HcQ is not an Information argument

HCQ doesn’t work? Not sure :

By now it should not surprise me that creationists will deny the mountains of evidence that demonstrate evolution to be true, yet cling to the slightest scrap of a scientific publication that seems to support something they want to be true.

It shouldn’t surprise me, but it still does.

Here are the facts:


[quote=“Giltil, post:212, topic:14233, full:true”]

So a “retrospective”, “monocentric” study (with no baseline data; I searched but couldn’t find) conducted from March - December 2020 is supposed to convince us that HCQ works in the face of several older and more recent large, randomized, double-blinded clinical trials which have shown no benefit for HCQ with or without azithromycin, zinc etcetera for Covid-19 treatment or prophylaxis? You certainly don’t know how medical research works.


This graphic in the Cochrane article you cited beautifully illustrates the state of evidence for and against HCQ usage.


There’s also the molecular data showing that SARS-CoV-2, unlike SARS-Cov-1, does not depend on the endosomal pathway affected by HCQ.

Metaphorically, the HCQ-sensitive pathway is the front door for 1 (explaining the robust inhibition of 1 seen in cultured cells in 2004), but just a back door for 2. HCQ would almost certainly (and only) have value as an adjunct in combination therapies that target 2’s front door.


HCQ may be as effective as M&M’s, but molnupiravir is looking very promising…

Merck pill breakthrough raises hopes of preventing COVID deaths

Mechanism of molnupiravir-induced SARS-CoV-2 mutagenesis


This study is poorly designed, Gilbert.

Raoult’s recent study is retrospective, which means it is subject to selection bias and to recall bias. It is better to look to the several large prospective, random-controlled trials on HcQ, don’t you think?

Those who want desperately to believe in HcQ’s ability to heal COVID are probably going to welcome this new Raoult study.

But that should be irrelevant. Those who value making decisions based on data from high-quality studies have no reason to stop relying upon the results from previously published prospective, random controlled trials (RCTs). The Cochrane Group, cited by @Faizal_Ali, lays out the data on HcQ gathered by RCTs quite well.

Meilleures salutations,
Christophe Falter


I think that’s being too nice. There was no random assignment. There was no blinding.

I would suggest that it was designed to show that HCQ works without caring whether it actually does.

1 Like

Gil, I advise you not to use this study as a basis for your personal health decisions. I see too many problems with the study, and it it far from conclusive that HcQ is actually an effective treatment.

The group receiving HcQ was younger and likely had fewer comorbidities (they didn’t have that data). It was a Day hospital, and patients receiving care may not represent the general population. The number of deaths was only 16, so there was very little power to control for other factors. This makes the statistical estimates very unstable. The statistical method used, logistic regression, cannot account for patients lost to follow up or different lengths of follow-up, which will bias results.

There IS a statistical difference between AZ and AZ+HcQ groups (I checked it myself) but that difference is not necessarily due to HcQ. The difference could be due to other reasons, like non-random treatment assignment, comorbidities, or bias in the data and analysis.

There are better statistical methods that could have been applied to this data (case-control matching, propensity score weighting, survival analysis), and I’m surprised this paper was accepted for publication without requiring those other analyses.


There is something genuinely interesting about the way that pseudo-epidemiology and pseudo-virology line up with creationism. I’m not entirely sure what to make of it. It’s clear that while some of this can be chalked up to basic inability to weigh evidence and understand scientific work, there’s more going on here: these beliefs would not “cluster” this way if there were not something weird happening. Nor would they correlate so well with beliefs about the last national elections here in the USA.

No really simple formulation, I think, can credibly account for the cause of all of it. The culture war is such an odd beast and it’s hard to say where it begins. I am just beginning, after doing some reading on the Russian revolution, to try to understand American reactions thereto, which seem to me to lie at the base of at least one of the causal chains involved here. But there really is no inherent reason why these beliefs should cluster so. There are lots of science-illiterate people who don’t think the earth is young, or that AIDS is caused by something other than HIV, or that the elections were stolen, or that Ivermectin in horse-deworming dosages is a great COVID treatment. And if they don’t think ONE of these things, they’re less likely to think the others. They may pursue all manner of nonsense. They may be having their chakras aligned right now, and I have at least one old friend who almost certainly is. But once one wanders into this particular fact-swamp, one seems to encounter every belief that thrives in it. I am sure I will always wonder why.


It doesn’t seem to be the case. See fig 3.

The authors acknowledge this point as a limitation of the study.

I disagree. I think that the data of this study weigh in favor of HCQ. But even if it was not the case, another important point revealed by the study is that Raoult’s treatment is safe. Why is it important ? Well, because it has been shown that stress is the second most important risk factor for covid19. This means that caring for patients by offering them a safe treatment will have the virtue of calming them, reducing their stress, and thus increasing their chances of survival.
Bottom line: most probably, Raoult has saved many lives.

Well, the first point I would like to make is that observational studies have a long track record of successes. The second point is that it has been shown that RCT and observational studies don’t differ significantly in their conclusions.
Also relevant to your point, the following passage taken from the author’s discussion:
RCTs were excluded because of lack of systematic biologi- cal diagnosis or absence of death (Supplementary Table 9). Only 1 randomized controlled trial with PCR-proven diagno- sis and at least 1 death was identified in the outpatient setting, probably because the sample size needed to identify a signifi- cant difference with sufficient power in mortality risk is dif- ficult to achieve with such a design in this context. This RCT [27] reported a non-significant but beneficial effect of HCQ (Table 3). However, new RCTs with medical quality crite- ria are not expected to change results since it has been shown that results from RCT and observational studies did not dif- fer significantly [37, 38].

What document are you quoting from, Gilbert?

Meilleures salutations,


1 Like

Raoult’s new study I posted here

So what? Acknowledging it doesn’t make the problem disappear.

That’s irrelevant because RCT studies are superior and we go with their conclusions when they disagree with observational ones. This is a matter of principle, not political convenience.

That’s objectively false, as shown for HCQ and in many, many other cases.


That’s just silly, because if HCQ works, it will work without death as the endpoint.

Again, if HCQ works or does not work in an inpatient setting, one should see the same thing in an outpatient setting.

That’s what those of us who do science describe as industrial-strength handwaving.


What? RCT were invented exactly because we know without them systematic and misleading biases affect the outcomes.

What’s next, double blinding makes no difference either?


August 23, 2021, American Journal of Medicine Editorial: Updates on hydroxychloroquine in prevention and treatment of COVID-19

Previously, we recommended a moratorium to healthcare providers concerning prescriptions of hydroxychloroquine. Since that time, no significant benefits have been found in the recent randomized evidence for post-exposure prophylaxis and among hospitalized patients. Regarding risk, hydroxychloroquine derived a reassuring safety profile from decades of prescriptions for autoimmune diseases of greater prevalence in younger and middle-aged women, whose risks of fatal outcomes due to QTc prolongations are very low. In contrast, the risks associated with COVID-19 are much higher because mortality rates for COVID-19 and the side effects of hydroxychloroquine are both highest in older patients and those with comorbidities, both of whom are predominantly men. The current totality of evidence more strongly supports our previous recommendations concerning the lack of efficacy and possible harm of hydroxychloroquine in the treatment and prevention of COVID-19.

After reviewing the various categories of side effects, this survey
Risk versus Benefit of Using Hydroxychloroquine to Treat Patients with COVID-19
concludes that HcQ is no benign placebo…

…we have summarized the available data pertaining to the adverse events associated with HCQ use, alone or in combination with azithromycin, in patients with COVID-19 in order to fully assess the risk versus benefit of treating COVID-19 patients with these agents. The results of this review lead us to conclude that the risk of adverse events associated with using hydroxychloroquine (with or without azithromycin) outweigh the potential clinical benefits and thus recommfend against its use in the treatment of COVID-19.

Clinical protocols have improved from the beginning of the pandemic, and while vaccination together with treatment - much improved from the onset of the pandemic - will not grant immortality, it provides protection against the likelihood of severe disease and the scythe of the reaper is not likely to be covid. This is what is saving untold thousands of lives.


Nevertheless, it seems there is no significant effect estimate differences between observational studies and RCTs


We’re not talking about the size of an effect, Gil, we’re talking about whether the effect exists at all. You are grossly misrepresenting these papers. From the former:

We excluded studies that compared randomized trials to non‐randomized trials. For example, we excluded studies that compared studies with random allocation to those with non‐random allocation or trials with adequate versus inadequate/unclear concealment of allocation.

Wouldn’t that exclude your application of these to the difference we’re arguing about here?