I’ve been meaning to write this post for months now, but have hesitated. Up until yesterday, that hesitation has always led to my abandoning the post and choosing a different topic. The reason for my hesitation is that the topic of this post happens to be a scientist whom, before the COVID-19 pandemic hit, I had long admired and whom, as a result of his publications, statements, and activities during the pandemic, I no longer do. I am, of course, referring to John Ioannidis, who first made a splash over 15 years ago, when he published what remains his most famous article, “Why Most Published Research Findings Are False”, leading me (and others) to point out how cranks and quacks have misused and abused Ioannidis’ work to “prove” that science is so unreliable that their quackery or antivaccine pseudoscience should be taken seriously. As you’ve probably already guessed, it was all nonsense. Over the years, Prof. Ioannidis’ work has inspired posts by nearly all of the bloggers who’ve been regulars here at SBM, including topics discussed in his papers such as:
- How popularity of a scientific topic might lead to more unreliable research
- The life cycle of translational research and how long it takes to go from idea in cell culture or a test tube to an actual treatment
- How often seemingly important findings in science don’t hold up in subsequent studies
- How everything we eat does and doesn’t cause cancer
Don’t get me wrong here. I don’t always agree with Ioannidis; for instance, I think he did exaggerate how often research is “wrong” and, in addition, took major issue with his argument that the NIH is so conservative that only the very “safest” projects are funded and that the “brave maverick scientists” who see “bolts out of the blue” to make great leaps in science tend not to be NIH funded. I’ll return to this particular paper at the end of the post, because now, in retrospect, I see it as a harbinger of Ioannidis’ activities during the COVID-19 pandemic that I missed at the time. Had I been more attuned now to what I had noticed then, it might have led me to be less surprised by Ioannidis’ behavior, in which he’s consistently downplayed the deadliness of COVID-19 and, in doing so, engaged in the same sort of scientific sloppiness that he had become known for skewering in other investigators. I’ll enumerate in more depth in a moment, but before I do please indulge me as I do a little exercise.
A little exercise, for perspective on COVID-19
I realize it’s hard to do now that, as of yesterday, the estimated toll in the US alone from COVID-19 is over 560,000 dead from close to 31 million cases; worldwide it’s nearly 2.8 million deaths out of 128 million cases, with all of those numbers almost certainly being significant undercounts of the true toll. Still, try to envision the situation a year ago. At that time, COVID-19 had only recently arrived in the US, the death toll was still “only” approximately 36,000 worldwide, and by the middle of March in the US there had only been 68 dead due to the disease. Try to envision the world many months before the peak of the latest surge in January, at which time as many as 4,000 people were dying each day in the US of COVID-19.
Before I take you back to those days, there is one useful exercise I like to do to provide perspective. Not unlike the one I did with Marty Makary’s claim that medical errors are the third leading cause of death in the US in order to show the extreme implausibility of his estimate, it’s just a quick check based on simple math and knowledge of basic statistics alone. (Unsurprisingly, Makary of late has joined the COVID-19 minimization crew, specifically claiming that we will have herd immunity by April, which, last check for me, is only three days away, with my own state of Michigan appearing to be in the middle of a new surge driven by more transmissible variants of the virus.) The entire population of the US is approximately 330 million people. That means that in a little over a year COVID-19 has killed approximately 0.17% of the entire population of the US, or one out of every 589 people. That’s an enormous toll. To give you even more perspective, let’s look at the five leading causes of death in the US every year. At number one, heart disease claims approximately 660,000 Americans a year, while cancer claims 600,000; accidents, 173,000; respiratory disease, 157,000; and strokes, 150,000. COVID-19 was easily the third leading cause of death in 2020 in the US, despite not exacting a large toll until three months into the year. It wasn’t even close.
Next, just as an intellectual exercise, assume that every single American has been infected with COVID-19. If that were true, then that would mean that the “hard” lower bound of the infection fatality rate (IFR) for SARS-CoV-2, the virus that causes the disease, is currently approximately 0.17%. (IFR is the fatality rate of all infections, including asymptomatic infections. The case fatality rate, or CFR, represents the fatality rate among people diagnosed with a disease.) Of course, nowhere near 100% of Americans have been infected. Now, if you accept as accurate published estimates that, as of yesterday, 31 million Americans have been infected with COVID-19 (again, it’s almost certainly a massive undercount), that would lead to an estimate of 1.8% as the upper bound of the IFR for COVID-19 in the US. Yes, that is a greater than ten-fold range, and the IFR is certainly much lower than 1.8%, but, again, we definitely have an absolute lower bound for the IFR of 0.17%, which is only continuing to grow, as we are still losing roughly 1,000 people a day to the virus. Thus the IFR is certainly at least double, if not triple 0.17% or more. And, yes, I know that IFR can change as the pandemic progresses. People who might have survived when hospitals were not overrun die when they are, for instance. However, simple statistics can show that claims that COVID-19 is “no more deadly than the flu” are utter nonsense.
So why did I just go through this exercise? Hang on, and you’ll see.
Enter John Ioannidis at the dawn of the pandemic
So why have I, after having procrastinated so much over writing this post, finally pulled the trigger and written it? It came in the context of a recent Twitter discussion about “silencing” of academics who advocate contrarian views about the pandemic and a new paper by Prof. Ioannidis published on Friday. An example of one thread that I saw last week follows:
2/n You know the arguments I mean. They usually go something like "[x famous academic] is being horribly silenced/faced the modern inquisition!"
And they usually come off as, well, nonsense
— Health Nerd (@GidMK) March 26, 2021
8/n I mean, Professor Ioannidis is, by all accounts, a lovely guy and a very dedicated researcher
But he's also enormously famous, and clearly in no danger of being "silenced" in any way any time soon
— Health Nerd (@GidMK) March 26, 2021
Which led a friend of the blog, who wrote an excellent book about critical thinking in medicine (reviewed by Harriet Hall) to bring up an example:
Two screenshots below:
The first from the stat news editorial defending JPAI.
The second a quote from JPAI.
Draw your own conclusions. pic.twitter.com/2RwHQV50Pa
— Jonathan Howard (@JHowardBrainMD) March 26, 2021
And, a couple of weeks before:
No. But you are defending a guy who did. Had we taken it seriously last year, maybe we could have tamed it. But so many were mislead into thinking it was no big deal. Those responsible for convincing the public it was a mild threat have no right to complain about lockdowns now. pic.twitter.com/o8nREq8KXa
— Jonathan Howard (@JHowardBrainMD) March 10, 2021
Which brings us back to John Ioannidis. Let’s go through a bit of background first.
Back in April 2020, Ioannidis was co-author of a much-criticized study, COVID-19 Antibody Seroprevalence in Santa Clara County, California. The study was originally published on the pre-print server medRxiv, which by then had become the go-to outlet to publish COVID-19 research before it had been fully peer-reviewed and made it into even the E-pub-ahead-of-print sections of journals. The reason at the time for using preprint servers, which have not been as prominent a year into the pandemic as they were then, was that SARS-CoV-2 and the pandemic were so new and the science was moving so fast that waiting for new papers to go through peer review before they saw the light of day seemed way too slow for the situation.
Ironically, this brings up a point and prediction made by Ioannidis over a year ago with which I most heartily agreed (and continue to agree), namely that governments and public health officials were making decisions without good data. There is no doubt that this was true then (and, to a lesser extent, remains true now). Even a year later, after a tsunami of COVID-19 studies, the science remains more confusing than we would like. Ioannidis also predicted a “once-in-a-century” data fiasco. Even though that was a pretty easy prediction in the early days of the pandemic, you have to give it to the man. He was right. It’s just ironic that Prof. Ioannidis ended up substantially contributing to the data fiasco.
It started when Ioannidis wrote:
If we assume that case fatality rate among individuals infected by SARS-CoV-2 is 0.3% in the general population — a mid-range guess from my Diamond Princess analysis — and that 1% of the U.S. population gets infected (about 3.3 million people), this would translate to about 10,000 deaths. This sounds like a huge number, but it is buried within the noise of the estimate of deaths from “influenza-like illness.” If we had not known about a new virus out there, and had not checked individuals with PCR tests, the number of total deaths due to “influenza-like illness” would not seem unusual this year. At most, we might have casually noted that flu this season seems to be a bit worse than average. The media coverage would have been less than for an NBA game between the two most indifferent teams.
Some worry that the 68 deaths from Covid-19 in the U.S. as of March 16 will increase exponentially to 680, 6,800, 68,000, 680,000 … along with similar catastrophic patterns around the globe. Is that a realistic scenario, or bad science fiction? How can we tell at what point such a curve might stop?
The wag in me can’t help but note that, as of today, the figure of 680,000 deaths used by Prof. Ioannidis as an appeal to ridicule of overblown warnings about COVID-19 death tolls is currently easily the closest to the death toll that we are, unfortunately, likely to see in the US before the pandemic is finally over. Hindsight aside, though, Ioannidis did have a point. If a disease spreads exponentially through an immunologically naïve population, how do you know where it will stop? He just guessed wrong, so very wrong! Normally, there would be no shame in being wrong in such an estimate. Lots of eminent scientists underestimated or overestimated COVID-19 in those very early days of the pandemic. In the fog of uncertainty over a new virus and the potential extent of its spread, coupled with how viruses can spread exponentially, it was very easy for even the best to be off by a factor of ten or even 50. It’s what Ioannidis did subsequently that became the problem. This began a month later, when Ioannidis and colleagues published his most (in)famous COVID-19 study.
Ioannidis’ study (which wasn’t published in The International Journal of Epidemiology until a month ago) examined the prevalence of antibodies to SARS-CoV-2 virus and estimated that between 2.5% and 4.2% of the residents of Santa Clara County had been infected with coronavirus as of early April 2020. That estimate was drastically larger, 50- to 85-fold larger, than had been estimated up to that point using swab testing. Of course, at the time there was a severe shortage of COVID-19 tests, which meant that the true prevalence of antibodies to the virus was likely considerably higher than the official counts at the time, but as much as 85-fold higher? Using this seroprevalence, Ioannidis and his colleagues concluded that the IFR of COVID-19 was around 0.17% (interval: 0.12%-0.2%), only somewhat higher than seasonal influenza at 0.1%. Predictably, this helped feed the narrative at the time that “COVID-19 is not much worse than the flu”, a narrative that still continues to this day.
Even at the time, even in the context of the extreme uncertainty of the early months of the pandemic, that figure seemed implausible, as Geoffrey Barber of WIRED noted not long after the study had been published and hit the national news, using mathematical reasoning similar to that of my little exercise above:
Skeptics have noted that the conclusions seem at odds with some basic math. In New York City, where more than 10,000 people, or about 0.1 percent of the population, have already died from Covid-19, this estimated fatality rate would mean nearly everyone in the city has already been infected. That’s unlikely, since the number of new cases, and deaths, is still mounting, fast. Others pointed to the Stanford group’s unusual use of Facebook recruitment, which may have drawn in people who were sick in February or March and couldn’t get a swab test to confirm it—the situation for a lot of people in Santa Clara County, an early Covid-19 hot zone where tests were initially scarce. That could have led to an oversampling of people who had antibodies to the virus. Others, noting that only 50 out of the 3,330 people tested, or 1.5 percent, actually tested positive, quibbled with the methods used to weight the sample, which skewed heavily white and female.
Others harshly criticized the study’s methodology, pointing out that, given the false positive rate of the then-existing antibody tests and how Ioannidis and crew had not taken the false positive rate into account properly, the study’s results could just as easily been consistent with close to zero infections in Santa Clara County during the time period of the study. Basically, critics pointed out that, due to the low prevalence of COVID-19 in April 2020, the small number of individuals included in the study, and what was known about the specificity of the COVID-19 antibody test, Ioannidis and his co-investigators could not rule out the possibility that the positive test results they got could all have been false positives.
It gets worse:
When the Stanford team — Drs. Jayanta Bhattacharya, John Ioannidis and Eran Bendavid — released the first draft of their Santa Clara County-based preprint, the news was stunning. The nation’s first study of its type, it found that the virus was astoundingly 50 to 85 times more prevalent than presumed. But that meant the death rate was far lower.
Yet the project raised eyebrows from the start.
Even before they started collecting data, the researchers openly questioned “stay at home” orders. Ioannidis wrote a provocative article arguing that if COVID-19 is less deadly, widespread restrictions “may be totally irrational.” A Wall Street Journal editorial by Bhattacharya and Bendavid was entitled “Is the Coronavirus as Deadly as They Say?” Bhattacharya revisited that theme in the Hoover Institution and Fox Nation program “Questioning Conventional Wisdom.”
When their preprint was published, its conclusions backed the trio’s policy arguments – and it was saddled with statistical problems.
It failed to describe key calculations and made at least five material mistakes, according to Will Fithian, assistant professor in UC Berkeley’s Department of Statistics. The population-weighted intervals in a table were miscalculated. The authors plugged the wrong interval into a formula. They made two math errors in executing that formula. And, misreading their test kit’s manufacturer insert, they used the wrong numbers for the antibody test’s specificity.
Oh, did I forget to mention that one of the future authors of the Great Barrington Declaration was a colleague of Ioannidis and co-author of the study? (More on that later.) True, Ioannidis and company did revise and republish the study in late April, but the revised study still estimated that the prevalence of COVID-19 in Santa Clara County at the low end of the original range, which was plenty high. Worse, in the interim, Ioannidis had shown up on Fox News, CNN, and a number of other media outlets promoting his team’s results and using them to cast doubt on the effectiveness of public health interventions and “lockdowns” to slow the spread of the virus.
Then there was this story in Buzzfeed News in July:
Stanford University scientist John Ioannidis has declared in study after study that the coronavirus is not that big of a threat, emboldening opponents of economic shutdowns — and infuriating critics who see fundamental errors in his work.
But even before the epidemiologist had any of that data in hand, he and an elite group of scientists tried to convince President Donald Trump that locking down the country would be the real danger.
In late March, as COVID-19 cases overran hospitals overseas, Ioannidis tried to organize a meeting at the White House where he and a small band of colleagues would caution the president against “shutting down the country for [a] very long time and jeopardizing so many lives in doing this,” according to a statement Ioannidis submitted on the group’s behalf. Their goal, the statement said, was “to both save more lives and avoid serious damage to the US economy using the most reliable data.”
Although the meeting did not happen, Ioannidis believed their message had reached the right people. Within a day of him sending it to the White House, Trump announced that he wanted the country reopened by Easter. “I think our ideas have inflitrated [sic] the White House regardless,” Ioannidis told his collaborators on March 28, in one of dozens of emails that BuzzFeed News obtained through public records requests.
As an aside, I can’t help but point out that, ironically, one member of this group was Dr. David Katz, an “integrative medicine” doctor formerly at Yale whose name should be familiar to longtime readers of this blog for his advocacy of a “more fluid concept of evidence than many of us have imbibed from our medical educations” that he uses to justify using homeopathy.
Dr. Katz aside, early in the pandemic, Prof. Ioannidis published shoddy research on COVID-19 that had raised the suspicion of having been politically motivated, or at the very least Prof. Ioannidis gave the impression of not having been as rigorous in his science as he had long demanded of other scientists. Then he got even worse and started to repeat COVID-19 conspiracy theories, such as the one that claims that COVID-19 death tolls are inflated by inappropriately blaming deaths on COVID-19 that were really due to comorbidities:
He claims- without evidence- that doctors are inflating covid deaths, filling out death certificates wrong. This is one of the most upsetting conspiracies about the pandemic. pic.twitter.com/6gcondnhMU
— Jonathan Howard (@JHowardBrainMD) March 28, 2021
This is what I like to refer to as the “6% gambit,” as the most extreme version of this conspiracy theory is that only 6% of the deaths attributed to COVID-19 were really due to COVID-19. It’s a conspiracy theory that relies on the person believing it having no clue how death certificates are normally filled out.
Why did Prof. Ioannidis consistently minimize COVID-19? Perhaps it was due to hubris, but it could also (or alternatively) been due to his political blindspots. Even so, I’d still say that there is no real shame in being wrong, if only one admits it frankly and tries to correct oneself. Unfortunately, with Prof. Ioannidis, that didn’t happen. Quite the opposite, in fact, as you will see.
“I never said that I knew that the death toll was going to be 10,000”
If there’s one characteristic Prof. Ioannidis has demonstrated during the pandemic, it’s that, whenever his COVID-19 prognostications are criticized or attacked, he tends to deny that he had made them and/or double down on similar arguments. For example, Dr. Howard (mentioned above), loves to compare these two statements:
Please. Ioannidis has underestimated Covid every step of the way and when his errors were pointed out he denied his own words. The scientific American authors similarly distorted his own words. 1/2 pic.twitter.com/XkhzHxJ17W
— Jonathan Howard (@JHowardBrainMD) March 24, 2021
This comes from an interview that Prof. Ioannidis did in Medscape in July, by which point there had been 132,000 deaths in the US:
You’ve criticized models for using faulty data in projecting the death toll. When the lockdown started there were only 60 deaths in the US. You projected 10,000 deaths using an IFR computed from infected passengers on Diamond princess cruise ship. Yet today there more than 132,000 deaths — the figure would likely have been even higher were it not for the social distancing/ lockdown we employed on March 16. Though the mortality numbers are still much lower than the doomsday predictions of Imperial college, they do make your projections overly optimistic.
I never said that I knew that the death toll was going to be “10,000 deaths in the US”. How could I, in a piece where the message was “we don’t know”! The 10,000 deaths in the US projection was meant to be in the most optimistic range of the spectrum and in the same piece I also described the most pessimistic end of the spectrum, 40 million deaths. The point I wanted to emphasize was the huge uncertainty.
Not exactly. First, Prof. Ioannidis was mixing apples and oranges here. His prediction of 10,000 deaths was for the US alone, while that 40 million estimate was for the entire world under a worst case scenario. More importantly, any fair reading of Prof. Ioannidis’ previous article should make it clear that he clearly favored the lowest end of his estimates. In fact, let’s look at his “worst case scenario” estimate in context:
In the most pessimistic scenario, which I do not espouse, if the new coronavirus infects 60% of the global population and 1% of the infected people die, that will translate into more than 40 million deaths globally, matching the 1918 influenza pandemic.
The vast majority of this hecatomb would be people with limited life expectancies. That’s in contrast to 1918, when many young people died.
One can only hope that, much like in 1918, life will continue. Conversely, with lockdowns of months, if not years, life largely stops, short-term and long-term consequences are entirely unknown, and billions, not just millions, of lives may be eventually at stake.
If we decide to jump off the cliff, we need some data to inform us about the rationale of such an action and the chances of landing somewhere safe.
Notice how he explicitly stated that he did not espouse that estimate. Also notice how he immediately pivoted to—you guessed it!—another comparison to the flu, emphasizing how COVID-19 tends to kill the old, not the young, and speculating about the “costs” of “lockdowns”. I also can’t help but note that we already have nearly three million deaths out of nearly 130 million cases worldwide. That’s a death rate of well over 1%. Even if the IFR is under 1%, we’d better hope that nowhere near 50-60% of the global population becomes infected.
Unfortunately, even now, in March 2021, Prof. Ioannidis appears to be sticking to his guns. He just published a paper arguing that the IFR from COVID-19 is actually 0.15%. Worse, he’s used the opportunity to settle scores with—of all people—a graduate student who has a large following on Twitter and has been highly critical of Prof. Ioannidis’ studies.
As a professor of biology named Carl Bergstrom noted:
In his latest paper about COVID infection fatality rates, John Ioannidis does not address the critiques from @GidMK, but instead engages in the most egregious gatekeeping that I have ever seen in a scientific paper. pic.twitter.com/P08sFIovD6
— Carl T. Bergstrom (@CT_Bergstrom) March 28, 2021
The condescension and hypocrisy here is mind-boggling. pic.twitter.com/ppaLt4nt9T
— Carl T. Bergstrom (@CT_Bergstrom) March 28, 2021
Anyway, you can read it for yourself. It's published in the journal for which Ioannidis previously served as Editor in Chief. https://t.co/oN8wFaKCbm
Therein John claims the IFR for COVID is 0.15%.
By official counts, 0.166% of the US population has already died of COVID. pic.twitter.com/NhEI15BhhI
— Carl T. Bergstrom (@CT_Bergstrom) March 28, 2021
Prof. Ioannidis’ study was published in The European Journal of Clinical Investigation on Friday, and, to be honest, strikes me more as score-settling than a dispassionate review article. In particular, Prof. Ioannidis seems preoccupied with settling a score with Gideon Meyerowitz-Katz, who in December published with Lea Merone in The International Journal of Infectious Diseases a systematic review and meta-analysis of COVID-19 IFRs that estimated an overall IFR of 0.68% (95% confidence interval 0.53%–0.82%), with a range from 0.17% to 1.7%. From my reading, it was a solid meta-analysis and review.
This is how Prof. Ioannidis responded in his new paper. I’m quoting such a long passage because, quite simply, I could not believe my eyes as I read this. This long attack is in an actual scientific paper, and it took up more than a page of the submitted manuscript:
The evaluations by ICCRT4 and Meyerowitz-Katz1 have multiple flaws as well as eligibility, design, and analytical choices that consistently lead to higher IFR estimates. This raises questions of technical competence and/or bias.
In multiple main media interviews and quotes Meyerowitz-Katz is presented professionally as an “epidemiologist”, but apparently he has not received yet a PhD degree as of this writing and he is still a student at the University of Woolongong in Australia. Neither he nor his co-author of the evaluation (apparently another PhD student) had published any peer-reviewed systematic review or meta-analysis on any topic prior to the pandemic. By the end of 2019, Meyerowitz-Katz had published 2 PubMed-indexed papers (both on diabetes) that had received 2 citations and 1 selfcitation in Scopus. Meyerowitz-Katz is very active also in Twitter through an account called Health Nerd (56,800 tweets as of January 19, 2021). The Twitter account has interesting, smart content with strong advocacy, often supporting worthy causes. The same account has also been generating tweetorial content reviewing various COVID-19 papers, including many critical/highly negative comments on my papers, e.g. on the IFR evaluation.5 For fairness, readers may consult these Twitter criticisms of Meyerowitz-Katz and of another prolific Twitter critic with highly similar views as Meyerowitz-Katz (Atomsk’s Sanakan [64,200 tweets as of January 19, 2021], self-described as “Christian; Science, Denialism Debunked, Philosophy, Manga, Death Metal, Pokémon, Immunology FTW; Fan of Bradford Hill + Richard Joyce”, also supporting several worthy causes, e.g. debunking denialism). The tweetorials have been posted in Pubpeer (https://pubpeer.com/publications/C2A5DD4ED8B5A0B13F63A47FEC143A). Comparison against the present manuscript may hopefully help knowledgeable readers generate an informed opinion as to the merits of arguments raised. I don’t have a personal Twitter account, but was alerted to the negative tweetorials by Meyerowitz-Katz several months ago. At that time, the name of the Twitter account owner was not obviously visible (the photo showed an unrecognizable figure with big glasses and a cat), but Meyerowitz-Katz seemed to use the Twitter account prolifically to promote his own work and criticize work contradicting his work. The identity of the Health Nerd Twitter account has become transparent now, since the owner has added a photo of him (wearing a T-shirt that writes “Trust me, I am an epidemiologist”). The identity of the reverberating Atomsk’s Sanakan Twitter account is still unclear (to me at least) and its relationship to Meyerowitz-Katz, if any, is unknown.
Overall, one potential explanation is that the flaws of the Meyerowitz-Katz evaluation may simply reflect lack of experience and technical expertise of otherwise well-intentioned and smart authors with a heightened sense of advocacy during a serious pandemic that represents undoubtedly a major crisis. It is well-known that most published systematic reviews and meta-analyses in the literature have substantial flaws anyhow. For students performing their first evidence synthesis ever, choosing a topic that requires advanced expertise due to unusual cross-design features, difficult methodological challenges and convoluted and often erratic data, a highly-flawed final product should not be surprising. Perusal of the voluminous Twitter comments of Health Nerd similarly demonstrates immediately the wonderful enthusiasm, but also lack of adequate expertise required to conduct such analyses in any rigorous way. Nevertheless, it is worrisome that trustworthy media like Scientific American and The Guardian have espoused Meyerowitz-Katz’s views and serious organizations may guide their planning based on a flawed paper. Meyerowitz-Katz is a columnist also at the American Council on Science and Health (https://www.acsh.org/profile/gideon-meyerowitz-katz), a pro-industry advocacy group. He reports no conflicts of interest.
I will note here that I’ve criticized ACSH on multiple occasions as being industry-associated astroturf, but this last attack by Prof. Ioannidis goes far beyond the pale. Yes, Meyerowitz-Katz is indeed listed on the ACSH website, but he also only wrote one article for the organization (on Red Bull and energy drinks, of all things), and that article is over three years old. This has to be the most egregious ad hominem that I’ve ever seen, and particularly odd given that industry or business bias would tend to lead one to minimize COVID-19, the better to attack public health interventions that interfere with re-opening businesses. As for the rest, I’ll let Prof. Ioannidis’ own words speak for themselves and wonder how such a prolonged attack on a graduate student by an eminent professor ever got through peer-review. Indeed, I’ve never seen such an attack in a peer-reviewed journal article ever, and I’m not alone:
My jaw is on the floor. I’ve never seen this kind of ad hominem in an academic paper ever, and it’s from a scientist who has so disgraced himself in this pandemic—and in this very paper with mathematically impossible claims—that the hubris is all the more mind-blowing. https://t.co/zGJISOyMeA
— Tara Haelle (@tarahaelle) March 28, 2021
I've seen a lot of "good ol' boys" BS in scientific publishing, but this may be the worst. Not only a bad paper, but it completely punches down at grad students who previously criticized the work. This never would have been published if it didn't have a Big Name attached. https://t.co/wgVBZZgHll
— Dr. Tara C. Smith (@aetiology) March 28, 2021
I have only just started reviewing papers.
But if I saw this this, I would advise the author(s) to immediately remove it. Ad hominem and gatekeeping have no place in scientific journals, regardless of how ‘famous’ the author is. https://t.co/kNcEBk5eMI
— Alastair McAlpine, MD (@AlastairMcA30) March 28, 2021
Then it was pointed out to me. Prof. Ioannidis was Editor-in-Chief of The European Journal of Clinical Investigation from 2010-2019. Moreover, unlike most articles of this type, Prof. Ioannidis is the sole author.
One of the targets of Prof. Ioannidis’ ire spent an entire Tweetorial dissecting the distortions in the article. It’s long, but I find its arguments compelling. To summarize, Atomsks Sanakan points out (and documents) that Prof. Ioannidis has a history of cherry picking studies with non-representative samples because studies using representative samples tend to result in an estimate of an IFR higher than his and incompatible with his message. Reading Prof. Ioannidis’ screed disguised as a review article, I couldn’t help but be frustrated at Prof. Ioannidis’ relative lack of detail in the methodology, in which, after a discussion of the search strategy used to locate review articles and a description of the data extracted from each one, he says simply, “Based on the above, the eligible evaluations were compared against each other with focus on features that may lead to bias and trying to decipher the direction of each bias.” What? What were the criteria to identify “features that may lead to bias”? Whenever that is done in studies, generally more than one person are required to evaluate for bias.
Prof. Ioannidis’ attack is rather ironic given that he himself repeatedly accuses his critics, in particular Meyerowitz-Katz, of “cherry picking” data, a phrase that I’ve never seen in an academic paper before, just as I had never seen a whole page in a manuscript devoted to attacking a graduate student and his Twitter account, while darkly insinuating undisclosed conflicts of interest. I also found this line of attack pretty ironic given that the question of undisclosed conflicts of interest has been credibly raised with respect to Prof. Ioannidis himself. Moreover, the level of rigor in this review article would, if it had been written by someone else, likely have earned Prof. Ioannidis’ justified ire.
The reaction to Prof. Ioannidis’ personal attacks on a graduate student, to his “punching down”, have been swift and intense:
Stanford Prof John Ioannidis responded to criticism of his fringe Covid views by trying to discredit a grad student in a scientific journal. So damaging to science as a whole & learning environment of Stanford in particular.@GidMK himself put it best.https://t.co/I6UrCmxt3Y https://t.co/K1Yl6QX3wO pic.twitter.com/bQWVXxppMj
— Mallory Harris (@malar0ne) March 28, 2021
— Greg Dore (@GregDore2) March 28, 2021
If you’re one of the world’s most famous professors, as Ioannidis is, in a tenured position at Stanford, & you are literally attacking the *physical appearance* of a graduate student, it’s probably a sign that the merits of your scientific arguments are just weak sauce. https://t.co/QIOKXrB6vH
— Prof. Gavin Yamey MD MPH (@GYamey) March 28, 2021
“But peer review is awesome—peer-reviewed journals make sure papers are rigorous & of the highest standards”
Um, in a new paper in a peer-reviewed journal, a world famous professor attacks a graduate student for not having a PhD, wearing big glasses & having a cat.
— Prof. Gavin Yamey MD MPH (@GYamey) March 28, 2021
This is one step above calling him a dumb dumb stupid head who smells bad.
— Jonathan Howard (@JHowardBrainMD) March 28, 2021
Peak desperation: Tenured Stanford professor defends his point by insulting & humiliating a grad student. https://t.co/rAffL4yNm8
— Amy Maxmen, PhD (@amymaxmen) March 28, 2021
Yeah, it is stunning. Particularly ironic to see the Ionnaidis supporters, who previously called for civility, continue to support him despite these inexcusable attacks
— Dan Freedman, DO (@dfreedman7) March 28, 2021
That last one really got me, given how defenders of Prof. Ioannidis decried criticism of his writings, studies, and appearances in the media as “public shaming” (Jeanne Lenzer and Shannon Brownlee, who also decried the “COVID science wars”, portraying Ioannidis as being “silenced”) and the intensity of the attacks as turning scientific disagreements into politics, as Dr. Vinay Prasad did. (Remember him?) Oddly enough, I have yet to see any of Prof. Ioannidis’ defenders take him to task for his own lack of civility in “punching down” on a graduate student.
The bottom line is that it’s not a good look for a world-famous scientist who publishes for the World Health Organization, has over a thousand publications listed in PubMed, and regularly shows up on cable news networks as an expert, to launch such a personal screed against a graduate student. Truly, Meyerowitz-Katz’s criticisms of his work must have gotten under Ioannidis’ skin something awful. The whole article reeks of, “How dare a graduate student with a Twitter account question me?”
It makes me wonder how he might react if this post were to come to his attention and annoy him sufficiently. Given that I have maybe 1/20 the number of publications that he does and have had much, much less of an effect on science than he has, I’m almost certainly beneath his notice. On the other hand, Meyerowitz-Katz is a graduate student and had only published one research paper and a whole bunch of Tweets criticizing Prof. Ioannidis’ work, and that clearly got under his skin to the point where Ioannidis used his pull at a journal to publish a hit piece on Meyerowitz-Katz.
John Ioannidis: Brave maverick
As much as I used to admire him, since the pandemic hit John Ioannidis has consistently disappointed me to an extreme degree. In the last year, my disappointment with Prof. Ioannidis has gotten to the point where it’s hard for me to avoid lumping him with the COVID-19 minimizers/deniers like those who published and continue to promote the Great Barrington Declaration, one of whom was his co-author on his infamous Santa Clara seroprevalence study. The Great Barrington Declaration, boiled down to its essence, asserted that COVID-19 is not dangerous to the vast majority of the population, leading its writers and signatories to conclude that governments should, in essence, let SARS-CoV-2, the coronavirus that causes the disease, run rampant through the population in order to achieve “natural herd immunity”, while putting in place measures designed to protect only those viewed as “at risk”, such as the elderly and those with significant co-morbidities. (Note that, at the time the Declaration was published, there was as yet no safe and effective vaccine against COVID-19, while now there are at least four.) Of course, as many noted, it is not possible to protect the vulnerable if COVID-19 is rampaging unchecked throughout the rest of the population. Also, as I noted when I wrote about it, the Great Barrington Declaration was the product of the American Institute for Economic Research, a right-wing, climate science-denying think tank, which recruited three ideologically—shall we say?—amenable scientists to sign on as authors of the declaration, which was basically, as I put it, “eugenics-adjacent” and full of misinformation and half-truths.
Moreover, I’m not the only one who’s now soured on Prof. Ioannidis. For example, Scientific American columnist John Horgan, someone with whom both Steve Novella and I have had disagreements based on his downplaying of skepticism in medicine with respect to homeopathy:
Optimism has also distorted my view of the coronavirus. Last March, I took heart from warnings by Stanford epidemiologist John Ioannidis that we might be overestimating the deadliness of the virus and hence overreacting to it. He predicted that the U.S. death toll might reach only 10,000 people, lower than the average annual toll of seasonal flu. I wanted Ioannidis to be right, and his analysis seemed plausible to me, but his prediction turned out to be wrong by more than an order of magnitude.
Horgan didn’t go quite far enough in his criticisms for my taste, but such is life.
Then there’s Alex Rubinsteyn:
“What a weird turn to see John Ioannidis pushing one of sloppiest studies in the deluge of Covid-19 papers,” Alex Rubinsteyn, an assistant professor of computational medicine and genetics at the University of North Carolina School of Medicine, wrote on Twitter. “If he weren’t an author I would expect [the study] to show up in one of his talks as a particularly potent cocktail of bad research practices.”
Then, of course, there are all the scientists on Twitter criticizing Prof. Ioannidis. In fairness, one has to acknowledge that there are things Prof. Ioannidis has argued that have some merit. His estimates of IFR were closer to the mark than some of the very high estimates early in the pandemic, but they were still off considerably in the other direction. He was not wrong about the poor quality of so much of the data and research on COVID-19; it’s just, in an amazing feat of lacking self-awareness, he himself contributed to it as well.
This brings me back to that discussion of Ioannidis’ paper claiming that the NIH is too conservative and that only conservative, “safe” science is funded. It was more than that, though. He claimed that the scientists on NIH study sections were no better than scientists not on NIH study sections. Before I get to that, though, I note that Ioannidis’ cardinal sin since the pandemic started is not to have been wrong, even repeatedly so. It’s been his extreme arrogance:
Instead, Ioannidis sounded sure of himself. He was right; the others had it wrong. He called out other research teams by name—Johns Hopkins, Imperial College London—to berate their findings as “astronomically wrong,” and “constantly dialed back to match reality.” Here he was, about to come out with an exciting and important finding—if he were right, it could change almost everything about how we deal with this virus—and he seemed unworried by the possibility that something might be amiss with the project.
If anyone should understand how the pressure to contribute to the science of the crisis might lead to flawed work and exaggerated claims, it ought to be Ioannidis, arguably the world’s most famous epidemiologist. Who knows? Perhaps like so many of us, he’s just stressed out by the whole damned thing. Maybe he’s just off his game.
The article from which this quote came dates back to May 2020. Now, eleven months later with the benefit of hindsight, I don’t think you can say that Ioannidis was “off his game”. With his attack on a graduate student, he’s continued to double down and, in fact, has even gone further than Freedman had previously described. That is what brings me back to my previous discussion of his article about those “safe” scientists at the NIH, with a funding process that he’d characterized as “conformity” and “mediocrity”. I wrote this over eight years ago:
In the end, as much as I admire Ioannidis, I think he’s off-base here. It’s not that I don’t agree that the NIH should try to find ways to fund more innovative research. However, Ioannidis’ approach to quantifying the problem seems to suffer from flaws in its very conception. In light of that, I can’t resist revisiting the discussion in my last post on the question of riskiness versus safety in research, and that’s a simple question: What’s the evidence that funding more risky research will result in better research and more treatments? We have lots of anecdotes of scientists whose ideas were later found to be validated and potentially game-changing who couldn’t get NIH funding, but how often does this really happen? As I’ve pointed out before, the vast majority of “wild” ideas are considered “wild” precisely because they are new and there is little good support for them. Once evidence accumulates to support them, they are no longer considered quite so “wild.” We know today that the scientists whose anecdotes of woe describing the depredations of the NIH were indeed onto something. How many more proposed ideas that seemed innovative at the time but ultimately went nowhere?
And my conclusion:
However, the assumption underlying Ioannidis’s analysis seems to be that there must be “bolts out of the blue” discovered by brilliant brave maverick scientists. It’s all very Randian at its heart. However, science is a collaborative enterprise, in which each scientist builds incrementally on the work of his or her predecessors. Bolts out of the blue are a good thing, but we can’t count on them, nor has anyone demonstrated that they are more likely to occur if the NIH funds “riskier research.” It’s equally likely that the end result would be a lot more dud research.
Maybe the problem with Prof. Ioannidis was there all along, and I just didn’t see it until the pandemic amplified it for all to see. He seems, dating back at least to 2012, have had the belief that conventional science is too “safe” and “conformist,” perhaps with a bit of a self-image of himself as being the “brave maverick doctor” or iconoclast. Maybe that’s why, during the pandemic, he was so easily drawn to being a “rebel” or a “contrarian,” whose findings bucked the existing consensus, and maybe that’s why he can’t give that up. After all, it’s happened to greater scientists than he. Moreover, Prof. Ioannidis seems to be an excellent cautionary tale at how being a critic doesn’t necessarily mean that you can do what’s being criticized that well. He’s very good at finding the flaws in studies, but his studies during the pandemic demonstrate that, when designing studies of his own, he’s prone to every bias and flaw that he criticizes in others.
In any event, I should go back and read some of Prof. Ioannidis’ old work in light of what I know about him now, with the realization that the pandemic has done me a favor. I wonder what I might find.