Article

2020-05-01 Science & Technology

Coronavirus antibody test with 99% success rate approved for use in Europe

Archived material is restricted to Rantburg regulars and members. If you need access email fred.pruitt=at=gmail.com with your nick to be added to the members list. There is no charge to join Rantburg as a member.

Posted by g(r)omgoru 2020-05-01 03:57|| || Front Page|| [6 views ] Top

#1 Cuban elections had 99%+ outcomes too.

Posted by M. Murcek 2020-05-01 05:31|| 2020-05-01 05:31|| Front Page Top

#2 Mister relevance.

Posted by g(r)omgoru 2020-05-01 05:43|| 2020-05-01 05:43|| Front Page Top

#3 Wasn't there a news post that the South Koreans found the reinfection numbers were false positives?

Posted by Procopius2k 2020-05-01 07:27|| 2020-05-01 07:27|| Front Page Top

#4 Another antibody test, P2k.

Posted by g(r)omgoru 2020-05-01 07:35|| 2020-05-01 07:35|| Front Page Top

#5 MM, IIRC, that's the same as California Mail In Ballots, true?

Posted by AlanC 2020-05-01 08:19|| 2020-05-01 08:19|| Front Page Top

#6 No, Cali mail in ballots are 110% certified.

Posted by M. Murcek 2020-05-01 08:37|| 2020-05-01 08:37|| Front Page Top

#7 So 1:100 tests have bogus results?
In 100,000 tested there will be 1000 infectors?

Posted by Skidmark 2020-05-01 14:24|| 2020-05-01 14:24|| Front Page Top

#8 #7 Nope.
(a) 0.01 is both False positive (F+) and False negative (F-), fractions, and it's unbelievably good. Most medical tests have F+, F- = 0.1 (10%).
(b) 1% F+ are not infective - they are not immune - there is a difference.
Finally, what's most useful is estimating fraction of immune in the population, Skid (the famous "herd immunity").
Denote True positive (T+).
And say, fraction 0 < x < 1 are immune. Then you run a test on a representative sample and get
(O+) = positive/total = (T+)x + (F+)(1-x) ==>
x = (O+ - F+)/(T+ - F+).
The results Lex always waves around are O+ and they mean zilch if F+ is large.

Posted by g(r)omgoru 2020-05-01 14:51|| 2020-05-01 14:51|| Front Page Top

#9 p.s. Suppose you use the test to determine who can go to work. So, suppose one F+ sneaks through. And he goes to work. Let us be generous and assume he's asymptomatic carrier (infectious). Whom is he going to infect? The chances of two F+ at the same place are (0.01)^2 = 0.0001*number of workers. And, even then, the other F+ may not be susceptible but rather another asymptomatic.

Posted by g(r)omgoru 2020-05-01 14:59|| 2020-05-01 14:59|| Front Page Top

23:58 Dron66046
23:58 Whiskey Mike
23:57 Whiskey Mike
23:56 Whiskey Mike
23:55 Whiskey Mike
23:55 Whiskey Mike
23:55 Dron66046
23:53 Whiskey Mike
23:52 Whiskey Mike
23:51 Whiskey Mike
23:46 Whiskey Mike
23:45 Dron66046
23:37 gorb
23:33 Marilyn Tojo7566
22:53 SteveS
22:23 Clem
22:21 Clem
22:13 Lex
22:11 Whiskey Mike
22:01 Whiskey Mike
22:00 Whiskey Mike
21:58 Whiskey Mike
21:57 Whiskey Mike
21:55 Whiskey Mike