The Scientific Community's Polymerase Chain Reaction
In response to an article I posted here previously (https://escipub.com/Articles/IRJPH/IRJPH-2021-08-1005.pdf?fbclid=IwAR3kHdPDrgIgXDgebZDAF_xfKB9oMdeF9yzNaGeeRyRr8gUO-4CTUfX9TQk
), more than one friend responded with links to compendiums of studies in
support of the efficacy of masks. For example, this one:
I committed to reading the first ten from this collection to
learn more and to make sure I wasn’t just plugging my ears and closing my eyes.
Here are my impressions of those 10 (which actually turned out to be over 20,
since I had to follow the deeper links to finally get to the raw data). This is
not to say my opinion is the be-all-end-all, but I do consider myself to be a
considerate, science-loving, non-moron and as such feel the need to explain how
someone can oppose universal masking while also being those things:
Study 1.) This was a summary of about a dozen handpicked studies,
coupled with a conclusion far more confident than any of the individual studies
could afford to claim. It was the building of a case. The fact that this
counted as part of the list is the epitome of what I find so annoying, adding
clout and influence to severely limited studies that don’t deserve it. Glossing
over the shortcomings which are at first directly acknowledged in the first-hand
sources themselves but that rarely survive the game of telephone when a piece
like this one does the work for us. Substituting overwhelmingly echoed consensus
for overwhelming evidence and calling it the same thing (much like the intent
of the list-gatherer in the first place). That’s not to say that consensus has
no value – it does. But consensus receives way too much respect as “the science”
and tends to bury how mixed, uncertain, marginal, and limited the original data
is. This was a perfect example.
When I got down to the actual studies this was building upon, I found common themes
of comparing case numbers from before and after a mask mandate. How does one
know that the difference was caused by the mask mandate and not the seasonal
way of things or the ebb and flow of saturation or other voluntary self-preserving
behaviors of individuals? Did they compare it to a neighboring locale without a
mask mandate? No. They compared it to their own predicted trajectories. Not
surprisingly, they found a difference.
Another theme I found here was extrapolation. What is the
value of a study that has people shouting through cones if in the trial they’re
entering a sterile room, receiving a sterile mask, hanging out with no one
else, and staying for 10 minutes tops? Proving that the theoretical mechanics
are initially sound is one thing. Extrapolating from that to claim that this will
make a long-term difference for the general public requires ignoring a nearly
infinite set of other variables. For this reason, a controlled study that takes
place in the real world should trump these findings immediately by virtue of
non-extrapolation.
Other shortcomings included the inability to separate masking
behavior from other actions that people who masked also reported taking. Or,
the ethics of the situation preventing scientists from designating a control
group. Or, the size/scope of an observation being so small that it might as
well be a story about that time you wore red and aced a test. These shortcomings
were explicitly acknowledged in the original studies, but they wash out as they
are curated here.
The most compelling study that I saw from this first
review was one from Thailand. It at least followed a large number of individuals
in a real world setting and allowed for a mix of self-reported behaviors
including those who reported not masking. However, only those who wore a mask at
all times were reported to see any difference (with a confidence interval of
improved odds reaching as low as .06). It found no benefit for occasional use,
and it acknowledged the collinearity problem of not being able to isolate one
protective behavior from other voluntary protective behaviors (imagine the behavior
of the kind of person who willingly masks up at all times). Still, if you
wanted to say that “doing all the things at all the times,” including not
getting close to anyone or speaking with anyone for more than a couple of minutes,
gives you better odds, I don’t think I could argue with that. I just think that
at this point if you are still doing all of the things in light of a vaccine
available to anyone who is at any kind of increase risk, we have a different view of the meaning of
life. More importantly, if you think that selective use of a mask in virus-rich
environments will help you keep a foot in both camps, this study does not
support your hope. (And number 7 might make you sad).
2) Another pre-chewed review substituting as another study, inflating the numbers. But
this time coupled with a collection of theoretical models to defend believing
more strongly than current available evidence suggests. If you read long
enough, this review acknowledges that if you include the phrase “RCT” (randomized
controlled trial) in your search, there is only weak evidence for a small effect
of masks against disease in general, while if you accept simple observational
studies the effect is reported to be higher (with the collinearity problem). Specifically,
for Covid 19, it acknowledges that at the time of its publishing, there were no
RCTs, one observational trial, and unclear evidence from other respiratory
illnesses. But hey, the models are compelling if you could only verify that all
of the assumptions that they rely on turn out to be true.
3) A laser-light video experiment. I already addressed the
problem with this. I don’t take issue with the theoretical mechanics being
demonstrated here – my beef is with the, to me, cartoonish jump in application.
I will add that we knew, many months ago, this particular virus is not confined
to droplets. And the virus lasts for 2 days on cloth. So when you’re yelling
through yesterday’s dried out mask in order to be heard by the guy across from
you, are you not launching an extra concentrated helping of yesterday’s virus? Sure,
you could change them out all the time, but do you think that someone who was
compelled to put one on in the first place is doing that?
4) More before-and-after modeling with nothing to compare to
but a projection of what “might” have happened without intervention. Following
these in real time, I knew full well that neighboring states were enjoying the
same rates of dips in cases without mandates. The frustration was mind numbing.
Nothing so lazy and anti-science as “I intervened, and things weren’t as bad as
I thought they were going to be without my intervention!”
5) A story about a sick guy who flew on a plane with a mask
and a cough. A testament to the efficacy of masks or to the air purification
systems on planes? Or just a guy on a plane with a mask and a cough who spoke
to no one and minded his own business?
6) Aaaaaand another review counting as another study. Also
acknowledging the low certainty of the conclusions if you look closely enough.
I also love how the author of the website list deliberately chose to leave the
word “could” out of his quote when selecting this to be part of the 49.
7) The construction of a model to defend a hypothesis that
masks are more effective in public than so far proven. It starts out by
acknowledging that “randomized clinical trials have shown inconsistent or
inconclusive results, with some studies reporting only a marginal benefit or no
effect of mask use.” It then builds a model centered around virus abundance to
explain the discrepancies away. The model postulates that masks are more
effective in a virus poor environment, and that when the environment becomes too
virus rich they lose the battle. Fine, although I consider that to be a point
in my favor. I don’t know why this gets to be counted as part of the 49 when
mask mandates are often specifically for virus-rich environments.
8) Again, combing through old studies and summarizing them
with your conclusion does not a new study make. Points for limiting their
search to randomized controlled trials, even though obviously none of the
trials were for Covid 19. The conclusion was also another amped up “if you do
all of the things early and well, then we think the odds are good. Unless you
work in a hospital, in which case only a respirator used at all times gives you
any better luck than going mask-free.” Soo...we’re supposed to believe that
these masks make a difference when we smush into a subway car or storm the
field at a football game when you just explicitly acknowledged that they aren’t
sufficient for moving around in a hospital?
9) This is nearly the same as number 3, only it branches out
to include all the different types of masks. Some fail the test of theoretical
mechanics even in the lab. But again, it’s just the lab.
10) Another commentary! Ugh! Acknowledging that there are
several reasons to question the efficacy of masks against covid 19, then moving
on to describe, theoretically, why we should discount those and mandate them
anyway. Dated April of 2020.
I came for real world trials to trump one of the largest
scale real-world observations with as close to a control group as you could get
under the circumstances. I got theories, models, arguments, and literature
reviews that amount to summarizing the best of what we could gleam from the poorly-related
past. I’m sorry, but this is a deliberate attempt to substitute the quality of
the science with the number of voices.
This review did somewhat alleviate my resentment of masks at church. There are who knows how many people out there still trying to do “all-the-things,” for whom church is the richest viral load they will encounter all week. I can refrain from deliberately bursting their bubble for the sake of enabling them to make it to church, though I would have appreciated the request as being framed that way. After that, if you need me, I’ll be somewhere like here:
On a similar note, but from a different source, consider the
following charts displaying what happened in the real world, region by region.
These came from a quiz. That everyone fails. Because we prefer to believe our behavior
has been justified.
Comments
Post a Comment