The Evolving Science of Face Masks and COVID-19

, ,

Since the Centers for Disease Control and Prevention began recommending that members of the public mask up last spring, additional research has backed the use of face masks to combat the coronavirus. While knowledge gaps still remain, experts agree that masks should be used — and increasingly, they are emphasizing the use of better masks.

In our last deep dive into the research on masks in April 2020, we explained that there were some lab studies that supported the idea that masks would be effective against the coronavirus, or SARS-CoV-2. But direct evidence that face masks prevent transmission of respiratory viruses in a community was limited.

In some ways, that basic takeaway hasn’t changed much. Benjamin Cowling, an epidemiologist at the University of Hong Kong, told us that “there is good mechanistic evidence from laboratory studies that masks should have an effect on transmission” and that “evidence from randomized trials has not been consistent with a large effect of masks on transmission, but has been consistent with a small effect of masks on transmission.”

“I continue to believe that mask use will reduce transmission, although mask use alone is not enough to prevent COVID from spreading in a community,” he said in an email.

Some experts remain skeptical that masks — at least the ones used by most people today — have much effect on transmission of SARS-CoV-2, although they still think people should wear them.

Lisa Brosseau, an industrial hygienist consulting with the University of Minnesota’s Center for Infectious Disease Research and Policy, doesn’t discount face coverings entirely, but cautions against relying on them. And she thinks people — especially workers potentially exposed for many hours while on the job — should have access to improved ventilation and other preventive measures, along with significantly better masks.

“We, as a society, have been wearing leaky masks,” said John Volckens, an aerosol scientist at Colorado State University who has studied mask performance in response to the COVID-19 pandemic. “And leaky masks provide much less protection than high-quality, good filter, well-fitting masks.”

The CDC issued new recommendations in February to encourage people to improve the fit and filtration of their masks. Drawing on the findings of lab experiments, the agency suggested layering a cloth mask over a disposable one — a form of “double-masking” — or using a mask fitter or brace, among other options.

Newer Research

Over the last year, additional research has generally supported the notion that face masks can reduce transmission of the virus, although proof is still lacking. 

Numerous lab studies, for example, show that masks can partially block exhaled respiratory droplets, which are thought to be the primary way the virus spreads — and may offer some protection to the wearer.

In one study, scientists at the CDC’s National Institute for Occupational Safety and Health tested a variety of face coverings for their ability to prevent the outward spread of particles from a simulated cough. N95 respirators performed the best — blocking 99% of the particles — while medical masks blocked 59% and a cloth mask blocked 51%. The only covering that failed to do much of anything was a face shield, which stopped just 2%. 

In another experiment, researchers in Japan evaluated how well different masks on two mannequins that faced one other reduced exposure to the coronavirus. Cotton or surgical masks on the mannequin releasing the virus cut the amount of exposure to the other by 50% or more. If only the exposed mannequin wore such a mask, the protective effect was smaller, but if both wore a mask, transmission decreased by 60% to 70%.

A few epidemiology studies also point to face masks helping to reduce spread of the coronavirus. According to a study of 124 households in Beijing, China, mask-wearing prior to when an infected family member fell ill was associated with a 79% lower risk of spreading the virus to others in the household. 

And in an analysis of more than 1,000 people in Thailand who had been in contact with someone with COVID-19, those who reported wearing a mask at all times during that contact were 77% less likely to become infected than those who did not wear masks.

Other indirect evidence comes from studies that have documented associations between self-reported mask wearing and control of the virus in a community, or the implementation of a mask mandate and a subsequent decline in COVID-19 cases or rates of hospitalization.

One study of 15 states and Washington, D.C., identified a 0.9% drop in the daily growth of COVID-19 cases in the first five days following the mask mandate, which grew to a 2.0% daily decrease after 21 days. 

In Kansas, which instituted a mask mandate in July but allowed counties to opt out, the number of new COVID-19 cases per capita fell by 6% in areas with mask mandates, but rose by 100% in those without.

Another study in the U.S. did not identify an impact of a mask mandate, but found that a 10% increase in self-reported mask usage was associated with a state being 3.5 times more likely to have control of its epidemic.

One potentially telling example of the power of masks comes from a report about two hair stylists in Missouri. Both had symptoms of COVID-19 but wore masks and are not known to have spread the virus to any of their clients, who also were masked. Of the 139 total customers, 67 agreed to be tested for SARS-CoV-2, and all came back negative.

Each of these studies has limitations. In the hairdresser study, for example, there’s no way to know if masks were the reason why none of the hairdressers’ clients subsequently developed COVID-19 — and it’s possible some of the people who declined to be tested did in fact contract the virus from those interactions.

Similarly, the studies reporting associations between mask use or mask mandates and better COVID-19 metrics also can’t show that masks necessarily drove or contributed to those improvements.

Dr. Roger Chou, a professor of medical informatics and clinical epidemiology at Oregon Health & Science University, told us that while such studies are suggestive of a benefit, it’s “difficult to determine causality from these studies” because it’s hard to control for other factors that impact infection rates, such as physical distancing or other public health measures.

The lab-based studies, too, can point to how and why masks would work to cut transmission of the coronavirus, but may not reflect the performance of masks of various quality against SARS-CoV-2 in the real world.

Still, for many scientists and public health agencies the data are convincing. Writing in the Journal of the American Medical Association, two CDC scientists described the evidence as “compelling.”

“When you put it all together,” Volckens said, “it’s really hard to refute that masks aren’t a vital piece of infection control.”

A man wears a double mask as he walks in Times Square on April 6 in New York City. (Photo by Kena Betancur/Getty Images)

Chou, who is also the director of his university’s Pacific Northwest Evidence-based Practice Center, has co-authored a series of updated rapid reviews of the evidence on face masks for preventing transmission of respiratory viruses, including the coronavirus. While his reviews have concluded that the traditional epidemiological evidence remains limited — similar to other reviews by the World Health Organization and the European Centre for Disease Prevention and Control — he nevertheless said there was a consensus that masks do work.

“I have pretty high confidence that masks help protect wearers against infection and also likely provide protection from source control,” Chou said, noting that despite differences in the way various agencies describe the evidence, he was unaware of any regional or country authority that doesn’t recommend masks at this point.

“[T]here may be some disagreements about when to wear masks and what types of masks, but whether to wear masks or not when you are in close proximity to other people is not really a question,” he said.

One important study that both Volckens and Chou pointed to is a randomized controlled trial of face masks in Denmark — the first and so far only such study conducted during the COVID-19 pandemic.

The trial tested whether giving free surgical masks to volunteers reduced the likelihood that they would contract SARS-CoV-2, as compared with a control group that was not given masks. It failed to identify a statistically significant effect in the mask group.

But as SciCheck has explained before, the study was only designed to detect a large effect of 50% or more — and the results suggested a smaller, 18% reduction in risk. (The study, notably, also didn’t try to measure whether masks helped prevent spread of the virus from wearers to others.)

“The estimate was imprecise and not statistically significant, but it was consistent with a protective effect,” said Chou. “This is not large, but for a disease like SARS-CoV-2 that spreads exponentially it is important from a public health perspective.”

Cowling, the University of Hong Kong epidemiologist, told us at the time that a 20% protection level is right around what he would expect, based on past research on influenza.

Volckens added that the blue surgical masks used in the study are quite leaky — and that if better masks had been worn, he suspects the results may have been more definitive. “They didn’t even use a great mask,” he said, “so it’s not surprising that they didn’t see a significant effect.”

Moving Toward Better Masks

Not everyone is convinced that face coverings have done much to limit the spread of the coronavirus. Brosseau, a retired certified industrial hygienist with expertise in respiratory protection against bioaerosols, is critical of many of the lab studies that have come out in the last year, which she says have not used rigorous methods. 

“Many of the studies that I’ve read that claim that they get very high filter efficiency on face coverings were not done correctly,” she said in a phone interview. “They’re done at low flow rates — the flow rates that people breathe at. But that’s actually not what you need to discriminate between a good filter and a bad filter.”

She’s also left unswayed by the epidemiological data, which are only correlational.

Brosseau’s primary concern is that people, particularly workers, are not aware of how deficient their face coverings may be, likening the usage of subpar masks to wearing a bikini outside in 10 degree weather.

“I don’t say don’t wear them, I just say don’t expect them to do much,” she said. “They are not as magical as everyone thinks they are.”

Cloth masks might offer some benefit for people running into a grocery store for five to 10 minutes, but for workers who are there for hours, Brosseau thinks they do little to capture or filter out the smaller respiratory particles, or aerosols, that can linger in the air and build up. That’s why she thinks more attention needs to be paid to better ventilation systems and better masks — ultimately, N95 respirators for workers and other standardized masks for the public.

On that front, ASTM International, an organization that publishes standards for personal protective equipment, released a standard for barrier face coverings in mid-February. Brosseau said it’s too early for companies to have made masks that conform to the standard, but is hopeful that it will lead to improved designs.

Volckens, the Colorado State aerosol scientist, doesn’t share Brosseau’s skepticism about face coverings, but agrees that it’s past time for people to be using better masks. 

“We haven’t been wearing enough good masks,” he said in a phone interview. “It really just comes down to two things: Masks that aren’t great filters and masks that are leaky. And by leaky, I mean they don’t fit to your face and so the air doesn’t go through the mask — the air goes around the mask.”

His experiments have revealed wide variability in mask performance, not only between different kinds of masks, but also in how a given mask fits or how much or how loudly someone speaks, which will vary from person to person.

As for the CDC’s new recommendations to “double mask” by placing a cotton mask or mask fitter over a medical mask, Volckens said the agency’s tests were “pretty compelling,” even if the scientists only tested a few masks and didn’t use human subjects.

“I think it confirms what many aerosol scientists have known for months, and that is that we’ve been wearing leaky masks,” he said. “Aerosols behave more like smoke. And no one would think that wearing one of those blue surgical masks would keep all the smoke that you would exhale if you were pulling off a cigarette.”

The idea behind the recommendation, he said, is not really that you need a second layer, but that you need something that holds down the disposable mask so it fits better.

Brosseau cautioned that adding too many layers can increase the breathing resistance — something that wasn’t measured in the study. That can make it harder for someone to breathe and more likely that a mask would leak on the side. Short of a respirator, she said adding a mask brace or fitter would be one of the better options, since that would provide a better seal without including a potentially problematic extra layer.

As more is learned, scientists say mask guidance could shift again in the future. Volckens, for example, said it’s not yet known how good is good enough when it comes to masks. And there are still many remaining unknowns about how the coronavirus spreads, including how much virus it takes to infect a person and how that might depend on particle size or the exact route the virus uses to enter the body. “Without that information,” he said, “it’s really hard to make risk-based decisions.”

Editor’s Note: does not accept advertising. We rely on grants and individual donations from people like you. Please consider a donation. Credit card donations may be made through our “Donate” page. If you prefer to give by check, send to:, Annenberg Public Policy Center, 202 S. 36th St., Philadelphia, PA 19104.

Leave a Reply

Your email address will not be published. Required fields are marked *