The link from @flyrodder that I was responding to about the IHS float mentioned this number.
Note that the point is not whether PCR tests are 70%, 80% or 90% accurate in the real world (I doubt that 70% is the whole story) but rather that whatever the number is, other tests can never be more accurate than PCR when the measure is taken to be agreement with the PCR test rather than rating PCR and alternative test against a common population of standard samples.
This is the skewed world we face when a test achieves incumbency and everyone is expected to report on consistency with the incumbent rather than the incumbent and pretender being compared against a common population of test cases whose positive/negative status is know though some independent process.
