Image: Markus Spiske


A fraudulent or at least flawed study has supported the treatment of a huge number of people with invermectin

Ivermectin is an antiparasitic medication used to treat many types of worms and similar illnesses. It's very safe, widely used around the world, and in most forms a useful medicine to have on hand if you think you've been exposed to contaminated human feces, or if you just need to disinfect your sheep.

However, there has been a lot of buzz about ivermectin for another reason. According to several groups and following the best practices around the world, as well as in some scientific studies, ivermectin is a silver bullet against COVID-19. And while there may be some doubt as to whether ivermectin works, with the World Health Organization recommending that it be used to treat COVID-19 only in the context of a clinical trial, there is also a lot of optimism about its use as a treatment. Half a dozen countries have officially promoted ivermectin as a COVID-19 drug, and it has likely been given to tens of millions of people around the world by now, with the prices skyrocketing as a result.

This is all a little worrying when you consider that we don't really know if ivermectin works to treat COVID-19, but it's understandable given that the most expensive treatments aren't always available and people need something cheap that they can at least try. in the case of people dying of infection. It's not ideal, but there are some signs that ivermectin works and it's very safe, so why not give it a try?

Except for one thing. It appears that one of the major human trials of ivermectin may be a sophisticated work of scientific fraud. It's hard to know even if the study actually took place. If true, this could mean that ivermectin has absolutely no benefit for COVID-19, and tens of millions of people around the world have been misled.

Buckle up, this is a bumpy road.


Before we get to the study itself, there's something important to understand here – one study alone rarely changes everything in medicine. The fact is that individual studies tend to be relatively small and therefore leave us with a degree of uncertainty that is problematic if we are talking about what medication you take to save your life from a deadly disease.

To solve the problem of many small trials, we conduct things called systematic reviews and meta-analyses. These are scientific research aggregations that group all known studies on a topic into one place and then combine them into a statistical model so we can see what the overall effect of a drug might be. Instead of a dozen small studies, we get one large aggregate estimate, which in theory is the final word on whether or not a treatment is effective.

The only problem with these analyzes is that if a single study has a large number of participants or a huge effect, it can turn the overall trend into something positive, even though the studies have generally not found any results. Now, usually this isn't a huge problem, but it does mean that sometimes you have a whole body of literature saying something works using the gold standard aggregation of many studies that is actually based on the results of a single survey.

And every now and then, that study turns out to be terribly flawed.


With that in mind, let's see our troubled study. In fact, I've written about this before – it's a pre-print, not yet accepted in a scientific journal, published in research square by a group of physicians from Egypt. The study is titled “Efficacy and Safety of Ivermectin for Treatment and Prophylaxis of the COVID-19 Pandemic” and is quite impressive in its raw value, with the authors recruiting 400 people with COVID-19 and 200 close contacts and randomly allocating them to receive ivermectin or a placebo. Surprisingly, the study found that people treated with ivermectin were 90% less likely to die than people given the placebo, which, if true, would make ivermectin the most incredibly effective treatment ever discovered in modern medicine.

Furthermore, as befits the largest single randomized trial to date of ivermectin for COVID-19, its impact is disproportionate in the literature. As I showed you before, excluding just this single piece of research from multiple meta-analytic models almost completely reverses their results. It is no exaggeration to say that this single piece of research is prompting almost all of the belief in the benefits that people currently ascribe to ivermectin.

However, even at first glance there are some problems. The authors used the wrong statistical tests for some of their results – for the technically minded, they report chi-square values ​​for continuous numerical data – and their methodology is full of loopholes. They report no allocation concealment, there are questions about whether there was an intention-to-treat protocol or whether people were switched between groups, and the randomization information is woefully inadequate. As a study, it seems very likely to be biased, making the results quite difficult to trust.

But this is perhaps unsurprising, given the possibility that the study never took place.

data discrepancies

The problems with paper they start with the protocol, but they certainly don't end there. In a truly spectacular story, a British masters student named Jack Lawrence was reading this article and noticed something strange – the entire introduction appears to have been plagiarized. It's actually pretty easy to confirm this - I've copied and pasted a few sentences from different paragraphs into Google and it's immediately apparent that most of the intro has been pulled from elsewhere online without attribution or acknowledgment.

Now that's a bad sign, but it barely scratches the surface of the issues here. We could perhaps forgive some plagiarism in a pre-print if he actually found a miracle cure, but there are worse problems in the text.

For example, there are numbers that are incredibly improbable, almost impossible. In table 4, the study shows the mean, standard deviation and intervals for recovery time in the study patients. The problem is that with a reported range of 9 to 25 days, a mean of 18, and a standard deviation of 8, there are very few number settings that would match this result. You can even calculate this using the sprite tool developed by clever fraud detectives James Heathers, Nick Brown, Jordan Anaya and Tim Van Der Zee – to average 18 days consistent with the other values. Most patients in this group would have had to stay in the hospital for 9 or 25 days exactly. Well, that might not be entirely impossible, but it's so weird that it raises very serious questions about the results of the assay itself.

Somehow it gets even worse. It turns out that the authors uploaded the actual data they used for the study to an online repository. Although the data is locked, our hero Jack Lawrence was able to guess the file's password – 1234 – and gain access to the patient-level anonymous information that the authors used to put together this article.

the data file is still online (and you can download it for $9 + tax) and examine it for yourself. I have a copy and it's amazing how obvious the flaws are, even at first glance. For example, the study reports that it got ethical approval and started on June 8, 2020, but in the data file uploaded by the authors to the preprints website, 1/3 of the people who died from COVID-19 were already dead when the researchers began recruiting their patients. Unless they were getting dead people to consent to participate in the trial, that's not really possible.

Furthermore, approximately 25% of the entire group of patients who were recruited for this supposedly prospective randomized trial appear to have been hospitalized even before the study began. begin, which is either a mind-boggling breach of ethics, or a very bad sign of potential fraud. Even worse, if you look at the values ​​for different patients, it appears that most of group 4 are simply clones each other, with the same or very similar initials, comorbidities, lymphocyte count, etc.

All these are classic and characteristic signs of scientific fraud – repeated segments (with minor changes) in the data are found in various scientific papers that were later proven to be fraudulent. Even if this isn't outright fraud, the ethical concerns of randomizing people into a clinical trial before ethical approval appears are enormous.

I won't go into all the discrepancies these data show. So much Jack lawrence as Nick Brown have examined the issues carefully, and I strongly recommend that you read both articles before proceeding. They explain, in minute detail, all the reasons why it seems quite unlikely that this study could have taken place as described.

I'll give you a second to read this. Believe me, it's worth it.

Okay, we're back. So the authors put online a raw dataset that is very similar to the results of the study they posted, but also obviously false. What does that mean?

Well, it's possible that this is a big mistake. Perhaps the authors created a spreadsheet that matched the study too closely, which was also made up of simulated information for some unknown reason, and then accidentally uploaded that instead of the real information. Maybe when they put the file online they somehow managed to copy and paste a dozen patients 5 times, overwriting the actual values ​​in the rush to get their study started.

Perhaps. It seems highly unlikely, especially given how closely most of the information in this spreadsheet is to the study results, but it's possible that this was a mistake made by otherwise credible researchers.

On the other hand, there is the very real possibility that parts or even all of the study will be fabricated. We don't know - maybe we'll never know for sure - but it doesn't seem far-fetched given the impossible numbers even within the manuscript itself.


Taken literally, this is very bad. This study has been viewed and downloaded over 100.000 times and, according to Google Scholar, has accumulated 30 citations since November 2020, when it was first posted. This means that this potentially fraudulent research was used to target treatment to patients, likely many people at this point.

But it gets worse. Much worse, in fact.

You see, a lot of the hype around ivermectin is not due to this single study but, as I said earlier, to the meta-analyses. two studies recent, in particular, drew a lot of attention, both claiming to show, with the highest quality of evidence, that ivermectin saves lives.

The problem is, if you look at those big, aggregated models, and you remove just this one study, ivermectin loses almost all of its claimed benefit. See the recent meta-analysis by Bryant et al.. that's been all over the news – they found a 62% reduction in lifetime risk for people who were treated with ivermectin, compared to controls, when they combined randomized controlled trials.

However, if you remove Elgazzar's article from your model and run it again, the benefit goes from 62% to 52% and largely loses statistical significance. There is no observed benefit in people with severe COVID-19, and the confidence intervals for people with mild and moderate illness become extremely wide.

Also, if you include another study which was published after the release of Bryant's meta-analysis, which found no benefit for ivermectin for death, the benefits seen in the model disappear completely. to another recent meta-analysis, simply deleting Elgazzar is enough to remove the buff entirely.

This is important. It means that if this study is fraudulent, it will have huge implications not just for the people who trusted it, but for all the research that included it in their analysis. Until there is a reasonable explanation for the numerous discrepancies in the data, let alone the implausible numbers reported in the study, any analysis that includes these results should be considered suspect. Given that this is currently the largest randomized clinical trial of ivermectin for COVID-19, and most analyzes so far have included it, this is a truly concerning situation for the literature as a whole.


The somewhat appalling reality is that there have been enormous numbers of people treated with ivermectin largely on the basis of a trial that, if not outright fraudulent, is so flawed that it should never have been used for any medical decision anyway. . Even if we ignore the false data loaded with the study, it's hard to ignore the myriad other problems the research demonstrates.

Where does this leave us with the question of whether ivermectin works for COVID-19? Well, first of all, once you exclude Elgazzar from the research pool, the current best evidence shows a fairly consistent lack of benefit. There are still one or two very positive small studies, but overall ivermectin does not seem to reduce the risk of dying from COVID-19.

now there are much larger rehearsals ongoing to answer this question with certainty, and I look forward to your results - as I said before, the main problem right now is that the evidence just isn't good enough to be sure anyway. However, if the largest single trial yet of a drug has been fraudulent, it's not a promising sign.

We are also left with a monumental reckoning. It shouldn't take a master's student/investigative journalist to look at a study to realize that it is potentially a fraud. This study was reviewed by dozens of scientists, including myself, and although I said it was extremely low quality, I didn't notice the issues with the data.

It has been more than six months since this study was posted online. By all indications, the results of this research have been used to treat thousands, or perhaps even millions, of people around the world. And yet, so far no one has noticed that most of the introduction [of the article] is plagiarized?

The scientific community and we noble defenders of truth have screwed up.

There's no nice, easy ending for me to wrap up here. The investigation into this study is ongoing and one day we may have the story in its entirety, but in the meantime the damage has been done.

The scientific enterprise likes to be excited by the idea that science is “self-correcting” – that flaws and errors are discovered through the very process of science. But that wasn't a mistake. Accredited experts have read this trial and rated it LOW risk of bias. Not once, but repeatedly. Doctors looked at it and saw a miracle cure, despite the problems the research has, even on cursory analysis.

I'm not sure this study is fraudulent, and we may never know. But we know that no one should ever have used it as evidence for anything. We will have to acknowledge the fact that a truly disgraceful survey, which may turn out to be a hoax, has been put online and used to direct the treatment of millions of people around the world.

And no one noticed until it was too late.

*Gideon Meyerowitz-Katz, an epidemiologist, is a PhD candidate at the University of Wollongong, Australia.

Translation: Sean Purdy e Leandro Tessler for the magazine science question.

See this link for all articles


  • About artificial ignoranceEugenio Bucci 15/06/2024 By EUGÊNIO BUCCI: Today, ignorance is not an uninhabited house, devoid of ideas, but a building full of disjointed nonsense, a goo of heavy density that occupies every space
  • Franz Kafka, libertarian spiritFranz Kafka, libertarian spirit 13/06/2024 By MICHAEL LÖWY: Notes on the occasion of the centenary of the death of the Czech writer
  • The society of dead historyclassroom similar to the one in usp history 16/06/2024 By ANTONIO SIMPLICIO DE ALMEIDA NETO: The subject of history was inserted into a generic area called Applied Human and Social Sciences and, finally, disappeared into the curricular drain
  • Letter to the presidentSquid 59mk,g 18/06/2024 By FRANCISCO ALVES, JOÃO DOS REIS SILVA JÚNIOR & VALDEMAR SGUISSARDI: “We completely agree with Your Excellency. when he states and reaffirms that 'Education is an investment, not an expense'”
  • A look at the 2024 federal strikelula haddad 20/06/2024 By IAEL DE SOUZA: A few months into government, Lula's electoral fraud was proven, accompanied by his “faithful henchman”, the Minister of Finance, Fernando Haddad
  • Strengthen PROIFESclassroom 54mf 15/06/2024 By GIL VICENTE REIS DE FIGUEIREDO: The attempt to cancel PROIFES and, at the same time, turn a blind eye to the errors of ANDES management is a disservice to the construction of a new representation scenario
  • PEC-65: independence or patrimonialism in the Central Bank?Campos Neto Trojan Horse 17/06/2024 By PEDRO PAULO ZAHLUTH BASTOS: What Roberto Campos Neto proposes is the constitutional amendment of free lunch for the future elite of the Central Bank
  • Chico Buarque, 80 years oldchico 19/06/2024 By ROGÉRIO RUFINO DE OLIVEIRA: The class struggle, universal, is particularized in the refinement of constructive intention, in the tone of proletarian proparoxytones
  • Volodymyr Zelensky's trapstar wars 15/06/2024 By HUGO DIONÍSIO: Whether Zelensky gets his glass full – the US entry into the war – or his glass half full – Europe’s entry into the war – either solution is devastating for our lives
  • The melancholic end of Estadãoabandoned cars 17/06/2024 By JULIAN RODRIGUES: Bad news: the almost sesquicentennial daily newspaper in São Paulo (and the best Brazilian newspaper) is rapidly declining