Just great, Zeynep. As with the Counters, we see that your vision for Insight goes beyond publishing to catalyzing a thoughtful-about-thought thought community. The principles you're talking about speak to not just a theory of knowledge, but a practice of knowledge; what the ubiquitous "how to spot misinformation" guides are groping for but fall so far short of achieving. Would love to sign up for the Zeynep 'Practice of Knowledge' forum - what are my chances?
I love "not just a theory of knowledge but a practice of knowledge." I think you've pinned one of the things that makes Zeynep's thinking so unique and valuable.
Full disclosure, I'm from MN and went to Luther College where Michael Osterholm is also from.
There is a bit of a take down on Michael which I think is appropriate (he said it and it did show western bias I thought) and unfair (he said other things at exactly the same time that were very good and actually he was one of the first experts in the USA to say it.)
Michael was skeptical as Zeynep points out, very early that unsymptomatic people would be spreading the disease.
That said, at precisely the same time he was one of the first in the US and loudest saying two important things: 1) the virus is likely airborne, "just by breathing", and 2) that super-spreaders were likely a key component of how the virus spreads. Both 1) and 2) were right on:
Jan 21st: "The thing we worry about as health officials is a thing called 'super spreading,' where we have certain individuals that are not just infectious but highly infectious."
Jan 24th: "The [novel] coronavirus is one that can be transferred quite readily by the respiratory tract—just breathing. So, it has the potential to spread quickly around the world. It also is one that a sizable portion of the population typically has severe illness and death."
Then on asymptomatic spread, this turned out to be wrong:
Jan 26th: “I seriously doubt that the Chinese public officials have any data supporting this statement,” said Michael Osterholm, director of the Center for Infectious Disease Research and Policy at the University of Minnesota. “I know of no evidence in 17 years of working with coronaviruses — SARS and MERS – where anyone has been found to be infectious during their incubation period.”
By Feb 24th Osterholm had reversed himself on his Jan 26th statement:
Feb 24th: "The coronavirus that causes COVID-19 seems to spread like influenza, through the air, person to person. Unlike Ebola, SARS and MERS, it can be transmitted by individuals before the onset of symptoms or even if they don’t become ill."
I agree that Michael could have and should have pointed out that his earlier guidance based on "all his years" of studying MERS and SARS, was wrong and arrogant Jan 26th. He should have been clear and should have been honest on the 24th that he got this piece wrong.
So in general I think Zeynep's take-down is well placed because Michael never, to my knowledge, confessed that he had it wrong on this very important topic at the beginning.
I do think it reflects badly on him that when he was right he trumpets his rightness and when he was wrong he said nothing that would let us understand he learned something. That is NOT science, that is P.R.
Hi Scott. He also kept insisting—in the summer!—that masks would produce a false sense of security and thus implying they were harmful. But you are right, he also got other things correct. But the interesting thing here is not just the lack of accountability (and I don't mind people getting things wrong: I am sure he's excellent as a doctor) but that infectious disease specialists in many other countries had none of these issues we had here. They had everything right, early on, because this wasn't that confusing or obscure. It's really something worth studying how we managed not to see what was in front of our noses *and* were being told to us by really excellent experts who had the right experience before (SARS) *and* were publishing so much!
Zeynep, I 100% agree with you. It is so worth studying. I am a data scientist, my company works on how to make strategic measures work, tying cause with effect. I have a passion for the subject you are on, absolutely love the question you are posing re: accountability, how to make that work better, so if there is anything I can do to help, please don't hesitate to shout.
Agree Zeynep. I was/am very angry at him with respect to the mask piece. You are a breath of fresh air! I love reading your stuff.
In terms of your theory of knowledge, you could Michael Osterholm as exhibit 1. A hard headed male who thinks he knows everything there is possible to know. Just ask him.
He decided early on that because health care workers were the front line workers, 100% of all n95 masks needed to be siphoned off. This was then done, even though the practice was n95 masks are disposable, meant to be disposed of sometimes 10 times a day. That premise which he set and concrete and shouted at everyone at high volume I found instructive.
The relative odds of giving vulnerable populations access to good masks was never part of his world view. What if the people most likely to end up in the ICU don't get sick to begin with?!
He then had gone on to say a whole set of cascading illogical things on this vector, only front line workers (read trained and using properly) shall have access to good quality PPE.
In his defense, I think unquestionably having a safe environement for front line health care workers had to be the top of the pyramid.
That said there were many even within health care settings who were getting much lower quality PPE than they deserved while doctors who are at the top of an abusive domination/submission pyramid were disposing of masks right and left rather than reusing in a safe way....He said little to nothing about that that I know of....
We are all flawed human beings, so we could take down Fauci too, and let's not even get started with Redfield, CDC.
The bad advice from experts was a stunning thing to behold. You have done the world a great service in my view showing the power of taking a cross disciplinary approach and in so doing, taking them all to task!...
I think this is an important article and would appreciate reading more about the theory of knowledge. I have long found it concerning that we pay more attention to what leaders say than "the facts on the ground" of what they do and don't do. Your discussion of considering the embarrassment factor surely applies to many actions in government and medicine.
One other thought with respect to the embarrassment factor- In a courtroom, one of the exceptions to the hearsay rule is "statements against interest"- ie, the notion that something said is more likely to be true if in saying it the speaker is speaking against their own interests. I think this is similar to the idea that if someone says something that is potentially embarrassing it is more likely to be true.
Same idea! Under the write conditions, what leaders say or do is quite indicative. But, as this post tries to explain, one has to think through the institutional and historical context. I did not have any insight into the medical situation in Wuhan—neither did the medical doctors or virologists making policy for this threat. In effect, though, understanding authoritarian regimes, especially China, was quite useful especially in January of 2020 when we did not have independent scientists on the ground.
I heard your interview on the Ezra Klein podcast yesterday - and was curious about your take on the "individual" versus "population" level. In my field of environmental health (and more generally public health overall) there seems to be a vast disconnect between clinicians and researchers in biomedical sciences who focus on individual outcomes (or individual mechanisms) versus those of us who work in prevention that focus on population outcomes. In fact, they are often completely dismissive -- and demand that any "benefit" or "risk" be "clinically significant" (i.e., to an individual patient) even though at the population level, a small change in benefit or risk may have large aggregate effects when multiplied by the size of the population. Masks, rapid at home tests, etc. are just the latest examples -- though perhaps their effects are magnified due to the exponential nature of infectious diseases, whereas these effects "linear" in most public health interventions (e.g., reducing sodium intake and cardiovascular diseases). Any thoughts about bridging this divide?
I wonder if there's also a difference between making the theoretical case for population-level impact of a certain intervention or practice as opposed to individual-level impact (e.g. masking), and making the *empirical* case -- the challenge of assessing and pinpointing this impact -- at each respective level, and how these are perceived.
For instance, it seems fairly intuitive and uncontroversial that while masks may individually protect the wearer and the person near them to some degree (especially with high-quality masks), their real power should come when everybody is wearing them collectively in the course of a range of daily activities given the interconnectedness and exponentiality of transmission. And yet, there are so many variables at the population level and so hard to definitely *prove* via RCT that widespread masking led to X outcome over Y -- so many other "holes" that could offset this -- that paradoxically the individual case for masks might come across as more compelling and affording greater agency ("I protect you, you protect me!"), even though it's arguably just as difficult to empirically verify a particular instance of this as for an entire country. After all, "We're all in this together" is belied by the outsized role of inequities in how people fare, and there are many examples of low correlation between a country's rate of mask-wearing and how well it is doing.
So, masks are clearly most effective as a universal societal practice - and yet, it might appear that they're not, or that it's impossible to tell. Maybe clinicians and researchers used to dealing with individual-level phenomena and direct cause-effect relationships are impatient with such ambiguity, and privilege the empirical evidence over the theoretical evidence.
This may be an odd question, but I was wondering if you could do a post about or direct me to the best resources you know of about how to become a better "systems thinker." I'm an undergrad from the US and would really appreciate any guidance--books, articles, links--that you could provide. I've found your writing invaluable throughout the pandemic and want to improve the way I think about complex problems. Thank you
I think there's a part of your metaepistemological approach that you're underestimating/leaving out: what sources you use to inform the process. From the very beginning you had a good understanding of what sources were informative. It seemed like you had them preidentified from teaching SARS, field work in Hong Kong etc.
It would be interesting to hear your process for identifying which sources to incorporate.
This is a tough one. Interestingly, though, even unreliable sources can be informative in context, the way China's denials were when later juxtaposed with their swift action. It gave us a sense that it had really, already spread widely.
It's amazing to see how many bright, flashing signals were missed even as bigger media houses started reporting on the outbreak after the lockdown started. Two days after the lockdown Wuhan officials said five million people had left the city before the lockdown started. Five Million! And millions were travelling inside and to abroad for the Chinese New Year.
After the Thailand case, on January 28 Reuters reported a tourist bus driver who took a group of tourists from Wuhan had the virus. He had not been to China.
Even after all this, even as late as March, countries like the US, UK and India were ignoring what was staring at them. Did they all think this was going to be like Sars? Or did the fact that Sars was contained fairly quickly made everyone complacent?
Yes, the SARS experience may have made countries complacent though I got the opposite lesson from that near-miss. We came so, so close to catastrophe there.
Last night I was starting a fire in our wood stove and the newspaper I pulled out was the one with the headline from March about our governor closing schools. It was weird to sit there and think about how much was new in those ensuing weeks. But I also remembered that my husband was on a business trip in London at the end of January 2020, and he called me and our immediate family stateside and said it would be a good idea to stock up on some non-perishables because stores there were already sold out of a lot of things and he figured we only had a few weeks before the U.S. started having to close down.
So, today on the evidence of just one positive test a large part of Perth Australia is clamped down tight by a Govt directive enacted within hours. We now know that's what is required to hope for containment in an *isolated* city. What hope, then, for those severe restrictions being put on in January last year? That's what was required then to have any chance of containment - multiple municipalities all acting in concert with the same severity of civil restrictions. That's where politics in the liberal democratic world couldn't even give that hope a chance, the virus was out & away under a brand-new order of transmissible routes. No amount of lone bravery speaking to the top level of administration could push "hard lockdown now" through the layers of legislation. Under the "consultative democracy" of PRC, though, containment was achievable.
Rather than puzzling for administrative levers to pull quickly next time, it may be more useful in the long run to figure out has gained from the inertia. Look at capital flows - is this all just zero-sum, or something else?
And now we know R0 is considerably higher than SARS of 2.2. Maybe higher than 3! Especially considering that recent epidemiological thinking has adapted the total US infections # to be above 100 million.
What is the role played by our certainty on previous knowledge? Working for a long time within the same subject might blind you to different behaviours.
That is an excellent question. I think, for example, a lot of early SARS evidence held (but we disregarded it). For example: seasonality, airborne, clustering. On the other hand, this one crucial thing changed, and we were very late to it. I am still amazed at how early Japanese scientists were to the whole package.
Just great, Zeynep. As with the Counters, we see that your vision for Insight goes beyond publishing to catalyzing a thoughtful-about-thought thought community. The principles you're talking about speak to not just a theory of knowledge, but a practice of knowledge; what the ubiquitous "how to spot misinformation" guides are groping for but fall so far short of achieving. Would love to sign up for the Zeynep 'Practice of Knowledge' forum - what are my chances?
Thank you! It's a process, that I am also learning as I go along, and trying to share as best I can!
I love "not just a theory of knowledge but a practice of knowledge." I think you've pinned one of the things that makes Zeynep's thinking so unique and valuable.
Full disclosure, I'm from MN and went to Luther College where Michael Osterholm is also from.
There is a bit of a take down on Michael which I think is appropriate (he said it and it did show western bias I thought) and unfair (he said other things at exactly the same time that were very good and actually he was one of the first experts in the USA to say it.)
Michael was skeptical as Zeynep points out, very early that unsymptomatic people would be spreading the disease.
That said, at precisely the same time he was one of the first in the US and loudest saying two important things: 1) the virus is likely airborne, "just by breathing", and 2) that super-spreaders were likely a key component of how the virus spreads. Both 1) and 2) were right on:
Jan 21st: "The thing we worry about as health officials is a thing called 'super spreading,' where we have certain individuals that are not just infectious but highly infectious."
Jan 24th: "The [novel] coronavirus is one that can be transferred quite readily by the respiratory tract—just breathing. So, it has the potential to spread quickly around the world. It also is one that a sizable portion of the population typically has severe illness and death."
Then on asymptomatic spread, this turned out to be wrong:
Jan 26th: “I seriously doubt that the Chinese public officials have any data supporting this statement,” said Michael Osterholm, director of the Center for Infectious Disease Research and Policy at the University of Minnesota. “I know of no evidence in 17 years of working with coronaviruses — SARS and MERS – where anyone has been found to be infectious during their incubation period.”
By Feb 24th Osterholm had reversed himself on his Jan 26th statement:
Feb 24th: "The coronavirus that causes COVID-19 seems to spread like influenza, through the air, person to person. Unlike Ebola, SARS and MERS, it can be transmitted by individuals before the onset of symptoms or even if they don’t become ill."
I agree that Michael could have and should have pointed out that his earlier guidance based on "all his years" of studying MERS and SARS, was wrong and arrogant Jan 26th. He should have been clear and should have been honest on the 24th that he got this piece wrong.
So in general I think Zeynep's take-down is well placed because Michael never, to my knowledge, confessed that he had it wrong on this very important topic at the beginning.
I do think it reflects badly on him that when he was right he trumpets his rightness and when he was wrong he said nothing that would let us understand he learned something. That is NOT science, that is P.R.
Hi Scott. He also kept insisting—in the summer!—that masks would produce a false sense of security and thus implying they were harmful. But you are right, he also got other things correct. But the interesting thing here is not just the lack of accountability (and I don't mind people getting things wrong: I am sure he's excellent as a doctor) but that infectious disease specialists in many other countries had none of these issues we had here. They had everything right, early on, because this wasn't that confusing or obscure. It's really something worth studying how we managed not to see what was in front of our noses *and* were being told to us by really excellent experts who had the right experience before (SARS) *and* were publishing so much!
Zeynep, I 100% agree with you. It is so worth studying. I am a data scientist, my company works on how to make strategic measures work, tying cause with effect. I have a passion for the subject you are on, absolutely love the question you are posing re: accountability, how to make that work better, so if there is anything I can do to help, please don't hesitate to shout.
Agree Zeynep. I was/am very angry at him with respect to the mask piece. You are a breath of fresh air! I love reading your stuff.
In terms of your theory of knowledge, you could Michael Osterholm as exhibit 1. A hard headed male who thinks he knows everything there is possible to know. Just ask him.
He decided early on that because health care workers were the front line workers, 100% of all n95 masks needed to be siphoned off. This was then done, even though the practice was n95 masks are disposable, meant to be disposed of sometimes 10 times a day. That premise which he set and concrete and shouted at everyone at high volume I found instructive.
The relative odds of giving vulnerable populations access to good masks was never part of his world view. What if the people most likely to end up in the ICU don't get sick to begin with?!
He then had gone on to say a whole set of cascading illogical things on this vector, only front line workers (read trained and using properly) shall have access to good quality PPE.
In his defense, I think unquestionably having a safe environement for front line health care workers had to be the top of the pyramid.
That said there were many even within health care settings who were getting much lower quality PPE than they deserved while doctors who are at the top of an abusive domination/submission pyramid were disposing of masks right and left rather than reusing in a safe way....He said little to nothing about that that I know of....
We are all flawed human beings, so we could take down Fauci too, and let's not even get started with Redfield, CDC.
The bad advice from experts was a stunning thing to behold. You have done the world a great service in my view showing the power of taking a cross disciplinary approach and in so doing, taking them all to task!...
I think this is an important article and would appreciate reading more about the theory of knowledge. I have long found it concerning that we pay more attention to what leaders say than "the facts on the ground" of what they do and don't do. Your discussion of considering the embarrassment factor surely applies to many actions in government and medicine.
One other thought with respect to the embarrassment factor- In a courtroom, one of the exceptions to the hearsay rule is "statements against interest"- ie, the notion that something said is more likely to be true if in saying it the speaker is speaking against their own interests. I think this is similar to the idea that if someone says something that is potentially embarrassing it is more likely to be true.
Same idea! Under the write conditions, what leaders say or do is quite indicative. But, as this post tries to explain, one has to think through the institutional and historical context. I did not have any insight into the medical situation in Wuhan—neither did the medical doctors or virologists making policy for this threat. In effect, though, understanding authoritarian regimes, especially China, was quite useful especially in January of 2020 when we did not have independent scientists on the ground.
I heard your interview on the Ezra Klein podcast yesterday - and was curious about your take on the "individual" versus "population" level. In my field of environmental health (and more generally public health overall) there seems to be a vast disconnect between clinicians and researchers in biomedical sciences who focus on individual outcomes (or individual mechanisms) versus those of us who work in prevention that focus on population outcomes. In fact, they are often completely dismissive -- and demand that any "benefit" or "risk" be "clinically significant" (i.e., to an individual patient) even though at the population level, a small change in benefit or risk may have large aggregate effects when multiplied by the size of the population. Masks, rapid at home tests, etc. are just the latest examples -- though perhaps their effects are magnified due to the exponential nature of infectious diseases, whereas these effects "linear" in most public health interventions (e.g., reducing sodium intake and cardiovascular diseases). Any thoughts about bridging this divide?
I wonder if there's also a difference between making the theoretical case for population-level impact of a certain intervention or practice as opposed to individual-level impact (e.g. masking), and making the *empirical* case -- the challenge of assessing and pinpointing this impact -- at each respective level, and how these are perceived.
For instance, it seems fairly intuitive and uncontroversial that while masks may individually protect the wearer and the person near them to some degree (especially with high-quality masks), their real power should come when everybody is wearing them collectively in the course of a range of daily activities given the interconnectedness and exponentiality of transmission. And yet, there are so many variables at the population level and so hard to definitely *prove* via RCT that widespread masking led to X outcome over Y -- so many other "holes" that could offset this -- that paradoxically the individual case for masks might come across as more compelling and affording greater agency ("I protect you, you protect me!"), even though it's arguably just as difficult to empirically verify a particular instance of this as for an entire country. After all, "We're all in this together" is belied by the outsized role of inequities in how people fare, and there are many examples of low correlation between a country's rate of mask-wearing and how well it is doing.
So, masks are clearly most effective as a universal societal practice - and yet, it might appear that they're not, or that it's impossible to tell. Maybe clinicians and researchers used to dealing with individual-level phenomena and direct cause-effect relationships are impatient with such ambiguity, and privilege the empirical evidence over the theoretical evidence.
This may be an odd question, but I was wondering if you could do a post about or direct me to the best resources you know of about how to become a better "systems thinker." I'm an undergrad from the US and would really appreciate any guidance--books, articles, links--that you could provide. I've found your writing invaluable throughout the pandemic and want to improve the way I think about complex problems. Thank you
I think there's a part of your metaepistemological approach that you're underestimating/leaving out: what sources you use to inform the process. From the very beginning you had a good understanding of what sources were informative. It seemed like you had them preidentified from teaching SARS, field work in Hong Kong etc.
It would be interesting to hear your process for identifying which sources to incorporate.
This is a tough one. Interestingly, though, even unreliable sources can be informative in context, the way China's denials were when later juxtaposed with their swift action. It gave us a sense that it had really, already spread widely.
True, reminds me a bit of your Gezi/CNN-penguin tweets back in the day
Exactly!
Overall love the article Zeynep. Excellent work!
Thank you!
It's amazing to see how many bright, flashing signals were missed even as bigger media houses started reporting on the outbreak after the lockdown started. Two days after the lockdown Wuhan officials said five million people had left the city before the lockdown started. Five Million! And millions were travelling inside and to abroad for the Chinese New Year.
After the Thailand case, on January 28 Reuters reported a tourist bus driver who took a group of tourists from Wuhan had the virus. He had not been to China.
Even after all this, even as late as March, countries like the US, UK and India were ignoring what was staring at them. Did they all think this was going to be like Sars? Or did the fact that Sars was contained fairly quickly made everyone complacent?
Yes, the SARS experience may have made countries complacent though I got the opposite lesson from that near-miss. We came so, so close to catastrophe there.
Last night I was starting a fire in our wood stove and the newspaper I pulled out was the one with the headline from March about our governor closing schools. It was weird to sit there and think about how much was new in those ensuing weeks. But I also remembered that my husband was on a business trip in London at the end of January 2020, and he called me and our immediate family stateside and said it would be a good idea to stock up on some non-perishables because stores there were already sold out of a lot of things and he figured we only had a few weeks before the U.S. started having to close down.
So, today on the evidence of just one positive test a large part of Perth Australia is clamped down tight by a Govt directive enacted within hours. We now know that's what is required to hope for containment in an *isolated* city. What hope, then, for those severe restrictions being put on in January last year? That's what was required then to have any chance of containment - multiple municipalities all acting in concert with the same severity of civil restrictions. That's where politics in the liberal democratic world couldn't even give that hope a chance, the virus was out & away under a brand-new order of transmissible routes. No amount of lone bravery speaking to the top level of administration could push "hard lockdown now" through the layers of legislation. Under the "consultative democracy" of PRC, though, containment was achievable.
Rather than puzzling for administrative levers to pull quickly next time, it may be more useful in the long run to figure out has gained from the inertia. Look at capital flows - is this all just zero-sum, or something else?
Also we cannot treat the collective as a palimpsest; the wrong messages take hold and cannot be easily substituted.
Indeed.
And now we know R0 is considerably higher than SARS of 2.2. Maybe higher than 3! Especially considering that recent epidemiological thinking has adapted the total US infections # to be above 100 million.
What is the role played by our certainty on previous knowledge? Working for a long time within the same subject might blind you to different behaviours.
That is an excellent question. I think, for example, a lot of early SARS evidence held (but we disregarded it). For example: seasonality, airborne, clustering. On the other hand, this one crucial thing changed, and we were very late to it. I am still amazed at how early Japanese scientists were to the whole package.