Lancet's May 2020 hydroxychloroquine COVID study — is essentially worthless — yet everyone is citing it anyway

© 2020 Peter Free

 

27 May 2020

 

 

Mental vacuity contagion

 

As a thematic follow-up to my blurb about the CDC's astonishing stupidity in mixing viral RNA and antibody test results, I comment today regarding similar brainlessness from another source that should know better.

 

This one revolves around The Lancet's retrospective study, which ostensibly cast deep shade upon using hydroxychloroquine as a COVID-19 medication:

 

 

We were unable to confirm a benefit of hydroxychloroquine or chloroquine, when used alone or with a macrolide, on in-hospital outcomes for COVID-19.

 

Each of these drug regimens was associated with decreased in-hospital survival and an increased frequency of ventricular arrhythmias when used for treatment of COVID-19.

 

© 2020 Mandeep R Mehra, Sapan S Desai, Frank Ruschitzka, and Amit N Patel, Hydroxychloroquine or chloroquine with or without a macrolide for treatment of COVID-19: a multinational registry analysis, The Lancet, DOI: https://doi.org/10.1016/S0140-6736(20)31180-6 (online first, 22 May 2020)

 

 

First, even the Lancet study's authors caution readers that

 

 

[A] cause-and-effect relationship between drug therapy and survival should not be inferred.

 

These data [also] do not apply to the use of any treatment regimen used in the ambulatory, out-of-hospital setting.

 

Randomised clinical trials will be required before any conclusion can be reached regarding benefit or harm of these agents in COVID-19 patients.

 

© 2020 Mandeep R Mehra, Sapan S Desai, Frank Ruschitzka, and Amit N Patel, Hydroxychloroquine or chloroquine with or without a macrolide for treatment of COVID-19: a multinational registry analysis, The Lancet, DOI: https://doi.org/10.1016/S0140-6736(20)31180-6 (online first, 22 May 2020) (at pages 8-9)

 

 

Nevertheless, the authors' cautionary advisement has not stopped our culture's plentiful supply of politicized airheads in over-extending the study's findings — exactly in the manner that the authors had warned against.

 

 

Evidently, Trump-hatred . . .

 

. . . and his promotion of hydroxychloroquine use in the coronavirus pandemic — trounces the application of common scientific sense.

 

 

Let's take a closer look at . . .

 

. . . what was methodologically questionable in this Lancet study.

 

Both these flaws are arguably inexcusable, when one considers the study's investigative purpose.

 

First

 

The study makes no mention of using hydroxychloroquine in conjunction with zinc supplementation.

 

That's odd. Getting zinc into virus-infected cells is the main reason that many physicians have been enthusiastic about trying hydroxychloroquine in the COVID context. Its method of action gives zinc a way to get into cells, so as to inhibit viral replication.

 

Second

 

The study makes no mention (at all) of when — with regard to the coronavirus infection's suspected, not hospitalized, onset — hydroxychloroquine dosing was started.

 

That is also a strange oversight. Hydroxychloroquine's proponents unanimously say that early-on dosing is mandatory, if the treatment is to benefit patients.

 

 

So — questionable investigative faith?

 

The Lancet group completely missed the two most basic points that underlie empirical uses of hydroxychloroquine that are receiving favorable attention in other countries.

 

The Lancet study's authors — apparently intentionally, given their lack of genuine discussion of both elements — ignored the two major reasons why a significant proportion of hydroxychloroquine-using medical professionals (around the world) turned to using that inexpensive generic drug in the first place.

 

 

Another major methodological flaw

 

This one points to the unpersuasive attempt (by the study's authors) to say that they statistically controlled for 10 potentially confounding variables:

 

 

age

 

sex

 

race or ethnicity

 

body-mass index

 

underlying cardiovascular disease and its risk factors

 

diabetes

 

underlying lung disease

 

smoking

 

immunosuppressed condition

 

and

 

baseline disease severity.

 

 

No disrespect is intended . . .

 

. . . but even a statistical genius could not pull that one off.

 

The study's sample size is too small (even with 96 032 patients) to statistically winnow the potential machinations that the paper's ten confounding variables potentially contributed.

 

In thinking about this methodologically, one has to be aware that these confounders would almost certainly be acting in unknown and unpredictably intertwined ways.

 

Yet despite that, the study's authors want us to believe that they achieved a statistically persuasive handle on such an immensely complicated mess.

 

 

This impossibly optimistic . . .

 

. . . 'we controlled for' assumption (by itself) should motivate tossing this study into a landfill.

 

One cannot claim to have done the methodologically impossible, and then expect other folks to trust one's reasoning.

 

A major part of the scientific process is understanding the phenomenological nature and interactive complexity of the precisely defined arena that one is operating in.

 

Researchers cannot just magically wave off variables and considerations that so obviously challenge the research integrity of the problem that they are investigating.

 

Are there other indications that this study may have been, similarly, too casually undertaken?

 

Yes.

 

 

The paper's problematic supplementary Table S3

 

French physician (Didier Raoult) — who stands behind his empirical use of hydroxychloroquine in treating COVID-19 — in spite of this study's purported findings — noticed a probable indicator of falsification or data-fudging in a table documenting the studied patient groups' demographics.

 

Dr. Didier drew attention to Table S3 — Summary Data by Continents. That table appears here, reached via the Lancet article's Supplementary Material link.

 

An examination of S3 indicates highly improbable percentage similarities, recorded with respect to demographic and medical traits, that (according to the Table) are surprisingly uniform in going from one continent to another.

 

Dr. Raoult highlighted only five of those in his tweet, here.

 

However, my eye is suspicious of the numerical evenness of the whole Table.

 

It is simply not likely that so many demographic and medical condition characteristics would be virtually identical, percentage-wise, when distributed between and among continents.

 

Even someone only casually associated with medicine and public health knows that physical, medical and ethnicity-associated traits vary noticeably from one culture, place or people to another.

 

If you don't believe me, do your own research. For instance, track varying smoker percentages from one place or culture to another. Those numbers are often relatively easily available.

 

Then, compare the numbers that your online research obtains with those that the "current smoker" line in Table S3 says that it found among its widely distributed hydroxychloroquine patients.

 

Smell a rat?

 

I might be able to accept such a smoothing of demographic-medical data for just one or two demographic characteristics. But seen across virtually the entirety of the Table S3, the harmony there recorded is too placidly smooth to trust.

 

I strongly suspect, with Dr. Didier, that these numbers have been distorted in some fashion.

 

 

I am not attributing evil intent to the study's authors

 

It is possible, by way of one easy hypothesis, that their attempt to mathematically control for the listed 10 potentially confounding variables bit them in the behind.

 

However, the fact that no one on the authorial team apparently noticed — or commented upon — S3's improbably smooth picture of intercontinental Reality is not confidence-generating.

 

Why should it take outsiders to notice glaring improbabilities like those recorded in Table S3?

 

 

The moral? — Lancet's 22 May 2020 hydroxychloroquine COVID study is essentially worthless

 

Few of hydroxychloroquine's proponents thought that the drug (with or without zinc) was going to be therapeutically effective, once patients were severely ill enough to present themselves at hospitals.

 

This means that the Lancet study essentially attacked a straw man. Presumably (one might infer) just to get itself into prestigious (and politicized) print.

 

And the media and people — who are now citing the Lancet study as presumptive proof of hydroxychloroquine's lack of efficacy against COVID-19 — are merely indicating the depths of their own inability to think scientifically.

 

Unfortunately for us, that sort of mental vacuity seems to be the cultural norm now.

 

Witness the CDC and FDA's serial bone-headedness during the same pandemic.

 

Are we living a 'retarded' society's attempt to suicide?