Thread:
Its pretty clear that we aren't doing enough testing.
But how much is "enough"? There is no easy way to assess.
So over the past few weeks, we've been trying to take different approaches to estimate this number.
The first question is -- enough for what?
1/n
Its pretty clear that we aren't doing enough testing.
But how much is "enough"? There is no easy way to assess.
So over the past few weeks, we've been trying to take different approaches to estimate this number.
The first question is -- enough for what?
1/n
What if I want to identify all COVID19 patients?
That's easy -- test all 320 million Americans (assuming a perfect test)
OK...that's impractical
What if I wanted to capture "most" patients. It begins by first estimating how many people have COVID today?
Our estimate....
2/n
That's easy -- test all 320 million Americans (assuming a perfect test)
OK...that's impractical
What if I wanted to capture "most" patients. It begins by first estimating how many people have COVID today?
Our estimate....
2/n
Our estimate is that there are about 150K+ new COVID19 infections in the US per day right now.
If we want to identify most COVID19 patients, we estimate probably need to be doing a couple of million tests today
3/n
If we want to identify most COVID19 patients, we estimate probably need to be doing a couple of million tests today
3/n
OK -- that's also impractical.
How about if we want to "open up" on May 1? Then what?
We did a deep dive, talked to dozens of folks, and tried to come up with estimates.
We estimate, using a variety of approaches, that it will take well over 500,000 tests a day
4/n
How about if we want to "open up" on May 1? Then what?
We did a deep dive, talked to dozens of folks, and tried to come up with estimates.
We estimate, using a variety of approaches, that it will take well over 500,000 tests a day
4/n
Read more about our approach -- its still in draft form and is not a "research paper" but an outlining of our approach and thinking.
globalepidemics.org/2020/04/18/why…
globalepidemics.org/2020/04/18/why…
We welcome feedback
One thing is clear to me -- we are likely underestimating the # of tests. May be a lot
We likely need millions per day as @paulmromer and others have said
500K is the floor. The bare minimum. I don't know how we open -- and stay open -- with less
Fin
One thing is clear to me -- we are likely underestimating the # of tests. May be a lot
We likely need millions per day as @paulmromer and others have said
500K is the floor. The bare minimum. I don't know how we open -- and stay open -- with less
Fin
