Edit 2: Updated 5/04/2020 as the US CDC has shifted their recommendations in light of new evidence to recommend the use of cloth face-masks in public situations. The World Health Organization has changed their stance slightly, although remain largely neutral on face mask wearing. The article has been updated to reflect the new evidence — this is still largely about the inconsistency addressed in point 3.
Edit: Updated 2/04/2020 as the lead author of the only RCT on cloth masks has described their use as a “last resort”, given the plausible evidence from this trial that they may in some cases increase the risk of infection and/or transmissibility of the virus from people who themselves are sick.
Note: at the outset, I will admit that I am primarily a chronic disease epidemiologist. While I am happy to use my expertise to explain the evidence to you, the final word as ever should come from the experts, such as your local health authority or the World Health Organization.
In times of coronavirus, nothing is more contentious than the question of masks. There are dozens of conflicting opinions online, from the poorly-researched yelling on twitter to the much more informed arguments on, well, also twitter, as well as numerous articles published in almost every media source in the world with an opinion.
We’ve heard everything from “you don’t need a mask” to the wonderful “they might be better than nothing” from the New York Times, which is a beautifully encouraging message about the evidence for mask use.
So it makes sense that most people are still confused. Will a mask protect you? Are they necessary for average people to wear in their day-to-day life? And, most importantly, if you don’t wear a mask, what should you be doing to help prevent the spread of the novel coronavirus?
One thing to mention at the outset — the word “mask” covers a range of sins. I’ll mostly be talking here about two specific types of mask: surgical masks (of the kind you often see worn incorrectly on TV shows) and N95/P2 respirators, which are secure tight-fitting devices that filter pretty much all air when used properly. In fact, I’ll almost entirely be talking about surgical masks, because there’s definitely no need for non-healthcare workers to be using respirators no matter how fancy they are — they need to be properly fitted, worn perfectly, and are mostly needed for specific medical procedures rather than your day-to-day existence.
Now, there is a lot of evidence out there, and it is extremely dense. The reality is that this question isn’t as simple as an easy yes or no, despite everyone’s fierce desire that it should be. But, since everyone needs some sort of take-home from the articles they read, here are the key points that I’ll be going through:
- Masks provide a barrier to droplets caused by coughs and sneezes
- Healthcare workers use masks because they care for people coughing and sneezing a lot
- Masks may prevent some transmission FROM people coughing and sneezing
- However, they PROBABLY DON’T help prevent transmission to people in non-healthcare environments
- Therefore, you probably don’t need a mask unless you’re sick yourself
As I said, this is all very complicated, so if you read nothing else have a look at the list above. The basic idea is that masks are absolutely protective for people who are in regular contact with respiratory droplets, such as healthcare workers caring for sick people. The problem is, out in the ‘real world’ that we all live in, the evidence is pretty clear that masks probably don’t do very much to prevent transmission. This is why the World Health Organization (WHO), and a number of other health authorities around the world, don’t recommend mask use for the general public — only those who are taking care of sick people.
So, all that being said, let’s dive into the evidence.
Hopefully we can all make it out of the comments section alive.
The first point — and the one that has everyone so mad at the WHO and CDC — is that masks absolutely do work to prevent infection in certain circumstances. This is the point that every angry person out there has been yelling for the last few weeks: “If doctors wear masks, why shouldn’t I?”
And it’s a fair question, because masks absolutely do prevent droplet-borne disease. The mechanism is very simple — people coughing and sneezing create a cloud of droplets in the air in front of them. Even speaking and singing creates such a cloud, although it may not be as big. This cloud is very short-lived — it usually only lasts for a fraction of a second — but can be breathed in and cause infections for people in the line of fire.
It’s worth noting that this doesn’t mean that these diseases are ‘airborne’, at least not in the colloquial sense. While there are pathogens that can remain viable in the air for some time — notoriously tuberculosis, which can remain infectious for hours in the air — most respiratory diseases similar to COVID-19 are spread through these tiny droplets that fall down or disappear after no more than a second or two.
All of this makes the basic concept of face masks very simple. If you can get infected by breathing in viral particles carried in droplets, you can potentially prevent the infection by wearing a physical barrier that will stop the droplets carrying the virus.
And there is certainly evidence that this works. One systematic review from 2015, for example, found 23 studies that used various methods to test whether masks prevent viral particles from getting through surgical and N95 masks, and found that although generally N95 masks were the best, surgical masks did an ok job as well.
Which brings us to the main conclusion of point 1:
There is ample evidence that masks prevent you breathing in respiratory droplets from someone coughing or sneezing directly at you, as is commonly seen in healthcare scenarios.
This brings us neatly to the second point, which concerns healthcare workers. These are people taking care of those who are constantly coughing and/or sneezing, because they are sick. It stands to reason that, if a mask prevents respiratory droplets from getting on your face — your mouth in particular — then wearing a mask when someone’s coughing droplets at you should prevent infections to a point.
And, if we look at the evidence, that’s precisely what we find.
Even outside of pandemics, healthcare workers are exposed to respiratory diseases, so there is actually quite a bit of evidence looking at the protective effect of masks. A systematic review in 2017 found that both surgical masks and respirators reduced the risk of bacterial and clinical infection — although interestingly not viral — by more than 40% for clinical staff. A more recent systematic review by the prestigious Cochrane collaboration found that both surgical masks and respirators are equally effective in clinical settings at reducing the risk of infection for professionals.
So, while the benefits for confirmed viral infections are not quite as obvious, there is certainly evidence that masks can help to prevent the spread of infectious diseases to healthcare workers. There’s even some evidence that they prevented the spread of SARS-CoV-1, which caused a pandemic in the early noughties, which makes it likely that they will work for SARS-CoV-2, the pathogen causing our current crisis.
The one caveat with this is that mask-wearing isn’t a behaviour in isolation — while the evidence shows masks are associated with reduced infection rates in healthcare workers, even in randomized controlled trials it is hard to tell if this is specifically because masks block particles or because when you wear a mask you do lots of other things as well. You have to clean your hands more often and more thoroughly, for example, and pay more attention to your hygiene.
However, even with this caveat, it is fair to sum up point 2 like this:
Masks are almost certainly effective at preventing healthcare workers from catching respiratory infections while caring for sick people.
Wearing A Mask When YOU Are Sick
The third point is arguably the most contentious, because there really isn’t much evidence out there to either back up or refute it. The argument that people who are sick should wear masks is mostly down to the theory, rather than actual gold-standard clinical trials.
Which is good to point out, because it’s worth noting that all of the other points we’re talking about here are supported by really good evidence. While some people will read different things into study results, it’s important to remember that this isn’t speculation — you can follow the links yourself and read up on all of these things. I encourage you to do so.
However, that being said, the rationale for people who are sick to wear masks is fairly simple. While I disagree with a lot of this article — you can’t just chuck out all the evidence you don’t like to focus on theory, that’s ridiculous — it does give you a very good overview of why a sick person wearing a mask might stop some infections.
Basically, if you wear a mask, and then cough or sneeze, you don’t spread your disgusting, virus-filled droplets everywhere. If everyone who had any symptoms did this, it might prevent some infections, as their nasty particles wouldn’t get all over our nice, clean stuff.
The challenge with this is that this isn’t necessarily the primary way that these diseases make their way around the community. What’s likely much more common — and, by most indications, the main way that COVID-19 is spread — is someone coughing into their hand and then touching surfaces, or coughing directly onto those surfaces, where the virus can live for a long time. Someone else then touches the surface, doesn’t wash their hands, and goes on to touch their face, which causes an infection. Arguably the best simple explainer of this phenomenon is from the TV show Scrubs, season 5 episode 12 “My Cabbage”.
Some people are pointing to countries like South Korea and Singapore, where mask-wearing is common, to try and support the argument for wearing a mask when you’re sick, because they controlled their epidemics very quickly. Unfortunately, this is a terrible argument, because those countries also did lots of other things that had nothing to do with masks. It also ignores other countries like China, where mask-wearing doesn’t appear to have done much for the spread of COVID-19, and Iceland, where the disease is being controlled impressively well without much mask wearing in the general population.
As ever, correlation does not equal causation.
That being said, point 3 can be reasonably summed up:
There is some rationale for mask-wearing for people who are currently experiencing symptoms of COVID-19, in particular coughing, even in the general population.
Masks Probably Don’t Prevent Infections In Community Settings
So, you’ve been reading along, nodding your head, but now you’ve arrived at the contentious part. The part that has caused more rage to be thrown at the CDC since they recommended against the consumption of raw milk.
Strap yourselves in, it’s going to be a bumpy ride.
We know there’s a plausible theoretical explanation for why masks should work. We know that healthcare workers who are trained properly in using masks and are exposed to lots of respiratory droplets are less likely to be infected if they wear masks. We have even seen why there’s some rationale for sick people in the community to wear masks, although the evidence is lacking.
So why does the WHO — and many other public health bodies worldwide — not recommend mask usage for the general public?
To answer this question, we have to look at what happens when you actually give people masks, teach them how to wear them, and see what happens. In other words, if we look at randomized controlled trials, where two equivalent groups of people are randomly allocated to either have masks at home or not during epidemic or pandemic disease (mostly influenza), and compared on whether they were more or less likely to be infected when given masks/training.
And when we look at these trials, the answer is fairly consistent: masks don’t appear to reduce the risk of household or community transmission of infectious respiratory diseases.
Let’s unpack this a bit. A number of trials have given people masks, trained them to use the masks properly, and seen whether these people are protected from respiratory infection compared to a control group with no masks. A systematic review from 2015 looked at this question, and found that if you look at everyone who was given a mask compared to those who weren’t, the masks didn’t help to prevent any infections at all. Similarly, the largest study so far, conducted here in Australia, found the same thing — the mask-wearing group had just as many infections as the control.
Sounds settled, right? Theory be damned, the evidence shows that masks don’t work!
Well, it’s still not quite that simple. In most of these studies, people didn’t always wear the masks that they were given. This makes sense, because masks are uncomfortable and wearing them all day is rarely going to be an easy ask of people, especially when they are at home. People are also terrible at mask etiquette, and regularly do it wrong even in healthcare settings where everyone has been trained.
So, some of these studies did what’s known as a per-protocol analysis. This is when you look only at a subgroup of your total figure to see if the intervention — masks — were effective for those with high compliance (i.e. people who wore their masks most of the time). By definition, this is of lower quality than the original trial, because you’ve lost the randomization that makes your study so strong to begin with.
And, while not entirely consistent, there seems to be some evidence from these per-protocol analyses that people who wear their masks religiously may have a reduced risk of infection, although the longer the incubation period of the disease, the smaller the benefit. The big study that I mentioned from Australia estimated that with a 2-day incubation period for influenza, the efficacy of mask-wearing for people who wore their masks a lot was roughly halved, which raises some questions about COVID-19, which has a 4–7 day incubation period itself.
Now, people will happily argue that they are the exception*, that they know how to use masks properly, and that they should therefore be the exception to the rule and wear one everywhere. The problem is that we have to use the best evidence, and the best evidence suggests that most people are, in fact, not outliers, and probably won’t be an exception to the rule.
Which brings us to our wonderfully messy conclusion for point 4:
Mask provision is probably ineffective at preventing the transmission of respiratory diseases such as COVID-19 in the community. There is some evidence that it might be helpful for those who are very strictly adherent, but this is of low quality and inconsistent among trials. Thus, current best evidence strongly argues against mask use in the general population.
You Probably Don’t Need A Mask
All of this brings us to our last point. Now, I know that everyone really wants to wear a mask. These are scary times, which makes us all need to do something. Some people have taken up exercising in their homes. Some have started gardens, or online forums to discuss COVID-19. Personally, I’ve been writing almost non-stop since the pandemic began, in between reading every article that comes out, frantically hoping for some sort of silver bullet against the virus.
In times like these even the most minor interventions with low evidence that, at best, will probably have very little impact on disease spread feels necessary. It might not work, but at least you’re doing something, right?
Unfortunately, it is rarely as simple as that.
Masks are extremely low-risk, but not entirely benign. The recent call for home-made masks, for example, is a bit worrying because there is some evidence that fabric masks can counter-intuitively increase your risk of infection.
More importantly than that, however, is the fact that there are only so many masks in the world right now. There are mask shortages across the globe, driven partly by health facilities gobbling them up, but also because consumers have pulled them off the shelves at an alarming rate. Everyone who buys a mask for their own use is potentially denying it to a healthcare worker who really needs it, with pretty much no benefit to themselves.
And this is why the WHO and other organizations don’t recommend mask use. There may be some benefit, but even if there is it’s likely to be very minor and only for a small subsection of the population who wear their mask all the time and adhere to proper infection control measures when they do. And while I know everyone think’s that they are special, the evidence quite strongly indicates that you probably are not.
So, to round off, here’s the summary of point 5:
Most people probably don’t need to wear masks to stop themselves getting sick from COVID-19. If you yourself are sick, it might be a good idea to put one on, particularly if you’re going outside, but given the limited mask stocks this is not true for people who are feeling fine and just want a bit of extra protection.
Wash Your Hands, Don’t Touch Your Face, Practice Social Distancing
Here we are at the end. The final part of the long-read where I round off the piece with wise words that leave you feeling all nice and glowy inside.
Except, it’s a global pandemic, so warm and glowy feelings are hard to come by.
But, while you probably don’t need a mask to help prevent COVID-19 from spreading**, there are absolutely things that you can do to protect yourself and others. I’ve written in more details about these elsewhere, but they basically boil down to three things: wash your hands, don’t touch your face, and practice social distancing. Work from home if you can. If you can’t, pretend that you have COVID-19 and every person you meet is your 90-year old grandmother.
Stay safe, stay home, and remember that we are all in this together.
*If you want to see examples of this, I am certain there will be dozens in the replies as soon as this is posted, so enjoy yourself with a quick look-see.
**Note: some people have suggested mass mask-wearing as a way to prevent asymptomatic spread of COVID-19. This is not entirely ludicrous, because there’s evidence that the disease can spread even from people with no symptoms, and given point 3 it seems somewhat reasonable to assume that this would provide some protection, but even then the effect is not for individuals per se and is likely to be very small.