Clearly biased review.
around "analysis" and "study" show the authors' bias.
Did not (and have not) contacted us.
Authors have never contacted us, questioning their interest in accurate
Incorrectly claims that we assume all
individuals in treatment countries are treated.
We don't. Anyone reading
as far as the second sentence of the paper will know this. For the analysis to
be useful, we only need to know that usage is significantly higher in
treatment countries, which is supported by hundreds of references.
Claims 100% adherence in RCTs.
that 100% adherence is required in an RCT, however imperfect adherence is
common. For example, adherence in the NEJM PEP study was 75%.
Claims cherry-picking on countries clearly not
in either group.
Authors did not read the paper. They claim cherry
picking for Brazil, Spain, and Italy. However, the study compares countries
that made clear decisions for the majority of their outbreak. Brazil, for
example, started usage relatively late and has increasing but very mixed use.
Extensive supporting references are provided in the Appendix. Authors are
correct that Indonesia should have been excluded (this has been
Incorrectly claims we stated HCQ was 100%
We never stated this, and it makes no sense. While at this
time, 100% of papers in certain categories present positive results, we very
clearly state that the nature and degree of these results varies widely.
Baseless claims of misprepresention of other
Authors provide no details of any incorrect claim in
our paper, while misrepresenting our study and other studies themselves. It is
true that a few studies have claims unsupported by the data, which we note.
Interestingly, authors here appear to believe that certain studies could not
have any errors, a standard which is apparently selectively applied to benefit
Authors cherry-pick 5 of over 70
Authors cherry-pick 5 studies to support their claim of
inefficacy, neglecting to mention the other 65+ studies, and neglecting to
mention that every one of the 5 they cite is discussed in detail in our paper
(and we reference all of the 65+ other studies as well).
This site has cherry-picked and misrepresented
research in the past.
We note this site has a history of cherry-picking
and misrepresenting research in this area, see [this
for example, where authors conveniently cherry pick less than one
third of studies supporting their conclusion, and claim flaws in studies not
supporting their conclusion, while ignoring flaws in studies supporting their
conclusion. For example, authors say one study "concluded that HCQ did not
prevent COVID-19 infection within four days after exposure", however the
treatment in that study (an Internet survey RCT) was not even administered
within four days - medication was actually shipped to patients, and shipping
was not done over the weekend. It appears unlikely that authors read the
studies they are using. In another example they note a trial of "mild to
moderate" patients showing no improvement, while neglecting to mention that
"mild to moderate" in this study refers to hospitalized patients with 14%
randomized in the ICU, and neglecting to mention the very late
Claims intense debate on safety.
debate continues on efficacy, the majority of those closely following the
research (on both sides) agree that early safety concerns (largely from the
retracted Lancet study), were unfounded and not supported by subsequent
trials, especially for early treatment.
Relies on Twitter personalities.
refer to extremely inaccurate analysis from an extremely biased Twitter
personality. See our response
Random claims about connections.
make random claims about connections with other Twitter accounts because they
were referenced as providing feedback, an implied ad hominem attack.
False analysis claims. "The authors
made all of their calculations using the sample size of 2.7 billion people,
based on the population sizes of countries in both groups, which invalidates
all their statistical analyses and the conclusions they drew from them."
Nothing about this is correct. The entire population is used in only one
instance, where we present the relative risk for the entire population in each
group (which is approximately the same as when averaging across countries).
Claims a sample size of 19.
of deaths alone in the countries analyzed is over 300,000.
Claims missing details with no examples.
However, this is a very simple analysis and we believe that all data and
details to reproduce the results are provided.
Incorrectly claims IFR should be used.
Authors apparently did not read the paper which explains why case statistics
are not reliable.
Incorrectly claims not a randomized clinical
Indeed, it's not a clinical trial, but we have never said it is.
Authors are confusing "clinical" with "controlled". It's a trial - a
medication is certainly being trialed. It's controlled - there is a control
group. It's randomized - the group a person gets is random and chosen in
advance, independent of their medical status or the membership of the other
group (in contrast to a retrospective observational study).
Irrelevent ecological fallacy
Claims ecological fallacy (you cannot directly infer the
properties of individuals from the average of a group), however the study does
not do this.
Claims "observational ecological
However this is a prospective study - assignments were done
in advance, a person's assigment was random, independent of their medical
status, and there is a control group.
Authors note the
demographic differences between countries, failing to note that we actually
adjust for these.
Claims "numerous large and robust clinical
trials which suggest no beneficial effect of hydroxychloroquine treatment in
While there are negative trials for late stage
treatment, all trials concerning early treatment are either positive or
This review appears to have
been done by people who are unfamiliar with the existing body of research, are
not interested enough in accuracy to even contact us, and rely on claims from
biased Twitter personalities rather than reading the paper.