COVID-19 – do masks help?
The simplest definition of culture is “The way we do things around here.” If you’ve ever had the privilege of visiting the far east – China, Japan, Hong Kong etc – you will have noticed that they do some things differently to western countries i.e. our cultures are different. One of our cultural differences, which has been starkly apparent during the COVID-19 outbreak, has been the wearing, or not, of masks.
Blowing one’s nose, or using a tissue, in public is considered the height of rudeness in the far east. Anyone showing the first sign of a cold (let alone flu, or something nastier) will don a face mask out of respect for other people. They are saying – I have some germs right now and I want to protect you. We don’t do this in the west. We are more likely to stare at someone wearing a mask thinking that it looks odd and then sneeze down a train carriage.
Dr Eric Westman dropped me an email with some very interesting videos on masks and virus protection and he asked if I could review the research on this topic. After spending an hour looking at the videos, I wanted to do exactly that…
The official positions on masks
The official positions from the World Health Organisation (WHO), the US Centres for Disease Control (CDC), the US Surgeon General and the UK Deputy Chief Medical officer are similar – you don’t need a mask. Some go even further and suggest that wearing a mask may do more harm than good. Others caution that masks should be saved for caregivers.
WHO: “If you do not have any physical symptoms, such as fever, cough or runny nose, you do not need to wear a medical mask. Masks alone can give you a false feeling of protection and can even be a source of infection when not used properly.” Dr April Baller WHO Health Emergencies Programme (Ref 1).
CDC: “If you are NOT sick: You do not need to wear a facemask unless you are caring for someone who is sick (and they are not able to wear a facemask). Facemasks may be in short supply and they should be saved for caregivers” (Ref 2).
US Surgeon General: “Seriously people – STOP BUYING MASKS! They are NOT effective in preventing general public from catching #Coronavirus, but if healthcare providers can’t get them to care for sick patients, it puts them and our communities at risk!” (Ref 3).
UK Deputy Chief Medical Officer (Dr Jenny Harries): “If a healthcare professional hasn’t advised you to wear a face mask, it’s usually quite a bad idea. People tend to leave them on, they contaminate the face mask and then wipe it over something. So it’s really not a good idea and doesn’t help” (Ref 4).
Eric sent me three videos:
1) A video posted on March 25th by Rachel Thomas and Jeremy Howard (Ref 5). The video is 35 minutes long. Much of it translates a video posted on March 14th by Petr Ludwig, a Czech speaker, writer and social media influencer. Both videos explain the campaign #Masks4All and some of the science behind it. The campaign #Masks4All can be seen here (Ref 6). Petr’s original video (with subtitles) can be seen here (Ref 7). The message from both videos is that masks do help, and we should make our own.
2) A video posted on March 28th by Petr Ludwig, featuring Adam Vojtech, the Minister of Health for the Czech Republic. This video is fewer than four minutes long and it is the one to watch (Ref 8). The message is simple and powerful. The Czech Republic has mandated that everyone who has to leave home must wear a mask. The adopted slogan is, by everyone wearing a mask, “I protect you and you protect me.”
3) The third video was one that Eric has done, showing how surgical room training can be helpful to everyone right now (Ref 9). It is so vital to avoid infection transmission in surgery that staff have standard methods for handling items. They know if something has been contaminated and they know how to identify and avoid such an item. We can learn from this.
The videos were powerful and contained scientific references, but they weren’t systematic reviews of the evidence, so let’s do a top-level one… (Note 10).
A systematic search of the evidence
Every time I review a topic, I use an academic database to search for all relevant articles on that subject. Sometimes a topic search will only elicit a handful of returns and you need to make do with what you can get. Other topics return hundreds, or thousands, of papers and you then need to be more selective.
The first search was to see if anything has been written on COVID-19 particularly. Searching COVID-19 and mask(s) returned two potentially relevant articles:
– Feng et al. Rational use of face masks in the COVID-19 pandemic. The Lancet. 2020 (Ref 11).
– Leung et al. Mask masking in the COVID-19 epidemic: people need guidance. The Lancet. 2020 (Ref 12).
The first article is relevant to our research question. The second article might be useful if we find that masks can help the general public.
The first article was just two pages long. It contained a useful table summarizing global positions on masks. It reported the WHO, US and UK positions, as I did above. It also summarized the recommendations on masks in China, Hong Kong, Singapore, Japan and Germany. Interestingly, Hong Kong was the only country strongly recommending masks for the general public (if symptomatic, or if on public transport, or if in crowded spaces). China graded people by risk of infection and advised accordingly. Japan thought that the effectiveness of wearing a face mask to protect from contracting viruses was limited. Germany quoted the WHO position verbatim.
The paper authors noted that recommendations varied across countries and called for more research to clarify a definitive position. They observed that “there is an essential distinction between absence of evidence and evidence of absence.” The article reported that “increase in use of face masks by the general public exacerbates the global supply shortage of face masks, with prices soaring, and risks supply constraints to frontline healthcare professionals.” This doesn’t answer our research question – we need to know if masks will help the general public. The consequences of the answer are a separate issue.
The second search didn’t limit articles to those written about COVID-19. It looked for any papers about masks and protection against viruses in general. Searching for “mask(s)” and “virus” returned 922 papers. Ticking “humans” brought this down to 686 – still too many to review. Ticking “systematic review” and/or “meta-analysis” brought the search down to 20 papers. Far more manageable and focusing on the best evidence available. i.e. not a trial here and there – but collations of all trials available.
Even with such a specific search, a number of papers always turn out to be irrelevant to the research question (e.g. the 20 included a number of papers on bronchiolitis in high-risk children). Additionally 5 out of the 20 papers were by Jefferson et al and were revisions of one of the papers focused on below.
Among the few remaining systematic reviews and meta-analyses, there were two Cochrane reports. Notwithstanding recent events within the Cochrane Foundation, the articles emanating from this organisation are considered to be the most independent and valuable evidence available.
There are two key papers to examine, therefore:
1) Jefferson et al. “Physical interventions to interrupt or reduce the spread of respiratory viruses. 2011” (Ref 13). (This was the most recent of the Jefferson papers).
2) Burch & Bunt. Can physical interventions help reduce the spread of respiratory viruses? 2020 (Ref 14).
Let us look at these two studies…
The Cochrane reviews
1) The 2011 Cochrane paper reviewed 67 studies (including randomised controlled trials, population studies, case control studies, and time series studies, in many different countries). The conclusion was that the following interventions can help – frequent handwashing and barriers to transmission such as isolation and wearing protective clothing (masks, gloves and gowns). The review found “no evidence that the more expensive, irritating and uncomfortable N95 respirators were superior to simple surgical masks.”
The key output of meta-analysis is a pooled odds ratio (or risk ratio or hazard ratio). This number tells us – looking at all the evidence that we can find, from all the studies available – what does the intervention do to the odds/risk of something happening compared with not doing the intervention? This measure tells us relative risk, not absolute risk, and so caution should be applied (Note 15).
This paper reported that only the case control studies were sufficiently homogenous (similar) to be used in meta-analysis. There were nine of these. Interestingly, all nine assessed the impact of public health measures to curb the spread of SARS (severe acute respiratory syndrome – another coronavirus) during February to June 2003 in China, Singapore and Vietnam. They are, therefore, very relevant to the examination of data for the current coronavirus, COVID-19. The nine studies included some specifically involving health care workers, as opposed to the general population, but the efficacy of each intervention can still be assessed.
The key diagram (replicated below) is on p127 of the 162-page Cochrane report. The following diagram reports the odds ratio/effect size for each of nine interventions. The number of studies and the number of participants contributing to each finding is also shown. The numbers in brackets after the effect size denote the 95% confidence interval. If the confidence interval does not include 1.0, the result is said to be significant and not a chance occurrence. All of the interventions achieved a significant result.
The paper reported that, using data from 7 studies and 3,216 participants, “simple mask wearing was highly effective (OR 0.32).” That means that those wearing a mask were about two thirds less likely to contract SARS as those not wearing a mask. That’s relative risk. Analysis 1.3 on p129 allows us to calculate absolute risk. Among those wearing a mask there were 268 cases (of virus) in 681 people. Among those not wearing a mask there were 1,573 cases in 2,535 people. That was an absolute difference of 39 cases in 100 vs 62 cases in 100 (23 cases in 100). That’s well worth the intervention.
While wearing an N95 respirator appeared to be even more effective (OR 0.17), this was based on just 3 studies, involving 817 people, and the upper limit of the confidence interval was above that of just mask wearing. Those washing hands frequently (defined in the paper as a minimum of 11 times daily) were approximately half as likely to contract SARS than those not doing this (OR 0.54). Wearing masks was more effective than frequent handwashing – both using relative risk (OR 0.32 vs 0.54) and absolute risk (39 cases in 100 vs 57 cases in 100 i.e. 18 cases in 100).
This Cochrane review supported wearing masks, along with many other measures, for reducing the risk of virus transmission.
2) The Burch and Bunt study was a revision of the Jefferson et al 2011 Cochrane review. It reiterated that most of the evidence came from case control studies and that findings suggested that “handwashing; wearing of masks, gowns, and/or gloves; and use of eye protection may reduce the spread of respiratory viruses.” The paper added that results were inconsistent across studies and that the evidence overall was seen as very low certainty. The best evidence (moderate certainty) was for handwashing plus masks. None of the meta-analysis numbers from Jefferson were updated and so the figures remained the same.
This Cochrane review thus also supported wearing masks, along with many other measures, for reducing the risk of virus transmission.
One other paper was worthy of review from this second search. A paper from 2017 was called “Effectiveness of masks and respirators against respiratory infections in healthcare workers: A Systematic Review and Meta-Analysis” (Ref 16). As the title states, this paper undertook a systematic review and meta-analysis of the benefit of masks and respirators against respiratory infections in healthcare workers. Meta-analysis of randomised controlled trials found evidence for a protective effect of masks and respirators against respiratory illness. Meta-analysis of observational studies found evidence for a protective effect of masks and respirators against SARS specifically.
Despite the fact that our research question is about masks and the general public, I reviewed this paper because I am not aware of any debate questioning the value of masks among healthcare workers.
The campaign #Masks4All is encouraging people to make and use cloth masks and to make the cloth masks look very different to health care worker masks, so there can be no accusations about the general public having taken masks needed for front line staff.
Eric has been researching in parallel and he helpfully found an article from 2015, which has been cited 23 times (a measure of impact) (Ref 17). This study, published in the BMJ, involved 1,607 hospital health care workers, working in high risk wards. The wards were randomized to medical masks, cloth masks or a control group (usual practice, which included mask wearing). Participants used the mask on every shift for 4 consecutive weeks.
The paper abstract (summary) claimed that the rates of all infection outcomes were highest in the cloth mask group. The relative risk was given as a large 13 (CI 1.69-100.07) for Influenza Like Illness (ILI). What the abstract didn’t say was that ILI was self-reported. When laboratory confirmed viruses were compared, there were no significant differences between actual viruses in the medical or cloth mask groups.
Putting this all together
The official advice from the WHO, US, UK and many other authorities is that masks don’t help and might possibly harm. Masks are claimed to be of possible harm in three ways i) by giving a false sense of security, ii) by transmitting infection if not used properly, and iii) by reducing masks available to health care workers.
i) This review does not find evidence to support (i) – any security afforded seems valid, not false.
ii) The BMJ article on cloth masks appeared to offer evidence for (ii), but this was found to be invalid and frankly disingenuous. If there are risks of repeat-use cloth masks, as opposed to throw-away surgical masks, these risks can be alleviated by making sufficiently robust masks, washing them appropriately (Ref 18) and avoiding the chance of the masks contaminating surfaces – as shown in Eric’s video.
iii) This review dismisses (iii) – save masks for care workers – as being irrelevant to the research question – can masks help? (iii) might be a public health strategy to protect limited supplies for health care workers. That would be understandable, but it doesn’t provide evidence that masks don’t work. On the contrary, it emphasizes their value. Health care prioritization can also be supported by the helpful online campaigns to encourage people to make their own masks.
As I was finishing this article and wondering why the official advice is as it is, there were signs that this might be about to change. Eric alerted me to a Wall Street Journal article reporting that the US is reconsidering its position following revised guidelines in some parts of Europe (Ref 19). Trust is vital at the current time and we need to be able to trust official advice. There are many things that we don’t know and for which we don’t have evidence, but on this topic we do and have had for some time. Clearly ‘saving masks for the frontline’ has been a factor and this was explicit in the Surgeon General’s tweet and in The Lancet paper “Rational use of face masks in the COVID-19 pandemic.” But that doesn’t excuse not following the evidence and not providing alternatives for the masses.
The bottom line is that there is evidence that wearing a mask reduces the risk of virus transmission. Perhaps the west needs to embrace the culture of the east and make mask wearing (and making) “the way we do things around here!”
* The official positions from the World Health Organisation (WHO), the US Centres for Disease Control (CDC), the US Surgeon General and the UK Deputy Chief Medical officer are similar – you don’t need a mask. Some go even further and suggest that wearing a mask may do more harm than good. Others caution that masks should be saved for caregivers.
* A campaign called #Masks4All has been gathering momentum in the past few days. The campaign message is that masks do help, and we should make our own.
* I searched the academic database for any articles about masks and the current virus of concern, COVID-19. One article in The Lancet noted the different advice in different countries and called for more research to clarify a definitive position.
* The search for masks and viruses generally produced so many academic papers that it was possible to focus on the best evidence available – systematic reviews and meta-analyses.
* Since 2007, the Cochrane Collaboration has been reviewing and updating evidence on interventions to reduce the spread of respiratory viruses. The interventions examined include hand washing, wearing masks, wearing gloves and more.
* The main Cochrane paper (2011) has quantified the value of these interventions. Mask wearing was found to have (statistically) significant value, as did the other interventions. The most recent Cochrane review (2020) supported the findings, although adding some caution about the strength of the evidence available.
* A systematic review and meta-analysis, which examined the effectiveness of masks against transmission of respiratory infections in healthcare workers, also found that masks were of benefit. I am not aware of any debate about the value of masks among healthcare workers – only among the general public.
* The need to save masks for healthcare workers appears to have been a key driver of the advice that the general public doesn’t need them.
* The evidence shows that wearing a mask protects against virus transmission. We can avoid any claimed harms by making our own masks (protect supplies for healthcare workers) and caring for them carefully (protect against contamination).
Ref 1: https://www.who.int/emergencies/diseases/novel-coronavirus-2019/advice-for-public/when-and-how-to-use-masks
Ref 2: https://www.cdc.gov/coronavirus/2019-ncov/prevent-getting-sick/prevention.html?
Ref 3: https://twitter.com/surgeon_general/status/1233725785283932160
Ref 4: Dr Jenny Harries with the Prime Minister, Boris Johnson, at 10 Downing Street, March 11th, 2020. https://twitter.com/10DowningStreet/status/1237760980450451456
Ref 5: https://www.youtube.com/watch?v=BoDwXwZXsDI&t=1072s
Ref 6: https://masks4all.org/story
Ref 7: https://www.youtube.com/watch?v=K5xy2n941jM&feature=youtu.be
Ref 8: https://www.youtube.com/watch?v=jZtEX2-n2Hc
Ref 9: https://www.youtube.com/watch?v=MZy-g9wgSiw
Note 10: My PhD was an application of systematic review and meta-analysis to the dietary fat guidelines and so I am very familiar with (and trained in) this technique. A full systematic review adheres to established methodology (PRISMA) and can take weeks/months, but a top-level systematic search can provide a good starting point to answer a research question.
Ref 11: Feng et al. Rational use of face masks in the COVID-19 pandemic. The Lancet. 2020 https://www.thelancet.com/journals/lanres/article/PIIS2213-2600(20)30134-X/fulltext
Ref 12: Leung et al. Mask masking in the COVID-19 epidemic: people need guidance. The Lancet. 2020. https://www.ncbi.nlm.nih.gov/pubmed/32142626
Ref 13: Jefferson et al. Physical interventions to interrupt or reduce the spread of respiratory viruses. Cochrane Database Syst Rev. 2011. https://www.ncbi.nlm.nih.gov/pubmed/21735402
Ref 14: Burch & Bunt. Can physical interventions help reduce the spread of respiratory viruses? 2020 https://www.cochranelibrary.com/cca/doi/10.1002/cca.2965/full
Note 15: If 1 in 1million people taking a drug have an event and 1.3 in 1 million people taking the placebo have an event, the relative risk difference is 30% (1.3-1.0). The absolute risk difference is 0.3 in 1 million. Big difference!
Ref 16: Offeddu et al. Effectiveness of Masks and Respirators Against Respiratory Infections in Healthcare Workers: A Systematic Review and Meta-Analysis. 2017. https://www.ncbi.nlm.nih.gov/pubmed/29140516
Ref 17: MacIntyre et al. A cluster randomised trial of cloth masks compared with medical masks in healthcare workers. BMJ Open. 2015. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4420971/pdf/bmjopen-2014-006577.pdf
Ref 18: I can’t find any academic papers on this, but a trusted consumer site suggests that using detergent and washing at 60’c or even 90’c will kill germs https://www.which.co.uk/reviews/washing-machines/article/washing-machine-temperature-guide
Ref 19: https://www.wsj.com/articles/u-s-reviews-guidance-on-masks-to-fight-coronavirus-as-europe-embraces-their-use-11585676543