasing.

But this isn’t the end of the story either because of an issue in the way that tests are reported. Christina Pagel, a professor at University College London, pointed out that the number of “test processed” almost certainly includes multiple tests per person – the same person being counted as one case but being tested at least twice. For example, in the week beginning July 16, around 545,000 pillar-two tests were processed, but only 240,000 people were actually tested.

We should really be dividing positive tests by the number of people tested to find out whether the proportion of positive testing people is increasing. The good news is that this data is available on the NHS test-and-trace report website. The bad news is that the data is only reported weekly and lags by around a week. Still, the data shows that the number of tests processed is increasing more quickly than the number of people tested.

When the number of weekly pillar one and pillar two cases are divided by the number of people tested, the proportion of people testing positive in England appears to fall until mid-July. However, it has shown consistent rises for the last two weeks (see the chart below)."
file-20200806-42413-641h2r.png
 

Mr. Tea

Shub-Niggurath, Please
Staff member
That article links to an ONS page, showing results of their ongoing survey, which is independent of the NHS and derives from sampling taken entirely at random. It therefore isn't vulnerable to accusations that it's simply showing the result of increased testing in a few areas where infection rates remain above average, because the percentage of positive tests is completely independent of the number of tests performed. And it shows a small but significant uptick since late July.

Granted, it's not as dramatic as the rise suggested by the overall number of positive tests, but it does show that the rise cannot be blamed entirely on increased testing and that the underlying infection rate is rising again.
 

Mr. Tea

Shub-Niggurath, Please
Staff member
I'm absolutely not arguing for another general lockdown, by the way. I think there was a good argument for that back in April and May, even if its implementation was piecemeal and far less effective than it could have been. It strikes me that targeted lockdowns of areas where infection rates are increasing - and by that I mean individual towns or even particular neighbourhoods of particular towns - would be the way to go, while the rest of the country goes about its business.
 

mixed_biscuits

_________________________
@Mr. Tea It's complicated: you can have the same pos tests/total tests ratio but also either pos_tests/distinct testees > pos_distinct_testees/distinct_testees or the other way around.

Depending on how data of people who test positive and also negative are processed, could also have anything between pos_distinct_testees/distinct_testees = 1 and pos_distinct_testees/distinct_testees = 0
 

mixed_biscuits

_________________________
Would need to tie each person's particular test results to them only (rather than just saying there are x repeat tests, as repeating negatives affect many of the summary statistics differently to repeating positives) and also decide how to cash out test results that contradict previous results.
 

Mr. Tea

Shub-Niggurath, Please
Staff member
Possibly so, I'd have to think about it.

But the important thing is that M-Day looks like it might be arriving a few days earlier than predicted.
 

HMGovt

Bamber Clatscoigne

The consultation covers:
  • authorising temporary supply of an unlicensed product
  • civil liability and immunity
  • expanding the workforce eligible to administer vaccinations
  • promoting vaccines
  • making provisions for wholesale dealing of vaccines
OK. Having unqualified people delivering an unlicensed vaccine on a mass scale. Has anybody heard of this consultation until today?
 

Mr. Tea

Shub-Niggurath, Please
Staff member
USA's official death toll stands at 182,400 as of now. With an averaged daily toll of 965, that gives us ~19 days till the magic number of 200,000 is reached. Given that the death rate is very slowly decreasing, and may be down to 900/day within a couple of weeks, I'm prepared to push the predicted date of what I'm calling "M-Day" out to September 18.
Two days early.
 

IdleRich

IdleRich
I thought that it wouldn't be until next year though? Whoever (surely not the same person who said that his friends believe all his theories cos the predictions are so accurate?) said that should probably apologise, re-think his or her model and reverse all previous statements on the topic.
And just generally be really embarrassed.
 

Mr. Tea

Shub-Niggurath, Please
Staff member
I thought that it wouldn't be until next year though? Whoever (surely not the same person who said that his friends believe all his theories cos the predictions are so accurate?) said that should probably apologise, re-think his or her model and reverse all previous statements on the topic.
And just generally be really embarrassed.
And probably never mention the subject again as long as he, or she, should live.
 

IdleRich

IdleRich
Ah this is interesting at least. A direct and simple prediction that can either be wrong or right. As long as you recognise worldometer or whatever... I can see a scenario where that ticks over 200k and then you say "Ah but you have to divide that number by 17" or something. But no that's unworthy of me.
Or not...
 

Mr. Tea

Shub-Niggurath, Please
Staff member
Got to wait for the definitive count as I dare say it will be recounted downwards like ours.
Excess deaths in the USA were over 240,000 for less than six months in the spring and summer.

Perhaps there will be a small downward adjustment. There could just as easily be an adjustment in the other direction. Either will be much smaller than the number who've died from it without being diagnosed.
 
Top