A recent study of mask effectiveness is hailed by two Oxford professors of evidence-based medicine in the following terms:
Yesterday marked the publication of a long-delayed trial in Denmark which hopes to answer that very question. The ‘Danmask-19 trial’ was conducted in the spring with over 3,000 participants, when the public were not being told to wear masks but other public health measures were in place. Unlike other studies looking at masks, the Danmask study was a randomised controlled trial – making it the highest quality scientific evidence.Heneghan & Jefferson, “Landmark Danish study shows face masks have no significant effect,” The Spectator, 19 Nov 2020.
After summarizing the research of the article (linked above), they summarize its conclusion:
And now that we have properly rigorous scientific research we can rely on, the evidence shows that wearing masks in the community does not significantly reduce the rates of infection.Ibid., emphasis mine.
However, this is the conclusion of the actual scientific paper:
The recommendation to wear surgical masks to supplement other public health measures did not reduce the SARS-CoV-2 infection rate among wearers by more than 50% in a community with modest infection rates, some degree of social distancing, and uncommon general mask use. The data were compatible with lesser degrees of self-protection.Bundgaard, et. al., 2020: https://www.acpjournals.org/doi/10.7326/M20-6817
A few lines previously, the authors explain the qualification of “by more than 50%” as follows:
Although the difference observed was not statistically significant, the 95% CIs are compatible with a 46% reduction to a 23% increase in infection.Ibid.
Now, how one logically revises the above, limited conclusion of the actual article—with all of its qualifiers, including the fact that general mask use was not in place and the reduction recorded within confidence intervals—into “the evidence shows that wearing masks in the community does not significantly reduce the rates of infection” seems possible only by way of a style of vague abstraction that trades on the blurred limits of a limited scientific experiment for the purported good of the reading public.
What is the danger? Coronavirus logic now extrapolates from such a revision, equivocating through loss of precision, and confidently but groundlessly concludes that the wearing of masks is not effective.
Indeed, the actual study—if one reads beyond the abstract—confirms this:
Although no statistically significant difference in SARS-CoV-2 incidence was observed, the 95% CIs are compatible with a possible 46% reduction to 23% increase in infection among mask wearers. These findings do offer evidence about the degree of protection mask wearers can anticipate in a setting where others are not wearing masks and where other public health measures, including social distancing, are in effect. The findings, however, should not be used to conclude that a recommendation for everyone to wear masks in the community would not be effective in reducing SARS-CoV-2 infections, because the trial did not test the role of masks in source control of SARS-CoV-2 infection.Ibid, my emphasis.
The truth, even more so than public health, is a common good. Critique of this critique, on any point where it falls short, is welcome.
7 thoughts on “Coronalogic”
Distinguo? Quaestio, puto. John, I don’t see that you are taking into account the significance. Data may be “compatible” with a good outcome, but if it is not significant, trained scientists using such techniques say “we have no proof.” Right?
And furthermore, with an ‘n’ of 3,000, if masks made a significant difference, you would think the significance would show up.
Yes, I agree. But that just does back to what I’m saying. The technical language doesn’t translate well into everyday language. The technical conclusion is that there is no proof within the limited parameters of the study. It doesn’t undermine other general recommendations, as the authors of the study themselves point out. The “Spectator” version of the conclusion doesn’t reflect this nuance well, it seems to me.
I suppose if it said “*one recent study* shows that wearing masks in the community does not significantly reduce the rates of infection,” that would be more accurate. I really wonder about a big study I found, a regression analysis across almost 200 countries, now that infection rates have soared again, but probably with no significant reduction in mask wearing. That showed a significant effect on infection rates, but perhaps it was simply a “post hoc” situation. Masks are beloved of a lot of people with a love that, it seems to me, exceeds what the evidence warrants. “Love of science” isn’t like that.
Thanks, Craig—my gripe was only with the way that the study was being reported, while not pointing out that the study hadn’t tested a population at large wearing masks. What is the other study you found?
I don’t know why that showed up as anonymous. It’s me, Craig White. Not ashamed of it.
Amplius et distinguo: The failure of the article it seems is to focus on ‘statistical significance’ and ignore ‘expected payoff.’ When we talk about mask safety we are interested in expected payoff for community masking, right? Even if something is not statistically significant, the expected payoffs could still justify the action. If the confidence interval allows for a 23% reduction in infection, then… wouldn’t that mean whatever value we place on infection rates would determine the social policy?