LA County is beginning a program of randomized serological Covid-19 tests, using a marketing agency's tool for selecting representative samples to administer 1,000 tests at drive-up windows.
latimes.com/california/sto…
1/
latimes.com/california/sto…
1/
They want to conduct this test fortnightly for three months, which will cost $1.1m.
It's a crucial step, one that will infuse our policy (health, economic, educational, etc) discussions with much-needed facts.
2/
It's a crucial step, one that will infuse our policy (health, economic, educational, etc) discussions with much-needed facts.
2/
Getting a baseline of infection rates versus people reporting (even mild) symptoms will help us get a better handle on the danger presented by the virus, correcting for the selection bias that attends only testing people known to have been exposed or presenting with symptoms.
3/
3/
It's true that we don't know if the presence of antibodies indicates immunity. There have been some reports of reinfections by people who had symptoms, tested positive, then tested negative (some even had multiple negative tests).
4/
4/
This is a separate, massive unknown that randomized testing won't necessarily help clarify. If it's true that recovering from coronavirus doesn't confer immunity, it will be very dire, because it means that we won't be able to attain herd immunity without a vaccine (~18 mos).
5/
5/
I am tentatively optimistic that this is not the case, based on analyses like this:
medium.com/wintoncentre/t…
Tldr: if you administer a LOT of tests, then even very rare errors will crop up (a 1-in-a-million error will happen about 5 times in 5m tests)
6/
medium.com/wintoncentre/t…
Tldr: if you administer a LOT of tests, then even very rare errors will crop up (a 1-in-a-million error will happen about 5 times in 5m tests)
6/
It seems, on balance, that it's more likely that the very unlikely event that someone had a malady other than coronavirus also had a false-positive coronavirus test, got better, then actually GOT coronavirus, than it is that recovering from covid doesn't confer immunity.
7/
7/
The former is very, very unlikely. My lay understanding of virology suggests that the latter is very, very, VERY unlikely. When a phenomenon is very, very, very, VERY widespread, both outcomes are possible.
8/
8/
(But also: motivated reasoning is a thing. All our lives would be much, much better if recovery conferred immunity, so it may be that I'm unconsciously overweighting that possibility for self-serving reasons).
eof/
eof/
