Thread by @michaelmina_lab

To detect #COVID19 before it spreads to others, we need frequent accessible testing.

We cannot detect pre/asymptomatic people before they spread if we do not test frequently.

A new article in @NEJM by ⁦@DanLarremore⁩, Roy Parker and Me

1/x nejm.org/doi/full/10.10…

An ideal screening test is one with high sensitivity.

During a pandemic of a fast moving virus that transmits asymptomatically, it is difficult to detect people before they transmit to others.

2/x

This far we have focused almost all of our screening efforts on the use of the very sensitive (and specific - a good thing) qPCR.

The qPCR meets the molecular needs of detecting this virus. It has an extraordinary sensitivity.

But it is extremely limited

3/x

Focusing on PCR as a screening method today necessarily equates to low frequency screening.

With a virus that transmits before people feels symptoms, low frequency testing equates to missing the infectious stage of most people’s infections. Catching them too late, or never.

4/x

So in the @NEJM piece, we suggest that the goal should shift away from the sensitivity of the test to detect molecules and towards the sensitivity of the testing program to find infectious people and filter them out before infecting others.

5/x

In other words, we want to focus on the *sensitivity of the testing regime* to detect and stop transmission, rather than focus on the analytical sensitivity of the test to find molecules in the relatively small number of samples that can be reasonably tested.

6/x

The PCR test is wonderful and has a terrific sensitivity. But it comes at the expense of being very limited.

In short, the best test in the world with the highest molecular sensitivity has a near zero % sensitivity to detect infectious people if it can barely be used.

7/x

If the world can create cheaper faster tests that can be produced and distributed to millions of people and used frequently, then the greater frequency more than makes up for the potentially lower molecular sensitivity of the frequent and accessible tests.

8/x

Ultimately this means that a low molecular sensitivity test that is used very frequently by many people will have a MUCH greater *sensitivity to catch infectious people* before they have a chance to spread virus to others.

To put this in perspective...

9/x

Collectively, in US, our screening programs based on high molecular sensitivity PCR likely catch less than <5% of infectious people in time to act and prevent them from spreading.

Thus, our PCR based testing has a <5% sensitivity to detect #COVID19 before it spreads.

10/x

This is why we must rethink the meaning of sensitivity of #COVID19 tests. We must move away from thinking just about the ability to detect molecules and towards the ability to detect infectious people.

The sensitivity of the testing program must be high - not the test.

@US_FDA