Read on Twitter
1/
TW, #Covid19uk case scenarios
IIUC
Here's the 2 HUGE problems with this for human beings.
The best many antibody tests can get in accuracy is high 90%s (97% is reasonable)
The 3% are antibodies too close to differentiate
So 3% of results SUGGESTING they're immune, but https://twitter.com/bbckamal/status/1245800546893774850
TW, #Covid19uk case scenarios
IIUC
Here's the 2 HUGE problems with this for human beings.
The best many antibody tests can get in accuracy is high 90%s (97% is reasonable)
The 3% are antibodies too close to differentiate
So 3% of results SUGGESTING they're immune, but https://twitter.com/bbckamal/status/1245800546893774850
2/
they are not immune.
The certificate or wristband does not confer immunity.
NOW, if a smaller fraction of the population has the antibodies this problem is amplified.
e.g. If circa 5% of the UK population (call it 3mn people) had #covid19 and we test 10% of the population
they are not immune.
The certificate or wristband does not confer immunity.
NOW, if a smaller fraction of the population has the antibodies this problem is amplified.
e.g. If circa 5% of the UK population (call it 3mn people) had #covid19 and we test 10% of the population
3/
(call that 6mn) to try to catch /everyone/ who has had it, because we've tested double the amount of people that actually had covid19 the inaccuracy total is doubled too (it's still 3%, but 3% of 6mn).
That could be ~180,000 people with certificates of immunity who are NOT
(call that 6mn) to try to catch /everyone/ who has had it, because we've tested double the amount of people that actually had covid19 the inaccuracy total is doubled too (it's still 3%, but 3% of 6mn).
That could be ~180,000 people with certificates of immunity who are NOT
4/
immune.
And this gets WORSE NOT BETTER.
The two ways to prevent this are
a) Only test people you're sure had covid19. Small, tight, fairly pointless numbers
or
b) Wait until a larger proportion of the population has been exposed, e.g. 30%, THEN test 10%
immune.
And this gets WORSE NOT BETTER.
The two ways to prevent this are
a) Only test people you're sure had covid19. Small, tight, fairly pointless numbers
or
b) Wait until a larger proportion of the population has been exposed, e.g. 30%, THEN test 10%
5/
In the original scenario where circa 5% of the population have had covid19, if (and data is patchy but this doesn't seem unreasonable) IF there's a 1% mortality rate, 1% of 3mn would be 30,000.
Covid has always been as-much or more about healthcare systems being overwhelmed
In the original scenario where circa 5% of the population have had covid19, if (and data is patchy but this doesn't seem unreasonable) IF there's a 1% mortality rate, 1% of 3mn would be 30,000.
Covid has always been as-much or more about healthcare systems being overwhelmed
6/
frankly the maths on that has always been horrifying AF and I'm not going to break it out here, this part is bad enough.
In the later scenario above where 30% of the population has had covid19, at a 1% mortality rate ...
And THIS was the path the tories chose, rather than
frankly the maths on that has always been horrifying AF and I'm not going to break it out here, this part is bad enough.
In the later scenario above where 30% of the population has had covid19, at a 1% mortality rate ...
And THIS was the path the tories chose, rather than
7/7
lockdown early, long term; work on a vaccine, on the supposition from unnamed human-behaviourists that people couldn't stick to it and *it would be bad for the economy*
I've spent a long time running numbers on covid19.
I hope I'm wrong with the numbers in this thread.
lockdown early, long term; work on a vaccine, on the supposition from unnamed human-behaviourists that people couldn't stick to it and *it would be bad for the economy*
I've spent a long time running numbers on covid19.
I hope I'm wrong with the numbers in this thread.
