|
Re: Coronavirus
The test used is pretty accurate with good sensitivity and specificity (sensitivity is correctly diagnosing a positive and specificity is correctly diagnosing a negative) I have seen numbers of 99.9% for both. That's pretty good for a clinical diagnostic test!
RT-PCR is quite a technically demanding technique but the failures are more likely to be false negatives than positives
BUT....
99.9% specificity means a false infection rate of 100 per 100,000. Even if COVID went away today, we would likely still see this background noise so at low infection rates, the effect of false positives is large.
Yesterday, we saw 4368 cases from 219723 tests. Even if we take the theoretically false positives out (220 cases) we are still a smidge over 4000 cases, up from 3105 or ~2900 when you remove theoretical false positives last week
|