COVID-19 Special Update: Evaluating the Value Proposition of Using a Cloth Face Mask to Slow COVID-19

I have not seen much solid science around cloth face masks, so I want to present a short apolitical view of what we know and what we don’t know.

The Center for Disease Control released a short publication a while back with a (mild?) recommendation to wear face masks, even homemade cloth ones, in public to prevent asymptomatic transmission. Here’s the recommendation.

As a sociologist would have anticipated, this recommendation seems to have been a flag that people with opposing views of society have each latched on to. I doubt any level of evidence will disabuse anyone of their position, but as someone who is very curious about the effect of a face mask (especially a homemade one), I’ll try to put together what is out there in one place as best as I can.

Some Representative Research into this Issue

First, I’ll editorialize and say that it’s surprising there’s not more research on PPE or the mechanisms of spread of a virus in general (see my assessment of the paper that seems to be the instigator for the 6 foot rule). Perhaps that will be one of the benefits coming out of COVID-19… It had been 100 years since a real outbreak and now the funders and researchers will focus on this issue more heavily.

• Viral load and outcome in SARS infection: The role of personal protective equipment in the emergency department (2006): This is a study with multiple citations that followed a SARS (not COVID-19) cluster of 16 people, 4 of whom were heath care workers who wore PPE. This is a good-faith attempt to evaluate PPE in SARS cases, but external factors like ventilation, duration of exposure, proximity, AGE, etc., that may have strong correlation with infection are admittedly not controlled. RESULTS: Their conclusion was that PPE, “including the N-95 mask, does not confer 100% protection against SARS infection, it seems that it may lessen the initial degree of exposure to the virus with a subsequently lower viral load in the upper respiratory tract. This may result in a milder disease clinically, less chance of secondary transmission, and a faster recovery.” TOD’S ANALYSIS: I think this is a well-run study that was extremely short of data or support (guessing they weren’t funded well) who did the best they could. The researchers indicate that their results indicate that N-95 respirators — though they don’t block 100% of the virus — are worthwhile to use and may at minimum speed the passing of the disease in the wearer. My reading as someone who teaches and practices statistics and data science for a living is that their results indicate that SOMETHING about the health care workers put them at lower risk. PPE may be the best guess at a causal element, but it may just be correlated with multiple other factors (one being my guess that the health care workers were many years younger than the patients) that was more causal.
• Protection by Face Masks against Influenza A(H1N1)pdm09 Virus on Trans-Pacific Passenger Aircraft (2009): This is another airplane study (done by China’s CDC) that doesn’t really come to a scientific conclusion on PPE, but I include it here to discuss the challenges of these kinds of studies. There is a control group in this study who is wearing masks, but the researchers admit that they don’t know the mix of N95 respirators to other types of masks. There’s also a large confounder in this study and that is because the researchers admit that there’s a possibility that H1N1 transmission started before their study started, on the flight before the one they conducted their experiment on. There’s a lot of good info in this study, but they don’t come to much of a conclusion on the effectiveness of PPE based on their experiment. However, this study has been cited and is interesting to read.
• Mathematical Modeling of the Effectiveness of Facemasks in Reducing the Spread of Novel Influenza A (H1N1) – 2010: This study has been cited a large number of times, so it’s research is influencing other research in the area of modeling the impact of PPE. Based on their simulations, they claim that “the results show that if N95 respirators are only 20% effective in reducing susceptibility and infectivity, only 10% of the population would have to wear them to reduce the number of influenza A (H1N1) cases by 20%. We can conclude from our model that, if worn properly, facemasks are an effective intervention strategy in reducing the spread of pandemic (H1N1) 2009.” TOD”S ANALYSIS: For someone longing for a mathematical look at this problem (virus transmission in the presence of a face mask) this is a great read. Some will claim that this is in-simulation and needs to be conducted on live subjects (but see the above for the difficulties thereof). Their results indicate that high-functioning N-95 Respirators can reduce transmission (see chart below) but that Surgical Masks (maybe just a small step better than our homemade cloth masks) don’t have any appreciable effect.

Another Summary of the Literature – by Illinois-Chicago School of Public Health

This article was published as a commentary by a scientist at the University of Minnesota’s Center for Infectious Disease Research and Policy and you can find it here. The authors (from the University of Illinois-Chicago School of Public Health) also include a large set of references on this particular topic (the effectiveness of different kinds of PPE in virology), which may be useful for anyone who wants to do their private research into this.

OVERVIEW: First off, this is a short summary of the applicable literature and is one of the few places comparing N95 respirators, Surgical Masks, and Cloth Masks. I’d recommend following the link and reading your article because they summarize it well. But in case you don’t want to do so, their takeaway is that cloth masks have no effect either against becoming infected or against infecting others. Surgical masks have a benefit, but about 1/5 of the benefit of a N95 respirator. The N95 respirator has a benefit both against becoming infected (“healthcare workers continuously wearing N95 FFRs were 54% less likely to experience respiratory viral infections than controls” at a 97% confidence level) and against infecting others but most patients cannot tolerate wearing a N95 for long periods at a time, so they’re not a good solution for source control.

Here’s the authors’ thoughts in summary, “Leaving aside the fact that they are ineffective, telling the public to wear cloth or surgical masks could be interpreted by some to mean that people are safe to stop isolating at home. It’s too late now for anything but stopping as much person-to-person interaction as possible.”

CONCLUSION

This is the best that I can come up with short of developing an unneeded dissertation. The masks have become a political signaling mechanism, which is unfortunate for our country. I understand both sides of the signalling mechanism and sympathize with both sides. I’d summarize by saying that some who have a more collective worldview of the world see the masks as visible evidence of care for others and compliance to the recommendations of authorities. Others who have a more individualistic worldview see the masks as a sign of a governmental mandate that violates their personal choices. I cannot find any evidence, however, of science that would support the important question of whether the cloth facemask recommendation is based upon real phenomenon regarding virus transmission and prevention thereof. Much of what I see indicates that the cloth facemasks (and surgical facemasks) have little effect. I suspect there are other studies that conclude otherwise, but there are so many papers and pre-prints out there… (update: I’ve found some new papers and am including them below as I find them)

UPDATES

Here are relevant papers I have found since publishing this:

• https://www.medrxiv.org/content/10.1101/2020.04.17.20069567v3 – This is a pre-print (not yet peer reviewed), but it is interesting. N95 respirators used to validate tests between 3M surgical masks and homemade cloth masks. There’s a wide range of effectiveness of homemade masks (between 38% and 96% of the 3M surgical mask baseline). Focus was on particle removal (less than 1 micron), but not necessarily viruses. They evaluated 10 different designs for home face masks, so this might be a valuable source for evaluating if your cloth face mask is the best available to you.
• Hamster Paper from Hong Kong University. Seen reports on this one, but can’t find the actual paper (maybe it’s being translated still?). Stay tuned.
• https://theconversation.com/masks-help-stop-the-spread-of-coronavirus-the-science-is-simple-and-im-one-of-100-experts-urging-governors-to-require-public-mask-wearing-138507 – Here’s an article written by a data scientist from UCSF (two cheers for data science in medicine!) with an associated pre-print arguing for masks to prevent asymptomatic transmission. Predicated on coronavirus transmission being largely through droplets, which doesn’t seem to be the prevailing wisdom with SARS or COVID-19 per papers I’ve collected. The first recommendation paragraph of the pre-print is as follows: “Our review of the literature offers evidence in favor of widespread mask use as source control to reduce community transmission: non-medical masks use materials that obstruct droplets of the necessary size; people are most infectious in the initial period post-infection, where it is common to have few or no symptoms (10–16); non-medical masks have been effective in reducing transmission of influenza; and places and time periods where mask usage is required or widespread have shown substantially lower community transmission.” I enjoyed the paper greatly, as the researchers did a very thorough literature search, but it seems like their paper doesn’t necessarily support their conclusion completely. Most of the papers they reference (and the ones I’m showing here) on influenza do not actually show that cloth masks are effective, especially for source control. It may well be true that “places and time periods where mask usage is required or widespread have shown substantially lower community transmission”. The masks are one of infinitely many variables and may simply be correlated with other factors in those societies (i.e., their form of government, the societal structures, economic factors) that may have been more causal of the lower community transmission. One doesn’t know, unless we can control for those variables, and there’s always a danger of trying to solve a problem by addressing the wrong variable. Extremely good paper, though, very thorough, and the main author’s website is solid too.