"You do not really understand something unless you can explain it to your grandmother." Albert Einstein So what I am seeking to compute from the cited data is, given that : 1. LF test is negative 2. LF test is negative and PCR test is positive what is the probability that someone is corona infected ?? "Actually it reminds me of my most interesting (so not that interesting) brush with stats that is completely off topic but I will relay it anyway. When I was working on the Sizewell B Public Inquiry there was a question which asked the CEGB something to the effect of what was the chance of the power station blowing up. The response was along the lines of 'so small it cannot be calculated'. So a paper was later submitted to the Inquiry with a title like "The chance of a fully loaded jumbo jet crashing on a fully occupied Wembley Stadium", just to prove that things with a really small chance of happening could be calculated." Off topic, but kindred nonetheless. There is a branch of mathematics called "utility theory" . What it does is associate a value (the "utility" ) with an event occurring. A positive number means good, negative means bad. So in the case of Sizewell B, probability theory may have given that scenario a low probability P, but the utility function U would have been a very big negative number. Probability calculations performed in the likes of "decision support systems" apply the utility as a multiplier which propagates thru.
Actually it was about the probability of infection if the LFT is positive and the follow up PCR is negative.
Statistics is actually a very useful branch of mathematics, but its usefulness is massively diminished by people who don't understand its limitations quoting results which don't demonstrate what they claim.
My mistake on the ordering. But to be clear, from the article you cited, the data only allows this to be computed : P(corona+ | (PCR- and LF+ ) ) There is no notion of "follow on" .
Agreed but the practical situation is that you are required to have the PCR test if the LFT is positive. The chance of someone genuinely having Covid at the time the LFT is done and genuinely not having it a few hours later when the PCR is done are probably small enough to be ignored.
Based on your source data : P(corona+ | LF- ) = 0.003022052 So given : 1. your LF test is negative, there is a 3 in 1000 chance that you are infected. 2. the probability that 1 in 100 of the population are deemed to be infected, #1 is not sufficiently low enough IMHO to use the LF test for anything
My point was about positive lateral flow test results. The LFT certainly can't confirm you are negative but is useful to trigger isolation if it comes out positive.
P(corona+ | LF+ ) = 0.96 So the summary is (for LF tests only) : - 4 in 100 will be incorrectly deemed to be infected - 3 in 1000 will be infected, deemed to be clean and escape the net This is "actionable data" . So from that ... 1. How should the healthcare system manage subsequent PCR testing ?? The primary drivers will be the following ratios ( PCR / LF) : - cost - time to acquire results - test throughputs - test capacity 2. For a "decision support" system , what utility do you place on those : - 4 in 100 who are incorrectly removed from society (work etc) - 3 in 1000 who will be roaming free
Most people I know working in care who have to test daily just use the liquid with the LFT to get a negative, they don't do the test properly because they don't want to lose money. Particularly if there is a risk of a false +ve.
That's another failure of Government policy. You shouldn't lose money from a positive test as you harm society more by not isolating. We can withstand the few false positives having to isolate....we've failed to learn the lesson that the thing that harms the economy is people actually being ill which means the right thing to do is to act early to prevent that.
As if I needed an excuse not to buy their over priced cheese on bread , makes me wonder what the Dominos marketing dept are on
That's a bit like saying burger King is a bit better than McDonald's...May well be true but I ain't buying either pile of ****e tbh
Ah, but Sky need the viewing figures over Christmas weekend, so try not to cough on each other too much
