John P. A. Ioannidis
John P. A. Ioannidis is a Greek-American physician-scientist, writer and Stanford University professor. He has made contributions to evidence-based medicine, epidemiology, and clinical research. He is particularly known for studying scientific research itself in the fields of medicine and metascience.
He is C.F. Rehnborg Chair in Disease Prevention, Professor of Medicine, of Epidemiology and Population Health, and (by courtesy) of Biomedical Data Science, and of Statistics; co-Director, Meta-Research Innovation Center at Stanford (METRICS).
Born in New York City in 1965 and raised in Athens, Greece. Valedictorian (1984) at Athens College; National Award of the Greek Mathematical Society (1984); MD (top rank of medical school class) from the National University of Athens in 1990; also received DSc in biopathology from the same institution.
Trained at Harvard and Tufts (internal medicine and infectious diseases), then held positions at NIH, Johns Hopkins and Tufts. Chaired the Department of Hygiene and Epidemiology, University of Ioannina Medical School in 1999-2010 while also holding adjunct professor positions at Harvard, Tufts, and Imperial College. Senior Advisor on Knowledge Integration at NCI/NIH (2012-6).
Served as President, Society for Research Synthesis Methodology, and editorial board member of many leading journals (including PLoS Medicine, Lancet, Annals of Internal Medicine, JNCI among others) and as Editor-in-Chief of the European Journal of Clinical Investigation (2010-2019). Delivered ~600 invited and honorary lectures.
Recipient of many awards (e.g. European Award for Excellence in Clinical Science , Medal for Distinguished Service, Teachers College, Columbia University , Chanchlani Global Health Award , Epiphany Science Courage Award , Einstein fellow ). Inducted in the Association of American Physicians (2009), European Academy of Cancer Sciences (2010) American Epidemiological Society (2015), European Academy of Sciences and Arts (2015), National Academy of Medicine (2018). Honorary titles from FORTH (2014) and Ioannina (2015), honorary doctorates from Rotterdam (2015), Athens (2017), Tilburg (2019), Edinburgh (2019, ceremony moved to 2021), Aristotle U Thessaloniki (2020, ceremony pmoved to 2021).
Multiple honorary lectureships/visiting professorships (Caltech, Oxford, LSHTM, Yale, U Utah, U Conn, UC Davis, U Penn, Wash U St. Louis, NIH among others). The PLoS Medicine paper on “Why most published research findings are false” has been the most-accessed article in the history of Public Library of Science (>3 million hits).
Author of 8 literary books in Greek, three of which were shortlisted for best book of the year Anagnostis awards. Brave Thinker scientist for 2010 according to Atlantic, “may be one of the most influential scientists alive”. Highly Cited Researcher according to Thomson Reuters in both Clinical Medicine and in Social Sciences.
Citation indices: h=207, m=9 per Google Scholar. Current citation rate: >5,000 new citations per month (among the 10 scientists worldwide who are currently the most commonly cited, perhaps also the currently most-cited physician).
The original interview is 46 minutes. I reduced it to 21 minutes. Like Dr Hodkinson, Professor Ioannidis this year has been censored by YouTube, and has like Dr Hodkinson received death threats. At one point the severity of which left him fearing for the life of his elderly mother.
Transcript of Clips
Each separate clip transcribed and time-stamped below.
John Ioannidis ➝ 00:00
So we have very little evidence on interventions that are used in large scale across entire populations, entire countries, much of the world in a setting of a pandemic like COVID-19. And this is, this is really concerning.
The evidence is – the evidence gap is replaced by other types of data, lots of modeling data, and some observational data. And without wishing to dismiss this type of information, we have to be very cautious because they do have lots of limitations.
Modeling is lots of fancy attempts. And I enjoy it myself, but it’s, unfortunately, making by default, a lot of assumptions and these assumptions may or may not hold true. And therefore you just don’t know much of the time whether the inferences of these models are trustworthy.
In fact, we have seen even within the first several months of the pandemic, that the vast majority of the modeling that has been done has not really been able to get realistic predictions and inferences.
Some of the modeling has led even to very misleading inferences because people trust too much these predictions, and they act based on them as if they were very strong evidence where actually they’re very weak evidence.
Now, some models also use more data than others. Some are more data-driven and less assumption-driven, but even those that are more data-driven, these data are observational. They occur in a very unstable environment of an evolving epidemic wave. They should be seen with a lot of caution and not easily extrapolated.
Clearly realizing that we don’t have good evidence is the first step. And it takes some brave people and some brave statements to really push that point that we don’t have evidence. You need to convince others who say, “I don’t care. I just need to do something because we have a serious situation s here and forget about evidence. I’m just going to act.”
But obviously this kind of works for a very acute situation, that is an acute disaster but for a condition that is likely to last for months and years, saying that we will just follow our gut feeling and these modeling and observational data as they come, I think is a recipe for disaster.
And I think that we have a very high risk of doing a lot of harm, unless we just say we don’t have evidence, let’s get evidence that is trustworthy as quickly as possible.
John Ioannidis ➝ 03:12
Unfortunately, a lot of evidence is not shared for various reasons, there’s a huge literature on publication bias and selective reporting, on why that happens on the incentives for sharing data, that sometimes are focusing more on sharing extravagant, extreme, significant results and lots of more modest, moderate results non-significant results do not get to be disseminated.
Sometimes there’s conflicts of interest. In the case of a pandemic, you have like a complete mess of very fast moving science peer review being subverted probably by zealots sometimes who want to accept papers that fit their worldview immediately, and not accept or destroy papers that are proving that their thoughts and ideas and ideologies are wrong.
So these biases that exist, and we have very well documented them along many, many decades across biomedicine and other scientific fields, when it comes to the urgent setting of a pandemic, I think that they can really get worse.
John Ioannidis ➝ 04:22
In Covid-19, media and even more social media really took control of science to a large extent and that’s very unfortunate because these are people who are not trained most of the time to really cover science.
We’ve seen a deficit of science journalism, even in the past. There were very few journalists who knew about science. Also many scientists in a way became media agents during the pandemic.
I know of many highly respectable colleagues who became social media animals. You know, social media are not necessarily a place to have balanced scientific discussion. They have the benefit of rapidity and of some sort of unhindered expression, which is great, but at the same time it creates a situation where science is shaped and approved or disapproved by large mobs rather than careful thinking, proofs, evidence.
You have people who can tweet away more than their opponents, who can harness a larger crowd of followers, and also journalists who can really shape a given reality of evidence, which may have very little to do sometimes with with real evidence.
John Ioannidis ➝ 05:58
Yes, that editorial I wrote to express my concern about using machineries that are not part of scientific method, but part of activism and advocacy to replace the scientific method. I’ve seen many scientists who somehow in their tweets and in their open letters, in their social engagement, they say ‘many people, many scientists agree with me and my opponents, they’re just fringe people. They belong to – there’re pseudo scientists’, even worse.
And, you know, pseudo scientist may include Nobel laureates and extremely talented scientists and some of the top methodologists in the world. So it creates a very explosive mix when you get scientists behaving as activists or advocates, and collecting signatures is probably one such example.
I’m perfectly fine with collections of signatures when it comes to ethical issues like people being fired, or, you know, some ethical issue in the community that needs to be settled with mobilization of opinion.
But saying that I have one scientific theory that I will prove it because I have collected enough signatures. This is completely ridiculous. I think that Copernicus would have collected one signature and he would have had probably thousands of “scientists” in his era saying that his heliocentric model was entirely wrong. You know, Einstein in 1905 again, you would have a single vote for his new theories. And if you were to ask everyone else, they would say they hadn’t heard about it, or that would be completely wrong and they would be wrong.
John Ioannidis ➝ 08:07
The third point is the involvement of politics which makes things even worse. I have felt that science has had kind of a difficult relationship to politics. Ideally, you want science to involve, to inform society and leadership in making decisions.
This is the ideal situation when good science based on good evidence is informing the decisions of our society. And obviously these decisions will be made in most circumstances by politicians or other types of leaders.
However, you may reach the situation where you may see that politics is subverting science, that politics is is taking control of science in different ways. And what we have seen in the Covid-19 setting, we have seen that scientific positions were espoused by specific entities in the political divide.
So if you were to claim that you believe in X, you would be immediately classified as conservative or progressive, or, you know, Trump supporter or Biden supporter. And again, I think this is completely ridiculous because you know, science is not dictated by politicians, by parties. It should not be an issue of political divide.
We all struggled to find the truth, to come as close to the truth as possible, to minimize error, to minimize bias, to get rid of existing biases. First of all, we need to detect them.
So having that whole political turmoil super-imposed on science really creates a very difficult situation. And obviously then you get all the methods of politics, including lobbying, including subversion, including smearing of opponents, including fake news, including social media campaigns, including distorted journalism, that are infused into science, and that can be devastating.
November 16, 2020
“Copernicus would have collected one signature and he would have had probably thousands of ‘scientists’ in his era saying that his heliocentric model was entirely wrong.”
I think science is very delicate. If you start treating it with such doses of poisons of that sort, science is dying and what you’re left with is just angry opinionated experts at each other’s throat which is not what science should be at at any point and by any means.
John Ioannidis ➝ 10:58
We have very little evidence that mandatory or enforced restraints are doing any good. Especially in the setting of, and ensuring fatigue, I don’t have evidence for that, but I can bet that it’s more likely that they will do more harm than good.
I think that they will create even more fatigue. I think that they will probably damage even more mental health and other adverse consequences of that whole setting.
And I think that these are things that we need to study because otherwise we end up in a situation where many governments and leaders are kind of announcing new measures every week, if not every day. And just following some non reliable indicators, like number of cases, which of course depend on how many people you test, and whenever they see a spike or any increase, they feel pressured to impose even more measures as if that would do anything. Although we have no evidence that all these minute variances of aggressive or more aggressive measures really would be helpful.
So I think that in principle, I’m not in favor of legal enforcement and policing. I think that I’m willing to get pushed back on this. And I would like to see pushback generated with some reliable evidence and ideally that could include randomized trials as we discussed at the very beginning.
But I’m worried that the balance of benefits and harms, we know so much about the potential harms in terms of absolute magnitude, even though they may not be accurate. We’re talking about 50% of the population becoming infested with anxiety or depression and 25% of young people having suicidal ideation.
November 16, 2020
“However, you may reach the situation where you may see that politics is subverting science, that politics is is taking control of science in different ways.
And what we have seen in the Covid-19 setting, we have seen that scientific positions were espoused by specific entities in the political divide.”
Even if this is like a 50% off and our estimates are not accurate, you would need a tremendous benefit of these legal enforcements to even match this, let alone all the other adverse consequences. So off the start, I think that these aggressive legal enforcements seemed very problematic to me.
John Ioannidis ➝ 13:30
I think that many people do feel insulted. They feel that their freedom is restrained, that their very fundamental rights are restrained. You know, what would a dictatorship do more? I don’t think that any dictatorship has managed to really restrained life as much as some of the lockdowns have by a large margin of difference.
So I’m not a conspirator theorist or anything. I think that the people who apply these measures are well intentioned. But I don’t think that they consider the horrible perception that many people get out of these measures. Now there’s other people who might feel that not taking measures is a sign of weakness of their leaders.
So one needs to balance these two perspectives, you know, people who say, you know, please shut down everything we’re dying. And you have the counter argument here that some of these leaderships feel that they need to respond to that quest for maximum measures. But I think that this is getting to the point of some sort of psychopathology, and we know that even under a normal circumstances, a very large share of the population has phobias for pathogens.
In the current setting, where we have a virus that has been trumpeted wide and right and left as the major pandemic of the century, these phobias become very dominant in the population. And I think that they can be devastating.
John Ioannidis ➝ 15:21
So the risk to the the vast majority of people is completely negligible. There are some people who have devastating risk, of course.
John Ioannidis ➝ 15:29
I think that as I mentioned at the very beginning, lots of the decisions about Covid-19 were based on modeling studies and their track record has not been very good. Their track record actually was not very good even in the past, even before Covid-19, but somehow the implications were not so serious.
You know, most people would not really realize that models were so wrong because the consequences of the modeling would not affect their lives. They will not affect them, their choices, their options.
In the current situation modeling has changed their lives in major ways. And I think that the failures of modeling have become very obvious and the consequences can be real devastating.
We know that modeling is one useful contributor to our understanding of evidence. So I’m not saying to completely discredit it. I think that we can work towards making it more transparent, more feasible to understand how it works so that other people can test and and reproduce its machinery.
Many models are black boxes that people will not even see how they work and why they get the inferences that they get.
Giving enough room for the uncertainty that is involved in the assumptions, in the data that are fed into the model. And therefore also in the output of the models.
November 16, 2020
“what would a dictatorship do more? I don’t think that any dictatorship has managed to really restrained life as much as some of the lockdowns have by a large margin of difference.”
I think many people leaders, politicians, policy makers, but also scientists in this pandemic showed that they had no clue how unreliable modeling is. They just took these numbers at face value. They looked at the plots, they have very nice colors and they looked very scary and they just said, well, this is science. No, this is not science.
This is one weak form of science, among the weakest probably that we have.
And I think that this was not realized. This led to a lot of confusion. It led to very wrong expectations. It led to wrong decisions. For example, when people were thinking that they would get 10 times more patients coming to the hospital, they sent already infected Covid-19 patients from the hospitals to nursing homes, and then they devastated the nursing homes.
John Ioannidis ➝ 18:01
Models in Covid-19 was not just another failure of modeling. This has been a very long track record of mostly failures. It was something that led to people dying, and I think this is really horrible and we should find ways to avoid it.
Unfortunately major institutions, universities, major journals, they still trust a lot of modeling that is very unreliable. They publish lots of it probably with very sketchy peer review among other models who are just willing to promote and self-promote this type of work.
I think it’s becoming kind of a mafia. And I’m not saying this to… I’m saying it more in a laughing way rather than in a derogatory way. But I think these major journals and institutions should realize that modelers are not what they seem to be.
John Ioannidis ➝ 18:56
I think that there’s personal responsibility for everyone. I think that the personal responsibilities is a feature of human life. And when it comes to professionals in public health, of course it takes an extra dimension, because there’s a sense that they can make a difference in, especially in crisis times, but even in normal times.
I think we need to see this with some temperance, however, and avoid becoming advocates that forget about evidence and they’re just out there to save the world. I think it’s wonderful to save lives, to save the world, but the path to error is usually led by good intentions.
And I worry that many people in public health are very weak in research methodology. They’re very not well-prepared in terms of understanding evidence and especially the limitations of evidence.
Many of them have very strong beliefs. And this means that often they go into war without really knowing what they’re doing.
What we have seen in Covid-19 is also that many people who were not even involved In public health took the role of public health professionals. We saw the major newspapers and media really flooded with people who had no clue about public health, let alone evidence and research methods, who really took up to themselves to save the world. And this is really scary.
We saw also lots of scientists who either have no training in public health and methods, or have some but it’s very incomplete and very unilateral. They also took very maximalistic positions.
November 16, 2020
“And I worry that many people in public health are very weak in research methodology…
Many of them have very strong beliefs. And this means that often they go into war without really knowing what they’re doing.”
I think this is again very scary because in cases of crisis, you need to have complimentarity. You need to have multiple people from different ways of life, with different expertise, with very strong methods training that would try to give a very balanced perspective.
And of course, act, I’m not saying not to act, but act in a responsible way that would maximize the ratio of benefit to harm. I don’t think that this has happened, unfortunately in the Covid-19 crisis.
- Dr Robert Endres — Lockdowns and Masks are not Evidence-Based. Vaccine is not Fully Tested.
- PCR Inventor Kary Mullis Talks About Anthony Fauci — “he doesn’t know anything really about anything”
- Reid Sheftall MD — Did China Play Coronavirus Card to Damage Adversarial Economies?
Source: ‘Entrevista a John Ioannidis: Health Policy Based in the Lack of Evidence’, https://www.youtube.com/watch?v=x4TiD4xw6Lk (Copyright Fundación Gaspar Casal, November 16, 2020)