[conspire] When to get a covid test, brains, and a weird case, and a cardiology vs. immunology joke

[Rick Moen]

Quoting Dire Red (deirdre at deirdre.net):

> _The Lancet_ and others were caught up in outright data fraud trying
> to sell snake oil in the form of chloroquine / hydroxychloroquine
> (which, to be clear, do work for some autoimmune diseases and malaria,
> and some side effects in some viruses) earlier this year:
> 
> https://www.sciencemag.org/news/2020/06/mysterious-company-s-coronavirus-papers-top-medical-journals-may-be-unraveling
> 
> The peer review process is not designed to catch data fraud that deep.

Exactly.  As you and I have discussed, it's very analogous to
(non-forensic) financial auditing of public companies:  The scope of
auditing -- my long-ago former profession -- involves an effort to 
verify the firm's financial condition and results of operations, but 
is by design is not promised to catch systemic financial fraud.

In the case of the May 'Hydroxychloroquine or chloroquine with or with
or without a macrolide for treatment of COVID-19' metastudy paper 
in _The Lancet_, it turns out to have been fatally flawed by bad data
from Chicago-based data-analytics company Surgisphere, which is why 
_The Lancet_ retracted the paper (and why _NEJM_ and others are combing
over other papers based on Surgisphere data).

(I summarised, above, as a 'tl;dr' aid for mailing list members.)

_All_ scientific findings are subject to review; some have more inherent
credibility than others.