Thread by @KatherineJWu

A few weeks ago, I heard scattered rumors about bizarre positives coming out of coronavirus testing programs at universities.

It didn't seem to be the coronavirus. It also wasn't contamination in the labs processing the tests. It was weirder. 1/

nytimes.com/2020/11/12/hea…

Researchers working with harmless, noninfectious genetic material from the virus (in the form of DNA) were testing positive, over and over again.

They weren't shoving their science up their noses. They were being careful, and doing great work.

The DNA clung to them anyway. 2/

If that DNA happened to overlap with the target of a coronavirus test, that quickly spelled trouble for some. The test picked up that "contaminating" DNA, and thought hey, this is exactly what I was looking for. Positive. 3/

This might sound pretty harmless. But a positive test has a lot of repercussions: isolation for you; quarantine for your close contacts; lost productivity; separation from family and friends. And, crucially, no more testing for you for the next 90 days, per CDC guidelines. 4/

That's frustrating, stressful, terrifying, and ultimately potentially dangerous. If a contamination positive isn't caught, people might be duped into thinking they're sick, then immune. What if they catch the actual virus in those 90 days? 5/

How would they know, and how would they protect the people around them? 6/

There's nothing in the CDC guidelines to specifically address this issue. That's caused some serious issues at some universities: People have been removed from testing programs for months. People questioned the testing process as a whole. 7/

To be abundantly clear: these contaminations are NOT, NOT, NOT something the general public should worry about. Amplicons are not flying all about, screwing up tests. 8/

And this is NOT a knock against molecular testing (tests that look for the coronavirus's genetic material). In fact, it's kind of a weirdly complimentary advertisement for how sensitive tests like PCR can be. The test DID ITS JOB. 9/

But this to me speaks to the importance of keeping lines of communication about testing open. We need to be flexible. We need to recognize that errors will happen. There will be unexpected hurdles and roadblocks. 10/

A potential solution for this one? Offering a second set of tests that looks for a different region of the coronavirus's genome, maybe one that doesn't overlap with the DNA being used in a research lab. Or doing some follow-up tests on the original patient sample. 11/

What is NOT a solution? Just issuing a blanket policy dismissing all of these cases as presumed "false" positives or mistakenly labeling them as all true positives. There needs to be evaluation on a case by case basis. A mistake in either direction could lead to an outbreak. 12/

I don't even like calling these "false" positives. They're technically not. Again, the test did its job. (This is actually type 3 error!) 13/

But there's another line of thought to this. When testing snafus happen, how can people best communicate them? A lot of researchers called for transparency. Talk about the issues, work together to resolve them. 14/

That way this does not seem like a "the sky is falling" disaster. It also opens up the possibility for more conversations on testing in general. All tests will have some flaws and drawbacks. 15/

But if people don't hear the right things about tests, and how to interpret them, and what results do and don't mean, they will be confused and frustrated. They might lose faith in tests. 16/

Certain tests out there ARE more prone to mistakes. To false positives and false negatives. Those are different than what happened here, but also crucial to discuss. What these errors have in common: They must be communicated. 17/

A lot of new tests are coming online, and being rolled out nationwide. That's an important endeavor. The country needs more tests. But this is an unprecedented situation, and it's crucial to be cautious. 18/

Thank you to @SBtotheDub and Rosemary She for weighing in on this crucial topic as outside experts. 19/

Thank you to members of the @geochurch lab who talked to me about these issues, and how they uncovered and addressed their own contamination: @LRR_M @GabrielFilsing1 @TimWannier (and @alex_mijalis who helped spearhead a project to track the DNA) 20/

@jenny_m_tam was a huge part of this effort too; the group even wrote a preprint! osf.io/9svjq/ 21/

Many others were instrumental in putting this piece together, not all of whom I can name. I'm grateful to every single one, and hope this starts some productive conversations. 22/22