Skip to content

Hydroxychloroquine update

Following up on our earlier post, James “not the cancer cure guy” Watson writes:

I [Watson] wanted to relay a few extra bits of information that have come to light over the weekend.

The study only has 4 authors which is weird for a global study in 96,000 patients (and no acknowledgements at the end of the paper). Studies like this in medicine usually would have 50-100 authors (often in some kind of collaborative group). The data come from the “Surgical Outcomes Collaborative”, which is in fact a company. The CEO (Sapan Desai) is the second author. One of the comments on the blog post is “I was surprised to see that the data have not been analyzed using a hierarchical model”. But not only do they not use hierarchical modelling and they do not appear to be adjusting by hospital/country, they also give almost no information about the different hospitals: which countries (just continent level), how the treated vs not treated are distributed across hospitals etc. A previous paper by the same group in NEJM says that they use data from UK hospitals (no private hospitals are treating COVID so must be from the NHS). Who is allowing some random company to use NHS data and publish with no acknowledgments. Another interesting sentence is about patient consent and ethical approval:

The data collection and analyses are deemed exempt from ethics review.

We emailed them to ask for the data, in particular to look at the dose effect which I think is key in understanding the results. They got back to us very quickly and said

Thanks for your email inquiry. Our data sharing agreements with the various governments, countries and hospitals do not allow us to share data unfortunately. I do wish you all the very best as you continue to perform trials since that is the stance we advocate. All we have said is to cease and desist the off label and unmonitored and uncontrolled use of such therapy in hospitalized patients.“

So unavailable data from unknown origins . . .

Another rather remarkable aspect is how beautifully uniform the aggregated data are across continents:

For example, smoking is almost between 9.4-10% in 6 continents. As they don’t tell us which countries are involved, hard to see how this matches known smoking prevalences. Antiviral use is 40.5, 40.4, 40.7, 40.2, 40.8, 38.4%. Remarkable! I didn’t realise that treatment was so well coordinated across the world. Diabetes and other co-morbidities don’t vary much either.

I [Watson] am not accusing the authors/data company of anything dodgy, but as they give almost no details about the study and “cannot share the data”, one has to look at things from a skeptical perspective.

Again, I have not looked into this at all. I’m sharing this because open data is a big deal. Right now, hydroxychloroquine is a big deal too. And we know from experience that Lancet can make mistakes. Peer review is nothing at all compared to open review.

The authors of the paper in question, or anyone else who knows more, should feel free to share information in the comments.


  1. What would cause such small variation across locations? In elections this would be evidence that the data was fabricated, although in this case there could be a more innocent explanation: perhaps they did some sort of matching to balance covariates across different regions? Is this explained somewhere in the paper?

    • Zhou Fang says:

      Maybe someone took the entire dataset and accidentally randomised it?

    • Anoneuoid says:

      What would cause such small variation across locations?

      Perhaps it was the “harmonizing” and “improve[ing] data integrity”?

      The Surgical Outcomes Collaborative (Surgisphere Corporation, Chicago, IL, USA) consists of de-identified data obtained by automated data extraction from inpatient and outpatient electronic health records, supply chain databases, and financial records. The registry uses a cloud-based health-care data analytics platform that includes specific modules for data acquisition, data warehousing, data analytics, and data reporting. A manual data entry process is used for quality assurance and validation to ensure that key missing values are kept to a minimum.


      The standardised Health Level Seven-compliant data dictionary used by the Collaborative serves as the focal point for all data acquisition and warehousing. Once this data dictionary is harmonised with electronic health record data, data acquisition is completed using automated interfaces to expedite data transfer and improve data integrity. Collection of a 100% sample from each health-care entity is validated against financial records and external databases to minimise selection bias.

      I find it difficult to believe that they have records about (eg smoking) for 100% of these patients. I think there is a lot of imputation going on to generate these numbers.

    • Okay, I think I figured it out. See the spreadsheet I created here:

      The final column (“Ratio”) is the ratio of how much variation you observe, relative to how much variation you would expect if every country had the same population fraction for that covariate.

      Note that for the covariates that aren’t treatment covariates, the ratio is close to 1, but for the treatment conditions it’s far from 1. So they probably matched on all of the non-treatment covariates but then reported that as the raw data. Seems like a weird choice and I don’t understand how you can do that without running out of data in some of the countries, but it convinces me that at least the data aren’t made up.

      • Carlos Ungil says:

        Good job! Presenting by mistake, or at least in a misleading way, processed data as raw could explain the mysterious homogeneity is some variables Not all of them as you noticed, only the “unused” ones like antiviral therapy which is suspiciously similar unlike the other therapy choices. I had the feeling that Australia (lower N) was showing a higher “sampling variability” but I didn’t connect all the dots.

      • Lu Chen says:

        Thanks for the analysis. It makes sense.

        If “they probably matched on all of the non-treatment covariates but then reported that as the raw data”, why did they perform another propensity score matching analysis in Suppl Table S7?

        match baseline by continent (Table S3)-> then match baseline by treatment (Table S7). This process will probably lose lots of data. Not to mention the disease severity parameters (e.g.,SpO2 & qSOFA) are insufficient for propensity score matching analysis. Why not directly match baseline by treatment from the beginning?

      • I find your explanation pretty convincing, but if you’re right, doesn’t it mean there was significantly more than 96,032 patients in the data before matching? I don’t see how to reconcile this with the way in which they describe the data in the paper, where they claim to have “included all patients hospitalised between Dec 20, 2019, and April 14, 2020, at hospitals participating in the registry and with PCR-confirmed COVID-19 infection” and state clearly there were 98,262 of them, from which they excluded 2,230 patients for various reasons having nothing to do with matching to balance covariates across different continents.

  2. Mike Roberts says:

    The NHS does indeed allow private companies access to data, subject to various public interest guidelines, GDPR, anonymisation etc.

    If they are using data from English hospitals then a record of the data release should be published here:

    I had a quick search for “surgi” in the March and April publications and couldn’t see anything, but it’s possible they used a different name. It’s also possible that they are using data from Wales or Scotland, and I’m not sure where the record of that would be published.

  3. Anoop says:

    I hope he writes a letter to Lancet with these concerns rather than just emailing you.

    • zbicyclist says:

      Me, too. But:

      1. By getting a posting here, he potentially finds out flaws in his reasoning or additional hints he’d want to add. So he gets to battle-test his arguments a bit.

      2. With all things COVID, time is important. I don’t know that getting a rapid response from the Lancet would be at all guaranteed.

  4. Andy Wang says:

    If I am following the method correctly in this Lancet article, the data included hospitalised patients between Dec 20th to Apr 14th. 5 Australian hospitals were included according to its appendix ( and mortality was recorded as 73 (12%) which I take it to mean N=73 which was 12% of the 609 from Australia.

    Of note, as of 3pm on Apr 14th, we only have 61 death in the whole of Australia (

    • Carlos Ungil says:

      I don’t think there is a contradiction, people will typically die several days after hospitalization. But it’s remarkable the coverage they have of Australian cases: as of April 12 there were 752 hospitalizations.$File/covid_19_australia_epidemiology_report_11__reporting_week_ending_23_59_aest_12_april_2020.pdf

      • Andy Wang says:


        But to be able to capture 73 out of 102 total death (as of today), they must have captured most of the major centres as well as nursing home clusters in Australia, I doubt any one centre in Australia has more than 10 deaths – but this is just my impression. As I understand it, majority of our deaths were from nursing home, and not all were admitted to a hospital.

        Also, to say the ethics wasn’t required was very unacceptable. They would have required at least their local ethics approval to be able to apply to our data custodians (separate hospitals) to release the data. One would expect 73 out of 102 these patient data to be potentially identifiable data. It is very unlikely that our hospitals would release patient data externally without ethics, and no acknowledgement for the 5 centres that contributed to the data. Process to extract data from our eMR systems isn’t straight forward and isn’t the same even between different government health districts. It would be very difficult without someone coordinating it in Australia and if they did, agreeing to it without at least getting acknowledged.

        They should have listed the hospitals they obtained the data from at the very least. Otherwise it looked very much like they might have fabricated Australian data. That makes me angry.

    • James Watson says:

      Hi Andy,
      Thanks for this comment, this is a really good point.
      This page has deaths in Australia broken down by state (near bottom of page)

      By April 22nd there were 74 deaths across 7 states. Thats two more than reported in The Lancet paper. However the paper says that they only have 5 hospitals contributing data in Australia (Table S1). I’m guessing that each hospital would be in it’s own state…

      • Ben says:

        “Australia” = continent, not country

      • Andy Wang says:

        Deaths were scattered around different hospitals and nursing homes in Australia. In NSW alone, you can track the location of the deaths daily on the NSW Health website, and if you look closely, they were scattered throughout 13 hospitals and 2 nursing homes. No hospitals reported more than 3 deaths during that period.

        I’ve made comments on PubPeer where all the links to the deaths in NSW can be found on the official NSW Health website if you are interested.

        Even if you count Australia as a continent (Australia + New Zealand, and if you want, plus Papua New Guinea and West Papua), the numbers still don’t add up with 5 hospitals.

        The method they described about automated data extraction from hospitals, or the method they described in their NEJM paper where this is being done automatically periodically, isn’t something our NSW Health eMR would allow or capable of without someone coordinating it. To add to complexity, NSW Health uses one eMR system for the Hospital, and ICU generally uses a different system. The complexity is enormous. The covariates they listed (smoking, comorbidities… etc) could not have been easily extracted automatically, they are usually being entered in as free text and needs text searching to identify them. You can’t not have local resource allocated to it to be able to extract those data automatically, at least at the Hospital level. Even though I cast this doubt, I can’t be sure they haven’t had prior arrangements to have this automated data extraction running. If they do, it’d involve quite a lot of resources to set it up and I doubt the 5 hospitals would agree to not be acknowledged for it.

        Anyway, unfortunately all these means we still don’t know whether hydroxycholorquine is or isn’t helpful or harmful for COVID-19. One step back rather than forward.

  5. Anoneuoid says:

    Too many smokers, too few people on hydroxychloroquine, too few people with low spO2. I’d bet, assuming this is real data, that 50-90% of these patients were hospitalized for something besides covid.

    • David Cohen says:

      I agree comletely. These patients are too young (median age 54) and too healthy (90% with O2 sat >=94% on admission) to be a real-world sample. For reference, the recent NYU paper on all-comers hospitalized had a median age of 64 and only 50% with an O2 sat of 50% or greater on admission.

      • Aakriti Gupta says:

        Also, proportion of hypertension is 25% and DM 12% from the US – much different than what we are seeing in NYC data – where HTN is more than 50% and DM is more than 30%.

      • Manu Paa says:

        2 ideas on that :
        – “Patients for whom one of the treatments of interest was initiated […] while they were on mechanical ventilation […] were excluded” (Lancet Summary) –> This exclusion factor could be a reason (not sure but seems plausible)
        – People with high risk of cardiac accident cannot be given Hydroxychloroquine and Azitromycine for safety reasons

    • It’s interesting that the experience of this disease outside the western developed nations seems to be very different… In Brazil and Mexico for example many younger people are getting sick:

      • Luca Beltrame says:

        Perhaps a naive question: but are these reports taking into account poor healthcare and hygienic conditions? I think these, like with that H1N1 outbreak in Mexico in 2009, have a very large impact.

      • Anoneuoid says:

        Washington Post has a really poor track record. That story seems to rely on real people instead of anonymous officials so is better than most of what they publish, but if still like to see a better description of the methods used to generate those numbers.

        I had some speculation but deleted it. I’d rather just see more info on where those numbers are coming from before devoting time to explaining it.

        • Meire says:

          I am Brazilian and I can’t trust the numbers here. First because many, if not most, of covid deaths related by media and even official departments, don’t match with the deaths certifications. Second, it was imposed in many states to label as covid anything looked as flu, without tests (they said they run out tests) and we know we have an H1N1 breakthrough. The numbers are even a hundred time more frequent on states that are against the president. Unfortunately, it is not about science but politics

  6. Anonymous says:

    As someone who grew up in Africa, I am naturally skeptical of smoking/ hyperlipidemia/ diabetes rates that are in the same ballpark as western countries. Even if he can’t disclose the data, can he at least point the to the sources

    • Anoneuoid says:

      Actually I was talking to a friend from Zimbabwe last night and he was saying he never even heard of diabetes until he came to the US.

      • Carlos Ungil says:

        Those hospitals (if they exist at all, who knows) are more likely to be in South Africa or Egypt than in Zimbabwe.

        The prevalence of diabetes (in the population 20-79) is higher in South Africa and Egypt than in the US or Europe.

        • Naglaa Arafa says:

          That’s true. The covid positive population are also rather younger than other countries (based on the scant data the ministry of health provides). However where are these sophisticated EMR systems they are talking about?

  7. Thorsten says:

    Like the comment from Watson !!!

    But who is he? Statistician, doctor,…?


  8. We have a better initiative: open data, muti-center Real-world evidence, international consortium with top hospitals

    Join us

  9. Dale Lehman says:

    Then, this appears: If you look more closely, you will find that this particular author is at least as suspect as the paper we are discussing. It is a sad state of affairs that we have come to – you can’t trust what is published and you can’t trust the critics. If there ever was a time to require release of data, isn’t it now? I realize there are many sides to this debate and no simple solution will cover all cases. But I’d propose the following: if you want to publish anything that can influence public policy, then you should be required to release the data.

    • Andrew says:


      Yeah, no kidding. The discourse is out of control. We can’t trust Lancet, PNAS, etc., and we can’t trust randos on the internet who pull out the big rhetoric whenever they are politically triggered.

      +1 on releasing the data. This won’t solve all the problems, as politically-motivated critics will still be able to throw mud no matter what, but it should help.

    • Anoneuoid says:

      It is a sad state of affairs that we have come to – you can’t trust what is published and you can’t trust the critics.

      There is nothing new about this. Science is based on the concept of universal distrust. Nullis in verba…

      If you were just taking people’s word for things before you were doing it wrong.

      • Martha (Smith) says:

        Wow — my education has been lacking! I’d never heard of Nullis in verba (motto of the Royal Society) before. Thanks for pointing this out.

        Also, +1 on Dale’s comment, “if you want to publish anything that can influence public policy, then you should be required to release the data.”

        • at an absolute minimum, you should be ABSOLUTELY required to release the code, since it’s always possible to release code without leaking personal data, and potentially if the data is sensitive, at least release some kind of wide variety of summary statistics on the data, things like means, stddev, quantiles, correlation coefficients, level curve plots for density estimates of bivariate distributions, etc etc

    • zbicyclist says:

      “this particular author is at least as suspect as the paper we are discussing”

      Do you mean Watson or Leo Goldstein (blog author) or whoever arkancide_is_real may be? (More than one answer can be chosen)

      Of course, prior to yesterday I would guess none of us had heard of any of these people.

      • Dale Lehman says:

        I was referring to Goldstein, although these days it could be anyone.

      • James Watson says:

        akancide_is_real is a friend and I can vouch for him. He has thoroughly investigated many crises, including how a vaccine for coronavirus will come from Bill Gates and could lead to implantation of microchips in all o fus. This is clearly the deep state trying to work against the free American people. We will not stand for this.

        • Joshua says:

          James –

          > , including how a vaccine for coronavirus will come from Bill Gates and could lead to implantation of microchips in all o fus.

          I thought you were joking – until I consulted with Ms. Google.


          The representative survey of 1,640 US adults by YouGov for Yahoo News found that half of respondent Americans who say Fox News is their primary television news source believe the conspiracy theory. It’s the largest group responding this way, followed by self-described Republicans and “Voted for Donald Trump in 2016” — 44% of both those groups said they believed the conspiracy theory was true. Twenty-six percent of respondent Republicans said it was false, and 31% said they weren’t sure.


          I wonder if this belief will have legs if a vaccine does get developed? I suspect and hope not!

          One question is whether/how Trump supporters will reconcile this belief with their fealty to Trump, assuming a vaccine does get developed, he’s still in office, and he continues to support the benefits of a vaccine.

          • Joshua says:

            BTW –

            I think the whole Fox News viewers is a confound. I think it’s social media where people get this, and there’s an overlap.

            • Joshua says:

              What is “it?”

            • dhogaza says:


              I missed the part where your article states that a microchip will be injected along with a vaccine.

              This is an effort to use vaccination as leverage to issue IDs to masses of poor people who don’t have them. An analogy might be how CA, for instance, allows one to register to vote when one applies for or renews a driver’s license.

              • Anoneuoid says:

                There is no conspiracy theory. It is just common sense. How would you implement a portable, persistent, and unique identifier to each person that contains private info like vaccination history?





                Here is an example “how it works”:

                The first idea is to store the private key for your “digital ID” on your phone. This is not going to work. Anyone who has dealt with cryptocurrency knows internet connected devices are not a safe place to store more than “bar money”, you definitely do not want all your identity info on there. Then they say the technology can work “without connected devices” without going into details.

              • Anoneuoid. There is no need for an injectable PIT tag in order to have good cryptographic personal ID info. Have a govt issued smart card that provides you with a certificate you store on your phone valid for 3 months. Every 3 months or so you download a new signed certificate from the smart card. The smart card expires after 10 years like a passport… that we don’t do this already is kind of amazing. The technology has been possible since the late 90’s

              • dhogaza says:



                That’s a good video which demonstrates, in principle, how blockchain based digital ID combined with biometrics (they show fingerprints) and public key encryption can store such data while maintaining privacy and personal control over the data.

                This doesn’t really fit in the the whole “microchip injected secretly with your vaccine so you can be monitored without your consent” argument being made by anti-vaxxers.

              • Anoneuoid says:

                Have a govt issued smart card that provides you with a certificate you store on your phone valid for 3 months.

                This doesn’t work for the same reasons I would not store significant amount of crypto on my phone. Once your info is accessed, the genie is out of the bottle.

                Also they want to use this in African villages, etc where a doctor visits the town every few years and there are no records of when this last time it happened was, who got vaccinated last time, etc.

                Regardless, calling the “vaccine with chip” plan a conspiracy theory is just ignorant of what is going on. Is that the plan for covid-19 specifically, I don’t know but definitely seems like a time stuff a lot of people don’t want gets rolled out (eg, PATRIOT ACT).

              • Anoneuoid says:

                This doesn’t really fit in the the whole “microchip injected secretly with your vaccine so you can be monitored without your consent” argument being made by anti-vaxxers.

                Why do you think it would be secret? You will be told it is happening.

              • Joshua says:

                Anoneuoid –

                > There is no conspiracy theory.

                There is no conspiracy theory, or there is no conspiracy?

                It seems quite apparent that there is a widespread conspiracy theory.

              • Joshua says:

                Anoneuoid –

                > Why do you think it would be secret? You will be told it is happening.

                It isn’t that he thinks it will be secret. It’s that a large segment of the public thinks there is a secret conspiracy to track billions of people.

                You’re making this more complicated than it needs to be, imo.

              • Anoneuoid says:

                You’re making this more complicated than it needs to be, imo.

                There are plans to implant a chip into your body along with vaccines that will hold info on your vaccination history. This dovetails with other plans for a “digital ID” funded by many of the same organizations behind the vaccines that would also work best as an implant.

                It is coming, but it is not going to be a secret when they inject you with it. I guess that is the only part the conspiracy theorists got wrong then?

              • Joshua says:

                Anoneuoid –

                > I guess that is the only part the conspiracy theorists got wrong then?

                It seems to me that you’re missing *a lot” of what is wrong about these conspiracy theories. And, the elements that are wrong are part of the fundamental basis of the conspiracy theories – not just piddling details.


                Some are accusing Chancellor Angela Merkel and pharmaceutical companies of inventing COVID-19 to impose dictator-ship like conditions on the country. In some instances, they referred to the social-distancing measures as a “social Holocaust.”


                There are widespread conspiracy theories about this, which comprise multiple elements that are wrong – in fact one might say not even wrong.

                It seems to me that you’re downplaying the impact of the wrong aspects of the conspiracy theories. I see them as potentially quite harmful.

                I’m having a hard time figuring out why you and I have such a different view on this.

                It’s actually a bit unsettling that you don’t seem to see difference between the factual aspects of what’s going on and the nutter aspects of the conspiracy theories.

              • Anoneuoid says:

                It seems to me that you’re missing *a lot” of what is wrong about these conspiracy theories.

                I’m talking about the plans to inject people with a chip that holds their personal info and inject people with something that stores info about vaccination history along with a vaccine. Please just stick to that instead of mixing it with random other claims you found on the internet that mention covid-19.

                It’s actually a bit unsettling that you don’t seem to see difference between the factual aspects of what’s going on and the nutter aspects of the conspiracy theories.

                It seems to me you are having trouble distinguishing between conspiracy theories and publicly known facts.

              • Andrew says:

                Enough, all! This blog is not the right place to debate conspiracy theories, or arguments about whether a theory is a conspiracy theory, etc.

              • Joshua says:

                Anoneuoid –

                Last comment on this – I think beyond this we’re at the point of diminishing resturns:

                > Please just stick to that instead of mixing it with random other claims you found on the internet that mention covid-19.

                The comment that started this thread was mine, about the unfounded nutter conspiracy theories about the CV vaccine.

                You responded to say that they weren’t conspiracy theories, and then started introducing some related elements/facts that are being distorted in the conspiracy theories. And now you want me to stick to those elements rather than the topic I raised?

                Again, I was referencing nutter conspiracy theories that have become widespread. Please stick to that topic. Unfortunately, mixing the two topics tends to exacerbate the problem presented by the nutter conspiracy theories.

                > It seems to me you are having trouble distinguishing between conspiracy theories and publicly known facts.

                Not at all. I have been pointing to the nutter conspiracy theories all along, and trying to get you to be clear about the distinction between those nutter conspiracy theories and the publicly known facts.

              • Joshua says:

                Apologies Andrew –

                I only saw your “enough” after I posted that comment.

    • Or be audited or at least subject to possible audits…

    • Why am I suspect? Besides, I don’t ask to believe my word. Every my statement either supported by a source, or by the text of the discussed paper. My post have been revised with more info.

      P.S. I don’t have an opinion on arcancide(s)

      • Dale Lehman says:

        I won’t let you get off the hook so easily. Your headline: “Anti-HCQ Paper in The Lancet Uses Fake Data” is pure hype, and an example of the kind of thing people on this blog take great efforts to try to combat. Sure there are issues with the data, and people are trying to discern whether errors were made, and if so, whether it was accidental or intentional. But you have declared a result – and without anything more than the suspicions that have been expressed about the data source. Do you know the data was fake? On what basis do you know this? Then, when I look at your track record (yes, reputations matter, another common theme on this blog), I see a pattern of hyped politically-motivated statements. You have the burden of proof if you want to be taken seriously, just as the authors of this paper we are questioning do.

        • CCC says:

          Dale, you have no clue what youa re talking about. I ese that you and everyone else at Columbia are part of this conspiracy to off set our country. You are profoundly anti American. Its a free country.

          • Dale Lehman says:

            I shouldn’t respond to trolls. But I have never worked at Columbia – never even stepped foot on the campus (though I grew up in NY). I’ve never been to Columbia, the country, either.

            • D Agrawal says:

              No one has ever been to “Columbia” the country. lol ;-)

              • confused says:

                Ironically, at one time “Columbia” was a somewhat poetic synonym for the US: that’s why DC is the “District of Columbia”.

                I wonder if confusion with “Colombia” is part of why this vanished… I think it’s largely aesthetic, though, “Columbia” being part of a pseudo-classical tradition which is much rarer now with a decline in interest in classics, Latin in education, etc.; combined with the appearance of “Uncle Sam” as a personification of the US in place of the personified “Columbia” sometimes seen in the 19th century.

        • I didn’t post it on this blog.

          The headline is shocking, because the events are shocking. A database with 96,000 COVID-19 patients from 671 hospitals all over the world cannot exist because of privacy laws. There, it comes from a company that existed only on paper just 3 months ago. Which doesn’t even advertise that as a database, but as an analytics platform. In March, Surgisphere also announced a bunch of diagnostic tools on its website, supposedly big data / AI driven, but actually consisting of nothing more than JS code.

          Of course it is fraud. I am shocked that such frauds exist.

  10. Robert Clark says:

    About the percentages being equal across continents, I believe they are only showing the numbers after “propensity score matching”. This is the method that attempts to equalize numbers between the treatment groups and the control group for trials that are not randomized controlled.

    I would like to have seen the raw numbers themselves, not just the ones after they have been “adjusted.”

    By the way, this review of the paper shows the authors did not successfully equalize the numbers for intubated patients, which could have severely skewed the data against HCQ:

    Hydroxychloroquine: When medical science starts to look like political science.
    May 23, 2020

    Robert Clark

    • Martha (Smith) says:

      Thanks for the link.
      What a mess we are in.

    • James Watson says:

      No, the main Table in the supplementary is showing all the data, aggregated by continent.

    • Robert Johnston says:

      Robert Clark: (1) James Watson is correct; the Propensity Score Matching is shown in the supplementary dataset, and was just a ‘check’ they did to confirm what they found in their primary analysis, the Cox regression on the data shown in the main part of the paper, which has the distributions by continent shown in the first part of the supplement. (2) Quay seems to have missed the important detail that they excluded from their study any patients for whom treatment was begun after mechanical ventilation, and they say that progression to mechanical ventilation was an output of the study.

  11. Dr. Wilhelm Herdering says:

    There is a report of turkey,
    that early treatment by Hydroxychloroquine dramatically reduces covid-19 pneumonia rate.
    If you look to the number of covid-19 deaths then you see, death rate is lower than in Germany.
    If one takes tamiflu against Influenca, one should it take in the first 48 h after getting symptoms. Perhaps it is the same with sars-cov-2 and hydroxychloroquine.
    Google: turkey says hydroxychloroquine dramatically reduces pneumonia rate

  12. Greg Nelson says:

    Perhaps send the original authors a script to run on their data. Could even ask them for the headers of their data file and read it in etc for them, so authors don’t have excuse to not run it.

  13. Mendel says:

    From the paper: “The Surgical Outcomes Collaborative (Surgisphere Corporation, Chicago, IL, USA) consists of de-identified data obtained by automated data extraction from inpatient and outpatient electronic health records, supply chain databases, and financial records.”
    In my imagination, this reads as “hacked into a number of hospital databases” or “has access to the data processing done for hospitals and similar institutions” because the data set is so comprehensive. They say that they have access to financial records and are using them: “Collection of a 100% sample from each health-care entity is validated against financial records and external databases to minimise selection bias.”
    Is it plausible that a study would get access to these kinds of records? Financial records (maybe billing data)?

    Perhaps the matter is such that the hospitals are outsourcing some of their data processing, and the contract contains a clause that they’re allowing the data processor the use of anonymised data for scientific purposes?

    It feels weird that financial records enter into a medical research paper as data.

  14. Genet Robin says:

    Hi, I’m writing from France. Just to be sure about the Lancet article, is that correct that the data details stop on April 14 th 2020 ? If it’s the case, I just remark that on Australia there was 61 deaths On April 14th and you report 73 deaths in your analysis. Curious isn’t it ?

  15. Varrick says:


    I’m working as a kind of a researcher with a doctor of medicine. I’m not a specialist and please take this comment as a elemental question. I’m always interested in HCQ discussion. The doctor let me know the Lancet page. And the first thing I thought strange was the word “macrolide” and clarithromycin. Clarithromycin?? As for the treatment of Covid19 I’ve never heard of it. I searched in the web and found the treatment for Chronic Lyme Disease in 2003. But nothing else.

    As everybody knows the French doctor Raoult used Azithromycin and HCQ for Covid19. Other doctors in USA did the same. In the Lancet page there is no discription of how many pacients received Azithromycin (alone with HCQ).

    A doctor in Long Island knew that Azithromycin can cause fatal arrhythmia. So he replaced Azithromycin with another anti-biotic…. Doxycycline. 38 out of 47 recovered. The result was good.

    The structural formula of Azithromycin and Clarithromycin is simmilar but not the same. The similarity of structure does not mean that of the result. The structural formula of Azithromycin and Doxycycline is completely different but the result was similar.

    Does anybody know why Clarithromycin appeared all of a sudden? Why not Azithromycin alone (with HCQ)

    Sorry for the rough English. I’m writing this from Japan.

  16. Hello,

    I wrote a short rebuttal of the big Lancet chloroquine paper using Real World Evidence from 90k patients (3 minutes read):

    In summary, 3 flaws:
    no reproducibility
    no data traceability
    no review transparency

    This Lancet paper is harming the emerging RWE-Big Data field in which I am involved (with the consortium COVIND

    Mostapha Benhenda
    COVIND consortium organizer, and data scientist at Melwy

    • Ava says:

      From your article:

      > I don’t trust Fox News about global warming, so why should I trust The Lancet about Trump-approved Chloroquine? Fake news all around.

      This is getting confusing. Are you putting Fox News and The Lancet on an equal scientific footing? I also don’t think we should evaluate research on the basis of how many fields the research is harming…

  17. Lu Chen says:

    Another confounding is the exclusion criteria. Fig. 1 shows that 1102 patients were excluded because they received CQ derivatives after intubation.

    This doesn’t look like a serious problem, but it is *not* a negligible number if we compare with the total number of intubated patient (1243+616+814+403=3067). So *26% (1102/(3067+1102)) intubated patients were excluded!

    This may sound nothing, but it may significantly imbalance the baseline between treatment and control group, because no one is excluded in the control group. For example, the observation can be completely changed if:

    1. selection bias towards those who are strong & young who can tolerate nasal feeding tube (intubation prohibits swallowing HCQ pills). –> treatment group is biased towards the elder after excluding youngsters.

    2. selection bias towards those who looks hopeless, so doctor was hoping that HCQ can save his/her life. –> treatment group is biased towards the healthier after excluding the serious ones.

    3. HCQ is a game-changer (or toxic drug) especially when a patient is intubated, compared to before intubation. The current exclusion criteria will worsen (or inflates) the observed survival, because it will selectively exclude the intubation-then-survive (or intubation-then-death) in the treatment group but not in the control group.

    • Dom says:

      Also, there is no minimal duration of treatment in the inclusion criteria.
      So the patient who was given an HCQ pill right before being intubated is included (which seems quite likely) and treatment then appears to be the cause of the ventilation outcome.

  18. vl says:

    The study says : “The mean length of stay in hospital was 9·1 days (SD 6·4), with an overall in-hospital mortality of 10 698 (11·1%) of 96 032. The use of other antivirals was recorded in 38 927 (40·5%) patients as treatment for COVID-19. The most common antivirals were lopinavir with ritonavir (12 304 [31·6%]), ribavirin (7904 [20·3%]), and oseltamivir (5101 [13·1%]). Combination therapy with more than one of these antiviral regimens was used for 6782 (17·4%) patients.”

    So, they had enough data to measure the effects of the other anti-viral treatments. What did they not do so ? believe that, should one of these treatments have appeared to be beneficial, they would have been happy to publish it.

    Also, it is widely accepted now that anti-viral drugs may only work at the onset of the disease, and that, at later stages, other approaches are required. Why do they not mention that ?

    Lastly, we know that most patients in this study are from the US. Is it possible that this study would include, in the control group, patient that have been tested and diagnosed, but could not afford to stay in the hospital and receive a treatment (because of no or poor medical cover), and as a consequence, may have died outside of the hospital, without their death being recorded ?

    We really need to see raw data.

  19. AB says:

    The “current smoking” row on it’s own seems very strange, 1) EHR data rarely has smoking data (nor BMI), and 2) it varies by less than a percentage point across studies.

    In this comparison of patients hospitalised with COVID-19, and, COVID-19 characteristics vary substantially across sites, making this table seem extremely suspect.

  20. I have initiated a public project to try to clarify this controversy.
    This is an Open-Data, Open-science project, so everyone is free to comment and contribute:

  21. AB says:

    The appendix of the NEJM paper by the same group is equally as troubling.

    Is the NEJM population the same as for the Lancet, but from an earlier cut of the data? Strange then the average age is 49.6 in North America, 49.0 in Europe, and 49.2 in the NEJM paper. Doubly confusing, so early covid patients were 49 (across all continents!) and now they are all 51 to 56….. 5.6%, 5.4% and 5.9% were smokers in the first paper, and now they are 10% smokers across all….

    From table 1 of the NEJM paper: “Race and ethnic group were reported by the patient”. So who is asking the questions? How does the patient’s response get into their EHR data, this surgisphere database? How does that not need review from an ethics committee?

  22. TS says:

    Regarding arrhythmias: • Of the 502 patients who experienced denovo arrhythmias, how many died who were also in the HCQ with a macrolide group? • How many of the 1479 nonsurvivors in the HCQ with a macrolide group (N= 6221), died from ventricular arrhythmia?

    Regarding baseline disease severity: The best pulse oximetry threshold for detecting hypoxia is 92%. (see ) How many of the non-survivors (N=10,698) had a baseline oxygen saturation threshold of 92%? Adjusting the Pulse oximeter saturations (SpO2) to <93% would differentiate the sick from the very sick.

    Regarding survival rates for HCQ with a macrolide: • 76% of patients who took HCQ with a macrolide survived (4,742 out of 6,221) • 5.6% of all survivors took HCQ with a macrolide (4,742 out of 85,334) Might these findings indicate further investigation into HCQ with a macrolide in an outpatient, ambulatory setting?

  23. John Dagan says:

    Deadly selection bias

    (1) Тhe physicians all over the world administered hydroxychloroquine more often to the patients who had co-morbidity and who were more seriously ill than the rest of the patients, i.e. that were more likely to die. Why? Because they thought that those patients needed some therapy that may increase their chances to get through the diseases. This was also logical, this is how we normally act in face of the seriously ill patient. My belief? Not at all. There is more to this. The atmosphere in France was against hydroxychloroquine and the use only in the patients with high risk (of dying) was encouraged (the agent was requisitioned already in January for these purposes!). The drug was not prescribed without bias already from January 2020! The situation in the world was probably also influenced by this “Zeitgeist”.

    French Arreté de 26 mars 2020:
    (2) The French government issued an “arêté” on 26th March, 2020 forbidding the administration of hydroxychloroquine outside hospitals and obliging the hospital physicians to administer it exclusively to the patients that are seriously attained by the Covid-19 and have some co-morbidity (and who had a priori increased chances to die).

  24. D Agrawal says:

    This information is not “statistical” but I think it is important nonetheless. I hope the moderators post it. It’s the contents of two emails I have sent to Roni Rabin of the New York Times.

    Dear Ms. Rabin,

    I read your article, “Scientists Question Validity of Major Hydroxychloroquine Study” with great interest. I think that it is very important to bring additional information to your attention, namely that James Watson and most, if not all the signatories on his open letter, have conflicts of interests, some of which are very serious.

    James Watson is a “senior scientist” (statistician) at MORU Tropical Health Network in Thailand. MORU is running a 40,000 person worldwide clinical trial for hydroxychloroquine that is funded by the Bill & Melinda Gates Foundation, Wellcome, and Mastercard.

    The signatories on this letter are not just random scientists who are concerned with authenticity of the study published in the Lancet. They are conflicted participants.

    For example:

    Dr Nomathemba Chandiwana (Research Clinician, University of the Witwatersrand, South Africa) – His university is part of the MORU trial

    From the article (link above), one can see that others directly involved in the MORU COPCOV trial signed the letter. They include: Michael Avidan, Laurence Lovat, and Ramani Moonesinghe – all chief investigators in the COPCOV trial. The COPCOV trial statistician, Dr Hakim-Moulay Dehbi, also signed the letter.

    This is along with 7 other signatories who work directly for MORU.

    Other signatories are associated with Wellcome, trial treatment centers, or Watson himself.

    Take Caroline Buckee and Aimee Taylor from Harvard. Both published an article with Watson recently. Buckee’s postdoc fellowship was funded by Sir Wellcome (see her Harvard profile).

    I could cite many more examples but I think you get the point. Check out more names for yourself. You will find what I have found. People need to know the whole story so they can decide for themselves. Perhaps none of these conflicts of interest change any facts. Maybe they do. In any case, I hope that the gray lady, who I trust, will point out that the author of this letter and its signatories are definitely biased.


    D. Agrawal

    Dear Ms. Rabin,

    It is unfortunate that the MORU COPCOV trial started one day after the Lancet results were published. From what I have read, the Lancet study does not make any recommendations regarding hydroxychloroquine for COVID-19 outpatient profylaxis, which is the focus of the COPCOV trial. It looked at hospitalized patients and has nothing to do with outpatient prophylaxis.


    D. Agrawal

  25. Geoff says:

    There are some other odd statistics in their data.
    The error for age in Australia is +/- 17.7 but with 100 times more data, the US is +/- 17.8.
    The BMI is identical for US and EU (I am taking an average of Spain and Italy as they are the largest groups in the study) but 2016 data suggests that EU is about 26 and US about 29.
    Smoking is also suspicious, Australia and EU both at 10% but 2016 data suggests 26.5% and AU 15%.
    I used for these stats.

    • Carlos Ungil says:

      > The error for age in Australia is +/- 17.7 but with 100 times more data, the US is +/- 17.8.

      That’s not an error, it’s a standard deviation which doesn’t really depend on N.

      > The BMI is identical for US and EU but 2016 data suggests that EU is about 26 and US about 29.

      BMI varies with age and the dependency may be quite different in the US and the EU. Hospitalized COVID-19 patients are not representative of the overall population and their age alone may already explain to some extent the difference you noticed.

      > Smoking is also suspicious, Australia and EU both at 10% but 2016 data suggests 26.5% and AU 15%.

      As mentioned by Jacob Steinhardt above ( ) it seems that for some rows the matched stats (not the raw data) is shown. Anyway, as I said before, the proportion of current smokers in hospitalized COVID-19 patients may also be quite different from the proportion of current smokers in the overall adult population.

  26. Geoff says:

    Hi Carlos,
    That isn’t a standard deviation, it is a confidence interval which is dependent on sqrt(N), I don’t know how that it possible that it ‘doesn’t really depend on N’.

    The only thing that I can think of is demographic/socio economic reasons.

    Agreed on the 2nd point, I have no idea about EU v US BMI for hospitalised cases but the population is very different. But unless the private company (that is potentially cherry picking data) clears up this and the more important questions, it is a valid question.

    I’ll have to look into the 3rd point more, you didn’t make the argument very clear.

    • Carlos Ungil says:

      You’re wrong on #1, those intervals are not intended to represent the error on the estimation of the of mean age of some hypothetical population. They represent the variability of the age in the data and that won’t change when more people with a similar age distribution is added. Maybe you would have preferred them to report [ age: 54.4 +/- 0.1 ] but that wouldn’t be very useful.

      Regarding #3, they released a revised table because, they explain, the original one was derived from a propensity score matched and weighted table. In the unadjusted raw summary data we see for current smoker 10.3% in Europe and 6.3% in Australia which is more in line with your expectations (it’s quite higher in Europe, even though in both cases the numbers are less than half than what you gave for the overall population).

      • Geoff says:

        Hi Carlos,
        Is that method of reporting variability of the age common in medical surveys? I am from a physical science/mathematics background mainly. I would have expected a measure of variance that reflected the size of the population of AU related to its size compared to other areas based on their size.

        Not that there is a ‘population’ in the AU study any more, the latest ‘revision’ has reduced the number of AU cases to 60 odd (down from 600+) with 3 moralities (down from 12).

        The only reasons for this that I can think of are is either creating dodgy data or obtaining the data through dodgy sources.

        I have no problem with the raw data for #3, it would be expected that older populations smoke less than my 2016 data, I did not expect 65 year olds to smoke the same as the overall population.

  27. Jacques Pollini says:

    If I may add a comment on the Mehra et al study:

    Mehra et al do not indicate the time lag between baseline data collection and treatment start. If we assume they did the baseline at the same time as they did the diagnosis, then that time lag can be as much as 48 hours. During these 48 hours, the state of some patients can aggravate. We know that aggravation can happen quickly with COVID. And since Hydroxychloroquine is often used for compassionate use, it is possible that patients that aggravate are given the HCQ more frequently than those who do not aggravate. This is a selection bias which introduces a factor of inverse causation. They are given HCQ because they aggravate, not the other way around. This would explain why the treatment group has more people in mechanical ventilation than the control group in spite of having similar baseline data. The NYC study ( faced the same problem and did a sensitivity analysis to see if results were different with a baseline data collected 24 hours later (they were not). Another option would be to introduce a new variable “aggravation 48 hours after baseline or during whatever is the average time lag between baseline and treatment”, and control for that variable, or include it in the criteria for propensity score matching. I am not a statistician but I hope you some statisticians will be able to look more closely at my argument. If this reverse causality exists and if Mehra et al did not do anything to control it, then the whole study collapses even for that reason alone. It would not even prove that HCQ is dangerous on severe COVID patients, even if their data set was correct. Mehra et al must answer these comments.

  28. Eric Penrose says:

    According to, major contributor and Surgisphere founder Sapan Desai faces – three – medical malpractice suits in Cook County from the end of last year.
    And, while I’m not promoting per se here, I did pursure analysis there – that is up to the payment bit that I should have predicted (from what I’ve read I wouldn’t recommend paying because of cancellation problems etc). Along the way it said ‘we have uncovered multiple criminal records matching the name Sapan Desai’. I believe I already had identified the correct Sapan Desai. I’m not living in the US but if anyone wants to look into this further via County records – that seems a good idea and highly relevant.

    Also when looking at before this – and while using a Russian located VPN – I was amazed to see a site administrator warning along, with related symbol, while doing a search there for ‘hydroxy’. Then with another vpn this was ok, then ok later with a Russian vpn a second time, but I’ve never seen anything like that accessing a webpage.

    I have the two screenshots that I can email to anyone interested.

  29. Eric Penrose says:

    that was Cook County, Illinois.

  30. İlke sipahi says:

    The current Table S3 at Lancet’s website is dramatically different. Why the difference ?

    The picture you posted is an earlier version or what ?

    Even more discrepancies in the new Table S3. In north America cuurent + past smokers is 28% and COPD is 3.3%. In south America it is 8.8% and 4.1%. Why? Smoking rate is one third in south but COPS is 25 % relatively higher???

    • Anonymous says:

      “An incorrect appendix table S3 was included, originally derived from a propensity score matched and weighted table developed during a preliminary analysis. The unadjusted raw summary data are now included.”

    • İlke sipahi says:

      Hmm on further review Lancet says ‘An incorrect appendix table S3 was included, originally derived from a propensity score matched and weighted table developed during a preliminary analysis. The unadjusted raw summary data are now included. There have been no changes to the findings of the paper. ‘ Australian patient number goes down from 609 to 63 and you say the findings do not change!
      So the old Table s3 has adjusted values for baseline characteristics which is weird. Baseline characteristics should always be real numbers and not derived data. Even the death rate is different in the 2 tables. You are either dead or alive. How can you change that with modeling? Also what kind of modeling can decrease the former smoker frequency from 18.7% to 5.2% in south America? You first say 187 patients in south America were past smokers but the real number is 670 !!! This paper should be retracted today.

  31. Stavros Kyriazis says:

    Congrats for Lancet and NEJM tremendous U-Turn!
    Also and for WHO flip and backflip…

  32. Sabbir Rahman says:

    Given the obvious problems with this study, it is hard to see how it passed peer review unless the journal itself was complicit, together with the lead author, in promoting the conclusion that HCQ is harmful. The fact that a competent body such as the WHO would immediately publicise the results so widely despite its obvious flaws also strongly suggests complicity.

    Of all the drugs undergoing trials it would be strange that there would be such a concerted effort to prevent HCQ from being investigated properly unless it was somehow considered a threat. That could only be the case if (i) it genuinely constituted a significant danger to the public – which seems highly unlikely as it is a well-known drug that has been administered safely for decades, or (ii) if HCQ is actually likely to be efficacious – and indeed a number of smaller scales studies have shown this to be the case when it is administered in combination with zinc supplements in an appropriate way.

    Now, the second of these two possibililties would clearly be to the considerable benefit of the general public in terms of saving lives, and the only reason that it could possibly be considered a threat by the parties concerned is if they were acting on behalf of entities who would be affected negatively should drugs such as HCQ be found to be effective and come into widespread use.

    It is fairly clear that the only parties that would be negatively effected by widespread use of HCQ as an effective treatment for COVID-19 would be the profit-driven pharmaceutical companies, organisations and high-net-worth individuals who stand to gain significantly from sale of newer drugs or vaccines which are still under patent. Effective treatments for COVID-19 through repurposing of low cost generics would pose a considerable financial threat to this promising source of income.

    The lead author of the article, Professor Mandeep Mehra, whose research is funded by pharmaceutical companies, certainly does not shy away from making his bias against the use of HCQ clear on his LinkedIn page, or that Bill Gates is one of his key influencers. In addition, besides contributions from individual goverments (no longer the US), the WHO receives its greatest funding contributions from the two largest pro-vaccine organisations, namely the Bill & Melinda Gates Foundation (which is now its largest single source of funding) and the GAVI Alliance – whose founding partners include the Bill & Melinda Gates Foundation, UNICEF, the WHO itself and the World Bank, and whose broader alliance includes the pharmaceutical industry.

    It does not require a great deal of critical thought to recognise the corruption that must be going on here. The pharmaceutical industry and the vaccine lobby appear to have leading academics, top journals and even global health organisations in their back pockets, and corporate and individual greed is being given priority over the preservation of human life.

    • Joshua says:

      Sabbir –

      > . That could only be the case if (i) it genuinely constituted a significant danger to the public – which seems highly unlikely as it is a well-known drug that has been administered safely for decades, or (ii) if HCQ is actually likely to be efficacious – and indeed a number of smaller scales studies have shown this to be the case when it is administered in combination with zinc supplements in an appropriate way.


      It is fairly clear that the only parties that would be negatively effected by widespread use of HCQ as an effective treatment for COVID-19 would be the profit-driven pharmaceutical companies, organisations and high-net-worth individuals who stand to gain significantly from sale of newer drugs or vaccines which are still under patent.


      Arguing from incredulity can often lead to fallacious arguments. Even if we assume that the conconclusions in the article is wrong, your comment fails to adrress obvious uncertainties – in contrast to your absolute certainty.

      HCQ has never been taken before by so many people who might be at risk for many reasons, not the least of which is having been infected by a virulent disease which is as yet poorly understood. Additionally, independently of any potential negative impact for people who don’t typically take the drug, over-prescription could lead to critical shortages for people who depend on the dog. Even if we assume some “motivated” reasoning behind the article’s conclusions, it’s not so easy to determine what those “motivations” might be.

      Assuming poor faith in others, and presuming intent and motivation, is a tricky business – and requires getting over a high bar of proof. Speculation is fine, but without more evidence it should only be done with careful qualifications. Otherwise you risk ending up undermining your own credibility.

      • Joshua says:

        Obviously, dog = drug.

        Anyway, now with the retraction it will be interesting to see what more is determined about motivations of the authors.

        • Sabbir Rahman says:

          Well, now that we have some more information, a significant clue as to what may be going on is the failure by Professor Mehra and the Lancet to disclose a clear conflict of interest, namely that Gilead Sciences currently has two ongoing trials of its COVID-19 drug Remdesivir being carried out at Brigham & Women’s Hospital, of which Mehra is a Director:

          Hydroxychloroquine is of course in direct competition with Remdesivir as a treatment for COVID-19 so this failure to disclose should immediately start ringing alarm bells. It clearly begs the question as to whether Gilead may have had any involvement in the preparation of the Lancet article or had perhaps instigated it.

          The other missing piece of the puzzle relates to the source of the raw data. Given that Surgisphere seems to have had insufficient resources to fabricate such a large dataset without additional help, it would be worthwhile looking into whether perhaps a pharmaceutical company – again Gilead being the natural first candidate given the circumstances – could have been the actual source of the data.

          • Dale Lehman says:

            So I thought this episode was finished – another case of research malpractice – discovered, exposed, and leaving us with questions of whether or not the journals would go any further than the (quick, by their standards) retractions. But you now raise a far more disturbing twist. I’ll admit to being naive. If Gilead is really behind all this, and Mehra did not declare the potential conflict of interest, then this seems to be far more serious and intentional than malpractice by a relatively minor player (Surgisphere). I’m leery of such conspiracy theories, but the potential for this one to be true is too serious to ignore. I’m not ready to buy into what you are theorizing, but it would be irresponsible to just ignore it. The consequences are serious and it would be sad if the investigation does not continue. Please continue to focus attention on unraveling this story.

          • Joshua says:

            Sabbir –

            Again, I don’t think there’s anything wrong with speculating…but conspiracy theories, and this really is in essence a conspiracy theory, require a plausibility test, IMO.

            So let’s break down a bit what you’re saying. Tell me if I got something wrong.

            You’re saying that some dozen? (more) people across a few different organizations, who obviously know that there are tens if not hundreds of ongoing clinical trials to establish the efficacy of HCQ that will be producing findings over the next couple of months, decided to fabricate data to show that HCQ has no effect so that they could profit from promoting an alternative treatment. Apparently they would believe that their one study would be sufficient to prevent use of HCQ over a different drug on the basis of one study. (There is some evidence that would suggest that in a sense – given the response by the WHO, France, etc. in response to their study – but there’s no indication that their study increased use of other drugs or that any such effect would be sustained over the outcomes of many other clinical trials.

            Keep in mind that these are people who are well-established in their careers, which no doubt are quite lucrative, and who stand to gain from the existence of an effective COVID-19 treatment (HCQ if it is effective) in that having such a treatment would help improve the economy more rapidly. In addition, each and every one of these people is quite willing to let tens? of thousands of people die or be very sick by not getting access to a potentially effective treatment, as a result of their agreement to fabricate data to undermine the perception of efficacy.

            And even further, they fabricate data to make HCQ look ineffective, but they don’t go so far as to fabricate data to indicate that the drug is dangerous – even though there do seem to be a lot of people who think that the drug does potentially have serious cardiac counter-indications, at least for people who are susceptible for arrhythmias.

            And they would do all of this knowing that there would be a risk that someone might discover what was going on, or that someone involved might reveal what was going on after a change of heart, and that if what they were doing was discovered it would likely mean economic ruin for themselves and the organizations they’ve no doubt invested in heavily in terms of their careers, if not also serious legal ramifications.

            Is that a fair description of what you think is going on?

            I have to say, it doesn’t seem impossible to me, at all (implausible conspiracies do happen), but if it is as you describe, it doesn’t seem very likely. Given the myriad other possible explanations, I think some caution would be prudent.

            • Sabbir Rahman says:

              Hi Josh – I honestly do not know what is really going on, and I am merely trying to come up with the most plausible explanation for the observations. If you feel that you have a better explanation then please be my guest.

              There are a number of questions that remain unanswered but what troubles me the most about all this is where the flawed data set could have come from. It clearly does not appear to have come from the hospitals as claimed by Desai, but neither does Surgisphere appear to have had the resources to have produced the data themselves.

              This would imply that the data must have originated from some other entity that has both the wherewithal to produce it, and a vested interested in discrediting HCQ as a treatment for COVID-19. Whatever the case, it is fairly clear from the speed of the retraction that the three authors concerned no longer want to have anything to do with it.

              I do not think that this constitutes some kind of conspiracy theory. Obviously it is not a pleasant conclusion to reach that there must be some kind of foul play going on, but unless there is some other better and more innocent explanation for the observed facts, it seems to be a relatively straightforward inference that is hard to avoid.

              Anyay, like others here I am just an interested third party trying to understand what is going on and probably like everyone else, I do not have the means to investigate this matter any further beyond searching for relevant information publicly available on the internet.

              I genuinely hope that someone does however get to the bottom of this and puts an end to it. If these kinds of unethical activities are allowed to continue unimpeded, who knows how many innocent people’s lives could be at stake.

            • Joshua says:

              Sabbir –

              Yah. If your speculation stands up, it’s incredibly serious.

              • Sabbir Rahman says:

                This mind-blowing interview with France’s former Health Minister has just been brought to my intention:


                May 24, 2020: Philippe Douste-Blazy, Cardiology MD, Former France Health Minister and 2017 candidate for Director at WHO, former Under-Secretary-General of the United Nations, reveals that in a recent 2020 Chattam House closed door meeting, both the editors of the Lancet and the New England Journal of Medicine stated their concerns about the criminal pressures of BigPharma on their publications. Things are so bad that it is not science any longer.

                Suddenly everything becomes much clearer. You might want to save the video before it is taken down.

              • Sabbir Rahman says:

                For your convenience, I have typed out the transcript of the English translation (apologies in advance for any mistakes or typos):

                Q: But it’s hard to understand why scientists would voluntarily give bias to studies…

                PD: Exactly! That’s the great question. That’s the great question we are all asking ourselves, finally, and you know those Chatham House lectures in London…

                Q: Remind us what this is all about?

                PD: This is extremely interesting. These are meetings that are completely behind closed doors, only with experts. No one can record, no one is taking any pictures, there’s no [inaudible]. It’s only between experts.

                Q: Top secret.

                PD: Top secret. But still, there was a meeting the other day, of the directors of scientfic journals, like The Lancet, The New England Journal of Medicine…

                Q: The Lancet, which is that journal which published this study we are talking about….

                PD: These are extraordinary journals: when it’s written in The Lancet, it’s “written in The Lancet”. So that’s why…. here, we’re talking about something very important, this discussion that happened. And it ended up leaked: The Lancet’s boss, Horton, said: “Now we are not going to be able to, basically, if this continues, publish any more clinical research data, because the pharmaceutical companies are so financially powerful today and are able to use such methodologies, as to have us accept papers which are apparently, methodologically perfect, but which, in reality, manage to conclude what they want to conclude…” This is very, very serious!

                Q: But what you are telling us is very serious!

                PD: Yes.

                Q: That would mean that it is the pharmaceutical companies that are putting pressure on, including financial pressure, I guess…

                PD: Yes, but I am not the one saying it….

                Q: …on the scientific results! But you understand, who can we trust any more today?

                PD: Indeed, that’s why I allow myself to tell you about it, because it is one of the greatest subjects… never anyone could have believed…. I have been doing research for 20 years in my life. I never thought the boss of The Lancet could say that. And the boss of the New England Journal of Medicine too. He even said it was “criminal”, the word was used by them. That is, if you will, when there is an outbreak like the COVID, in reality, there are people… us, we see ‘mortality’, when you are a doctor or ourself, you see ‘suffering’. And there are people who see ‘dollars’, that’s it.

              • Andrew says:


                I have no idea if that is a real quote, as it’s a report of a second-hand conversation, but if it’s real . . . . all I can say is that if the Lancet editor thinks these are “papers which are apparently, methodologically perfect,” then he really should forget about peer review entirely. He should just judge articles on their potential importance, forget about correctness, and let post-publication peer review sort it all out.

              • Sabbir Rahman says:

                Hi Andrew (there is no “Reply” button under your comment, so apologies if this appears in an odd place) –

                Yes, as you say this is a second hand report so I personally would not attach too much importance to that specific word – I am guessing that “acceptable” would more likely have been used in the original conversation than “perfect”.

                The key point, of course, if his report is indeed true, is that the journals are being bullied by the pharmaceutical companies to publish studies which they know have been intentionally falsified by them in order to promote their own agendas.

                This puts the reliability of the entire academic literature into question as well as any policy decisions made based upon the information published therein.

              • Martha (Smith) says:

                The last words were striking — if I recall them correctly: “Where we see suffering, they [big Pharma] see money.”

      • Sabbir Rahman says:

        Thank you Josh for your response.

        First of all, let me make clear that like you, I do not know how effective HCQ is when used in various combinations. However, it is perhaps possible that you are unaware of quite how large a number of people around the world have been taking, and indeed continue to take HCQ despite the WHO’s warnings against its use. If it were not showing considerable efficacy, I seriously doubt that so many countries would have insisted on continuing to use it.

        For example, Prof Mehra on his LinkedIn page “likes” the comment by his Harvard Medical School colleague Haider Warriach praising the article Mehra et al just published in JAMA by his entitled “Prescription Fill Patterns for Commonly Used Drugs During the COVID-19 Pandemic in the United States” where they show that there were a staggering 483,425 excess fills of HCQ prescriptions during the 10-week period from 16 February to 25 April 2020 as compared with the previous year. Note that Mehra also “likes” Daniel Goldstein’s comment that “WHO halting enrollment of patients into the HCQ/CQ arm of the large SOLIDARITY trial based on potential harm data published on 5/22/20 in Lancet by Mandeep Mehra, MD FRCP et al.”, so it is clear that he was rather pleased with this outcome of his paper, and it is fairly safe to assume that he is not pleased by the very large number of people currently taking HCQ, which is why he was so keen to highlight this fact in his JAMA paper:

        And that is just in the US. In India, “all healthcare workers in hospitals and some frontline staff” dealing with COVID-19 continue to be administered with HCQ, despite WHO’s advice, because they are seeing “no harm” and potential benefits from its use as a prophylactic:

        In Brazil, which currently has the highest growth rate in COVID-19 cases worldwide, they are also continuing to recommend HCQ as the default treatment in defiance of the WHO, though the matter has now become highly politicised there:

        Turkey “has made significant progress in treating coronavirus patients in the early stages of the disease with the controversial malaria drug hydroxychloroquine”, with the proportion of coronoavirus cases registered with pneumonia dropping from 60% on 24th March to 19.5% on 6th April. Needless to say, Turkey has also insisted on continuing the use of HCQ despite the WHO’s concerns:

        In South Korea, HCQ has been one of the preferred treatments according to physician guidlines since at least February. In a retrospective study by Samsung Medical Centran and PNU of the use of HCQ as a prophylactic, of 184 patients and 21 care workers who had been exposed to COVID-19 at a long-term care hospital where massive COVID-19 infections were reported, 100% of those treated tested negative for the virus after a 14-day quarantine period. Amusingly, they were not able to prove that HCQ is effective for the prevention of COVID-19,” as there was no adequate control group:

        In Algeria, Dr. Mohamed Bekkat, a member of the Scientific Committee monitoring the evolution of the COVID-19 pandemic, said that, “We have treated thousands of cases with this drug with great success to date. And we have not noted any adverse reactions”, and “We have not recorded any deaths related to the use of (hydroxy)chloroquine”. Algeria is continuing to use HCQ despite the WHO’s decision to suspend clinical trials:

        It is the same story in Morocco:–20200528-0008.html

        Honduras has also been using HCQ as their default treatment for COVID-19. According to an article appearing on 5th May, the Honduran health minister ‘insists that patients who were treated with hydroxychloroquine in four different hospitals have responded satisfactorily, and that doctors have been using this treatment plan for over a month now’, and has stated that, “Hydroxychloroquine is used to treat malaria and will not cause harm to anyone, because doctors take extra care when treating patients with heart problems”:

        El Salvador had also been using HCQ as the default treatment until the WHO recommendation against it was announced. Rather interestingly, El Salvador’s Bukele questions why world leaders are being told to take HCQ [one wonders by who, exactly?] while the public is being warned away from it:

        Russia continues to stand by use of HCQ, and according to the Health Ministry, “domestic experience indicates the validity of the use of hydroxychloroquine when it is prescribed in certain groups of patients with COVID-19 in low doses”:

        Although some countries such as Italy, Belgium, France, Portugal, Egypt Tunisia, Colombia, Chile, Cape Verde, Albania and Bosnia have suspended use of HCQ on the basis of the WHO’s recommendations based upon the dubious Lancet study, many other countries that have had positive experiences with HCQ have retained it on their the list of approved drugs for the treatment of COVID-19 or for clinical trials despite the WHO’s recommendations against it:

        Are you starting to see a pattern here or shall I go on? I had started making a list of other countries continuing to treat patients with HCQ, but it quickly became apparent that the list was going to be a rather long one.

        Just as an indication of the extent to which HCQ is being tested or applied, India, which is world’s the largest manufacturer of HCQ and many other generic drugs, has exported large amounts of HCQ to at least 55 other countries, including the US, Mauritius, Seychelles, Afghanistan, Bhutan, Bangladesh Nepal, Maldives, Sri Lanka, Myanmar, Zambia, Dominican Republic, Madagascar, Uganda, Burkina Faso, Niger, Mali Congo, Egypt, Armenia, Kazakhstan, Ecuador, Jamaica, Syria, Ukraine, Chad, Zimbabwe, France, Jordan, Kenya, Netherlands, Nigeria, Oman, Peru, Philippines, Russia, Slovenia, South Africa, Sri Lanka, Tanzania, the United Arab Emirates, Uzbekistan, Urugway, Columbia, Algeria Bahamas, Mauritius and the United Kingdom:

        Perhaps all of the above evidence is merely “anecdotal”, but I think at some point one needs to draw the line and apply just a little common sense, and acknowledge that, despite the lack of clinical trials to prove it formally, HCQ has been showing considerable efficacy on the ground for some time in many countries worldwide, and as a consequence of these positive experiences, continues to be administered.

        • Sabbir Rahman says:

          I would also add that if the above information can be scraped from the internet by an amateur sleuth in a matter of a few hours, then it is stands to reason that the WHO will have a far more comprehensive understanding of the extent to which HCQ is being used worldwide and its efficacy.

          It is therefore not feasible that the WHO could be unaware that the outcomes of the Lancet study quite blatantly contradict the results being reported to them by the many countries countries actively administering HCQ, and therefore also rather hard to come up with a rational ethical justification for their blind acceptance of the conclusions of that study leading to their recommendation that clinical trials of HCQ be stopped.

          The matter seems so clear that even the inevitable defense of simultaneous “incompetence” on the part of both the Lancet and the WHO cannot satisfactory explain their troubling behaviour, and given that so many lives are at stake, a detailed investigation in to what really happened is certainly warranted.

        • dhogaza says:

          Sabbir Rahan

          “If it were not showing considerable efficacy, I seriously doubt that so many countries would have insisted on continuing to use it.”

          Hmmm … doctors still prescribe antibiotics to people who have viral infections.

        • Joshua says:

          Sabbir –

          I just saw those comments now – otherwise I would have responded sooner.

          I was actually familiar with most of the information that you provided (I’ve been following the issue a bit). I agree that the extent of the ongoing use suggests that there is a perception of efficacy among a lot of practitioners.

          But I will also note that you listed only the evidence on one side. As one example, they’ve stopped using HCQ as a standard protocol in many places in NY, and I rather doubt they would have done so had they seen clear benefit.

          But it seems to me that the evidence on both sides as yet is rather slim, at best, and often the evidence has been improperly leveraged despite its weaknesses by advocates. Even the findings of the RECOVERY trial hace their limitations despite its ample air and the anticipation that it would provide clear answers.

          As far as I can tell questionable behavior with respect to the available evidence exists on both sides of the issue.

          And again, part of the issue with its usage for COVID 19 is that obviously, it has never been used in the context of that particular disease until quite recently – and as such its use over many years in other contexts is only of somewhat limited value. Likewise the safety issues of much of its use recently in the COVID 19 context is also of only limited value, as the information on how it had been used (say dosage) on which patients (how ill were they, what comorbidities did they have) is not well controlled. And largely it has been used with very ill patients where if it did present risks, it would be rather hard to identify given everything that is going on with those patients. I’m sure that you agree that with all the potential biasing factors in both directions, and factors such as the placebo effect on both doctors and patients, it can be very hard to reliably assess the efficacy of a medication absent carefully controlled analysis.

Leave a Reply