Shouldn't we optimize for a higher number of false positives? In the end, sending someone in quarantine for 2 weeks without them being infected is probably less dangerous than a false negative. So making tests overly sensitive would seem more applicable to me here.