The folks over at The Oregonian have published a news story about a new scientific study, purportedly showing how awesomely effective masks are when people are forced to wear them. Today we are going to read that paper and explain what it actually shows us.
Perhaps I should also clarify: pro-mask people think masks are awesomely effective at preventing infection (the thing masks don’t do), not that they’re awesomely effective at crushing the spirit of six year old children (the thing masks actually do).
[The main office of The Oregonian is in Portland and in defense of the rest of the state of Oregon, if you are ever traveling outside of Portland everyone you meet will explain that Portland isn’t really part of Oregon. Being filled with useless, yet amazingly self-righteous, ex-Californians, Portland does unfortunately dominate politics in the state. The has resulted in the election of what would be recognized as the most useless governor in the country (Kate Brown) - if anyone actually cared about what happens in Oregon.]
The subject of masks has become very politicized, and this sort of study gets Psycho Kate really worked up because she can use it to justify masking your goldfish or making everyone face west when they exhale.
Here in Sanityville we reviewed the scientific studies on masks in excruciating detail1 and already know masks don’t do anything (at least, anything positive).
Quite suspiciously, there has been very little discussion in media about randomized controlled trials (RCTs), normally considered the gold standard of evidence in the field of medicine. It seems like lots of Kate’s sociopathic friends want to force us to wear masks, but the RCTs have consistently failed to show that community mask wearing stops, or even slows, rates of infection.
The study described in the article is still in pre-print (so not yet finalized) but just like Crazy Kate taking liberties with the Oregon Constitution, the paper’s authors seem to have taken some liberties with the concept of cause and effect. Let’s start with the news article:
Do face masks work against COVID-19? Here’s what new study of 44 countries found (Oregon Live2)
The article uses a lot of words like “association”, “could”, and “suggesting” in describing the results of the study.
During the COVID-19 pandemic’s early days, it wasn’t crystal clear how well face masks were preventing coronavirus infections or severe disease among wearers. Now, a new study reveals an association between mask policies and reduced COVID-19 deaths long before vaccines were added to the picture.
Oh, it wasn’t clear how well they were working? Not if they were working? I guess we already know the author’s bias.
Overall, increases in daily deaths were “significantly lower” in countries that enforced mask policies, suggesting face coverings did and do offer an additional layer of protection that “could prevent unnecessary COVID-19 deaths,” researchers said in a news release posted Dec. 15.
It doesn’t really sound like they’re promising anything exceptional here - the data is just making suggestions, like what kind of suit you should wear to the prom, or which Gubernatorial candidate you shouldn’t vote for if you enjoy breathing.
Journalism doesn’t pay nowadays so they have to cut costs where they can. The Oregonian staff is only about six people (and three beavers) so this story came from a news service, and their story appears to be a repackaging of this press release:
New study associates mask laws to reduction of COVID-19 deaths (American Journal of Preventive Medicine3)
The quotes in the article seem to be taken directly from the press release, which means we aren’t reading the words of the overworked staff at The Oregonian but rather a summation from the journal that is publishing the study.
A study of the impact of national face mask laws on COVID-19 mortality in 44 countries with a combined population of nearly one billion people found that, over time, the increase in COVID-19 related deaths was significantly slower in countries that imposed mask laws compared to countries that did not.
Quick side trip because I couldn’t pass this up
Before getting into the mask stuff, let’s take a quick look at this interesting bit that didn’t make it into the story printed in The Oregonian that they obtained from Tribune News Service when the news service summarized the press release from the Managing Editor of American Journal of Preventive Medicine.
Yes, this is really how these things work. The “news” is often just a reprinting of talking points from someone’s press release. It’s truly fortunate that we can count on PR people never, ever having any kind of agenda.
Moving on…
The average COVID-19 mortality per million population was 48.40 in countries with face mask policies and 288.54 in countries without face mask policies, and this was significantly greater compared to countries with mask mandates.
Someone please tell Kooky Kate that 48.4 deaths per million is just under 0.005% and 288.54 per million is just under 0.03%, over a 3.5 month period. Apparently not a lot of people in these countries died from COVID-19 in those few months.
And the total of 2,167,664 cases (not deaths) from COVID-19 during the study period sounds like a lot until you divide by the nearly 1 billion (911,446,220) total population for the countries in the study. It’s about 0.24%.
Okay, time to stop looking at the story about the press release about the paper, and instead just look at the paper.
Modeling COVID-19 Mortality Across 44 Countries: Face Covering May Reduce Deaths (Motallebi et al. 20214)
Well, we didn’t exactly start strong here. The phrase "may reduce deaths” is not specifically a guarantee of anything. Recall that in the article in Oregon Live describing the study we were told “…it wasn’t crystal clear how well face masks were preventing coronavirus infections or severe disease among wearers.” That’s quite the leap from “may reduce deaths.”
Getting back to the actual study:
This retrospective cohort study included 44 countries in Asia and Europe. COVID-19 mortality changes per million population (outcome) over time between countries with and without a face mask policy (exposure) were compared.
So we’re looking back at the data (retrospective) on mortality in different countries. The authors looked at data from a period of about three and a half months:
The study period spanned from February 15, 2020, which was the date of the first confirmed COVID-19 death within the targeted countries, to May 31, 2020. This period was objectively selected as all countries, both with and without mask policies, had homogeneously implemented restriction of movement and gathering.
This is the very beginning of COVID-19, with only a small number of deaths attributed to the disease at this point but with Kate the Lightweight already drooling over the potential for unneeded exercising of government authority.
The study countries had already implemented mask mandates, and all of them had already restricted movement and gathering, as of the first confirmed COVID-19 death in those countries.
There is an important caveat to this study - it was not designed to determine cause and effect. From the section on limitations of the study:
Regarding limitations, first, this model is not a causal model.
What? Could you repeat that?
Regarding limitations, first, this model is not a causal model.
One more time for intellectually challenged Governors, please?
Regarding limitations, first, this model is not a causal model.
So the paper shows an association between masks and mortality, but not that masks are the cause of the differences in mortality.
This study supports evidence in favor of face mask mandates as a strategy to decrease COVID-19 mortality. The results reveal a significant association between public face covering policy and COVID-19 mortality rate reduction in 27 countries with face mask policies.
Okay, so this study just supports other evidence that masks reduce mortality.
Now is a good time to bring up something in economics called Spurious Correlation5. This is when there is a correlation between two things that are actually unrelated. From the useful people at Investopedia:
In statistics, a spurious correlation (or spuriousness) refers to a connection between two variables that appears to be causal but is not. With spurious correlation, any observed dependencies between variables are merely due to chance or are both related to some unseen confounder.
For example, Hemline Theory6 says that the stock market follows the length of women’s dresses. Skirt hems are up when the market is up, and down when the market is down. Economists love finding things like this in their data - it’s their version of having a sense of humor.
The point here is obvious: the association doesn’t establish cause and effect, and we’ll need to look at some of that other evidence that masks reduce mortality.
The authors given several references for this other evidence and they go into specifics on a couple, by Eikenberry (2020) and Rosenstrom (2020):
Focusing on the public compliance with mask mandates, there are variations across different countries. For instance, in April 2020, Eikenberry and colleagues estimated that near-universal (80%) adoption of moderately (50%) effective masks could reduce COVID-19 mortality rates in New York and Washington by 17%–45% and 24%–65%, respectively. Rosentrrom et al. also suggest a significant decrease in COVID-19 prevalence and mortality if masks are adopted by 70% of the public.
The authors couldn’t find data on the rate of public compliance with mask mandates to include in their model, but are still super confident masks will help.
Consequently, these could not be applied in this model, but it can be assumed that even 50% public compliance results in a significant difference as suggested by Rosentrrom [sic] and colleagues.
(Always be suspicious when someone begins a statement with the words “but it can be assumed…”)
The authors are assuming, based on the work of other researchers, that there is a cause and effect relationship between mask wearing and mortality - and since their data shows a correlation between those two things, it supports that idea.
But is that really what these other studies proved?
Well, no. The two named references are papers that contain mathematical models of the potential impact of masks, but they have to insert an assumed mask effectiveness.
To mask or not to mask: Modeling the potential for face mask use by the general public to curtail the COVID-19 pandemic (Eikenberry et al. 20207)
The creators of this model tell us they have to make two assumptions - that the masks being worn have some level of efficiency both inward and outward. They’re pretty clear about it, too.
We assume that some fraction of the general population wears masks with uniform inward efficiency (i.e., primary protection against catching disease) of (εi), and outward efficiency (i.e., source control/protection against transmitting disease) of (ε0).
In fact, they assume that the probability of getting infected is directly proportional to the percentage filtering provided by the mask.
We assume a roughly linear relationship between the overall filtering efficiency of a mask and clinical efficiency in terms of either inward efficiency (i.e., effect on εi) or outward efficiency (ε0), based on Brienen et al. (2010).
Again with the “We assume…” One small problem though - this has never been proved.
There have been numerous randomized controlled trials of masks, most of which see no benefit. In a few cases there is a small (although not statistically significant) benefit, and in a few cases the mask wearing group does worse than the control (although again usually not statistically significant).
(Someone might want to mention that last bit to Kate “Mini-Mussolini” Brown. Wearing masks for no good reason can increase rates of illness - something for which we have actual evidence.)
Eikenberry et al. are basing their assumption of effectiveness on another paper by Brienen et al8. This paper was reviewed in another post9, and one of the inputs to their model is also an assumption of mask effectiveness. If the actual effectiveness is zero, the reduction in infections is zero.
Dear Governor Nitwit, here’s the important part:
Someone (Brienen et al.) published a model in which they assume masks are effective in order to make their estimates, this model was then used as a reference in another paper (Eikenberry et al.) to justify the same assumption in their model, and this second paper is used as a reference in the Motallebi paper as evidence of a cause and effect relationship.
That’s not how evidence of causality works.
The second reference from the Motallebi paper is also not helpful in the matter of causality:
High-Quality Masks Reduce COVID-19 Infections and Deaths in the US (Rosenstrom et al. 202010)
To reiterate, models don’t prove things, they predict things. For example:
Introduction of mask use with 50% efficacy worn by 50% of individuals reduces the cumulative infection attack rate (IAR) by 27%, the peak prevalence by 49%, and populationwide mortality by 29%.
So if 50% of the people wear masks, and those masks are 50% effective at preventing infection, then mortality for the whole population will be much lower. If. If if if.
But no RCT has ever shown anything approaching 50% efficacy. As mentioned earlier, actual efficacy is somewhere between a tiny reduction and a tiny increase in infections, but most likely zero.
Because the authors of this paper have no idea what the actual effectiveness of masks is, they just model a range of options:
We varied the likelihood of individuals wearing masks from 0-100% in steps of 20% (mask adherence) and considered 25% to 90% mask-related reduction in viral transmission (mask efficacy).
The modelers, in their defense, are clear that they are using assumptions here. These are predictive models based on a series of assumptions, and if those assumptions are wrong then the models are not helpful. It’s as simple as that.
So let’s go back to the original paper
The logical chain in the Motallebi paper is essentially this:
These models predict that if masks work, then there will be lower mortality.
We measured lower mortality in places where people wear masks.
Therefore the lower mortality was due to the masks.
The models use an assumption of mask effectiveness to calculate their reduction in morbidity and mortality, and the authors of this paper, seeing an association between masks and morbidity/mortality, have declared a cause and effect relationship.
Dueling health systems
As a thought experiment let’s compare two hypothetical groups:
The Chicken House Solar Generating Plant
The Palo Verde Alpaca Ranch
The solar plant, since it’s run by antisocial chickens, has no healthcare system at all. No one gets vaccinated, no one gets medical care, and of course no one wears a mask. The employee cafeteria does serve free omelettes, which is either a nice bonus or a little creepy, depending on your feelings about chicken cannibalism.
The ranch is run by cute, fuzzy alpacas. At the ranch everyone has great health insurance, exercises every day, and the little Alpaca Mussolini forces everyone to wear genuine alpaca wool masks all the time.
Which one do we expect to have lower mortality? Most likely the ranch - we all get that having a healthcare system will put them ahead. And the association between the masks and the mortality rate doesn’t prove anything by itself.
Another side note, this time about U.S. data
The absence of the U.S. from this analysis could be very significant. In the U.S. there is essentially one health system (things like hospitals, insurance, etc. work the same in every state). And 18.4% of Americans are part of Medicare11, giving them essentially the exact same healthcare.
But because things like mask mandates occur at the state level, this creates a situation where places with very similar demographics (filled with Americans) and health systems have different mask policies. One would think the U.S. would provide some of the best comparative data available.
Extremely simple summary for CNN anchors, CDC interns, and Governors of Oregon:
Someone published a paper showing a correlation between a policy and mortality rate.
There is no evidence in the paper that the policy has a cause and effect relationship with the mortality rate. The paper says this specifically.
The references given for the assumption of a cause and effect relationship also do not prove a cause and effect relationship. They are mathematical models that assume the relationship.
Kate Brown is useless and should resign. That’s only partially about masks, and mostly about her similarity to an Italian dictator.
Epilogue
The press release also included an interesting quote from the lead author:
“Across variants, vaccines may reduce mortality but not necessarily morbidity, and face masks continue to protect against both,” lead study investigator Dr. Sahar Motallebi of the department of social medicine and global health at Lund University in Sweden, said in the release.
Wait, vaccines reduce mortality (death) but not morbidity (illness)? That’s exactly the opposite of what the news around here has been telling us. But don’t worry, this study claims that masks “continue to protect against both.”
These “scientists” all need to get their stories straight.
Article from Oregon Live:
https://www.oregonlive.com/coronavirus/2021/12/do-face-masks-work-against-covid-19-heres-what-new-study-of-44-countries-found.html?utm_medium=social&utm_source=twitter&utm_campaign=oregonian_sf
Press release from American Journal of Preventive Medicine:
New study associates mask laws to reduction of COVID-19 deaths
https://els-jbs-prod-cdn.jbs.elsevierhealth.com/pb/assets/raw/Health%20Advance/journals/amepre/AJPM_Apr2022_PR_Motallebi_FINAL.pdf
Sahar Motallebi, Rex C.Y. Cheung, Babak Mohit, Shahram Shahabi, Amir Alishahi Tabriz, Banafsheh Sadeghi, Syamak Moattari. Modeling COVID-19 Mortality Across 44 Countries: Face Covering May Reduce Deaths. American Journal of Preventive Medicine, 2021; DOI: 10.1016/j.amepre.2021.09.019
https://www.ajpmonline.org/article/S0749-3797(21)00557-2/fulltext#relatedArticles
Spurious Correlation:
https://www.investopedia.com/terms/s/spurious_correlation.asp
Archived copy:
https://web.archive.org/web/20211222190843/https://www.investopedia.com/terms/s/spurious_correlation.asp
Hemline theory:
https://www.nasdaq.com/glossary/h/hemline-theory
Archived copy:
https://web.archive.org/web/20211222191203/https://www.nasdaq.com/glossary/h/hemline-theory
Steffen E. Eikenberry, Marina Mancuso, Enahoro Iboi, Tin Phan, Keenan Eikenberry, Yang Kuang, Eric Kostelich, Abba B. Gumel, To mask or not to mask: Modeling the potential for face mask use by the general public to curtail the COVID-19 pandemic, Infectious Disease Modelling, Volume 5, 2020, Pages 293-308, ISSN 2468-0427, https://doi.org/10.1016/j.idm.2020.04.001.
https://www.sciencedirect.com/science/article/pii/S2468042720300117?via%3Dihub
Brienen NC, Timen A, Wallinga J, van Steenbergen JE, Teunis PF. The effect of mask use on the spread of influenza during a pandemic. Risk Anal. 2010;30(8):1210-1218. doi:10.1111/j.1539-6924.2010.01428.x
https://pubmed.ncbi.nlm.nih.gov/20497389/
High-Quality Masks Reduce COVID-19 Infections and Deaths in the US. Erik Rosenstrom, Buse Eylul Oruc, Nathaniel Hupert, Julie Ivy, Pinar Keskinocak, Maria E. Mayorga, Julie L. Swann
medRxiv 2020.09.27.20199737; doi: https://doi.org/10.1101/2020.09.27.20199737
https://www.medrxiv.org/content/10.1101/2020.09.27.20199737v2
Americans on Medicare:
https://www.statista.com/topics/1167/medicare/#dossierKeyfigures
Archived copy:
https://archive.vn/aFdxp