They talk about potential false negative, but should have talked about potential false positive from sample cross contamination, because this isolated case seems in contradiction from everything we know about the french contamination history, a researcher that follow closely the phylogeny of the virus seems very dubious about that one. Now we need a sequence of the genome and serological test of the patient, the family and co-workers.
Also this is not the first hospital to do that in France, IHU Méditerannée Infection, from Raoult and Chloroquine fame did that at the beginning of the epidemic in China, they tested 2500 samples from several month ago and found absolutely nothing.
French contamination history? I'm sure the French have modern microbiology and virology techniques...
That said, I do agree that a seismological test of the patient would be more definitive, but sometimes you cannot contact the patient due to privacy concerns/ IRB rules. I don't know how French hospitals deal with patient data, but if it's anything like America, you are highly unlikely to be able to do those sort of things without additional approval.
I suspect they propose to detonate lots of small explosions at regular time intervals and at regular spacings on the surface of the patient, and then measure the responses using seismographs attached elsewhere to the patient, like what the oil companies do to produce maps of subsurface geology.
A microbiology researcher I know (one who is a former resident of France, no less) informs me that cross-contamination is a real risk, and that when it comes to tests involving PCR the level of paranoia involved to guard against this is not trivial.
I would really like to see them confirm this case with antibody testing.
Even with antibody testing, it would still be possible that the sample from December was a false positive, and the patient later contracted CoVID-19 and developed antibodies, but it would increase the certainty a great deal. If the person has antibodies and does not report having had an additional bout of CoVID-19-like symptoms after the initial disease, then this would be much more certain.
It is not. An article from last month put the first case found by contact tracers to the beginning of January, at least two weeks before the first official "imported case". Sad that the full report hasn't been published:
Paywalled: https://www.lemonde.fr/societe/article/2020/04/08/coronaviru...
Also this is not the first hospital to do that in France, IHU Méditerannée Infection, from Raoult and Chloroquine fame did that at the beginning of the epidemic in China, they tested 2500 samples from several month ago and found absolutely nothing.