A sensitive instrument is one that can pick up and amplify small signals that others miss, and no doubt this "sensitivity" has been an integral part of what has made your writing as successful as it has been. Getting overwhelmed more easily than you might otherwise be is the other side of that tradeoff, and it's great that you're not letting the tough side of that get in the way of giving things the response you know they're due.
Thanks for taking the time to respond to him, even though it hasn't been fun. I doubt I'm the only one that notices and appreciates the effort.
> Dr. Ioannidis is known to be unusually rigorous and this is part of his pro-rigor crusade
I'm not familiar with all his work, but I think he deserves a reputation for self-promotion more than a reputation for rigor. I think his most famous essay is "Why Most Published Research Findings Are False", and I think this paper is notable for its sensational headline much more than for its careful analysis.
We have done an intensive study of viral load dynamics in our phase 2 platform trial. Ivermectin got dropped pretty quickly (pre-specified futility rule was met). In contrast, known effective antivirals such as casirivimab/imdevimab and remdesivir showed an effect on viral load dynamics pretty quickly.
Alexandros blocked me on Twitter long ago just for repeatedly asking whether there was one good RCT supporting ivermectin. He apparently runs a startup, yet has the time to write a series of posts that literally total over 86,000 words (a sizeable non-fiction book!) just to respond to ONE blog post from Scott. And he did this over the space of a few months.
In other words, he has been engaged in a 24/7 side hustle--writing a full non-fiction book in a few months -- and moreover, it was all on a very narrow question about whether one particular blogger was mean about a particular drug that supposedly is a miracle cure for Covid (even though no rigorous study has ever shown that). Odd choice of priorities.
And despite aiming towards rationality, he is keen to write thousands of words nitpicking at the tiniest flaws in any negative ivermectin study, even while defending much worse studies or websites that defend ivermectin. Blatant double standard.
This analysis is great but I'd like to pay more attention to what our prior should be that some existing medication has a substantial benefit in treating a new disease when there isn't a strong theoretical basis supporting that use. My sense is that our prior should be very low. I mean, just ask yourself how much you'd be willing to bet about a large rigorous study showing a 10% drop in deaths from a random drug.
All the argument as to what direction the studies on ivermectin point is great and all, but even in the best case the evidence isn't very strong. I mean, it's not unusual in the slightest in medicine to see some initially suggestive studies only for the effect to turn out to be a mirage.
So while I think pinning down just how much evidence we have for ivermectin is important insofar as it should inform us whether to bother investing in yet more studies the idea that this should be enough to convince you to take the bet against the strongly theoretically supported side-effects (plus a reasonable prior that many drugs have them) seems like a big stretch.
I must admit that I scrolled straight to the conclusions, and even that felt like a strain at this stage of the ivermectin conversation. The graph of likely false positive results at the end of the post was quite interesting, and I admire Scott's dedication to this topic. Otherwise, I'm hoping we can all move on to other topics!
You seem to be missing the positive results in the MedinCell trial and the new paper on hemagglutination showing that IVM both prevents and reverses RBC clotting associated with poor outcomes.
I agree with these conclusions a lot more than those from your original post.
One quibble: what does "clinically significant effects" mean? Is the burden "better than a placebo"? "Better than Paxlovid"? "Better than no placebo"? The differences in efficacy between those thresholds are large enough to drive a truck through.
I think there might be a typo in the phrase "not for excess mortality in the control group" in the Cadegiani et al section
If they were hypothetically going to condemn him for giving unhelpful drugs to people, wouldn't that involve them complaining about excess mortality in the treatment group, not the control group?
Outside of the technical arguments, you deserve credit for taking the high road in responding to Alexandros.
If nothing else, his bad faith characterizations of you made it clear that he's not very strictly bound by the available evidence. No one is perfectly in control all the time, especially when there's strong disagreement - but getting as far out over his skis as he did in impugning your motivations is a strong tell for "motivated reasoning." Unless he's got mind-reading skills which he hasn't yet explicated.
But the bottom line is that accountability is the key - and you've demonstrated accountability here.
I doubt Ivermectin works, happy to be convinced with stronger evidence.
What I found most alarming was how politicized Ivermectin became, most glaringly with the CDC, pretending as if Ivermectin was only for livestock with their infamous "you are not a horse" tweet [1] which may (?) have encouraged people like Dr Jason McElyea to make up fictitious stories about people "clogging up hospital beds because they are OD'ing on Ivermectin" [2]
Meanwhile, the CDC suggests ivermectin off-label to treat lice, even though there are plenty of FDA approved OTC and prescription treatments [3]. I quote:
"Given as a tablet in mass drug administrations, oral ivermectin has been used extensively and safely for over two decades in many countries to treat filarial worm infections. Although not FDA-approved for the treatment of lice, ivermectin tablets given in a single oral dose of 200 micrograms/kg or 400 micrograms/kg repeated in 9-10 days has been shown effective against head lice"
I think many engineers would quibble with the characterization of Elon Musk as an engineer, let alone a brilliant one. He does seem to have had a knack for a) picking investments in engineering companies to buy* or start, and b) hiring good engineers to work for him. But those are not engineering.
* One of his main mistakes with Twitter seems to be his belief that it's an engineering company, rather than a media company.
"My usual heuristic is that when studies contradict, I trust bigger studies, more professionally done studies, and (as a tiebreaker) negative studies - so I leaned towards the studies finding no effect."
The debate really comes down to the validity of that heuristic. Trusting bigger, more professionally done studies would be a great heuristic if we assume corporate money does not corrupt scientific research. But this is of course a false assumption. I trust small studies done by doctors trying to treat patients in a third world country over large, professionally done studies brought to you by Pfizer.
I believe BigPharma is completely corrupt, and countless civil lawsuits and corporate criminal prosecutions bear this out. These corporations also corrupt the regulators, universities, and contractors who rely directly or indirectly on their patronage. In short, I rely on the Upton Sinclair heuristic - “It is difficult to get a man to understand something, when his salary depends on his not understanding it.”
Leaning towards negative studies would be a good heuristic in the case of patented and thus potentially lucrative drugs, but not in the case of generic drugs threatening a BigPharma income stream. Based on the foregoing, I don't have a strong belief that Ivermectin is effective as a Covid treatment, but I do lean in that direction based on the Upton Sinclair heuristic. You can go into all the studies and present critiques and counter-critiques until the cows come home, but it really just comes down to which heuristic you apply.
@Scott Are you planning on adding Biber to your mistakes page? Anything where you have to write "I was egregiously wrong, and I apologize to the study authors and to you." probably qualifies. I'm not writing this as any kind of attack; I just want to see that page remain up to date because I think it is one of the keystones of your credibility: I can trust you because you will tell me when you've been wrong (and not just in the depths of some post that I might not have read, though that's already an impressive and unfortunately rare step).
I feel like this is the best post that could have been written about a debate so full of sound and fury (and in many dimensions, not signifying nothing). At one point I felt like my eyes were bleeding from the point by point depths of discussion. At another I was blessed by the novel exposure to funnel plots and their interesting and likely significance. And at another by the graph that doesn't think it's about publication bias. At all points by the way this all works as a parable about how to debate and how well meaning people can perhaps go wrong and what it looks like to keep doing it
I was familiar with Alexandros' arguments before your original post and I respect his general capacity and attitude. I always doubted ivermectin but I was interested in the debate and nowhere else did I encounter actual debate. You two win by default, but also by a mile. It's quite understandable that someone with a high personal stake in something so high stakes made some personal remarks - but it would be worthy if he took them back
Nothing much to say about ivermectin at this point, though I still find it amazing how ivermectin was able to get 100s of times more interest and noise than fluvoxamine, which was probably the better hypothesis among the possible repurposing candidates, and was being researched around the same time period.
In any case, I find it weird that you would call your own mistake "offensive". I know these days being offended is kind of fashionable, but what on earth is offensive about a honest mistake in the middle of a best attempt at an objective take on a complicated and hot topic?
> would I bet even money that they'd see "statistically significant" 10% reduction in mortality? Thinking about it as hard as I can, If I felt I could trust the scientists to not pull any dirty tricks, I think I still would, yes.
I am impressed to see an answer that's not a panacea complex.
This is so different than what IVM fans were saying in 2021. Bret and Pierre were talking about using IVM to drive covid to extinction on Rogan that year. Pierre was going further saying it's likely to be a general purpose antiviral.
If people had said "take this in reasonable doses and maybe it shaves a few days off" the 2021 discussions would have went so much more smoothly.
For all who could use a little comic relief at this point, here are 3 DALL-E images of the Ivermectin debate. The prompt was "two men having an intense debate about a medication for intestinal parasites."
Good heavens - what a waste of time on balderdash and trying to play it through disagreements on statistics.
Where the hell were you when the Pfizer and Moderna info had to be forced by a judge to release it within 75 years ?
You can say what you believe and I will continue to believe what I/ we my spouse and 7 members of a Vet’s coffee group have experienced.
We have been on prophylaxis dose of IVM for nearly three years, surrounded by Covid victims in our daily lives, close contact with positive testing individuals.
A coffee group meeting every MWF every week for many years - 7 members have all gone thru their 3-4 x Covid infections and all have treated with IVM at proper 0.2 - 0.4 - 0.6 mg/kg doses as per flccc.net protocols excepting they took the 0.6 mg when having Covid to get over it but then stopped the regular dosage after getting well with a 5 - 7. day course of therapy, only to gain another Covid infection several months again later.
You can browbeat and try stats to kill off IVM but we will continue our usage for prophylaxis for as long as the idiocy continues.
You would be better served and maybe be accepted more readily if you chose to use the fraudulent results
Of Pfizer, Moderna et.al 60 day trials which more align your thoughts.
You keep your results and we will keep ours and obviously never the twain shall meet.
But we know our friends and fellow IVM users will be around to meet again and again as against the bullshit you keep trying to kill it off while ignoring the works of RCT’s with proper dosage and protocols.
Given that ivermectin was so deeply plugged into the larger COVID conspiracy nexus, in which the pandemic was some combination of (Chinese bioweapon/not that dangerous/media invention/5G side effect/easily treated with horse medicine) done to (usher in world government/make money for pfizer/hurt Trump's campaign), I'm finding it incredibly hard to even care about the debate now that the pandemic is functionally over basically everywhere except China, which is already a semi-totalitarian state anyway.
Your initial post on this in Nov 2021 was a revelation. Thank you very much. This is far outside my field, and yours was an exhaustive review of the evidence to date on ivermectin (such as it were), as well as a summary of Dr. Bitterman’s fascinating hypothesis. I shared it with many colleagues. What a great read.
Well done here to respond to one of your strongest and most committed critics on this file. You’re not looking for pats-on-back but should get some anyway. Not every day you see mea culpa’s as you’ve done here. It’s a strong show of character.
On the science of ivermectin, however, it boils down to a simple first principle for me: the scientific method. The burden of proof rests with those who propose ivermectin to be of benefit. Until such time that they can provide causal proof of “benefit” via an RCT that reaches statistical (and also hopefully, clinical) significance, the only solution is to accept the null and reject their “hypothesis”. Anything else is just noise.
I’d also add that the requisite standard of proof needs to be met for any and all potential “use-cases” of ivermectin, such that there is no conflating btw prophylaxis and treatment; or of minor vs severe illness; early vs late; on O2 vs not; the list goes on. “Proving” a treatment “works” is a tough slog. It’s no wonder the ivermectin crowd shied away from it. Which is too bad, since globally they had the time to do many poor studies that were never going to be sufficient to prove causation, whereas that time should and could have been more gainfully spent on protocols that at least had a hope of answering the question. And of course, the later and typically western RCT studies showed no benefit.
As for the “wager”, that is silly talk best reserved for religion than for medicine (as you rightfully suggest). We don’t prescribe things that are useless just because they are safe. Your critic’s “wager” belies his religious fervour for ivermectin, in the absence of (appropriate) evidence.
It’s people like your good self that kept air in the conversation, many of us knew something wasn’t right and while it was obvious you would be attacked I commend you for standing your ground.
> But it got results p = 0.001, and you would need a thousand studies in file drawers to produce one like that. Weird.
I'm not sure if I'm teaching you to suck eggs here, but a p-value doesn't report the probability that the two groups in a trial have the same effect in real life, but rather the probability that the observed differences are due to chance. So a far more common reason to see a p-value like this conditional on ivermectin having no effect would be some subtle error of trial design like a loss of blinding or statistical error in the analysis.
It seems entirely plausible to me that an equivalent study that made a coinflip statistical error leading to p=0.999 wouldn't get published, so you only need to worry about the existence of one hypothetical unpublished trial, not thousands of them
1. Thank you for the follow-up. Yes, in 2023 all work on ivermectin vs. covid looks pointless - but a lot generalises - and in Germany, Ivermectin was never a big thing in the public discourse to begin with (though at least Austrian pharmacies were sold-out, at times.)
2. I still feel your post was at least A- by content, not C. But sure, standards differ. This is the No1 substack on science - actually on rta - but still no habilitation. And no infamous NYT/WaPo.
3. The Carl Sagan/Pinker thing: Stuff must be true, but truth must not be boring. Your original post was not just very, very educative, it was hilarious. I was literally lol when reading and smiling for weeks after. Although on first clicking it, I doubted I would finish that post (that crazy worm-medicine? Dozens of studies reviewed? Meh. ) I would never have read this follow-up, if the original post had bored me.
Completely off-topic, but I'm curious how a name like "Meyerowitz" comes about. -owitz is obviously the Slavic patronymic marker -ovich, rendered into German where the ch sound doesn't exist. Patronymics are based on the father's personal name.
So the suggestion appears to be that someone named Meyer had a son in a Slavic country who took an ordinary patronymic, that patronymic was somehow reinterpreted as a family name, and then at some point someone from the family moved to Germany. (Or, possibly more likely, someone of an unknown name started a dynasty in... say, Russia, they used normal Russian patronymics alongside their weird Jewish names the whole time they were there, and then eventually someone whose dad had been named Meyer moved to Germany, where everyone forgot about his family name and reassigned his patronymic to serve that purpose?)
Does anyone know which of those events likely happened in what order, and when/where the "Meyer" spelling is likely to have been introduced? I see on wikipedia that Meyer Lansky was originally named "Maier", which I assume can only have been pronounced exactly the same way as "Meyer".
Or am I interpreting "Meyerowitz-Katz" as a two-parent surname when in fact it's a completely normal patronymic accompanied by a completely normal surname, hyphenated for no particular reason?
Very minor point, this sentence: "Ivermectin was originally an anti-parasitic-worm medication before being repurposed to fight COVID, and everyone agrees it is very good at this," is a bit confusingly written, because it's not clear what 'this' refers to. Later parts of the text clarify, but maybe you want to tweak the sentence in case someone decides to quote you out of context again.
Thanks for writing this up! I've actually mildly adjusted the other way after reading this, to be perfectly honest (and rather unexpected going in), but I think it's a good sign for rigorous reporting.
"Doesn’t this mean that the medical establishment wants to blame Cadegiani both for giving drugs that don’t work, and for not giving them to enough people?"
The main points were that he had ethics approval for a small study, on mild/moderate patients, in a single hospital in the Brazilian capital, and somehow the study ended up being conducted in several hospitals, in critical patients, in the Brazilian Amazon. Also, the consent forms handled to patients were not the same as the ones initially approved, and didn't contain many passages about patients rights.
This study had a 50% mortality rate in the control group, with a significant amount of hepatic and renal failure. The author says they were in the placebo group, but that wouldn't have been known as the deaths were happening, as it was a blind study and there was no effort on the part of the author to request unblinding to check if maybe the renal and hepatic toxicity was because of the maybe renal and hepatotoxic drugs. When requested, the author did not provide any data to provide the traceability of the medication, so it can't be ruled out that the placebo group was actually receiving medication. The "Independent data monitoring comitee" in the study was under the supervision of someone on the payroll of the pharmaceutical company making proxalutamide
So, either the study was properly blinded and he decided to not report excess Kinsley and failure deaths which could presumably be linked to the new, barely-tested drug, which would be a crime, or he knew that the placebo group had an overwhelming amount of excess deaths and didn't try to prematurely end the study, causing willful harm to the patients, which also would be a crime
I haven't been able to find any news on the subject since the police raid on the author's house last August, but this is unsurprising given the slow pace of the Brazilian justice system
" The whole “repurpose existing drugs against COVID” idea seems to have been a big wash. "
I would paraphrase as:
"Because !IVM, and perhaps !FLV , then the space of all existing drugs & compounds prior to the advent of COVID at any dosage regimen are very unlikely to show any safe efficacy against COVID morbidity or mortality."
Would you mind characterizing the statement further?
As always, I appreciate Scott giving a hearing to folks with different perspectives, and I especially appreciate his integrity in transparently engaging with his mistakes and with points won by the other side. That said, it is hard for me to see the potential value in this discourse, for two main reasons:
1. I don't think that litigating and dissecting the reported details of 1-2 dozen small positive studies has scientific value. If there's anything we can learn from The Control Group is Out Of Control (https://slatestarcodex.com/2014/04/28/the-control-group-is-out-of-control/), it's that the epistemic value of scientific studies is much more limited, conditional, and contextual than we would like. Deep dives into the reported protocol don't necessarily help. Shoveling a bunch of small studies into a meta-analysis doesn't necessarily help, if their internal issues combined with the file-drawer effect are sufficiently severe. Genuine epistemic progress simply requires better studies with larger samples, more reliable protocols, etc. I am not aware of any instance in which applying overwhelming brainpower to small studies has produced scientific progress. It seems to me that, by the time we get around to the Conclusions on this post, Scott has basically come back to his 2014 SSC post. We know this style of science doesn't work because it can be leveraged to prove anything, no matter how nonsensical.
2. I place low trust in Alexandros as an epistemic agent based on heuristics I consider reliable. Alexandros' posts are saturated with emotional, dramatic, adversarial language, and he spends a visibly disproportionate amount of time litigating points of very marginal relevance. I have seen this pattern frequently online, and it is a strong signal of low-quality thinking. The fatal temptation of polemicists is to become obsessed with winning at the expense of truth. In their sprawling, point-by-point refutations of their opponents, it quickly becomes impossible for a third party to discern what the original point of the discussion was, or how the points currently being discussed tie back to it. It's hard for me to say what I learned from this post, except that even Scott, with his extraordinary thinking and writing powers, cannot turn a discussion of this type into something illuminating. I'm glad Scott emerged from the morass with some lessons, but it's hard for me to say the same.
I think it is time to invoke Shaw's adage about wrestling in the mud with certain hooved animals and move on.
The Biber et al. study looks interesting and is the one most likely to shift my opinion towards "maybe ivermectin does have a specifically anti-covid/anti-viral effect as distinct from killing off parasites which enables immune system to do better fighting the virus", but it's still "early treatment of non-hospitalized patients with mild COVID-19", so that tells us nothing about "can it work for severe cases?"
But I am solid on Flavio A. Cadegiani, M.D., Ph.D., M.Sc. and his alleged gold medals don't impress me much:
"His superlative achievements include amongst the fastest PhD obtained in the history of Federal University of São Paulo (7 months) and concurrent gold medals in Mathematics, Chemistry and Physics Brazilian Olympics in his adolescence."
I remain to be convinced about getting a PhD in seven months, and there seem to be a lot of different Brazilian Olympiads so without knowing which ones he entered, it's impossible to say if he did win what he claims, or if the particular competition was ranked highly.
As to his covid treatments, he didn't simply stick to ivermectin - he was pushing for an anti-androgen as another miracle cure, except unhappily that study got pulled for bad methodology (and he seems to have gone for the anti-androgen *after* the ivermectin work, so not reliant on ivermectin as being the sole single cure):
"Violations of medical ethics and human rights committed in a trial of an experimental drug touted by President Jair Bolsonaro as a cure for covid-19 were the worst in Brazil’s history, the country’s research regulator has said.
The clinical trial of proxalutamide “disrespected almost the entire protocol” and may have contributed to the deaths of as many as 200 people, said the National Health Council (CNS), which oversees clinical research in Brazil. Some of those people were not adequately informed of the risks they were undertaking in the trial, and some did not know that they were taking part in one, it said.
Brazil’s attorney general is investigating the possible violations of medical ethics and human rights on the recommendation of the National Research Ethics Commission (CONEP), which forms part of the CNS. The trial’s principal investigator, Flavio Cadegiani, was identified in October along with 68 others by a parliamentary inquiry into Brazil’s management of the pandemic as having committed “crimes against humanity.”
Proxalutamide is an anti-androgen that blocks the function of some male hormones. It is being tested by China’s Kintor Pharmaceuticals for prostate cancer, breast cancer, and other uses. It has not been approved for use in any country but was talked up by Bolsonaro as a treatment for covid-19. Shortly after recovering from covid-19 he asked why it had not been approved for use and promised to make it available to all of Brazil."
There's also the "conflict of interest" disclosure about Dr. Cadegiani, et al. on the ivermectin study:
"Lucy Kerris is a paid consultant for both Vitamedic, an ivermectin manufacturer, and is co-founder, as well as acting as a paid consultant, for Médicos Pela Vida (MPV), an organization that promotes ivermectin as a treatment for COVID-19 and discourages COVID-19 vaccination. Flavio A. Cadegiani was a paid consultant (USD 1,600.00) for Vitamedic, an ivermectin manufacturer. Dr. Cadegiani is a founding member of the Front Line COVID-19 Critical Care Alliance (FLCCC), an organization that promotes ivermectin as a treatment for COVID-19. Pierre Kory is the President and Chief Medical Officer of the Front Line COVID-19 Critical Care Alliance (FLCCC), an organization that promotes ivermectin as a treatment for COVID-19 and discourages COVID-19 vaccination. Dr. Kory reports receiving payments from FLCCC. In February of 2022, Dr. Kory opened a private telehealth fee-based service to evaluate and treat patients with acute COVID, long-haul COVID, and post-vaccination syndromes. Dr. Kory expresses personal points of view against COVID-19 vaccination. Jennifer A. Hibberd is a co-founder of the Canadian Covid Care Alliance and World Council for Health, both of which discourage vaccination and encourage ivermectin as a treatment for COVID-19. Juan J. Chamie-Quintero is a contributor to the Front Line COVID-19 Critical Care Alliance (FLCCC), lists the FLCCC as his employer on his LinkedIn page, and expresses personal points of view against COVID-19 vaccination."
He may well be an expert in sports medicine (though it would seem easier to be an expert in a field where you're writing the majority of the works) but I don't think he is trustworthy re: "don't get vaccinated, take my miracle cure instead":
"Dr. Cadegiani is the only author of the sole book in Overtraining Syndrome, the prevailing sport-related disease among amateur and professional athletes. He is also responsible for approximately 70% of the articles published in the field in the world in the last 05 years, and reviewer for more than 90% of the manuscripts in the field."
"In response to this section, Alexandros stresses that he is not necessarily saying Borody et al is incorrect or challenging my decision to leave it out." -> probably meant to be: [..]Borody et al is correct[..]
"[..] authorize it, it (according to Carvallo) it is [..]" -> one less "it"
> Alexandros has grappled with this same mystery, and taken a different route, Alexandros has wrestled with these same problems, but solved it differently.
Scott, I do not need to tell you that you are an extremely talented writer. Why do you waste your precious skill and time answering to somebody without the credentials?
The TOGETHER trial, published in NEJM 2022 was negative. Done. This is the one observation that refutes the hypothesis that Ivermectin decreases hospitalization rate in patients with Covid-19. And it was done in Brazil, one of the countries with the highest prevalence of Strongyloides.
Much appreciated. I have a good impression of Alexandros's efforts in much of this and feel your post added further insight and nuance, regardless of one's prior.
A side of this I'd like your take on:
For me it seems not at all clear that Ivermectin would have an effect today. For one thing, Pierre Kory and many associated with him shouted a preliminary alert when delta was appearing, that it seemed ivermectin might not be nearly as effective on that variant and that all prior indications of likely ivermectin effectiveness should not be trusted absent new data from Delta and onwards.
1) Off the top of your head, does it seem probably to you that ivermectin was initially effective and a missed opportunity on the original strain but lost most/all of its advantage after variants?
2) Whether ivermectin is effective or not, seems very murky to me and ultimately of much less interest than whether our institutions acted appropriately by the data available. Much clearer than whether ivermectin is effective, I have a strong sense that there was a significant signal with low cost and risk that institutions were failing to explore and in many cases discrediting in deceitful ways. Do you have any thoughts on how appropriately any early indications were acted upon?
"The whole “repurpose existing drugs against COVID” idea seems to have been a big wash."
Dexamethasone is cheap and repurposed and is still one of the primary treatments for hospitalized covid patients requiring high-flow oxygen. This was confirmed with an early RCT from the RECOVERY trial.
is there an "undeniable unbreachable fortress of evidence" for the Pfizer vaccines?
Why is that the standard for IVermectin but not for the newfangled, non-immunizing treatments that causes heart attacks but are still mandatory. (i.e. Pfizer/Moderna vaccines)?
The result is claimed to be strongly positive for early treatment. Does it contradict any "high quality" trials? I think ACTIV6 and TOGETHER were later interventions. The devil is in the details, and Alexandros' takedown of ACTIV6 was convincing, especially as regards the study authors' mixing up of Bayesian confidence (ideally 1) and p-values (ideally 0). https://doyourownresearch.substack.com/p/activ-6-trial-ivermectin-scientists
FWIW, I also respectfully asked the corresponding author about a seeming inconsistency in the paper (by email from a top .edu address). No answer.
>I will repeat that my strong objection, is that you wrote " this is not how you control group, @#!% you". I therefore pointed to stat news to support my case that, yes, this can indeed be how you control group. That's all. In the article I even noted that this aversion towards disrespect to elders may even be a cultural difference between us. To be clear, if I were making a case for ivermectin, I would not be relying on this study as my starting point.
Okay, but I have to ask, what *is* Alexandros making the case for? Is he trying to make a case for ivermectin, or simply saying that you're being insufficiently respectful to the people who study it? Because "This methodology exists and is used in a couple of specific cases that have nothing to do with ivermectin" is irrelevant for the first, but a 21-part essay titled "The Potemkin Argument" seems a little overboard if all he wants to say is that Scott is sometimes snarkier than he should be.
In general, Alexandros's responses give me the impression of "JAQ-ing off" - switching back and forth between "Here's the case for Ivermectin" and "I'm not making an argument for ivermectin, just pointing out some methodological issues on the anti-ivermectin side" depending on how well each individual part of the argument holds up. But if you start off by comparing your opponents to Russian propaganda, then I'm going to be a bit dubious that you're approaching this in a spirit of collegial truth-seeking.
This is impressively detail oriented. It also maybe loses sight of the framing of this whole discussion.
Marinos' arguments are merely the motte to Weinstein's bailey.
Marinos argues that we should simply take a chance on ivermectin treatment, because it's harmless and it seems like it maybe works, if you squint at the data just right and disregard all the largest trials as corrupt.
Weinstein argued that we could stop the pandemic by using ivermectin, that the suppression of ivermectin proves that the institutions are captured by big pharma. Along with many others, he argued that you shouldn't get vaccinated because vaccines are unsafe and ivermectin is at least as effective.
Marinos, sites like ivmmeta, and no shortage of other Substack writers offer just enough rigor for that argument to be plausible.
I originally took an interest in ivermectin because some of my friends and family bought that argument. They did not get vaccinated, they caught covid, they bought hydroxychloroquine and ivermectin from some telehealth providers. None of them died. But they got plenty sick and didn't recover quickly. It sure didn't seem like a miracle cure.
In that case, we've moved on from Omura's wager -- "if it's even 10% effective, ivermectin could save so many lives". We've moved to, "is it as effective as the covid vaccines?"
As of 2021, the FLCCC recommended ivermectin as the primary treatment for covid and briefly suggested you talk to your doctor about maybe getting vaccinated, almost as an after thought. As of today, I believe they recommend not to get vaccinated at all. In fact, Dr Kory now recommends against vaccinating children for any diseases, not just covid:
And, of course, Omura's wager was always a forced binary, so the motte itself is hard to defend.
Scott explained this well in "Pascallian Medicine" -- you could make the same argument for taking any and every reasonably safe drug when you get covid. Maybe one of them helps? Who's to say otherwise? Why not take 20 drugs?
Had I gotten covid in 2020, I would have gone ahead and taken vitamins, zinc, aspirin, low risk things like that, despite having low certainty that they would help. But I never understood why ivermectin moved to the top of the list of things to experimentally take. What made it any better than other speculative treatments, like fluvoxamine or metformin or bromhexine or camostat? Just the popularity on Youtube?
If ivmmeta's methodology for evaluating drugs is correct, why do we even need ivermectin? They think quercetin works just as well and melatonin is close.
Where were all the podcasts telling people to take quercetin and melatonin? Wouldn't those save more lives?
I suppose the government can't suppress those drugs, so you can't make a conspiracy theory out of any of them. And you'll get a lot less Substack subscriptions that way.
Remdesivir was a repurposed drug that proved to be highly effective (but expensive and impractical, since it only worked well if injected early enough)
I still wonder if the related drug GS-441524 could have also worked as an oral version. I'm not sure, maybe you can't hit a high enough concentration that way.
And I wonder if there are drugs we missed entirely, by not testing them in large enough trials.
I think there are still a lot of good questions.
One is why "early treatment" supporters haven't pivoted to metformin:
I think the answer is easy: Most of them are selling blogs and videos, so they need to talk about popular topics. Vaccine and ivermectin conspiracies will remain the product for as long as it's the most lucrative thing to talk about.
It also raises more important questions, like: what's the bar for getting the FDA to actually care about a generic drug? Would a second trial confirming COVID-OUT's metformin results be enough? Can that be done in under a year? Is it possible to still get statistical significance now that the virus isn't killing as many people?
In general, it seems like we need a breadth first search for treatments that isn't biased to approve a patented product over a generic. And we need that search to work a whole lot faster, if it's going to save any lives in the next pandemic.
Right now we have one process that works pretty well for patented drugs. And we have some sort of chaotic process that throws lots of attention towards whatever's popular on Youtube. Is there some third approach which could find cheap treatments faster?
Quick question. All I wish to know about ivermectin- is the risk/reward ratio worth taking? You appear to be conceding a sizable amount of your early analysis to rather embarrassing (your words. I'd call it credibility suicide, but I'll go with your description) yet somehow claiming a stronger argument on your own behalf as a result..
This reads more like a last ditch ego stroke to satisfy your own fanboys than a refutation of his work. He has put in the effort and done the math. I would not judge him to be pro ivermectin, simply because he proved you wrong.
At the core of this debate, there seems to be no disagreement. You were mistaken on multiple points. There is no value in your concession or apology if the end result is merely to assuage your own guilt for failure to analyze the situation beyond the popular media tropes. .
You have defended the popular belief, poorly, and when eviscerated for your inability to reach a conclusion which matches the facts, you've responded with even less facts, yet still declared victory.
I have heard of you as an example of the rationalist community, but I think you may have left some letters off the self assessed title. Clearly this is rationalization and nothing more. Title yourselves appropriately, please.
Also a couple direct criticisms-
"In contrast, the pro-ivermectin side did an amazing PR job. "
You mean the side that didn't have a billion dollar advertising subsidy to promote itself and was maligned by every institution, agency, and media conglomerate with access to a broadcasting. antenna won the pr war on this one? On one side you have billions of dollars of efforts and an infrastructure built on concepts of Total Information Control and Awareness and the other side have Joe Rogan and Joe did an amazing job? That seems rational and reasonable.
"I give myself a C+ for results but an A for effort. I wish other people would do this so I could stay in my lane of sniping at people with bad opinions about antidepressants."
If everyone gets a trophy? Sure. You got it big guy. Take 5 gold stars for your work, as well, On the other hand, for actual value to the species in this situation, let's await a more objective judgement if you don't mind. Going out on a limb, I'd guess this is your swan song and very last bit off influence on the public discourse, but I have been wrong before and certainly will be again.
First off, I can't say how deeply I appreciate you correcting on the meta-analysis function (which I believe I railed at you about, sorry), if everyone had the guts to correct their math post-factum the world would be a better place.
Personally I still see the problem here being that medicine, using current trial approaches and statistical methodology, can't prove or disprove any intervention really works for anything, it got stuck at penicilin and smallpox vaccines, everything beyond that is a bunch of interesting drugs (some of which you can try, and might work) that have no real clear effect.
You can't study single digit effects on covid because it's either very fuzzy (time to negative PCR-T, symptom reduction style endpoint) or very niche (mortality from a disease that doesn't kill that many people, and almost always kills in conjunction with a host of other morbidities). Nor is this a good thing to study, we should either aim for drugs that work so well the endpoint, drugs should cure things, and we had drugs that cured things. But when they stopped working we redefined the endpoints instead of redefining drugs.
The proof for this is to look at any life expectancy numbers, the US added 8 years since the 1960 (10 if you discount covid), compared to 22 in the 60 years before that point -- in spite of <points at all the problems in the first half of the 20th century> -- We got vaccines for smallpox, hep a & b and polio, we figured out antibitoics --- then the well ran dry, and we've been grasping at straws that one can adjust away with better food or removal of pollutants or sewage.
This kind of debate is the thing that happens when you've given up on hopes of curing disease and moved the marker to 2nd or 3rd degree aggregated statistics.
I have followed this since the start. In many respects, "can't see the wood for the trees" comes to mind. I am pro ivermectin but it is not available to me here in the UK so For prophylaxis I take Multivitamins, C, D3, zinc AND Quercetin as the ionophore. (Mercola Qurcetin articles and other studies show benefit).Got cold like symptoms once in 3 years. Never masked, never tested with PCR so I do not know if I had Covid. However, I now have parosmia so smell and taste ruined. Onion, garlic and certain toiletries are disgusting to me now. Meanwhile, not a hysterical Alex Jones type analysis....
https://21stcenturywire.com/2023/02/01/covid-19-vaccine-a-military-response-to-a-public-health-threat/ covers the framework and mechanism building on Katherine Watt and Sasha Latypova work around the bioweapon. Long but thorough. Far more pressing than a spat about a suppressed Nobel prize winning, cheap, safe and effective molecule that has been around for decades. Remember, EUAs predicated on no treatments available. Dont get me started on HCQ and the murderous Oxford overdosing trial. Dont worry after all is over we can Build Back Better. I mean us, not Schwab, Tedros, Gates, Fink and their cronies.
Lurker here who carries out systematic review and meta-analyses in academia on a daily basis. Publication bias and funnel plots are better described as the study of "small study biases" than "publication bias" (but the original name has stayed and it would be hard to change it right now...).
Publication bias is one of many explanations, but typically the wrong one. The most likely explanation for assymetry is not that studies carried out suddenly dissapear. Instead, this assimmetry is more easily explained more typically related to p-hacking or other QRPs (Questionable research practices), or more generally poor experimental designs.
It is far easier to shift an effect through p-hacking /QRPs/poor experimental designs in a noisy small study and shift a negative result to more positive one, than in a less noisy and bigger one. Hence, the assimetry does not come from not publishing studies, but from tweaking small negative ones into positive and then publishing them.
This is so in general, but even more likely in this case (where it was aligned with conflicts of interest and political motivations). In this regard, the fourth argument (which is the only one specific to these set of studies, and not against funnel plots in general) is basically baseless.
Cue 37-part series from Alexandros explaining why Scott is Betraying Rationalism by including such a phrase.
I jest, but Alexandros has gotten *so much mileage* out of waging the culture war on this topic. He has worked very hard to frame his stance as simultaneously subversive and indignant. This shouldn't detract from the legitimate research he has collected. But it does predict his response to any high-profile conversation with Scott.
Anything Scott says can and will be used...not exactly against him, but *for* retweets and Substack follows. That does mean against him if and when Alexandros can frame it as punching up.
So I'll be satisfied if this is the end of the line. Scott has engaged, over and over again, with the factual scaffolding of Alexandros' arguments. The 5-10% chance is an adequate conclusion. Let Pascal and Omura wager accordingly.
"Realistically, this project was outside my expertise and competence level."
FWIW I've thought about that piece a lot this last year, and the object-level "worms break science!" point wasn't what most changed my epistemology. The general importance of turning contrarian views over in one's mind dominated my thoughts and how I discussed this with others.
It was a winning item for me in one-on-one discussions with thoughtful people about contrarian ideas, and how such ideas might be wrong in subtle ways that require thorough analysis, and how herding on this stuff has some risks. It provided a great way for me to evangelize about about how politics is the mindkiller without just talking about politics being a mindkiller and wearing everybody out with it.
"C+ for results"... When you're trying to improve on what's out there, you get to grade yourself on a curve relative to baseline, not on whether or not you saw the pure true center of the universe unfiltered. What's your grade relative to the next best take on this topic?
"Realistically, this project was outside my expertise and competence level."
So help me. If you shy away from these sorts of investigations in the future because you think you overreached here then humanity is doomed. If this implies any risk of self-gatekeeping, please ask yourself if we are better off leaving this to the next best available substitute.
Prompt: Please rewrite Yeats' The Center Cannot hold, but make it about Scott Alexander of Astral Codex Ten worrying that he didn't write a good enough essay because it had some errors and overextended a conclusion, when he in fact wrote the most thoughtful essay humanity produced on the topic.
Short rule of thumb for me. If Alex Jones is on the side of something, it isn't worth even considering. There is too much going on in life
A sensitive instrument is one that can pick up and amplify small signals that others miss, and no doubt this "sensitivity" has been an integral part of what has made your writing as successful as it has been. Getting overwhelmed more easily than you might otherwise be is the other side of that tradeoff, and it's great that you're not letting the tough side of that get in the way of giving things the response you know they're due.
Thanks for taking the time to respond to him, even though it hasn't been fun. I doubt I'm the only one that notices and appreciates the effort.
> Dr. Ioannidis is known to be unusually rigorous and this is part of his pro-rigor crusade
I'm not familiar with all his work, but I think he deserves a reputation for self-promotion more than a reputation for rigor. I think his most famous essay is "Why Most Published Research Findings Are False", and I think this paper is notable for its sensational headline much more than for its careful analysis.
https://en.wikipedia.org/wiki/Why_Most_Published_Research_Findings_Are_False#Reception
typo: contraian
this (thankfully) isn't a typical post, but the extremely detailed, technical analysis of a minor current thing from 2 years ago is peak ACX energy
pedantic typo patrol: "...equal number of studies on both size."
We have done an intensive study of viral load dynamics in our phase 2 platform trial. Ivermectin got dropped pretty quickly (pre-specified futility rule was met). In contrast, known effective antivirals such as casirivimab/imdevimab and remdesivir showed an effect on viral load dynamics pretty quickly.
https://www.medrxiv.org/content/10.1101/2022.07.15.22277570v1
https://www.medrxiv.org/content/10.1101/2022.10.17.22281161v1
Alexandros blocked me on Twitter long ago just for repeatedly asking whether there was one good RCT supporting ivermectin. He apparently runs a startup, yet has the time to write a series of posts that literally total over 86,000 words (a sizeable non-fiction book!) just to respond to ONE blog post from Scott. And he did this over the space of a few months.
In other words, he has been engaged in a 24/7 side hustle--writing a full non-fiction book in a few months -- and moreover, it was all on a very narrow question about whether one particular blogger was mean about a particular drug that supposedly is a miracle cure for Covid (even though no rigorous study has ever shown that). Odd choice of priorities.
And despite aiming towards rationality, he is keen to write thousands of words nitpicking at the tiniest flaws in any negative ivermectin study, even while defending much worse studies or websites that defend ivermectin. Blatant double standard.
This analysis is great but I'd like to pay more attention to what our prior should be that some existing medication has a substantial benefit in treating a new disease when there isn't a strong theoretical basis supporting that use. My sense is that our prior should be very low. I mean, just ask yourself how much you'd be willing to bet about a large rigorous study showing a 10% drop in deaths from a random drug.
All the argument as to what direction the studies on ivermectin point is great and all, but even in the best case the evidence isn't very strong. I mean, it's not unusual in the slightest in medicine to see some initially suggestive studies only for the effect to turn out to be a mirage.
So while I think pinning down just how much evidence we have for ivermectin is important insofar as it should inform us whether to bother investing in yet more studies the idea that this should be enough to convince you to take the bet against the strongly theoretically supported side-effects (plus a reasonable prior that many drugs have them) seems like a big stretch.
I must admit that I scrolled straight to the conclusions, and even that felt like a strain at this stage of the ivermectin conversation. The graph of likely false positive results at the end of the post was quite interesting, and I admire Scott's dedication to this topic. Otherwise, I'm hoping we can all move on to other topics!
You seem to be missing the positive results in the MedinCell trial and the new paper on hemagglutination showing that IVM both prevents and reverses RBC clotting associated with poor outcomes.
I agree with these conclusions a lot more than those from your original post.
One quibble: what does "clinically significant effects" mean? Is the burden "better than a placebo"? "Better than Paxlovid"? "Better than no placebo"? The differences in efficacy between those thresholds are large enough to drive a truck through.
I think there might be a typo in the phrase "not for excess mortality in the control group" in the Cadegiani et al section
If they were hypothetically going to condemn him for giving unhelpful drugs to people, wouldn't that involve them complaining about excess mortality in the treatment group, not the control group?
Outside of the technical arguments, you deserve credit for taking the high road in responding to Alexandros.
If nothing else, his bad faith characterizations of you made it clear that he's not very strictly bound by the available evidence. No one is perfectly in control all the time, especially when there's strong disagreement - but getting as far out over his skis as he did in impugning your motivations is a strong tell for "motivated reasoning." Unless he's got mind-reading skills which he hasn't yet explicated.
But the bottom line is that accountability is the key - and you've demonstrated accountability here.
I doubt Ivermectin works, happy to be convinced with stronger evidence.
What I found most alarming was how politicized Ivermectin became, most glaringly with the CDC, pretending as if Ivermectin was only for livestock with their infamous "you are not a horse" tweet [1] which may (?) have encouraged people like Dr Jason McElyea to make up fictitious stories about people "clogging up hospital beds because they are OD'ing on Ivermectin" [2]
Meanwhile, the CDC suggests ivermectin off-label to treat lice, even though there are plenty of FDA approved OTC and prescription treatments [3]. I quote:
"Given as a tablet in mass drug administrations, oral ivermectin has been used extensively and safely for over two decades in many countries to treat filarial worm infections. Although not FDA-approved for the treatment of lice, ivermectin tablets given in a single oral dose of 200 micrograms/kg or 400 micrograms/kg repeated in 9-10 days has been shown effective against head lice"
_____
[1] https://twitter.com/us_fda/status/1429050070243192839
[2] https://reason.com/2021/09/06/ivermectin-overdoses-oklahoma-hospitals-rolling-stone-hoax/
[3] https://www.cdc.gov/parasites/lice/head/treatment.html
I think many engineers would quibble with the characterization of Elon Musk as an engineer, let alone a brilliant one. He does seem to have had a knack for a) picking investments in engineering companies to buy* or start, and b) hiring good engineers to work for him. But those are not engineering.
* One of his main mistakes with Twitter seems to be his belief that it's an engineering company, rather than a media company.
> …but am generally not impressed.
Same. AM’s Twitter is mostly a dumpster. He’s clearly trying to ride Ivermectin into the spotlight.
Yet I still find these posts valuable due diligence. Thanks.
"My usual heuristic is that when studies contradict, I trust bigger studies, more professionally done studies, and (as a tiebreaker) negative studies - so I leaned towards the studies finding no effect."
The debate really comes down to the validity of that heuristic. Trusting bigger, more professionally done studies would be a great heuristic if we assume corporate money does not corrupt scientific research. But this is of course a false assumption. I trust small studies done by doctors trying to treat patients in a third world country over large, professionally done studies brought to you by Pfizer.
I believe BigPharma is completely corrupt, and countless civil lawsuits and corporate criminal prosecutions bear this out. These corporations also corrupt the regulators, universities, and contractors who rely directly or indirectly on their patronage. In short, I rely on the Upton Sinclair heuristic - “It is difficult to get a man to understand something, when his salary depends on his not understanding it.”
Leaning towards negative studies would be a good heuristic in the case of patented and thus potentially lucrative drugs, but not in the case of generic drugs threatening a BigPharma income stream. Based on the foregoing, I don't have a strong belief that Ivermectin is effective as a Covid treatment, but I do lean in that direction based on the Upton Sinclair heuristic. You can go into all the studies and present critiques and counter-critiques until the cows come home, but it really just comes down to which heuristic you apply.
@Scott Are you planning on adding Biber to your mistakes page? Anything where you have to write "I was egregiously wrong, and I apologize to the study authors and to you." probably qualifies. I'm not writing this as any kind of attack; I just want to see that page remain up to date because I think it is one of the keystones of your credibility: I can trust you because you will tell me when you've been wrong (and not just in the depths of some post that I might not have read, though that's already an impressive and unfortunately rare step).
I feel like this is the best post that could have been written about a debate so full of sound and fury (and in many dimensions, not signifying nothing). At one point I felt like my eyes were bleeding from the point by point depths of discussion. At another I was blessed by the novel exposure to funnel plots and their interesting and likely significance. And at another by the graph that doesn't think it's about publication bias. At all points by the way this all works as a parable about how to debate and how well meaning people can perhaps go wrong and what it looks like to keep doing it
I was familiar with Alexandros' arguments before your original post and I respect his general capacity and attitude. I always doubted ivermectin but I was interested in the debate and nowhere else did I encounter actual debate. You two win by default, but also by a mile. It's quite understandable that someone with a high personal stake in something so high stakes made some personal remarks - but it would be worthy if he took them back
Nothing much to say about ivermectin at this point, though I still find it amazing how ivermectin was able to get 100s of times more interest and noise than fluvoxamine, which was probably the better hypothesis among the possible repurposing candidates, and was being researched around the same time period.
In any case, I find it weird that you would call your own mistake "offensive". I know these days being offended is kind of fashionable, but what on earth is offensive about a honest mistake in the middle of a best attempt at an objective take on a complicated and hot topic?
> would I bet even money that they'd see "statistically significant" 10% reduction in mortality? Thinking about it as hard as I can, If I felt I could trust the scientists to not pull any dirty tricks, I think I still would, yes.
I am impressed to see an answer that's not a panacea complex.
This is so different than what IVM fans were saying in 2021. Bret and Pierre were talking about using IVM to drive covid to extinction on Rogan that year. Pierre was going further saying it's likely to be a general purpose antiviral.
If people had said "take this in reasonable doses and maybe it shaves a few days off" the 2021 discussions would have went so much more smoothly.
"I wish other people would do this so I could stay in my lane of sniping at people with bad opinions about antidepressants."
I'm interested in reading this sniping about antidepressants. Any links to SSC/ACX posts on that topic?
For all who could use a little comic relief at this point, here are 3 DALL-E images of the Ivermectin debate. The prompt was "two men having an intense debate about a medication for intestinal parasites."
steampunk: https://i.imgur.com/VlmdSZX.png
surrealist: https://i.imgur.com/iItysT1.png
And, Scott, especially for you:
stained glass: https://i.imgur.com/mKcsP8J.png
Might look elegant as a dining room window.
Good heavens - what a waste of time on balderdash and trying to play it through disagreements on statistics.
Where the hell were you when the Pfizer and Moderna info had to be forced by a judge to release it within 75 years ?
You can say what you believe and I will continue to believe what I/ we my spouse and 7 members of a Vet’s coffee group have experienced.
We have been on prophylaxis dose of IVM for nearly three years, surrounded by Covid victims in our daily lives, close contact with positive testing individuals.
A coffee group meeting every MWF every week for many years - 7 members have all gone thru their 3-4 x Covid infections and all have treated with IVM at proper 0.2 - 0.4 - 0.6 mg/kg doses as per flccc.net protocols excepting they took the 0.6 mg when having Covid to get over it but then stopped the regular dosage after getting well with a 5 - 7. day course of therapy, only to gain another Covid infection several months again later.
You can browbeat and try stats to kill off IVM but we will continue our usage for prophylaxis for as long as the idiocy continues.
You would be better served and maybe be accepted more readily if you chose to use the fraudulent results
Of Pfizer, Moderna et.al 60 day trials which more align your thoughts.
You keep your results and we will keep ours and obviously never the twain shall meet.
But we know our friends and fellow IVM users will be around to meet again and again as against the bullshit you keep trying to kill it off while ignoring the works of RCT’s with proper dosage and protocols.
Reality > RCT's
I think the Sillicon Valley echo chamber got the best of you
Given that ivermectin was so deeply plugged into the larger COVID conspiracy nexus, in which the pandemic was some combination of (Chinese bioweapon/not that dangerous/media invention/5G side effect/easily treated with horse medicine) done to (usher in world government/make money for pfizer/hurt Trump's campaign), I'm finding it incredibly hard to even care about the debate now that the pandemic is functionally over basically everywhere except China, which is already a semi-totalitarian state anyway.
Your initial post on this in Nov 2021 was a revelation. Thank you very much. This is far outside my field, and yours was an exhaustive review of the evidence to date on ivermectin (such as it were), as well as a summary of Dr. Bitterman’s fascinating hypothesis. I shared it with many colleagues. What a great read.
Well done here to respond to one of your strongest and most committed critics on this file. You’re not looking for pats-on-back but should get some anyway. Not every day you see mea culpa’s as you’ve done here. It’s a strong show of character.
On the science of ivermectin, however, it boils down to a simple first principle for me: the scientific method. The burden of proof rests with those who propose ivermectin to be of benefit. Until such time that they can provide causal proof of “benefit” via an RCT that reaches statistical (and also hopefully, clinical) significance, the only solution is to accept the null and reject their “hypothesis”. Anything else is just noise.
I’d also add that the requisite standard of proof needs to be met for any and all potential “use-cases” of ivermectin, such that there is no conflating btw prophylaxis and treatment; or of minor vs severe illness; early vs late; on O2 vs not; the list goes on. “Proving” a treatment “works” is a tough slog. It’s no wonder the ivermectin crowd shied away from it. Which is too bad, since globally they had the time to do many poor studies that were never going to be sufficient to prove causation, whereas that time should and could have been more gainfully spent on protocols that at least had a hope of answering the question. And of course, the later and typically western RCT studies showed no benefit.
As for the “wager”, that is silly talk best reserved for religion than for medicine (as you rightfully suggest). We don’t prescribe things that are useless just because they are safe. Your critic’s “wager” belies his religious fervour for ivermectin, in the absence of (appropriate) evidence.
It’s people like your good self that kept air in the conversation, many of us knew something wasn’t right and while it was obvious you would be attacked I commend you for standing your ground.
> But it got results p = 0.001, and you would need a thousand studies in file drawers to produce one like that. Weird.
I'm not sure if I'm teaching you to suck eggs here, but a p-value doesn't report the probability that the two groups in a trial have the same effect in real life, but rather the probability that the observed differences are due to chance. So a far more common reason to see a p-value like this conditional on ivermectin having no effect would be some subtle error of trial design like a loss of blinding or statistical error in the analysis.
It seems entirely plausible to me that an equivalent study that made a coinflip statistical error leading to p=0.999 wouldn't get published, so you only need to worry about the existence of one hypothetical unpublished trial, not thousands of them
1. Thank you for the follow-up. Yes, in 2023 all work on ivermectin vs. covid looks pointless - but a lot generalises - and in Germany, Ivermectin was never a big thing in the public discourse to begin with (though at least Austrian pharmacies were sold-out, at times.)
2. I still feel your post was at least A- by content, not C. But sure, standards differ. This is the No1 substack on science - actually on rta - but still no habilitation. And no infamous NYT/WaPo.
3. The Carl Sagan/Pinker thing: Stuff must be true, but truth must not be boring. Your original post was not just very, very educative, it was hilarious. I was literally lol when reading and smiling for weeks after. Although on first clicking it, I doubted I would finish that post (that crazy worm-medicine? Dozens of studies reviewed? Meh. ) I would never have read this follow-up, if the original post had bored me.
> Gideon Meyerowitz-Katz
Completely off-topic, but I'm curious how a name like "Meyerowitz" comes about. -owitz is obviously the Slavic patronymic marker -ovich, rendered into German where the ch sound doesn't exist. Patronymics are based on the father's personal name.
So the suggestion appears to be that someone named Meyer had a son in a Slavic country who took an ordinary patronymic, that patronymic was somehow reinterpreted as a family name, and then at some point someone from the family moved to Germany. (Or, possibly more likely, someone of an unknown name started a dynasty in... say, Russia, they used normal Russian patronymics alongside their weird Jewish names the whole time they were there, and then eventually someone whose dad had been named Meyer moved to Germany, where everyone forgot about his family name and reassigned his patronymic to serve that purpose?)
Does anyone know which of those events likely happened in what order, and when/where the "Meyer" spelling is likely to have been introduced? I see on wikipedia that Meyer Lansky was originally named "Maier", which I assume can only have been pronounced exactly the same way as "Meyer".
Or am I interpreting "Meyerowitz-Katz" as a two-parent surname when in fact it's a completely normal patronymic accompanied by a completely normal surname, hyphenated for no particular reason?
Very minor point, this sentence: "Ivermectin was originally an anti-parasitic-worm medication before being repurposed to fight COVID, and everyone agrees it is very good at this," is a bit confusingly written, because it's not clear what 'this' refers to. Later parts of the text clarify, but maybe you want to tweak the sentence in case someone decides to quote you out of context again.
Thanks for writing this up! I've actually mildly adjusted the other way after reading this, to be perfectly honest (and rather unexpected going in), but I think it's a good sign for rigorous reporting.
"Doesn’t this mean that the medical establishment wants to blame Cadegiani both for giving drugs that don’t work, and for not giving them to enough people?"
There are several pieces of context missing in this interpretation; here is a public note by the Brazilian Research Ethics Comitee https://conselho.saude.gov.br/ultimas-noticias-cns/2095-nota-publica-cns-elucida-a-sociedade-brasileira-fatos-sobre-estudo-irregular-com-proxalutamida (in Portuguese, but Google translate works well)
The main points were that he had ethics approval for a small study, on mild/moderate patients, in a single hospital in the Brazilian capital, and somehow the study ended up being conducted in several hospitals, in critical patients, in the Brazilian Amazon. Also, the consent forms handled to patients were not the same as the ones initially approved, and didn't contain many passages about patients rights.
This study had a 50% mortality rate in the control group, with a significant amount of hepatic and renal failure. The author says they were in the placebo group, but that wouldn't have been known as the deaths were happening, as it was a blind study and there was no effort on the part of the author to request unblinding to check if maybe the renal and hepatic toxicity was because of the maybe renal and hepatotoxic drugs. When requested, the author did not provide any data to provide the traceability of the medication, so it can't be ruled out that the placebo group was actually receiving medication. The "Independent data monitoring comitee" in the study was under the supervision of someone on the payroll of the pharmaceutical company making proxalutamide
So, either the study was properly blinded and he decided to not report excess Kinsley and failure deaths which could presumably be linked to the new, barely-tested drug, which would be a crime, or he knew that the placebo group had an overwhelming amount of excess deaths and didn't try to prematurely end the study, causing willful harm to the patients, which also would be a crime
I haven't been able to find any news on the subject since the police raid on the author's house last August, but this is unsurprising given the slow pace of the Brazilian justice system
" The whole “repurpose existing drugs against COVID” idea seems to have been a big wash. "
I would paraphrase as:
"Because !IVM, and perhaps !FLV , then the space of all existing drugs & compounds prior to the advent of COVID at any dosage regimen are very unlikely to show any safe efficacy against COVID morbidity or mortality."
Would you mind characterizing the statement further?
As always, I appreciate Scott giving a hearing to folks with different perspectives, and I especially appreciate his integrity in transparently engaging with his mistakes and with points won by the other side. That said, it is hard for me to see the potential value in this discourse, for two main reasons:
1. I don't think that litigating and dissecting the reported details of 1-2 dozen small positive studies has scientific value. If there's anything we can learn from The Control Group is Out Of Control (https://slatestarcodex.com/2014/04/28/the-control-group-is-out-of-control/), it's that the epistemic value of scientific studies is much more limited, conditional, and contextual than we would like. Deep dives into the reported protocol don't necessarily help. Shoveling a bunch of small studies into a meta-analysis doesn't necessarily help, if their internal issues combined with the file-drawer effect are sufficiently severe. Genuine epistemic progress simply requires better studies with larger samples, more reliable protocols, etc. I am not aware of any instance in which applying overwhelming brainpower to small studies has produced scientific progress. It seems to me that, by the time we get around to the Conclusions on this post, Scott has basically come back to his 2014 SSC post. We know this style of science doesn't work because it can be leveraged to prove anything, no matter how nonsensical.
2. I place low trust in Alexandros as an epistemic agent based on heuristics I consider reliable. Alexandros' posts are saturated with emotional, dramatic, adversarial language, and he spends a visibly disproportionate amount of time litigating points of very marginal relevance. I have seen this pattern frequently online, and it is a strong signal of low-quality thinking. The fatal temptation of polemicists is to become obsessed with winning at the expense of truth. In their sprawling, point-by-point refutations of their opponents, it quickly becomes impossible for a third party to discern what the original point of the discussion was, or how the points currently being discussed tie back to it. It's hard for me to say what I learned from this post, except that even Scott, with his extraordinary thinking and writing powers, cannot turn a discussion of this type into something illuminating. I'm glad Scott emerged from the morass with some lessons, but it's hard for me to say the same.
I think it is time to invoke Shaw's adage about wrestling in the mud with certain hooved animals and move on.
The Biber et al. study looks interesting and is the one most likely to shift my opinion towards "maybe ivermectin does have a specifically anti-covid/anti-viral effect as distinct from killing off parasites which enables immune system to do better fighting the virus", but it's still "early treatment of non-hospitalized patients with mild COVID-19", so that tells us nothing about "can it work for severe cases?"
But I am solid on Flavio A. Cadegiani, M.D., Ph.D., M.Sc. and his alleged gold medals don't impress me much:
"His superlative achievements include amongst the fastest PhD obtained in the history of Federal University of São Paulo (7 months) and concurrent gold medals in Mathematics, Chemistry and Physics Brazilian Olympics in his adolescence."
I remain to be convinced about getting a PhD in seven months, and there seem to be a lot of different Brazilian Olympiads so without knowing which ones he entered, it's impossible to say if he did win what he claims, or if the particular competition was ranked highly.
As to his covid treatments, he didn't simply stick to ivermectin - he was pushing for an anti-androgen as another miracle cure, except unhappily that study got pulled for bad methodology (and he seems to have gone for the anti-androgen *after* the ivermectin work, so not reliant on ivermectin as being the sole single cure):
https://retractionwatch.com/2022/06/10/researcher-attacks-journal-for-retracting-his-paper-on-covid-19-drug/
https://www.science.org/content/article/too-good-be-true-doubts-swirl-around-trial-saw-77-reduction-covid-19-mortality
https://www.bmj.com/content/375/bmj.n2819
"Violations of medical ethics and human rights committed in a trial of an experimental drug touted by President Jair Bolsonaro as a cure for covid-19 were the worst in Brazil’s history, the country’s research regulator has said.
The clinical trial of proxalutamide “disrespected almost the entire protocol” and may have contributed to the deaths of as many as 200 people, said the National Health Council (CNS), which oversees clinical research in Brazil. Some of those people were not adequately informed of the risks they were undertaking in the trial, and some did not know that they were taking part in one, it said.
Brazil’s attorney general is investigating the possible violations of medical ethics and human rights on the recommendation of the National Research Ethics Commission (CONEP), which forms part of the CNS. The trial’s principal investigator, Flavio Cadegiani, was identified in October along with 68 others by a parliamentary inquiry into Brazil’s management of the pandemic as having committed “crimes against humanity.”
Proxalutamide is an anti-androgen that blocks the function of some male hormones. It is being tested by China’s Kintor Pharmaceuticals for prostate cancer, breast cancer, and other uses. It has not been approved for use in any country but was talked up by Bolsonaro as a treatment for covid-19. Shortly after recovering from covid-19 he asked why it had not been approved for use and promised to make it available to all of Brazil."
There's also the "conflict of interest" disclosure about Dr. Cadegiani, et al. on the ivermectin study:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9525042/
"Lucy Kerris is a paid consultant for both Vitamedic, an ivermectin manufacturer, and is co-founder, as well as acting as a paid consultant, for Médicos Pela Vida (MPV), an organization that promotes ivermectin as a treatment for COVID-19 and discourages COVID-19 vaccination. Flavio A. Cadegiani was a paid consultant (USD 1,600.00) for Vitamedic, an ivermectin manufacturer. Dr. Cadegiani is a founding member of the Front Line COVID-19 Critical Care Alliance (FLCCC), an organization that promotes ivermectin as a treatment for COVID-19. Pierre Kory is the President and Chief Medical Officer of the Front Line COVID-19 Critical Care Alliance (FLCCC), an organization that promotes ivermectin as a treatment for COVID-19 and discourages COVID-19 vaccination. Dr. Kory reports receiving payments from FLCCC. In February of 2022, Dr. Kory opened a private telehealth fee-based service to evaluate and treat patients with acute COVID, long-haul COVID, and post-vaccination syndromes. Dr. Kory expresses personal points of view against COVID-19 vaccination. Jennifer A. Hibberd is a co-founder of the Canadian Covid Care Alliance and World Council for Health, both of which discourage vaccination and encourage ivermectin as a treatment for COVID-19. Juan J. Chamie-Quintero is a contributor to the Front Line COVID-19 Critical Care Alliance (FLCCC), lists the FLCCC as his employer on his LinkedIn page, and expresses personal points of view against COVID-19 vaccination."
He may well be an expert in sports medicine (though it would seem easier to be an expert in a field where you're writing the majority of the works) but I don't think he is trustworthy re: "don't get vaccinated, take my miracle cure instead":
"Dr. Cadegiani is the only author of the sole book in Overtraining Syndrome, the prevailing sport-related disease among amateur and professional athletes. He is also responsible for approximately 70% of the articles published in the field in the world in the last 05 years, and reviewer for more than 90% of the manuscripts in the field."
>This is a good place to note that I very poor memory of what I was thinking two years ago
Unless that's a joke I don't get you're missing a "have".
i am buying ivermectin from india and recived today.
Potential mistake:
"many drugs change secondary endpoints but fail to change the things we care."
Should there be an "about" at the end?
"In response to this section, Alexandros stresses that he is not necessarily saying Borody et al is incorrect or challenging my decision to leave it out." -> probably meant to be: [..]Borody et al is correct[..]
"[..] authorize it, it (according to Carvallo) it is [..]" -> one less "it"
> Alexandros has grappled with this same mystery, and taken a different route, Alexandros has wrestled with these same problems, but solved it differently.
I think this is a typo?
Scott, I do not need to tell you that you are an extremely talented writer. Why do you waste your precious skill and time answering to somebody without the credentials?
The TOGETHER trial, published in NEJM 2022 was negative. Done. This is the one observation that refutes the hypothesis that Ivermectin decreases hospitalization rate in patients with Covid-19. And it was done in Brazil, one of the countries with the highest prevalence of Strongyloides.
Much appreciated. I have a good impression of Alexandros's efforts in much of this and feel your post added further insight and nuance, regardless of one's prior.
A side of this I'd like your take on:
For me it seems not at all clear that Ivermectin would have an effect today. For one thing, Pierre Kory and many associated with him shouted a preliminary alert when delta was appearing, that it seemed ivermectin might not be nearly as effective on that variant and that all prior indications of likely ivermectin effectiveness should not be trusted absent new data from Delta and onwards.
1) Off the top of your head, does it seem probably to you that ivermectin was initially effective and a missed opportunity on the original strain but lost most/all of its advantage after variants?
2) Whether ivermectin is effective or not, seems very murky to me and ultimately of much less interest than whether our institutions acted appropriately by the data available. Much clearer than whether ivermectin is effective, I have a strong sense that there was a significant signal with low cost and risk that institutions were failing to explore and in many cases discrediting in deceitful ways. Do you have any thoughts on how appropriately any early indications were acted upon?
"The whole “repurpose existing drugs against COVID” idea seems to have been a big wash."
Dexamethasone is cheap and repurposed and is still one of the primary treatments for hospitalized covid patients requiring high-flow oxygen. This was confirmed with an early RCT from the RECOVERY trial.
https://en.wikipedia.org/wiki/COVID-19_drug_repurposing_research#Dexamethasone
https://www.covid19treatmentguidelines.nih.gov/management/clinical-management-of-adults/hospitalized-adults--therapeutic-management/
is there an "undeniable unbreachable fortress of evidence" for the Pfizer vaccines?
Why is that the standard for IVermectin but not for the newfangled, non-immunizing treatments that causes heart attacks but are still mandatory. (i.e. Pfizer/Moderna vaccines)?
What are the odds that the recent MedinCell PR is true vs. bogus? I'd say 50/50. https://www.clinicaltrialsarena.com/news/medincell-covid-19-prevention-trial/
The result is claimed to be strongly positive for early treatment. Does it contradict any "high quality" trials? I think ACTIV6 and TOGETHER were later interventions. The devil is in the details, and Alexandros' takedown of ACTIV6 was convincing, especially as regards the study authors' mixing up of Bayesian confidence (ideally 1) and p-values (ideally 0). https://doyourownresearch.substack.com/p/activ-6-trial-ivermectin-scientists
FWIW, I also respectfully asked the corresponding author about a seeming inconsistency in the paper (by email from a top .edu address). No answer.
>I will repeat that my strong objection, is that you wrote " this is not how you control group, @#!% you". I therefore pointed to stat news to support my case that, yes, this can indeed be how you control group. That's all. In the article I even noted that this aversion towards disrespect to elders may even be a cultural difference between us. To be clear, if I were making a case for ivermectin, I would not be relying on this study as my starting point.
Okay, but I have to ask, what *is* Alexandros making the case for? Is he trying to make a case for ivermectin, or simply saying that you're being insufficiently respectful to the people who study it? Because "This methodology exists and is used in a couple of specific cases that have nothing to do with ivermectin" is irrelevant for the first, but a 21-part essay titled "The Potemkin Argument" seems a little overboard if all he wants to say is that Scott is sometimes snarkier than he should be.
In general, Alexandros's responses give me the impression of "JAQ-ing off" - switching back and forth between "Here's the case for Ivermectin" and "I'm not making an argument for ivermectin, just pointing out some methodological issues on the anti-ivermectin side" depending on how well each individual part of the argument holds up. But if you start off by comparing your opponents to Russian propaganda, then I'm going to be a bit dubious that you're approaching this in a spirit of collegial truth-seeking.
This is impressively detail oriented. It also maybe loses sight of the framing of this whole discussion.
Marinos' arguments are merely the motte to Weinstein's bailey.
Marinos argues that we should simply take a chance on ivermectin treatment, because it's harmless and it seems like it maybe works, if you squint at the data just right and disregard all the largest trials as corrupt.
Weinstein argued that we could stop the pandemic by using ivermectin, that the suppression of ivermectin proves that the institutions are captured by big pharma. Along with many others, he argued that you shouldn't get vaccinated because vaccines are unsafe and ivermectin is at least as effective.
Marinos, sites like ivmmeta, and no shortage of other Substack writers offer just enough rigor for that argument to be plausible.
I originally took an interest in ivermectin because some of my friends and family bought that argument. They did not get vaccinated, they caught covid, they bought hydroxychloroquine and ivermectin from some telehealth providers. None of them died. But they got plenty sick and didn't recover quickly. It sure didn't seem like a miracle cure.
In that case, we've moved on from Omura's wager -- "if it's even 10% effective, ivermectin could save so many lives". We've moved to, "is it as effective as the covid vaccines?"
As of 2021, the FLCCC recommended ivermectin as the primary treatment for covid and briefly suggested you talk to your doctor about maybe getting vaccinated, almost as an after thought. As of today, I believe they recommend not to get vaccinated at all. In fact, Dr Kory now recommends against vaccinating children for any diseases, not just covid:
https://twitter.com/PierreKory/status/1606830454006587393?cxt=HHwWgoC9lZvyzcwsAAAA
And, of course, Omura's wager was always a forced binary, so the motte itself is hard to defend.
Scott explained this well in "Pascallian Medicine" -- you could make the same argument for taking any and every reasonably safe drug when you get covid. Maybe one of them helps? Who's to say otherwise? Why not take 20 drugs?
Had I gotten covid in 2020, I would have gone ahead and taken vitamins, zinc, aspirin, low risk things like that, despite having low certainty that they would help. But I never understood why ivermectin moved to the top of the list of things to experimentally take. What made it any better than other speculative treatments, like fluvoxamine or metformin or bromhexine or camostat? Just the popularity on Youtube?
If ivmmeta's methodology for evaluating drugs is correct, why do we even need ivermectin? They think quercetin works just as well and melatonin is close.
Where were all the podcasts telling people to take quercetin and melatonin? Wouldn't those save more lives?
I suppose the government can't suppress those drugs, so you can't make a conspiracy theory out of any of them. And you'll get a lot less Substack subscriptions that way.
Re: "The whole “repurpose existing drugs against COVID” idea seems to have been a big wash."
COVID-OUT seemed to report good results for metformin, with both a 40% reduction in ER visits and long covid rates:
https://www.nejm.org/doi/full/10.1056/NEJMoa2201662
https://www.medrxiv.org/content/10.1101/2022.12.21.22283753v1
Remdesivir was a repurposed drug that proved to be highly effective (but expensive and impractical, since it only worked well if injected early enough)
I still wonder if the related drug GS-441524 could have also worked as an oral version. I'm not sure, maybe you can't hit a high enough concentration that way.
And I wonder if there are drugs we missed entirely, by not testing them in large enough trials.
I think there are still a lot of good questions.
One is why "early treatment" supporters haven't pivoted to metformin:
https://medium.com/microbial-instincts/ivermectin-failed-in-a-high-dose-trial-but-metformin-might-help-treat-covid-4599720b0c22
I think the answer is easy: Most of them are selling blogs and videos, so they need to talk about popular topics. Vaccine and ivermectin conspiracies will remain the product for as long as it's the most lucrative thing to talk about.
It also raises more important questions, like: what's the bar for getting the FDA to actually care about a generic drug? Would a second trial confirming COVID-OUT's metformin results be enough? Can that be done in under a year? Is it possible to still get statistical significance now that the virus isn't killing as many people?
In general, it seems like we need a breadth first search for treatments that isn't biased to approve a patented product over a generic. And we need that search to work a whole lot faster, if it's going to save any lives in the next pandemic.
Right now we have one process that works pretty well for patented drugs. And we have some sort of chaotic process that throws lots of attention towards whatever's popular on Youtube. Is there some third approach which could find cheap treatments faster?
Quick question. All I wish to know about ivermectin- is the risk/reward ratio worth taking? You appear to be conceding a sizable amount of your early analysis to rather embarrassing (your words. I'd call it credibility suicide, but I'll go with your description) yet somehow claiming a stronger argument on your own behalf as a result..
This reads more like a last ditch ego stroke to satisfy your own fanboys than a refutation of his work. He has put in the effort and done the math. I would not judge him to be pro ivermectin, simply because he proved you wrong.
At the core of this debate, there seems to be no disagreement. You were mistaken on multiple points. There is no value in your concession or apology if the end result is merely to assuage your own guilt for failure to analyze the situation beyond the popular media tropes. .
You have defended the popular belief, poorly, and when eviscerated for your inability to reach a conclusion which matches the facts, you've responded with even less facts, yet still declared victory.
I have heard of you as an example of the rationalist community, but I think you may have left some letters off the self assessed title. Clearly this is rationalization and nothing more. Title yourselves appropriately, please.
Also a couple direct criticisms-
"In contrast, the pro-ivermectin side did an amazing PR job. "
You mean the side that didn't have a billion dollar advertising subsidy to promote itself and was maligned by every institution, agency, and media conglomerate with access to a broadcasting. antenna won the pr war on this one? On one side you have billions of dollars of efforts and an infrastructure built on concepts of Total Information Control and Awareness and the other side have Joe Rogan and Joe did an amazing job? That seems rational and reasonable.
"I give myself a C+ for results but an A for effort. I wish other people would do this so I could stay in my lane of sniping at people with bad opinions about antidepressants."
If everyone gets a trophy? Sure. You got it big guy. Take 5 gold stars for your work, as well, On the other hand, for actual value to the species in this situation, let's await a more objective judgement if you don't mind. Going out on a limb, I'd guess this is your swan song and very last bit off influence on the public discourse, but I have been wrong before and certainly will be again.
Does this change your original conclusion of "85%-90% confidence" that ivermectin doesn't work?
First off, I can't say how deeply I appreciate you correcting on the meta-analysis function (which I believe I railed at you about, sorry), if everyone had the guts to correct their math post-factum the world would be a better place.
Personally I still see the problem here being that medicine, using current trial approaches and statistical methodology, can't prove or disprove any intervention really works for anything, it got stuck at penicilin and smallpox vaccines, everything beyond that is a bunch of interesting drugs (some of which you can try, and might work) that have no real clear effect.
You can't study single digit effects on covid because it's either very fuzzy (time to negative PCR-T, symptom reduction style endpoint) or very niche (mortality from a disease that doesn't kill that many people, and almost always kills in conjunction with a host of other morbidities). Nor is this a good thing to study, we should either aim for drugs that work so well the endpoint, drugs should cure things, and we had drugs that cured things. But when they stopped working we redefined the endpoints instead of redefining drugs.
The proof for this is to look at any life expectancy numbers, the US added 8 years since the 1960 (10 if you discount covid), compared to 22 in the 60 years before that point -- in spite of <points at all the problems in the first half of the 20th century> -- We got vaccines for smallpox, hep a & b and polio, we figured out antibitoics --- then the well ran dry, and we've been grasping at straws that one can adjust away with better food or removal of pollutants or sewage.
This kind of debate is the thing that happens when you've given up on hopes of curing disease and moved the marker to 2nd or 3rd degree aggregated statistics.
I have followed this since the start. In many respects, "can't see the wood for the trees" comes to mind. I am pro ivermectin but it is not available to me here in the UK so For prophylaxis I take Multivitamins, C, D3, zinc AND Quercetin as the ionophore. (Mercola Qurcetin articles and other studies show benefit).Got cold like symptoms once in 3 years. Never masked, never tested with PCR so I do not know if I had Covid. However, I now have parosmia so smell and taste ruined. Onion, garlic and certain toiletries are disgusting to me now. Meanwhile, not a hysterical Alex Jones type analysis....
https://21stcenturywire.com/2023/02/01/covid-19-vaccine-a-military-response-to-a-public-health-threat/ covers the framework and mechanism building on Katherine Watt and Sasha Latypova work around the bioweapon. Long but thorough. Far more pressing than a spat about a suppressed Nobel prize winning, cheap, safe and effective molecule that has been around for decades. Remember, EUAs predicated on no treatments available. Dont get me started on HCQ and the murderous Oxford overdosing trial. Dont worry after all is over we can Build Back Better. I mean us, not Schwab, Tedros, Gates, Fink and their cronies.
Lurker here who carries out systematic review and meta-analyses in academia on a daily basis. Publication bias and funnel plots are better described as the study of "small study biases" than "publication bias" (but the original name has stayed and it would be hard to change it right now...).
Publication bias is one of many explanations, but typically the wrong one. The most likely explanation for assymetry is not that studies carried out suddenly dissapear. Instead, this assimmetry is more easily explained more typically related to p-hacking or other QRPs (Questionable research practices), or more generally poor experimental designs.
It is far easier to shift an effect through p-hacking /QRPs/poor experimental designs in a noisy small study and shift a negative result to more positive one, than in a less noisy and bigger one. Hence, the assimetry does not come from not publishing studies, but from tweaking small negative ones into positive and then publishing them.
This is so in general, but even more likely in this case (where it was aligned with conflicts of interest and political motivations). In this regard, the fourth argument (which is the only one specific to these set of studies, and not against funnel plots in general) is basically baseless.
This entire discourse has driven me to find refuge in poetry. Take it away, Alexander Pope:
A little learning is a dangerous thing;
Drink deep, or taste not the Pierian spring:
There shallow draughts intoxicate the brain,
And drinking largely sobers us again.
Fired at first sight with what the Muse imparts,
In fearless youth we tempt the heights of Arts;
While from the bounded level of our mind
Short views we take, nor see the lengths behind,
But, more advanced, behold with strange surprise
New distant scenes of endless science rise!
So pleased at first the towering Alps we try,
Mount o’er the vales, and seem to tread the sky;
The eternal snows appear already past,
And the first clouds and mountains seem the last;
But those attained, we tremble to survey
The growing labours of the lengthened way;
The increasing prospect tires our wandering eyes,
Hills peep o’er hills, and Alps on Alps arise!
Does anybody else find it funny that there was a whole debate over the statistical validity of funnel plots while letting p-values slide?
> gestalt vibes of untrustworthiness
Cue 37-part series from Alexandros explaining why Scott is Betraying Rationalism by including such a phrase.
I jest, but Alexandros has gotten *so much mileage* out of waging the culture war on this topic. He has worked very hard to frame his stance as simultaneously subversive and indignant. This shouldn't detract from the legitimate research he has collected. But it does predict his response to any high-profile conversation with Scott.
Anything Scott says can and will be used...not exactly against him, but *for* retweets and Substack follows. That does mean against him if and when Alexandros can frame it as punching up.
So I'll be satisfied if this is the end of the line. Scott has engaged, over and over again, with the factual scaffolding of Alexandros' arguments. The 5-10% chance is an adequate conclusion. Let Pascal and Omura wager accordingly.
"Realistically, this project was outside my expertise and competence level."
FWIW I've thought about that piece a lot this last year, and the object-level "worms break science!" point wasn't what most changed my epistemology. The general importance of turning contrarian views over in one's mind dominated my thoughts and how I discussed this with others.
It was a winning item for me in one-on-one discussions with thoughtful people about contrarian ideas, and how such ideas might be wrong in subtle ways that require thorough analysis, and how herding on this stuff has some risks. It provided a great way for me to evangelize about about how politics is the mindkiller without just talking about politics being a mindkiller and wearing everybody out with it.
"C+ for results"... When you're trying to improve on what's out there, you get to grade yourself on a curve relative to baseline, not on whether or not you saw the pure true center of the universe unfiltered. What's your grade relative to the next best take on this topic?
"Realistically, this project was outside my expertise and competence level."
So help me. If you shy away from these sorts of investigations in the future because you think you overreached here then humanity is doomed. If this implies any risk of self-gatekeeping, please ask yourself if we are better off leaving this to the next best available substitute.
Prompt: Please rewrite Yeats' The Center Cannot hold, but make it about Scott Alexander of Astral Codex Ten worrying that he didn't write a good enough essay because it had some errors and overextended a conclusion, when he in fact wrote the most thoughtful essay humanity produced on the topic.
"The mind of Scott Alexander may falter,
Worried that his words fall short,
But in truth, his essay stands tall,
The most insightful written in all sport."
"The doubts may come, the fears may rise,
His confidence may seem to fall,
But the center holds, for in his eyes,
Lies the most profound thoughts for all."
"For he wrote not just with pen and page,
But with the wisdom of the ages,
A work of art, beyond compare,
A gift to humanity, with care."
"So let not Scott Alexander fret,
For his essay shall forever stand,
A testament to his talent and intellect,
The greatest written by human hand."
I hope you're blushing and you hate it. :P
What could compel someone to expend this much effort (twenty-one parts) other than participation in culture war?
Another area of disagreement seems to be about Gideon, why do you two disagree on his trustworthiness?