My personal summary of #TWiV with @michaelmina_lab in 2 pictures: How we test, and how we should. Essentially we have over-fussed to get the highest sensitivity & specificity in #covid19 tests...without worrying about what patients we want to be sensitive for.Let's fix that fast! 
Sensitivity is needed when ppl start producing *lots of virus*: that's when they infect others. Don't care if we miss low-dose cases, so we really only need a "low-quality" test (counterintuitive!). The key: at <$1/test we could test daily in rich (level 4 @Gapminder) countries.
If today we miss 75%-95% of #covid19 cases with our expensive, sparsely used tests,even a 100% sensitive test only gives us 5-25% of true cases. By testing everyone, even with a really crappy test that only catches 75% of cases, we would increase by factor 3-15 how many we catch.
I still need to think through the false positive story on this, that wasn't discussed by @michaelmina_lab on #TWiV, but the testing approach got me really excited. Happy to hear your thoughts on this!
This video is also a very good summary of #TWiV 640 which explains it in more detail with nicer graphic than me. They don't emphasize the typical delay between viral load and symptom onset as much, but show some actual data around sensitivity:
