Trevor BedfordVerified account

Today's 210k cases is a new record. If trend continues as it has, it's likely that ~3500 of these individuals will succumb to their disease and these deaths reported later in the month. 4/4https://twitter.com/COVID19Tracking/status/1334661580064239616 …

The US is reporting over 2000 deaths per day from Dec 1 and I believe will do so consistently throughout December based on daily case loads above 120k starting early November. 1/4https://twitter.com/trvrb/status/1326404894954283008 …

Follow up #2: This doesn't mean that COVID-19 was completely absent from the US in January 2020, just that prevalence at that time was exceptionally low. Finding 0/3600 COVID+ acute respiratory specimens doesn't square with theoretical 2% ELISA positivity in Dec.

Follow up #1: Also, a reminder that we at the @seattleflustudy PCR tested 3600 samples from individuals with acute respiratory illness collected in January 2020 from Seattle and found zero positives for COVID-19. This is a much more specific assay.https://twitter.com/trvrb/status/1249414295042965504 …

The other angle to consider is that if we're supposed to believe that 2.0% of random blood donors in Dec 2019 are COVID+ this would translate to millions of infections in the population at large, in which case we would have noticed due to people dying in large numbers. 10/10

The authors highlight the study's limitation due to "potential cross reactivity with human common coronavirus infection" in the paper's discussion, but it unfortunately didn't make it into the @WSJ story (https://www.wsj.com/articles/covid-19-likely-in-u-s-in-mid-december-2019-cdc-scientists-report-11606782449 …). 9/10

It seems highly likely to me that the 39 "positives" from Dec 13 to Dec 16 reported by Basavaraju et al are due to cross-reactivity from recent seasonal coronavirus infection. It would just take a slight decrease of assay specificity to ~98% to explain this outcome. 8/10

However, there is ample reason to expect that individuals recently recovered from seasonal coronavirus infection will have more cross-reactivity to SARS-CoV-2 than random healthy adults. In fact this can be seen in this paper by Freeman et al (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7239067/ …). 5/10

The authors tested 1912 blood samples collected between Dec 13 and Dec 16 2019 and observed 39 positives (2.0%). A Fisher's Exact Test comparing 3/519 to 39/1912 is narrowly significant with p = 0.02. 4/10

The ELISA used by the authors has a stated specificity of 99.3% and the authors tested 519 "true negative" blood samples collected from 2016 to 2019 from healthy adults and suspected hanta virus patients and observed 3 false positives (0.6%) matching this specificity. 3/10

The authors do a careful serological investigation, but it necessarily suffers from testing a large number of samples with an assay that is not perfectly specific. 2/10

I don't think that this study by Basavaraju et al from @CDCgov can be taken as evidence that #COVID19 was circulating in the US in December 2019. 1/10 https://academic.oup.com/cid/advance-article/doi/10.1093/cid/ciaa1785/6012472 …

Reporting of cases and deaths will likely be down over Thanksgiving weekend. We'll have to wait until the week of the 30th to have a good sense of where things are headed. 10/10

This method projects that the huge ongoing case loads will translate to an average of over 2500 deaths per day in just two weeks, even assuming no further increases in circulation. 9/10

Although Rt is generally moving in a good direction, I don't know how rapidly societal measures can bring Rt below 1 across the US. 7/10