Yves here. Many of you may recall that Yaneer Bar-Yam and Nassim Nicholas Taleb, who are two of the four authors of a paper on the considerable and wide-ranging analytical/statistical defects in the infamous anti-mask Cochrane Report. were, along with Jospeh Norman, very early to warn (January 26, 2020) that Covid-19 had the potential to become a pandemic. The possibility of a Seriously Bad fat-tailed outcome meant aggressive action was the rational response. The fact that Covid-19 was a pandemic was not acknowledged by health officials until March. Norman, Bar-Yam and Taleb stressed:
Together, these observations lead to the necessity of a precautionary approach to current and potential pandemic outbreaks that must include constraining mobility patterns in the early stages of an outbreak, especially when little is known about the true parameters of the pathogen.
It will cost something to reduce mobility in the short term, but to fail do so will eventually cost everything—if not from this event, then one in the future
Biomedical scientist GM similarly had argued it would have been possible to contain the spread of the wild type Covid, which was much less transmissible than later variants, if officials had acted aggressively and early when the total number of infected was less than enormous and contract tracing also would have been viable. But that small window for effective action was lost.
As we see regularly, far too many have rubbished another important risk-reduction strategy, masking. The fact that wearing a maks to prevent contagion is considered polite in Japan and Southeast Asia and is separately often used as a response to poor air quality likely goes a long way in explaining much lower Covid infection rates in those regions versus, say, the Anglosphere. Lambert looks to be on to something when he keeps muttering about democidal elites….
By KLG, who has held research and academic positions in three US medical schools since 1995 and is currently Professor of Biochemistry and Associate Dean. He has performed and directed research on protein structure, function, and evolution; cell adhesion and motility; the mechanism of viral fusion proteins; and assembly of the vertebrate heart. He has served on national review panels of both public and private funding agencies, and his research and that of his students has been funded by the American Heart Association, American Cancer Society, and National Institutes of Health.
Forty years ago when I was an apprentice scientist, the HIV/AIDS epidemic brought out much of the best in biomedical science: Good thinking, cooperation, and collaboration. It also brought out some of the worst, as competing research groups raced to discover the cause of a horrific disease that was relentlessly killing healthy young men, mostly in the Global North, which is where such things get noticed first. Genuine progress was made fairly rapidly even so, and every week Nature, Science, and Cell contained the latest research which we read one after another in the lab. The latest HIV/AIDS evidence was a staple of Journal Clubs across the world.
Nevertheless, politics and so-called “special interests” got in the way, as they always seem to. But within a few years HIV was identified as the cause of AIDS. It is only a slight exaggeration to say that today the only factor standing in the way of worldwide use of effective anti-HIV therapy is money. How to prevent the spread of HIV through behavioral interventions is not a matter of dispute. If one wants to read the “other side of the AIDS story,” Celia Farber’s Serious Adverse Events: An Uncensored History of AIDS has recently been re-issued by Chelsea Green. A new copy sits on my table, and I look forward to re-reading it. There could be valuable lessons in there somewhere that are valid today. If so, I will let you know.
When what became known as COVID-19 appeared in Wuhan in late 2019, the cause of this respiratory disease was identified within weeks (this progress can be easily measured in days if one so desires). Such has been the acceleration of data accumulation in modern molecular biology and medicine. What took years when I was that young apprentice can now be accomplished in hours. Of course, that SARS-CoV-2 had precedents in the original SARS and MERS outbreaks helped. Which brings us to the question, “How do you stop the spread of a respiratory disease?” Although there was much hygiene theatre during the early days of COVID-19, that SARS-CoV-2 is airborne was obvious from the beginning of the current pandemic. This does not mean that frequent hand washing isn’t a good idea, fomites do indeed spread contagion. And when something airborne is spreading and can cause disease, masks are likely to work. Even relatively inefficient masks such as blue surgical masks can work at low viral loads when everyone wears them, another social and behavioral intervention that restrains an epidemic. Examples from the primary literature supporting mask usage rather than the secondary review literature are considered below.
Although this seems to have been largely forgotten, the use of masks during a pandemic has a an old history, as shown in a remarkable photograph of spectators in Grant Field in Atlanta during the 1918 football game between Georgia Tech and Furman. And masks were not the only intervention recommended:
At Tech, in fact, masks weren’t the only precaution taken. The Oct. 10 edition of the (Atlanta) Constitution apprised readers of another measure ordered by army medical authorities at the school to prevent the spread of the contagion at football games – no cheerleading…“Cheering is too much like sneezing: if it is to be done in these days of influenza, it should be done through a handkerchief, and a cheer through a handkerchief would not be worth doing,” writer J.H. McKee reported. “So there will be no cheerleading.”
All in all, eminent good sense! Did the masks work? No one can know today, but the engineering approach to prevention of the spread of a putative airborne contagion was an obvious intervention, at what was to become a leading engineering institution. However, as noted in this article by Ken Sugiura, local politicians demurred. Bans on gatherings were soon discontinued, and the death rate from the so-called Spanish Flu was about the same in Atlanta as everywhere else. That result might have been different.
Early in the current pandemic several well done theoretical and practical studies showed that masks are effective in preventing spread of disease, but this was not and has not been generally appreciated. The reasons are many and varied. The Cochrane Study on Physical interventions to interrupt or reduce the spread of respiratory viruses (Review), published on 30 January 2023, has received much attention. The interpretation of the Cochrane Review was that “masks do not work.”
Is this true? “No” is the conclusion of the paper Quantitative Errors in the Cochrane review on Physical interventions to interrupt or reduce the spread of respiratory viruses, recently published by Yaneer Bar-Yam, Jonathan M. Samet, Alexander F. Siegenfield, and Nassim N. Taleb (thanks to LS for sending this to me).
Before going further, it is important to note this paper is a preprint, which means it has not been peer-reviewed [1]. I usually wait until a paper has been published after peer review, but for me this is an exception that proves the rule for several reasons. The paper is well written, and the evidence supports its conclusions. Plus, the subject is very topical. The consensus among the public and our political and healthcare authorities is that COVID-19 is basically over. While COVID deaths have declined even as recordkeeping has become hit and mostly miss, this is also due to better clinical management of the disease. Given that long COVID is common and that current vaccines prevent neither disease nor its transmission, COVID-19 is still here, perhaps to stay for a very long time. Thus, it behooves all of us to avoid it. Masks, distancing, and improvements in ventilation and airflow, each an engineering approach, work, whatever the continuing biological and epidemiological course of COVID-19. So will antivirals, which to date have had an undistinguished impact on the pandemic.
This paper is also written by four scientists who have the expertise to interpret the Cochrane Review in detail. Still, arguments from authority can be misleading or frankly disingenuous. However, those with demonstrated authority in the subject, in the absence of conflicts of interest, should not be ignored, especially in a continuing crisis. I am not familiar with the work of the first three authors, but Nassim N. Taleb has demonstrated repeatedly that he is qualified to dig into the details and assumptions of the Cochrane Study. Yes, he can be controversial. But so are Michael Mann and James Hansen and Richard Lewontin, who have also been proven correct in their areas of expertise. Of the former, I plan to contribute something soon in this space. Of scientists (frequently biologists) whose expertise is not “infinitely” extensible, a note is here [2]. Scientism is not so nebulous a concept when authority in one discipline does not actually translate intelligently and intelligibly to another. This is frequently the case.
The article by Bar-Yam et al. notes that even though masks and respirators have proven their utility as an engineering solution to the problem of airborne pathogens and contaminants:
(S)tudies on their adoption over the last several decades in both clinical trials and observational studies have not provided as clear an understanding. Here we show that the standard analytical equations used in the analysis of these studies do not accurately represent the random variables impacting study results. By correcting these equations, it is demonstrated that conclusions drawn from these studies are heavily biased and uncertain, providing little useful information. Despite these limitations, we show that when outcome measures are properly analyzed, existing results consistently point to the benefit of N95 respirators over medical masks, and masking over its absence.
The mathematical expressions used in this paper are complex, and it has been a long time since I used anything remotely similar. But after several weeks with them, off and on, they do make sense to me. The authors show that:
(T)he analyses of randomized clinical trials are missing six things: (1) propagation of uncertainty from improperly neglected random variables, (2) compounding of effects due to unaccounted transmission and infection of non-study participants, (3) invalid correspondence between study question and design reflected in variables used, (4) analysis of significance—the meaning of the results through their implications for health, (5)invalid categorization of data associated with study conditions, and (6) clear definitions and characterization of adverse effects. Using two recent reports—a trial and a systematic review and meta-analysis, we show that studies comparing N95 respirators and surgical masks, while interpreted as showing equivalency, are compatible with a substantial benefit of N95s.
These are strong conclusions and I await formal publication and subsequent discussion in the literature. Conflation of surgical masks with N95 respirators has been common in the underlying trials, with inexact comparisons among who was wearing which masks where. These things matter. Other sources of data and reasoning from the engineering perspective have also been largely ignored in the typical mask trial. These include several strong papers from early in the pandemic that were mentioned above (all are open access) briefly summarized here:
(1) Quantitative modeling of the impact of facemasks and associated leakage on the airborne transmission of SARS-CoV-2: The results from different scenarios show that all the modelled facemasks provide a significantly higher protection when used as a source control rather than as a respiratory protection. FFP (filtering facepieces) masks have a higher filtration efficiency than surgical or community masks and provide a better protection if they are fitted accordingly to minimize the leakages.
(2) Unmasking the mask studies: why the effectiveness of surgical masks in preventing respiratory infections has been underestimated (Siegenfeld, Bar-Yam, and Taleb are authors of this paper): When the adherence to mask usage guidelines is taken into account, the empirical evidence indicates that masks prevent disease transmission: all studies we analysed that did not find surgical masks to be effective were under-powered to such an extent that even if masks were 100% effective, the studies in question would still have been unlikely to find a statistically significant effect.
(3) Face masks effectively limit the probability of SARS-CoV-2 transmission: We show that variations in mask efficacy can be explained by different regimes of virus abundance and are related to population-average infection probability and reproduction number. For SARS-CoV-2, the viral load of infectious individuals can vary by orders of magnitude (factors of 10). We find that most environments and contacts are under conditions of low virus abundance (virus-limited), where surgical masks are effective at preventing virus spread. More-advanced masks and other protective equipment are required in potentially virus-rich indoor environments, including medical centers and hospitals. Masks are particularly effective in combination with other preventive measures like ventilation and distancing.
(4) And from a paper published just after the first lockdowns of the pandemic (April 2020), Respiratory virus shedding in exhaled breath and efficacy of face masks: We identified seasonal human coronaviruses, influenza viruses and rhinoviruses in exhaled breath and coughs of children and adults with acute respiratory illness. Surgical face masks significantly reduced detection of influenza virus RNA in respiratory droplets and coronavirus RNA in aerosols, with a trend toward reduced detection of coronavirus RNA in respiratory droplets (note the double emphasis about aerosols, in May 2020). Our results indicate that surgical face masks could prevent transmission of human coronaviruses and influenza viruses from symptomatic individuals.
None of these papers is perfect, but few scientific papers are. The only “perfect” paper I have ever read or taught was a single page (pdf) published 70 years ago that led to a Nobel Prize nine years later (thanks to The Rev Kev for finding this free version in a comment here). Each of these four reports is a peer-reviewed paper in the primary literatureand each has a clear result: Masks prevent transmission of respiratory viruses. Other papers in the primary literature show the same thing. No randomized clinical trials of masks were ever needed, except to pad the bibliographies of the authors who published the papers. It follows that no large-scale, secondary review by Cochrane of the more than 60 papers (with more excluded from the analysis) of masks was ever necessary. But this is the way of Evidence-Based Medicine. Those of us in the biomedical science and medical education communities can think better and do better. And doing better is not HICPAC! HICPAC is Big Biomedicine, where something other than biomedical science, health, and wellbeing always dictates public practice and private and especially “non-profit” gain.
After the Cochrane Report was criticized in the media the authors denied ever saying that masks do not work. As the indispensable Naomi Oreskes explained a few weeks ago in a short piece at Scientific American:
The group’s report was published by Cochrane, an organization that collects databases and periodically issues “systematic” reviews of scientific evidence relevant to health care. This year it published a paper addressing the efficacy of physical interventions to slow the spread of respiratory illness such as COVID. The authors determined that wearing surgical masks “probably makes little or no difference” and that the value of N95 masks is “very uncertain.”
The typical journalistic shorthand was the cause of some of this controversy, led unsurprisingly by Bret Stephens of The New York Times. As Oreskes points out, “the report did make clear that its conclusions were about the quality and capaciousness of available evidence, which the authors felt were insufficient to prove that masking was effective…still the authors were also uncertain about that uncertainty, stating that their confidence in their conclusion was ‘low to moderate.’” To the contrary, the quality of the evidence that masks work is high and more than capacious enough. But masks do require effort and imply, among other things, that interior air quality must be addressed to prevent SARS-CoV-2 (and other respiratory pathogen) transmission. That would be expensive.
Oreskes also notes “the study’s lead author, Tom Jefferson of the University of Oxford (also the academic home of William MacAskill, of Effective Altruism and Longtermism fame) promoted the misleading interpretation. When asked about different kinds of masks, including N95s, he declared, ‘Makes no difference—none of it’… he called mask mandates scientifically baseless…Jefferson has claimed that COVID policies were “evidence-free.” This highlights a second problem, which is “the classic error of conflating absence of evidence with evidence of absence. The Cochrane finding was not that masking didn’t work but that scientists lacked sufficient evidence of sufficient quality to conclude that they worked. Jefferson erased that distinction, in effect arguing that because the authors couldn’t prove that masks did work, one could say that they didn’t work. That’s just wrong.”
Yes, it is wrong. And this approach always ignores evidence on the ground. This is the original sin of Evidence-Based Medicine and Big Biomedicine. Again, from Oreskes:
In fact, there is strong evidence that masks do work to prevent the spread of respiratory illness. It just doesn’t come from RCTs. It comes from Kansas. In July 2020 the governor of Kansas issued an executive order requiring masks in public places. Just a few weeks earlier, however, the legislature had passed a bill authorizing counties to opt out of any statewide provision. In the months that followed, COVID rates decreased in all 24 counties with mask mandates and continued to increase in 81 other counties that opted out of them.
This evidence is “anecdotal” and thus unscientific, not to mention lost down the memory hole. Although such evidence fits with one proven engineering solution to COVID-19, it is ignored due to what Oreskes has called “‘methodological fetishism,’ which occurs when scientists fixate on a preferred methodology and dismiss studies that don’t follow it…By dogmatically insisting on a particular definition of (often merely statistical) rigor, scientists in the past have landed on wrong answers more than once.” The original eugenics movement made possible by the statistics of Francis Galton, Karl Pearson, and R.A. Fisher comes to mind. And they/we will continue to do so as long as an obscurantist scientism fueled by conventional frequentist statistics reigns over science, in a world in which statistical significance very often has no clinical or other relevance whatsoever. Correlation coefficients and p-values often obscure more than they reveal.
We must and we can do better. Sooner rather than later.
Notes
[1] Peer review is a fraught practice these days and is no guarantee, previously covered here. But it does matter, and science will either return to something of a golden age or continue its slipshod descent into pay-to-publish-peer-reviewed irrelevance.
[2] As a biologist, I have often read the work of others who have had large impacts in areas outside of their expertise. The late E.O. Wilson is perhaps the most well-known among these eminent scientists whose impacts outside of his expertise have not withstood scrutiny. Wilson was unquestionably a towering scientist. He knew ants! His The Theory of Island Biogeography, which was written with Robert MacArthur, is justifiably a classic that among other things introduced r/K selection theory [many offspring/minor parental effort (rat) versus few offspring/major parental investment (primate)]. From this work to Sociobiology: The New Synthesis, On Human Nature, and Genes, Mind, and Culture is a long stretch. The tiresome Sociobiology Debate lingers, and the 50th anniversary (2025) of Wilson’s great book but for the final chapter, should be interesting. On Human Nature is still sometimes read. Genes, Mind, and Culture: The Coevolutionary Process deservedly sank without a trace. Wilson’s appreciation of human impacts on the ecosphere has been important, but the impression left by this work is that he was more interested in cataloging biodiversity that will be lost in the current great extinction rather than preventing it. Darwin’s four major works in one volume edited by Wilson is outstanding and the one source for Charles Darwin in content and presentation. Consilience was a disjointed mess and not an improvement on the original work of a previous polymath, W. H. Whewell. Wendell Berry took Wilson to task on consilience in an interesting response. Other evidence regarding Wilson’s interests and putative motivations regarding sociobiology (high NYRB paywall) have recently come to light.
