The massive confusion created by the media and the announcement of SalivaDirect for #COVID19 testing needs to be cleared up.

Yesterday I wrote a long tweet thread about it. The continued frenzy today tells me that extreme confusion persists.

Here goes:

1/

First I do not want to take anything away from SalivaDirect and the efforts of @YaleSPH and @NathanGrubaugh et al for getting this approved through EUA - no small feat!

For clarity, particularly for decision makers, The hype should be greatly tempered.

2/

The SalivaDirect assay is a High complexity CLIA laboratory qPCR lab based assay that resembles the first CDC assay in March, performed on manual loaded qPCR instruments.

It is not a high throughout or rapid test. It is a regular qPCR test like the CDC test.

3/

A very technically savvy lab can put it onto programmable liquid handling robots and make it more high throughout, yes. But this is very technically demanding, requires much more expertise than most CLIA labs have. Very few clinical laboratories will have this ability.

4/

If going through the trouble of setting up robots, this could already be done with the CDC assay. But when we did this w CDC assay, we found saliva was very difficult for the robots. And so we abondoned saliva early bc we recognized the need for high capacity - I.e robots.

5/

I worry that this same problem will persist as the SalivaDorect approach doesn’t solve inconsistencies in saliva or difficulties with this material on the robots.

6/

The SalivaDirect did remove a potentially reagent limiting step - RNA extraction (at some cost in sensitivity). But again, if doing this in high capacity on automated robots, the RNA extraction w regular CDC assay was already available and very cheap (ie magmax beads).

7/

So this is a good step to remove but has very marginal gains.

The SalivaDirect media says the assay will cost $5 or so. But this is reagents only. Regular CDC assay is nearly the same cost. (Primers/probe from IDT + rna extract about $5/test reagents cost).

8/

The real omission is that the cost of regular manual qPCR like SalivaDirect or CDC assay isn’t the reagents, it’s everything else that comes with setting up a high complexity CLIA lab and running thousands of samples. And these manual tests are quickly bottlenecked

9/

In addition, the SalivaDirect is not a rapid test. The reports say 3 hours or so. This is very similar again to original CDC assay (SalivaDirect is essentially the CDC assay with the removal of RNA extraction and performed on saliva)

Why the comparisons to the CDC assay?

10/

Because it’s been around and this manual type of PCR that both assays use has proved very very very difficult to scale. The bottlenecks in the US have not been the collection method - saliva versus self collected nose swab - they have been in the lab.

11/

It cannot be stressed enough that another manual PCR test a la CDC assay is not going to be a major game changer. The costs are ~$3 cheaper w out RNA extraction, yes. But that’s small potatoes. Saliva is nice - but may not be great in high throughout from our experience

12/

This thread is not meant in any way to diminish a new protocol like SalivaDirect from coming on line. But the hype has caused major confusion. Few labs will see a major benefit moving to this and those that can use this are likely already doing the CDC assay.

13/

While we need incremental improvements where possible - like SalivaDirect - For *laboratory based* diagnostics, we need major leaps that move away from manual qPCR (even if on robots - which we do use) and that can greatly enhance throughput like @ginkgo is aiming for.

14/14

Just one additional point - although I have been advocating a lot for distributed at home tests, my hospital life is as a medical director in a high complexity CLIA lab. I am not anti CLIA or qPCR. Just really see their downsides when used for wide scale public health testing.