Status
Not open for further replies.

Mr. Tea

Let's Talk About Ceps
Anyway, do you (@mixed_biscuits) accept that simply increasing the test rate in a certain area isn't going to falsely trigger a new outbreak warning for anyone with the capacity to understand the difference between an absolute number and a percentage?
 

mixed_biscuits

_________________________
Their published stats do not take into account the false positive rate; I think they're just ignoring that issue, which is why people like Prof Heneghan are bringing it to our attention.

The spikes that motivated those local lockdowns were in the tens of cases per area - nothing significant.
 

vimothy

yurp
It doesn't mean that they ARE all false.... but it means that it is quite possible.
(I'm not saying that this is what is happening but that is the case IF the false positive rate is indeed 1 percent).
it just means that the ratio of positives to all outcomes is the same as the ratio of false positives to FP and true negatives. you can get that first number in a variety of different ways. mixed_biscuits is implying that true postives and false negatives are both zero. in other words the test is completely useless bc the disease does not exist. which might be true (idk), but it's not enough to point to the equality of these two numbers to establish that.
 

IdleRich

IdleRich
it just means that the ratio of positives to all outcomes is the same as the ratio of false positives to FP and true negatives. you can get that first number in a variety of different ways. mixed_biscuits is implying that true postives and false negatives are both zero. in other words the test is completely useless bc the disease does not exist. which might be true (idk), but it's not enough to point to the equality of these two numbers to establish that.
No it doesn't establish that but it means it is possible.
 
Is prevelance that low or is it gliding around undetected though? The large scale outbreaks in meat packing plants and mushroom gaffs (Ireland admittedly) have to come from somewhere

Or is it coming from imports like the Brazilian meat in China.

Good for speculating this virus to be fair
 

Mr. Tea

Let's Talk About Ceps
The spikes that motivated those local lockdowns were in the tens of cases per area - nothing significant.
Current rate of daily new cases for the UK is just over 1/100k. Blackburn's latest figure is 80/100k for the last week, or more than 11 times higher.


I'd call that "significant"!
 

mixed_biscuits

_________________________
Is prevelance that low or is it gliding around undetected though? The large scale outbreaks in meat packing plants and mushroom gaffs (Ireland admittedly) have to come from somewhere

Or is it coming from imports like the Brazilian meat in China.

Good for speculating this virus to be fair

Heneghan says that with current prevalence, we wouldn't notice the virus if we weren't looking for it so I take it they are looking for it quite assiduously
 

mixed_biscuits

_________________________
Current rate of daily new cases for the UK is just over 1/100k. Blackburn's latest figure is 80/100k for the last week, or more than 11 times higher.


I'd call that "significant"!

Significant in the colloquial sense of 'important' - both data are v low
 

vimothy

yurp
This is what is at play: https://en.wikipedia.org/wiki/Base_rate_fallacy

A 1% false-positive rate wouldn't be too much of a problem given high prevalence but it is a huge problem if prevalence is very low, as it is now in the UK.
what actually _is_ the test sensitivity for covid tests?

"A systematic review of the accuracy of covid-19 tests reported false negative rates of between 2% and 29% (equating to sensitivity of 71-98%)"


sensitivity > 70% and potentially as high as 98% according to the study referenced here
 
Status
Not open for further replies.
Top