I posted this thread earlier because there's a serious challenge ahead: now that antigen tests are being used on a huge scale (WHO bought 120M; Abbott is producing ~50M/month), key Q is: what to do w/ result?
CDC says: + antigen test = "presumptive case".
https://wwwn.cdc.gov/nndss/conditions/coronavirus-disease-2019-covid-19/case-definition/2020/ https://twitter.com/DiseaseEcology/status/1310735831666024448

Understandably, this makes some health care professionals uncertain what to do next, especially if confirmatory PCR test takes days to get results back. Do they treat it like a confirmed case, try to get case to isolate & trace their contacts?
If few false +, no problem!

But if false +s are relatively common (even 1-2% would count as relatively common) & antigen tests are used for frequent screening of asymptomatic people, this would lead to large number of isolated/quarantined people, awaiting PCR tests. Here's an example...

My university is using 2x/week screening of ~2000 people to keep transmission limited (rationale: fig from
https://www.medrxiv.org/content/10.1101/2020.06.22.20136309v3). If prob false + = 1%, this would result in 0.01*4000 tests/week = 40 false +/week (plus their contacts: 5 each=200) needing isolation/quarantine.

Same logic applies to a skilled nursing facility or jail or hospital that performs same # of tests. If you have quick confirmatory PCR testing (or as @michaelmina_lab suggests, a 2nd antigen test for another part of virus protein), no problem. If not, this creates big headache.

Other well-known issue is lower sensitivity for RNA than PCR. No problem if you test more frequently to catch high viral loads when you are most infectious ( https://twitter.com/DiseaseEcology/status/1271281847754846211), but more tests = more opportunities for false +s. So, it's a bit of a spiral.

Would love to see Abbott & other antigen test companies recognize this challenge & build in set of confirmatory tests into pack of regular tests as @michaelmina_lab explains: https://twitter.com/michaelmina_lab/status/1297231253041709057

Otherwise this testing approach, which is exactly what is needed in principle, won't provide much help, especially in low transmission settings & especially for asymptomatic screening where prevalence is often <1%.