The Powerline blog, which I have great respect for, has a writer who insists on periodically writing on CV-19 even though he has no ability to understand research quality or results. Yesterday this person put up a post saying “Masks Work”. One of the two studies cited for the proposition was the Kansas mask study, which another reader pointed out was completely debunked and the writer had to acknowledge this in an update. But the other study is even worse. It is here, (Mask Study) and I believe I posted on it when it came out. The study uses data from last spring, ending in May. The Powerline blog post author insisted in a separate email that it uses data from the fall, but I took this link directly from his post and you can read in the abstract right up front when the data is from. The authors were trying to ascertain whether lockdowns and/or mask mandates had an effect on spread. So you see the first issue right up front–looking at a mask mandate, not mask wearing behavior. The first is pretty meaningless, the second is what you need to study.
The researchers used the period March 1 through only May 18 and only the 15 states hardest hit at that time. So only the spring wave states. Can you say seasonal/goegraphic and time period bias. Let’s also get this out of the way upfront, at this point in the epidemic, little testing was being done and the amount and conditions of testing were highly variable among states. So looking at cases is inherently screwed. The data is all over the place, just look at the table, but even when data is widely scattered, you can claim there is correlation if you use the right formula and that is what they did. But the uncertainty level is as great as the so-called effect. And they leave no lag for the supposed effect of the mandates on cases, the assumption is it occurs instantly. Now what is really useful is to go look at the epidemic curves in these states. They all look very similar and are pretty classically shaped, regardless of mask mandates or not. This study reminds of the garbage preprint that purported to show an effect of mask mandates by county and then had to be withdrawn when a few weeks later the case waves completely destroyed their analysis. Somehow that study never got updated either. Really, this is a completely useless study and if extended through the fall and winter would reveal zero affect of mask mandates. One hilarious note was that the authors said that New York and New Jersey had the greatest reduction in cases after they initiated their mask mandates. Those two states have among the highest death rates in the entire world.
And of course, no mention anywhere in the post of the Danish study, which remains the only randomized study of use of masks in the community to slow CV-19 spread and which showed no impact. Only someone who completely fails to understand research methodology would use the study cited to support the notion that masks work. Everyone else has given up on these facile analyses because they don’t show an impact. And as we are here in Minnesota, I would like to note that not once yet since our beloved Governor ordered the mask mandate has Minnesota had a several week run of cases in which there were fewer than in a comparable period before the mandate. In fact, notwithstanding 90% plus mask compliance, Minnesota somehow had a massive case surge in October/November.
A couple of other notes on masks. Germany has had one of the strictest mask mandates, including that people must wear medical masks. Germany is going into full, complete lockdown, closing every business, because of case spread. Masks simply do not work to stop the spread in the community.
And a medical professional reader writes to tell me that every day he sees people with contact dermatitis from mask-wearing and he has had to send several to the ER. So don’t tell me they don’t have downsides, the government will just do everything it can to hide them.