Thread by @MackayIM

What is a false positive coronavirus test? How is it possible? abc.net.au/news/2020-06-0…
"Usually due to a lab being contaminated with the virus or with product from a previous test"
-or clinic
-also sometimes due to quirks of the test; which is why we try to validate well

And to clarify
"Certainly at the moment there was no active infection present, but what may have happened is that it was the end of a previous infection — the test may have detected a little bit of that going on,"
- certainly is too strong a sentence starter.

Just to be clear, I don't have any inside info on the autopsy. It was reported that the deceased "did not have the disease after all" (news.com.au/lifestyle/heal…)

Also...
"it is also the first time PCR is being used [for coronavirus detection] in real-time — PCR is being put to the test."
-that real-time RT-PCR is being asked to determine if virus in clinically well people & in samples prior to symptoms, at such scale & with so many tests

And why do we think a false positive is so unlikely? Usually for a false (+) result to occur, you need to have some virus or previous test product get into the sample at some point in the process; during collection or handling or once in the lab

But AFAWK there are no cases in Blackwater or Emerald or pretty much anywhere in Qld at the moment. So where could any virus or positive test product come from?
If not another case, then perhaps positive control material used in the test? That depends on the test used.

Point-of-care commercial tests often have controls built in; other kits may have a control reagent that is specific to a control tube while in-house methods could use virus, or synthetic templates that could (relatively) more easily spillover-but they also have greater expertise

I'm not sure the Emerald lab is doing any in-house testing though.
Virus from culture then? Hard no. That requires a specialised lab, none of which are located in regional Qld areas AFAIK.

So I remain of the opinion that either this initial test had some rare foible/issue that threw up a false positive, or it was a true positive - *late or resolved* - infection, that couldn't be repeated due to the person being deceased.

But no-one else was acutely positive in the town. So..
I wonder if serology was done. I don't imagine the town would not be too happy to have more testing done at this stage though.
Just some thoughts anyway.

According to a new article in the Courier Mail newspaper (also quote form me), the test was a GeneXpert (cepheid.com/coronavirus, cepheid.com/Package%20Inse…), cartridge-based (2 target & internal controls), point-of-care, rapid nucleic acid test.

POS and NEG control runs are recommended for each new lot/shipment of the cartridges are received, a new operator starts or there are problems. They are run separately (different from in-house and some other molecular tests)-control material added to a distinct cartridge and run

The POS control material (seracare.com/AccuPlex-SARSC…) is non-SARS-CoV-2 (non-replicating recombinant heat-treated Sindbis virus), non-infectious & contains the target regions of interest. It could be a potential source of contamination if it got into the same cartridge as the sample

If so though, it should produce a solid POS result, not a very late (?40 cycles) one as occurred here.

Not a lot of validation done on this kit & mostly synthetic or cultured material, not real-world samples.
cepheid.widen.net/s/rkmhdbdm7o
But (also small size which may not pick up foibles..if any exist) reports like this suggest it does generally work well IRL sciencedirect.com/science/articl…