I'm glad to see mention in this article of the #1 issue with antigen test such as Abbott BinaxNOW (the one the fed. govt. just bought 150M of): 98.5% specificit

I'm glad to see mention in this article of the #1 issue with antigen test such as Abbott BinaxNOW (the one the fed. govt. just bought 150M of): 98.5% specificity = 1.5% false positives. Test a school of 1200 with an actual 0% prevalence and you'll get about 20 false positives. This will look like an outbreak, and cause chaos. Anyone responsible for server infrastructure knows what happens when you get a bunch of false alerts from your monitoring service - you send those emails to your spam filter [i.e. assume every positive is false] or turn off monitoring altogether [i.e. stop testing]. Think about 20X false positives per day. Also if people are self-testing, they'll assume their test is a false positive and just won't report the results. Confirmatory lab testing could avoid a full 10-day quarantine but the time to get an RT-PCR test and then get the result (24-72 hours when the system isn't overloaded) means several days of disruption for those 20 kids (/ faculty / staff). I think this essentially kills this concept, assuming other antigen tests have similar specificity. The Quidel antigen test has had a couple of high-profile false positive incidents already, so I suspect that is the case.