It depends on the goals of the test.
For our family of four, having two of us take the Roche test with over 99% specificity - and having them both return negative - provides me with some reasonable assurance about the unlikelihood that our family was exposed. I mean, sure, general epistemological limitations and the necessarily-acknowledged possibility for/improbability of two false-negatives can draw those results into question.
But unless someone wants to explain to me otherwise, it's good enough for me to support some decisions.
If the goal is to form public policy, however,