COVID-19: Do Masks Work?

by Raywat Deonandan, PhD
Epidemiologist & Associate Professor
University of Ottawa
(I add my credentials to these COVID-19 blog posts in case they get shared. I want readers to know that my opinion is supposedly an educated and informed one)

As I write this, much of the Northern hemisphere has been struck with a wave of multiple childhood respiratory diseases, COVID-19, RSV, and influenza prime among them. Pediatric hospitals are being overrun. Right now in Ontario, we are beyond 100% capacity, with (I’m told) some ICU beds doubling up kids:

One proposal to help slow the traffic into the ERs is to reinstate masking in most, if not all, public indoor settings. This possibility has triggered that tired debate of (a) whether masks work at all, and (b) whether they are harmful at all.

In this post, I will summarize the evidence for each of those questions. Want a quick preview? Of course masks work. And no, they have almost zero harms. If you still object to them, that is your right. But be aware that that objection is based on ideology or comfort, not on the scientific merits of masking.

What Does “Do They Work” Mean?

Let’s first begin with the question, what does it mean when we say, “Masks work”?  I think a lot of people think in terms of absolutes, especially when it comes to individual experiences.  “I wore a mask and I still got COVID.”  Sure, but what about *most* people who wore masks? Most will be well protected, but not perfectly so. Everything has a failure rate.

“My kid wears a mask to school, but takes it off to each lunch. So what’s the point?”  This is a common complaint. The point is that for a significant portion of your child’s exposure time, he or she is well protected. Not perfectly so, but to some degree.

The problem is that people visualize masks like scuba gear, and respiratory viruses like the surrounding ocean. If you have to take off your scuba gear for meaningful durations of time, you’ll drown, so what’s the point of packing scuba gear anyway?

Or it’s like a bullet-proof vest while you’re walking through a storm of bullet fire. If you have to briefly take off your vest to, I dunno, eat a sandwich (with your nipples?), then you’re going to get shot and killed. So why even bother with the vest?

It’s a faulty analogy. Mask wearing is not a panacea. Each moment you wear a mask, you reduce the probability of (a) transmitting infection to others and (b) receiving infection yourself. That reduction is never to zero. And the longer you stay in a high exposure environment, the more likely the mask will prove ineffective.  Consider this handy chart, compiled in the pre-Omicron era, so the duration times might not apply now, but the lesson still holds, that both the quality of the mask, who’s wearing it, and duration spent exposed matter:


So how then do we define success in masking, if it’s not for preventing personal illness?  Well, success is slowing community transmission.

So even if you personally got sick, that sucks, but did the overall levels of transmission in the community fall because enough people wore a good enough mask enough times?

In other words…. repeat after me… don’t let perfect be the enemy of good enough.

Are Masks Harmful?

Despite much pontificating and theorizing on this subject, there is zero evidence of any measurable harm caused by masks to either individuals or populations.

One review concluded, “In otherwise healthy individuals, wearing masks, even for an extended period of time, does not produce any clinically relevant changes… [in gas exchange].” They did point out, though, that people prone to headaches have a slightly higher risk of headaches with prolonged mask use.

Another concluded, “surgical and cloth masks did not significantly compromise ventilation and oxygen supplies in healthy individuals and may, therefore, be considered as not harmful.”

Other experimental studies showed, for example, that while wearing a mask, people have a harder time identifying faces. Well duh. That’s why bank robbers wear masks. But there was no evidence that this impaired the ability to identify faces once the masks were removed. So any arguments about impaired cognitive development of children is speculative and not backed by any experimental or observational evidence.

It’s also kind of silly since children see unmasked faces at home and on TV regularly.

One real negative consequence, though, is the tendency for masks to trigger acne and related conditions among those prone to them. And for those who must see mouths to enable lip reading, for example, masks can be an impediment. So for specific circumstances, exceptions and accommodations must be applied.

So, Do Masks Work?

There is a mountain of both observational and experimental evidence showing that masks work to slow and prevent COVID-19 infection and transmission, and that mask mandates work to increase mask wearing.

(a) Evidence that Masks Prevent Infection and Transmission

The number of laboratory and simulation studies showing the high effectiveness of various types of masks is considerable. But most people are only convinced by population-level investigations. While there are numerous studies to this effect, I will only mention two of them:

A case-control (652 cases and 1,176 controls), test-negative (i.e. observational) study from the end of 2021 in California saw that wearing an N95/KN95 respirator or wearing a surgical mask was associated with lower adjusted odds of a positive test result compared with not wearing a mask. Wearing a cloth mask also had a good effect, but the finding was not significant. In short, those who didn’t wear masks had more than double the risk of testing positive for COVID. The following graphic summarizes their finding:


Possibly the most important study is a cluster-randomized trial in Bangladesh, published earlier this year. They looked at 160,000 to 170,000 people in each trial arm. The intervention arm sought to actively encourage mask wearing, while the control arm did not. The bottom line is that the incidence of reported COVID-like symptoms (since the cost of testing everyone would be too onerous) decreased by 11.6%.


(b) Evidence that Mask Mandates Work to Reduce Community Transmission

Recently, a bombshell study out of Boston showed that “the lifting of masking requirements was associated with an additional 44.9 COVID-19 cases per 1000 students and staff during the 15 weeks after the statewide masking policy was rescinded.” In other words, the mask mandate in schools was likely holding back infection at a very high rate.  As with all observational studies, this is correlational, so a causal link can be questioned; but it seems likely.

A nationwide analysis of masking mandates in the USA concluded that, “the daily case incidence per 100,000 people in masked counties compared with unmasked counties declined by 25% at four weeks, 35% at six weeks, and 18%” across all six weeks post-mandate.

In its own review from earlier this year, Public Health Ontario concluded, “the implementation of mask mandates has been associated with statistically significant reductions in COVID-19 case growth, hospitalizations and deaths.”

Whereas, a Danish RCT found that simply recommending mask use in an environment where no recommendation was otherwise offered, did not result in a statistically significant reduction in COVID cases. Clearly, whether recommendations can work as well as mandates is a cultural issue. Some populations, like Japan, comply readily with public health recommendations. Other societies require the force of law to reach compliance levels necessary to enact measurable public health effects.

It would be nice if simply asking people to drive safely were enough. But unfortunately, many people require the threat of speeding tickets and fines to keep our highways safe. Same goes for things like mask wearing.


If Masks Work, Why Do Some Say They Don’t?

Beyond ideological reasons to doubt the effectiveness of masks, there are in my opinion two scientific reasons that those opposed to community masking consistently cite.

The first is the Cochrane review titled, “Do physical measures such as hand-washing or wearing masks stop or slow down the spread of respiratory viruses?” which concludes, “We are uncertain whether wearing masks or N95/P2 respirators helps to slow the spread of respiratory viruses.”

But that review was conducted in the Fall of 2020, well before the flood of new studies were published. It is based almost exclusively on influenza transmission, not COVID, which actually might matter a great deal, due to the differential transmission dynamics of the two diseases. Importantly, “we are uncertain” in no way means, “masks don’t work.”

Indeed, subsequent  and more targeted systematic reviews, like “Face Mask Use in the Community for Reducing the Spread of COVID-19: A Systematic Review” concluded, “the findings…  support the use of face masks in a community setting,” based on their computation that masking would reduce the reproduction number of the epidemic.

You will recall that the reproduction number is the average number of new infections created by a current infection. The goal is to reduce that number to one or less, using a variety of tools, including vaccination, distancing, and mask wearing.

The second argument that anti-maskers use is to cite the aforementioned Bangladeshi study, the one that found an 11.6% reduction in COVID symptoms. Their position is that 11.6% is piddling and not worth the effort.

Nuh-uh. 11.6% can be –and often is– the difference between exponential growth and a contained epidemic.  11.6% is nothing to sneeze at. Pun intended.

It’s also worth noting that in the Bangladesh trial, the control arm had 13.3% mask usage, while the intervention arm had 42.3% usage. That’s only a difference of 29 percentage points. So an 11.6% outcome difference is pretty impressive considering that the control and intervention arms weren’t so different.



Masks work. Of course they do. They work as imperfect personal protection and as highly impactful community protection.

It would be nice if strong recommendations resulted in sufficient mask compliance. But absent that, the evidence clearly shows that mask mandates result in greater compliance.

With respect to the current crisis of overloaded pediatric hospitals, the goal of mask wearing is not to solve the problem outright (though that would be nice). Rather, it’s to slow the traffic into the Emergency Rooms, so that when your child needs care, they can get it.

An 11.6% reduction in cases would certainly help.


UPDATE: I’d like to add links to two other treatises on mask effectiveness from highly competent statisticians:

  1. The science behind masks and their use” by Bill Comeau
  2. The Masks Masquerade” by Nassim Nicholas Taleb