Read on Twitter
Unfortunately, this is not good reporting from NPR.
It takes an extraordinary claim — that SARS-CoV-2 was circulating widely in the US in December 2019 and January 2020 — and presents it as fact.
Spoiler: it's probably not fact at all. https://www.npr.org/sections/coronavirus-live-updates/2020/12/01/940395651/coronavirus-was-in-u-s-weeks-earlier-than-previously-known-study-says?utm_campaign=storyshare&utm_source=twitter.com&utm_medium=social
It takes an extraordinary claim — that SARS-CoV-2 was circulating widely in the US in December 2019 and January 2020 — and presents it as fact.
Spoiler: it's probably not fact at all. https://www.npr.org/sections/coronavirus-live-updates/2020/12/01/940395651/coronavirus-was-in-u-s-weeks-earlier-than-previously-known-study-says?utm_campaign=storyshare&utm_source=twitter.com&utm_medium=social
https://www.npr.org/sections/coronavirus-live-updates/2020/12/01/940395651/coronavirus-was-in-u-s-weeks-earlier-than-previously-known-study-says?utm_campaign=storyshare&utm_source=twitter.com&utm_medium=social
The core the problem has to do with the way that science is presented in the popular press. Even a peer-reviewed study in a good journal rarely establishes something unexpected with absolute certainty.
And it definitely doesn't overturn mountains of evidence to the contrary.
And it definitely doesn't overturn mountains of evidence to the contrary.
The problem is, this study is presented as if it establishes fact in isolation, without bringing in the extensive context into which this paper was published.
A key mistake is that there is no effort to interview or get perspective from scientists not involved with the research.
A key mistake is that there is no effort to interview or get perspective from scientists not involved with the research.
The context? We have a huge number of samples — thousands, if not tens of thousands by now — of PCR-negative samples from that same time period without a single positive.
Yet this paper used antibody tests to uncover nearly 100 purported positives in a sample of seven thousand.
Yet this paper used antibody tests to uncover nearly 100 purported positives in a sample of seven thousand.
We have well over 3000 North American genomes for the virus; they tell a story consistent with a mid-to-late January introduction, not a December one.
(The new paper does not provide genomes, only antibody test results).
(The new paper does not provide genomes, only antibody test results).
Less compelling, but we also did not see the wave of pneumonias of unknown etiology that you would expect had there been a December introduction.
So what happened? Why does this paper find an incidence of over 1% in blood donation samples taken from Dec 13 2019 to Jan 17 2020?
We don't know for certain, but one reasonably hypothesis is cross-reactivity with one of the other, human-endemic coronaviruses.
We don't know for certain, but one reasonably hypothesis is cross-reactivity with one of the other, human-endemic coronaviruses.
Indeed this seems far more likely, in light of the broader context described here.
I'm not an expert in this specific area, but @trvrb is. Here are his thoughts. https://twitter.com/trvrb/status/1333647437869633537
I'm not an expert in this specific area, but @trvrb is. Here are his thoughts. https://twitter.com/trvrb/status/1333647437869633537
This is the context that @NPR absolutely should included before publishing the story they did.
In general (and even for studies with conclusions that are likely correct), it should be a red flag if a science news piece quotes only the press release and the authors of a study.
In general (and even for studies with conclusions that are likely correct), it should be a red flag if a science news piece quotes only the press release and the authors of a study.
