I did not read this in detail (I did read it, but I didn't work out the math, etc, which is what a proper commentary would require), but it's worth pointing out

I did not read this in detail (I did read it, but I didn't work out the math, etc, which is what a proper commentary would require), but it's worth pointing out that it's extremely easy to set up an analysis in a way that makes it highly unlikely that you'll achieve statistical significance, whether accidentally or on purpose or some combination of the two. Typically there's zero incentive for a researcher to do so, since nobody publishes "we failed to reject the null hypothesis" as a breakthrough result, but it's a really common mistake. Smooth out one too many variables, bucket broadly, do too many pointless comparisons and let the Bonferroni correction push your significance threshold to the moon, etc. It's arguably easier to slip into accidental insignificance than to p-hack or any of the other well known stats shenanigans that make most observational studies worthless. This is the kind of result where I do worry a lot about that, since it is one of a very few topics where a null result might still be newsworthy, and where the result grossly defies common sense (if people are really staying home, pretty much any model of disease spread would predict less spread).