Nicholas R. Longrich, PhD, Samuel K. Sheppard, PhD
From the Department of Biology and Biochemistry, University of Bath, Bath, BA2 7AY, United Kingdom
author for correspondence: email@example.com
[NOTE: this is a preprint version of a paper submitted for review on March 29, 2020. If you have suggestions for references to add or constructive criticism, please feel free.]
The US and UK governments advise against the use of masks by the public to fight the ongoing Coronavirus Disease 19 (COVID-19) pandemic, and so does the World Health Organization 1. But could they be wrong? The governments of China, South Korea, Hong Kong, and Taiwan 1 all recommend that the public wear masks to slow the spread of the coronavirus. China has ramped up production of facemasks, converting Foxconn factories that once made iPhones to make masks. Taiwan has also ramped up the production of facemasks, and prohibited their export. Both approaches can’t be right. Increasingly, advice against the use of face masks has been questioned 2, including by the head of China’s CDC3.
Common sense, scientific studies, and most of all the success of Asian countries in fighting the coronavirus suggest that masks may make a major difference. There are fewer scientific studies available to guide decision making than we would like, and the evidence is not always clear-cut. However, decision-making in a crisis requires that decisions are made in the absence of perfect clarity. What is clear is that the exponential mathematics of epidemics suggest that even if masks are of limited benefit in reducing infection rates, they could make a large difference over time, potentially slowing the pace of the epidemic, limiting its spread, saving lives, and ultimately allowing countries to restart the economies that their citizens depend on for their livelihoods.
Masks protect you from others, others from you
It would seem sensible to assume that any barrier between two people’s airways would reduce the probability of an airborne virus being transmitted between them. Masks worn by infected people catch some fraction of the virus-laden respiratory droplets that are exhaled during breathing and coughing. Perhaps just as importantly, breathing through a mask slows and deflects air as it is exhaled, potentially reducing the distance viral droplets are carried as aerosols. As an experiment, one can hold some flour in one’s hand, then give it a puff of air, or cough: it flies everywhere. When the same is done with a cotton T-shirt over the mouth, no matter how hard you try, it’s difficult to move more than a tiny fraction of the flour. In slowing the air released from your mouth, a mask limits the ability of your breath to disperse particles.
Meanwhile, masks worn by infected people catch some fraction of the virus they would otherwise inhale. If both infected and uninfected people wear masks, these effects multiply. For example, hypothetically, if an infected person’s mask reduces the amount of virus spread by 75%, and the uninfected person’s mask reduces it by another 75%, then the total reduction of the virus spread would be 94%.
It remains possible that this reduction will not be enough to prevent infection. However, masks could still help protect people who become infected, because dosage matters. Lower dosing of virus means the infection takes longer to build up, giving the immune system more time to mount a response. Higher viral dosage gives the virus a head-start in its race against the immune system, leading to a more dangerous, rapid course of infection. This is exemplified in animal models. For example, mice exposed to lower doses of inluenza virus get less ill 4 than those exposed to high doses – which became more ill and suffer more lung damage. Similarly, in chickens exposed to avian influenza, the higher the initial dosing, the faster the birds become sick and die 5.
The immune system fights viruses, like a farmer trying to remove weeds from his field. How difficult those weeds will be to remove depends a lot on how many weed seeds there are. 1000 seeds might not pose a major challenge 1,000,000 will make it far more difficult to control the weeds. Similarly, even where masks fail to prevent infection, by lowering the initial dose of virus they could make the difference between mild symptoms and a severe illness that requires hospitalization in already over-stretched health care systems.
Masks reduce viral spreading in laboratory and real world settings
Experimental laboratory studies suggest that surgical masks can reduce the inhalation of viral particles an average of tenfold 6, with some masks providing less protection, and some providing more. In the real world, several studies suggest that masks provide protection against the spread of airborne infections, specifically influenza. A study of a hospital in Germany found that after staff began wearing masks continuously, the number of patients contracting influenza in the hospital declined by almost half 7. 7. Furthermore, in a study of German households, use of masks and hand-washing helped prevent the spread of influenza between family members 8. 8. Strikingly, a study of university students found that masks plus hand sanitizer were able to reduce the spread of influenza by up to 75% 9.
Studies of Severe Acute Respiratory Syndrome (SARS) are likely to be particularly informative for understanding COVID-19 because both have a common origin as bat-derived coronaviruses 10, 11, and share broadly similar epidemiology. Within a hospital setting, consistent and proper mask use has been shown to be crucial in preventing transmission of SARS coronavirus to hospital personnel 12. Critically, two separate studies found that frequent mask use in public spaces during the SARS epidemic was associated with a lower risk of infection by the public 13, with one concluding that masks were strongly protective. The authors concluded that “…mask use lowered the risk for disease…” supported community use as a control strategy 14. This study found that intermittent use of masks was associated with a 60% reduction in risk and always using a mask was associated with a 70% reduction in risk of acquiring SARS 14. A systematic review of all retrospective studies of the SARS epidemic concluded that masks were effective in preventing the acquisition of SARS15.
It is hard to overstate the implications of these findings in the context of the current COVID-19 pandemic. The reproductive number of the virus determines the course of an epidemic. When the virus’ reproductive number falls below 1, i.e. each infected person passes the virus on to an average of less than one other person, the epidemic slows, then dies out. A number of estimates have been published for COVID-19s basic reproductive rate, R0, with a mean of 3.28 and a median of 2.79 16. If it’s spread could be reduced by 70% by the universal use of masks, its reproductive rate would become 0.837-0.984. The effective reproductive rate would drop below zero, and the epidemic would cease. This is a simplistic scenario, but could actually underestimate the effectiveness of masks, because it only assumes reduction in the rate at which people catch the virus, not the rate of transmission.
Figure 1. A simple model showing exponential growth in an uncontained outbreak over time (generation time = 7 days, R0 = 2.5) and with small reductions in the reproductive rate R. Interventions with small effects on transmission, especially if applied early, can have a large effect on total number of cases.
Models suggest masks could work to control pandemics
It remains possible that masks may have a more limited benefit, perhaps because they are not as effective as suggested by some studies, because people fail to use them effectively, or because of shortages of effective masks such as surgical masks and N-95s. To understand the potential effectiveness of masks it is important to consider them in the context of the impact of surprisingly small reductions in viral transmission rates. For example, consider how epidemics grow exponentially. Allowed to spread, one case of Covid-19 becomes 2.5 (assuming for this model and R0 of 2.5), each case causing 2.5 more, and so on. Over the course of 15 reproductive cycles, each taking 7 days, or about 3 months in total, one case can become 2.5 x 2.5 x 25… or 2.5^15 = 931,323 cases (Fig. 1).
Suppose that the use of masks cut the growth rate by just 10%. Each person infects 2.25 others, who infect 2.25 others, etc. Over 15 cycles, 2.25^15 = 191,751 cases. An 80% reduction. Understanding this exponential growth explains how the virus caught the world by surprise even as the pandemic was monitored in real time. But another aspect of exponential growth is that small decreases in the exponent greatly slow growth. A 10% increase in the exponent can have a massive effect, but even a limited intervention, with a 10% decrease over time, will pay large dividends (Fig. 1).
These are simple models, but more sophisticated modeling studies show large scale use of masks could mitigate, even suppress pandemics. A 2010 study found that above a certain threshold, widespread use of effective masks in the population was able to reduce the reproductive number (R) of an influenza virus below 1, at which point the pandemic would stop 17. If masks were highly effective (well-designed, used properly and consistently), then widespread use of masks could stop a flu pandemic if used by 50% of people. If masks were less effective, then they would stop the pandemic only when used by more than half of the population; if masks were highly ineffective, then they would flatten the curve of the epidemic, but not stop it 17.17. All of these scenarios would be extremely beneficial in the current outbreak.
A second study 18 found similar results in modeling pandemic influenza. The model found that the pandemic was highly sensitive to the proportion of people wearing masks. Even if 25% or 50% of the people wore masks, the number of cases is drastically reduced 18 and flattens the pandemic’s curve. Critically, masks are far more effective when implemented early, than implemented late in the course of an epidemic 18. The study also found that it was important for both infected and uninfected individuals to wear masks 18. These studies support the results of simple calculations: use of masks can make a major difference.
Real world experience suggests masks work in pandemics
Perhaps the most compelling evidence of the potential effectiveness of masks in the fight against COVID-19 comes from the real world. Specifically, countries that are controlling coronavirus epidemics- China, South Korea, Hong Kong, Taiwan, Vietnam, Singapore, Kuwait, and Japan- use masks (Fig. 2).
Correlation is not always causation. In theory, masks could be a signal of an effective pandemic response- an aggressive willingness by the government and public to do everything possible to control the outbreak- rather than a direct cause of suppression. Yet the diversity of these countries, and their responses, argues against such an interpretation. China, South Korea, Taiwan, Vietnam and Singapore differ greatly in their political organization, ranging from communism to democracies, and in their level of economic development. And strikingly, these countries also differ in their suppression strategies. China implemented a lockdown of Wuhan, shut down industry nationwide, implemented temperature checks and social distancing, tested extensively— and employed masks. Korea responded with aggressive testing and contact tracing—and masks. Japan has done far less extensive testing than Korea, but shut down schools and large gatherings— and used masks. The pandemic management strategies used by these countries far more diverse than has been appreciated. Arguably one of the few things all of these successes share is the widespread use of masks. And on the other hand, one common factor shared by the pandemic suppression strategies of the US, Canada, the UK and Europe is the decisionnot to use mask on a large scale, and to discourage the use of masks by the public. These patterns suggest that masks may be an important element of the most successful suppression strategies, but we currently lack the evidence to prove this hypothesis.
Figure 2. Western countries (US, Canada, UK, Europe; yellow, red and orange) versus countries and territories using masks as part of official or de facto government policy (China, South Korea, Japan, Hong Kong, Taiwan, Vietnam, Thailand, Kuwait, in blues).
What kind of mask? Surgical masks as good as N-95s, and improvised masks are better than nothing.
If masks are helpful in preventing the spread of respiratory disease, then which masks are effective? Surprisingly, surgical masks appear to be as effective as respirators in preventing infection. One study randomly assigned surgical masks or N95 respirators to nurses 19; both got influenza at about the same rate. A study of a hospital in Singapore during the 2009 Swine Flu outbreak showed the same 20. Another study tried to isolate influenza virus from infected patients wearing a surgical mask or an N95. Both were equally effective in stopping viruses 21. A recent study randomly assigned health care personnel to wear respirators or surgical masks. Those using respirators fell ill at a slightly higher rate than those using surgical masks (8.2% versus 7.2%) but the difference was not significant 22. Similarly, a systematic review found “limited evidence” that respirators were superior to surgical masks during the SARS epidemic 15.
This has important implications given current shortages. Respirators and surgical masks are in short supply, but surgical masks are cheaper and simpler, which should make it easier to accelerate production. Surgical masks are more comfortable and so compliance might be better, improving effectiveness.
If simpler masks can be effective, this raises the possibility of using improvised masks, homemade, or made in factories, to fill the gap. Would improvised cloth masks work? Research into the effectiveness of cloth masks is limited 23. The research that has been done suggests that homemade masks are inferior to surgical masks, but that they are better than nothing. One laboratory study found a homemade mask was half as effective as a surgical mask in filtering particles 24. Another laboratory study found that homemade masks made from a variety of materials stopped virus aerosols, but not as effectively as surgical masks 25. A surgical mask stopped 90% of viral aerosol particles, a dish towel, 72%, linen, 62%, and a cotton T-shirt, 51% 25. Improvised masks made out of cotton fabric would presumably perform similar to a T-shirt or linen, letting through about 3-5 times as many viral aerosol particles. The only real-world study of cloth masks found that they were less effective than medical masks, consistent with would is expected from the laboratory studies 26. Unfortunately, this study failed to have a proper control – a no-mask group – and so it cannot be used to argue that cloth masks don’t work, because the alternative – no masks – wasn’t evaluated.
Finally, another alternative would be the use of non-medical masks. Although not specifically researched in the context of preventing viral transmission, studies of non-medical masks, including dust respirators and cycle masks, showed that all of the dust respirators evaluated were superior to surgical masks in filtering fine particles, and two of the three cycling masks were comparable to a surgical mask in performance 27.
Clearly, improvised or non-medical masks should only be used when access to N-95 respirators or surgical masks is impossible. However, the speed and spread of the current pandemic have created a widespread shortage of respirators and medical masks. This suggests the need to implement value engineering 28 of the sort that dealt with materials shortages in World War II, and specifically to identify ways to produce filtration comparable to that of medical masks with cheaper, more easily sourced materials and production techniques, and also to find new ways to sterilize, reuse, and/or recycle masks 29.
Arguments against masks don’t hold up
The public doesn’t need them because doctors and nurses do.
The argument that the public doesn’t need masks because doctors and nurses do is logically inconsistent 30. Both can’t be true: masks can’t work for doctors and nurses and be vital to protect them, but fail to be useful to the public. A more logical argument is that doctors and nurses need masks more than the public. This may well be true, and it is of the upmost importance to protect frontline public health and care workers. However, this argument assumes that wearing masks is a purely defensive act. In fact, wearing a mask protects not only the wearer against infection, but also reduces the potential for onward spread from an infected individual. A mask stopping a chain of infection or a super-spreading event, could save many lives, members of the public and also doctors, nurses, and heart attack victims and cancer patients who might otherwise die if health care systems are overwhelmed. Particularly if worn by people who are otherwise likely to either catch or spread the virus, public use of masks will reduce the number of people coming into hospitals. The argument that no one else needs masks isn’t backed up by modeling or real-world evidence. Clearly, ensuring access to masks by health care workers must remain a priority, but it makes sense to issue them to others as well.
Masks are only needed to prevent infected people from spreading the virus. Another argument is that masks are only needed by infected people. The CDC officially recommends using masks by infected persons to prevent influenza and has made the same recommendation for coronavirus. The problem with COVID-19 is that it is difficult to know who is infected. Around half the people spreading show no symptoms, and do not know they are infected 31. This is the point of the lockdown strategy; if it was known who was infected, these individuals could be isolated. Because we don’t, we are forced to act as if everyone is. The same logic can and should apply to masks. In an uncontrolled pandemic, it is logical to act as if everyone is infected; everyone wears a mask. It’s an inconvenience, but less of an inconvenience than shutting down shops and pubs, canceling events, staying away from work, and forced quarantine.
We don’t know enough to act. Admittedly, there are many unanswered questions about COVID-19 and masks, and more studies are needed. But it has been argued that we can’t really do anything because we know so little. This isn’t logical. If one wakes up in the middle of the night to find the house burning down, there are a number of unknowns: how did it start? What’s on fire? How fast is it spreading? Where is everyone? Despite these uncertainties, one immediately takes action. The same applies in a pandemic; public health workers must act with incomplete data, with gaps and inaccuracies in their intelligence, but never the less must make difficult, life-or-death decisions, because failure to make a decision is often fatal. In a crisis situation, one does not wait for studies.
Asymmetrical gambles. Masks provide a potential upside, with limited downside. Widespread use of masks might help a little, or a lot, but they are unlikely to do much damage. Fears that masks would inspire ‘false confidence’ in their protection are exaggerated. Individuals wearing masks now tend to be those most aggressive in protecting themselves, their families, and society. Far from inspiring risk-taking, masks are likely to send a powerful social signal that the threat is real, and that we need to rapidly and collectively change our behavior.
Masks provide a high potential for an upside with no risk, and assuming they can be produced on a large scale, limited cost. Even assuming it is a gamble, it is arguably a rational gamble to take. The situation is much like Pascal’s Wager. Pascal argued that we should believe in God because if He existed, you’d go to heaven, and if not, one had nothing to lose by believing. Similarly, strategies with uncertain upsides and no downside are worth pursuing, particularly if there is a relatively inexpensive opportunity to help reduce the spread of COVID-19. The alternative, to wait for certainty, would seem a poor strategy to addressing an existential pandemic threat. That being said, as we have laid out above, the evidence for the efficacy of masks is far stronger than has been appreciated.
Strong evidence and arguments exist in favor of the widespread use of facemasks by the public. The principle behind facemasks- they reduce the amount of virus exhaled by infected people, and inhaled by uninfected- suggest that they should be a primary tool in combating a respiratory virus. Scientific research, including experimental studies, retrospective studies of the SARS epidemic, and modeling studies, suggests that they can be effective. Most importantly, the experience of countries that have used masks against SARS and the current coronavirus pandemic imply that they are likely to be highly effective when used by the public. The current shortage of medical masks is the only valid argument against their widespread use. More effort needs to be made to increase the production of facemasks, to find alternatives to surgical masks and respirators, and alternative materials and methods to manufacture facemasks on a large scale. Until this is possible, the use of homemade and improvised facemasks should be explored to protect individuals, and the public. We need more information on the effectiveness and proper design of facemasks, and emergency funding for research into this and other strategies to combat the coronavirus pandemic are needed to allow for informed decision making.
Box 1. Recommendations.
- Undertake research into the effectiveness, design, and use of masks
Make funds immediately available to scientists, engineers, and doctors. The granting process should follow a model in which funding rates are high, and approval is rapid. Following the DARPA strategy, large numbers of small grants should be made available, with successive rounds of funding contingent on success.
- Research should focus on the following
- effectiveness of masks
- design of effective masks, including alternative, easily sourced materials
- sterilization and recycling of masks
- feasibility of reusable masks
- modeling studies for rational strategies for optimal mask allocation and usage to maximize lives saved per mask.
- Increase production of masks
- Convert existing domestic factories to mass production of masks.
- Encourage small-scale manufacture and sale of masks at home and by small businesses, assuming effective designs and materials are available
- create partnerships in which developed countries invest in the production of masks abroad where costs are lower, in return for guaranteed access to masks
- Promote public awareness of masks
- campaign to promote public awareness of proper use
- campaign to change public perception of masks, from something viewed as strange or suspicious, to an act of solidarity and community support
- Feng S, Shen C, Xia N, Song W, Fan M, Cowling BJ. Rational use of face masks in the COVID-19 pandemic. The Lancet Respiratory Medicine 2020.
- Servick K. Would everyone wearing face masks help us slow the pandemic? Science 2020.
- Cohen J. Not wearing masks to protect against coronavirus is a ‘big mistake,’ top Chinese scientist says. Science 2020.
- Marois I, Cloutier A, Garneau É, Richter MV. Initial infectious dose dictates the innate, adaptive, and memory responses to influenza in the respiratory tract. Journal of leukocyte biology 2012;92:107-21.
- Gowthaman V, Vanamayya P, Nagarajan S, et al. Influence of Dose of Inocula on Outcome of Clinical Disease in Highly Pathogenic Avian Influenza (H5N1) Infections—An Experimental Study. Avian diseases 2010;54:576-80.
- Booth CM, Clayton M, Crook B, Gawn J. Effectiveness of surgical masks against influenza bioaerosols. Journal of Hospital Infection 2013;84:22-6.
- Ambrosch A, Rockmann F. Effect of two-step hygiene management on the prevention of nosocomial influenza in a season with high influenza activity. Journal of Hospital Infection 2016;94:143-9.
- Suess T, Remschmidt C, Schink SB, et al. The role of facemasks and hand hygiene in the prevention of influenza transmission in households: results from a cluster randomised trial; Berlin, Germany, 2009-2011. BMC infectious diseases 2012;12:26.
- Aiello AE, Perez V, Coulborn RM, Davis BM, Uddin M, Monto AS. Facemasks, hand hygiene, and influenza among young adults: a randomized intervention trial. PloS one 2012;7.
- Hu B, Zeng L-P, Yang X-L, et al. Discovery of a rich gene pool of bat SARS-related coronaviruses provides new insights into the origin of SARS coronavirus. PLoS pathogens 2017;13.
- Zhou P, Yang X-L, Wang X-G, et al. A pneumonia outbreak associated with a new coronavirus of probable bat origin. Nature 2020:1-4.
- Nishiyama A, Wakasugi N, Kirikae T, Quy T, Ha LD, Ban VV. Risk factors for SARS infection within hospitals in Hanoi, Vietnam. Jpn J Infect Dis 2008;61:388-90.
- Lau JT, Tsui H, Lau M, Yang X. SARS transmission, risk factors, and prevention in Hong Kong. Emerging infectious diseases 2004;10:587.
- Wu J, Xu F, Zhou W, et al. Risk factors for SARS among persons without known contact with SARS patients, Beijing, China. Emerging infectious diseases 2004;10:210.
- Jefferson T, Foxlee R, Del Mar C, et al. Physical interventions to interrupt or reduce the spread of respiratory viruses: systematic review. Bmj 2008;336:77-80.
- Liu Y, Gayle AA, Wilder-Smith A, Rocklöv J. The reproductive number of COVID-19 is higher compared to SARS coronavirus. Journal of travel medicine 2020.
- Brienen NC, Timen A, Wallinga J, Van Steenbergen JE, Teunis PF. The effect of mask use on the spread of influenza during a pandemic. Risk Analysis: An International Journal 2010;30:1210-8.
- Tracht SM, Del Valle SY, Hyman JM. Mathematical modeling of the effectiveness of facemasks in reducing the spread of novel influenza A (H1N1). PloS one 2010;5.
- Loeb M, Dafoe N, Mahony J, et al. Surgical mask vs N95 respirator for preventing influenza among health care workers: a randomized trial. Jama 2009;302:1865-71.
- Ang B, Poh BF, Win MK, Chow A. Surgical masks for protection of health care personnel against pandemic novel swine-origin influenza A (H1N1)-2009: results from an observational study. Clinical infectious diseases 2010;50:1011-4.
- Johnson D, Druce JD, Birch C, Grayson ML. A quantitative assessment of the efficacy of surgical and N95 masks to filter influenza virus in patients with acute influenza infection. Clinical Infectious Diseases 2009;49:275-7.
- Radonovich LJ, Simberkoff MS, Bessesen MT, et al. N95 respirators vs medical masks for preventing influenza among health care personnel: a randomized clinical trial. Jama 2019;322:824-33.
- Chughtai AA, Seale H, MacIntyre CR. Use of cloth masks in the practice of infection control—evidence and policy gaps. Int J Infect Control 2013;9.
- van der Sande M, Teunis P, Sabel R. Professional and home-made face masks reduce exposure to respiratory infections among the general population. PLoS One 2008;3.
- Davies A, Thompson K-A, Giri K, Kafatos G, Walker J, Bennett A. Testing the efficacy of homemade masks: would they protect in an influenza pandemic? Disaster medicine and public health preparedness 2013;7:413-8.
- MacIntyre CR, Seale H, Dung TC, et al. A cluster randomised trial of cloth masks compared with medical masks in healthcare workers. BMJ open 2015;5:e006577.
- Langrish JP, Mills NL, Chan JK, et al. Beneficial cardiovascular effects of reducing exposure to particulate air pollution with a simple facemask. Particle and fibre toxicology 2009;6:8.
- Miles LD. Techniques of value analysis and engineering: Miles Value Foundation; 2015.
- Bauchner H, Fontanarosa PB, Livingston EH. Conserving Supply of Personal Protective Equipment—A Call for Ideas. JAMA 2020.
- Tuekci Z. Why Telling People They Don’t Need Masks Backfired. The New York Times 2002 March 17.
- Li R, Pei S, Chen B, et al. Substantial undocumented infection facilitates the rapid dissemination of novel coronavirus (SARS-CoV2). Science 2020.