Wearing Masks is Effective

I’d like to highlight an interesting finding in a research paper documenting just how effective mask wearing is for reducing the spread of COVID-19. The paper cleverly noted that in New York City, social distancing first happened without masks, and only later, on April 17th, were masks mandated. The authors looked at the rate of change of COVID-19 before the mask mandate and after the mandate. They strengthen their case by contrasting what happened in NYC with what happened in the rest of the US.

It’s all about slope

The authors fit three different lines to the growth or drop in new COVID-19 cases. The important thing about such “fit lines” is that their slope shows how fast the cases are growing or shrinking. Below, I have put the slope in parenthesis at the end of each region’s description. The lines in New York city were fit when cases were dropping which is why the numbers are negative and then more negative.

• NYC cases in the 25 days before the mask mandate in NYC, but when stay at home orders were in place (-39)

• NYC cases in the 21 days after the mask mandate in NYC, when stay at home orders were still in place (-106)

• US cases (without NYC cases) in the 35 days in which almost all of the US was staying at home (70)

Masks caused NYC cases to drop faster

Prior to April 17th, cases in New York City were dropping through the use of social distancing alone. Right around April 17th, there is a distinct change in the rate of new COVID-19 cases. After people started wearing masks on April 17th, the cases started dropping faster. The speed of the decrease almost tripled with mask wearing. This contrasts with the rest of the United States, which saw no distinct change in the rate of new infections.

Conclusion: Wear a Mask!

This paper provides compelling data about the importance of wearing masks. The data isn’t airtight, as I explain below, but it is ingenious and I think that it is fairly strong. I haven’t assembled it all, but I get the sense that the evidence is building that extensive mask wearing can really shape the trajectory of this epidemic.

I’m going to write another post (in the Action section) discussing all the issues around wearing a mask: when, how, and what type. Until then, note that a mask is only effective when worn.

Appendix: Potential weaknesses in the analysis and possible rebuttals

What follows are several possible weaknesses in the study & my thoughts on each:

• Some other event may have occurred on April 17th in NYC to change the infection rate.

  • I currently do not know of any event in that timeframe as significant as the order to wear masks. If you know of any other potential cause of the larger drop in cases after the 17th, please let me know.

• The CDC actually started recommending wearing masks on April 3rd. Since that is the case, then the effects of mask wearing should have started happening prior to April 17th.

  • I am going to check in with some friends in NYC to see how much effect the April 17th order had.

• The paper assumes that the rest of the US did not start using masks when, because of the CDC directive, mask wearing became more common in fact.

  • It is my current impression that the majority of the US does not take mask wearing as seriously as NYC currently does.

• The y-axis of the plots are linear but should be logarithmic. The authors of this paper fitted a linear line to data that is exponential in nature. This is poor analysis.

  • I believe that if the data were properly fitted it would only increase the effect observed.