Thread by @DrAndreKalil

Should CDC recommend #COVID19 rapid antigen test for asymptomatic isolation exit?
R antigen test:
Picks up 2/100 who are contagious
Allows 3/100 to exit isolation still contagious
No help in 95/100
⬇️
Antigen test has no utility to exit asymptomatic isolation and may harm
⬇️
🧵1

FDA: Antigen test has a 60% false-negative rate in asymptomatic ppl at the infection start (screening)

Likely >60% at the infection end (isolation exit) → due to lower expected viral replication

But let’s use 60% assuming test will not lose performance at isolation exit
🧵2

CDC: 90-95% will not have any more replicable (no transmission) #Omicron at day 5

Asymptomatic at the infection end is even less contagious and

Vaccinated ppl are less likely to transmit

Thus the actual no transmission rate is >95%, but let’s use 95%
🧵3

If 5/100 (5%) are still contagious at day 5/asymptomatic, and antigen test misses 60% of contagious ppl, then only 2 positive/100 will be detected (assuming antigen test never gives false-positive result), while 3 positive/100 will wrongly exit isolation still contagious
🧵4

Rapid antigen test leads to more mistaken isolation exits than correct picking up of contagious ppl, which may cause harm:
Contagious ppl with false-negative antigen test may have false sense of security and less incentive to wear mask, unknowingly further spreading Omicron
🧵5

What about serial antigen tests?
Serial antigen tests at the end will be useless because different from infection start when the replication is rising, at the infection end the replication is falling, so the chance of catching live virus in a 2nd test will be even smaller
🧵6

What about negative predictive value instead of false-negative rate/sensitivity?
The higher the virus circulation-like now
⬇️
The higher the #Omicron prevalence
⬇️
The higher the pretest probability
⬇️
The lower the negative predictive value➡️antigen test performance worsens
🧵7

To be clear, I am not talking about testing at infection start in symptomatic ppl (diagnostic testing),
a situation in which rapid antigen tests are proven to be useful and
should be widely distributed and accessible worldwide
🧵8

I am talking about testing at infection end in asymptomatic ppl (screening for isolation exit)

By any basic scientific standards, current rapid antigen tests have no screening utility to exit asymptomatic isolation and may cause harm
🧵9