Face masks and coverings for the general public

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 1

26 JUNE 2020

Face masks and coverings for the general public:

Behavioural knowledge, eectiveness of cloth

coverings and public messaging

This rapid review of the science of the eectiveness of dierent face mask types and

coverings and behavioural adherence is from the Royal Society and the British Academy

to assist in the understanding of COVID-19.

This paper is a pre-print and has not been subject to formal peer-review.

SUMMARY KEY POINTS

• Cloth face coverings are eective in reducing source virus transmission, i.e., outward protection of others, when they

are of optimal material and construction (high grade cotton, hybrid and multilayer) and ﬁtted correctly and for source

protection of the wearer.

• Socio-behavioural factors are vital to understanding public adherence to wearing face masks and coverings, including

public understanding of virus transmission, risk perception, trust, altruism, individual traits, perceived barriers.

• Face masks and coverings cannot be seen in isolation but are part of ‘policy packages’ and it is imperative to review

interrelated non-pharmaceutical interventions in tandem including hand hygiene, sanitizers and social distancing when

maintaining the 2 metre or 1 metre+ distancing rule is not possible.

• Consistent and eective public messaging is vital to public adherence of wearing face masks and coverings. Conﬂicting

policy advice generates confusion and lack of compliance. Populations without a previous history of mask wearing have

rapidly adopted face coverings during the COVID-19 period.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 2

Executive summary

• Cloth face masks and coverings for the general public are

eective in improving: i) source protection, i.e., reduced

virus transmission from the wearer when they are of

optimal material and construction and ﬁtted correctly; and ii)

wearer protection, i.e., reduced rate of infection of those

who wear them.

• Optimal cloth face coverings are made from speciﬁc material

(e.g., high grade cotton), hybrid and multilayer constructions

(e.g., silk-cotton) and need to be ﬁtted correctly.

• Many countries implemented a policy requiring the general

public to wear face masks and coverings in all public places

by mid-March 2020.

• Countries with no previous history of wearing face masks

and coverings amongst the general public rapidly adopted

usage such as in Italy (83.4%), the United States (65.8%) and

Spain (63.8%) by the end of April 2020.

• A systematic review isolated key socio-behavioural factors

to understanding public adherence to wearing face masks

and coverings, namely:

• public understanding of virus transmission, including

ecacy of source versus wearer protection, diagnostic

uncertainty and inability to self-diagnose.

• risk perception, individuals’ underestimation of health

risks and perception that protection is only relevant for

vulnerable groups, or outside of their proximity.

• previous national pandemic experience resulting in

rapid response and socio-political systems, allowing for

more or less coordinated action and public trust.

• individual characteristics, such as younger people and

men having a lower threat perception and compliance

with interventions.

• perceived barriers, lack of supply of surgical masks and

perceived competition with medical resources, resource

constraints to obtain coverings, comfort and ﬁt.

• Consistent and eective public messaging is vital with non-

pharmaceutical interventions more eectively seen as part

of ‘policy packages’ to acknowledge:

• interventions as interrelated, to be reviewed in tandem

with face masks and coverings related to hand hygiene,

sanitizers and social distancing when maintaining the 2

metre or 1 metre+ distancing rule is not possible.

• public communications must be clear, consistent and

transparent with inconsistent, premature, alarmist

information or that without a clear source raising

scepticism and lowering compliance.

Conclusion

In England face masks and coverings for the general public

in public places have not been mandated beyond public

transport and hospitals. Wearing a face mask or covering

in the UK has had very low uptake (~25%, late April 2020).

The lack of clear recommendations for the general public

and low uptake of wearing face masks and coverings may

be attributed to: (i) over-reliance on an evidence-based

medicine approach and assertion that evidence was weak

due to few conclusive RCT (randomised controlled trial)

results in community settings, discounting high quality

non-RCT evidence. There have been no clinical trials of

coughing into your elbow, social distancing and quarantine,

yet these measures are seen as eective and have been

widely adopted; (ii) inconsistent and changing advice from

supranational organisations (WHO, ECDC) and other nations

with variation in policy even within the UK; (iii) concern over

the applicability of ﬁndings across multiple settings (health

care versus general public, other pandemics and countries),

yet many ‘lessons learned’ from previous pandemics,

including public wearing of face masks and coverings,

repeat themselves during COVID-19; and, (iv) mix of supply

concerns of PPE shortages of surgical face masks with

recommendations for face mask and covering wearing for

general public.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 3

Contents

1. Introduction and motivation 4

2. Eectiveness of cloth face coverings 4

2.1 Background and existing knowledge: health care settings of surgical masks and respirators 4

2.2 Evidence of eectiveness of public wearing of masks and coverings in community-based

studies 5

2.3 Meta-analysis of cloth and paper masks of protection of wearers in health care settings 7

2.4 Mask types and outcome as risk of infection 6

2.5 Data analysis 8

2.6 Meta-analysis results: Cotton masks associated with infection reduction 8

2.7 Face masks for protection of others: Eectiveness of cloth face masks varies

by fabric type, mask construction and gaps 9

3. International face mask and covering policies 10

3.1 Varied and changing nature of policy information 10

3.2 Face mask requirements and recommendations: An international comparison 11

3.3 Location of mask wearing policy 11

4. Behavioural factors related to face mask adherence: Systematic literature review 13

4.1 Data collection systematic review 13

4.2 Public understanding of virus transmission 13

4.3 Risk perception: perceived likelihood of infection and perceived beneﬁts 14

4.4 Previous national experience with pandemics, socio-political systems and trust

in government and science 15

4.5 Individual characteristics: Vulnerability, compliance and discrimination 16

4.6 Perceived barriers: supply concerns, resource constraints and comfort 17

5. Public adherence to face mask and coverings, relationship with other interventions

and importance of clear and consistent public messaging 18

5.1 Adherence to face mask wearing: a cross-national comparison 18

5.2 Adherence to other non-pharmaceutical interventions: face masks and coverings

in perspective 19

5.3 Package policies: face mask usage, physical distancing rules and hand sanitizer 20

5.4 The importance of communications, clear and consistent public messaging 21

6. Conclusion 23

Appendices 25

References 34

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 4

1. Introduction and motivation

As many countries evaluate changes in non-pharmaceutical

interventions to counter the spread of COVID-19,

considerable focus has been given to the use of face masks

and coverings (see Box 1) and related interventions. Next

to hand washing and social distancing, face masks and

coverings are one of the most of widely adopted non-

pharmaceutical interventions for reducing the transmission

of respiratory infections. As outlined in the Royal Society’s

DELVE report on face masks for the general public, their

review concluded that asymptomatic and presymptomatic

individuals are infectious, respiratory droplets are a major

mode of transmission and face masks reduce droplet

dispersal

. Face masks and coverings that are made of

optimal material and have a good ﬁt can provide protection

for both the wearer, but also those around them. For this

reason, many nations have already introduced them, such

as the United States in early April 2020

, with the WHO

recommending their use by the general public in early June

along with international experts urgently calling for their

introduction for the public during the COVID-19 crisis

At the time of writing, they have not been adopted for the

general public in England and various other countries and

there appears to be several gaps in our existing knowledge

on the subject of face masks and coverings. This report

contributes the following to inform current knowledge

and policy-making on face masks and coverings. First, we

present existing knowledge about the eectiveness of

cloth masks and face coverings, a meta-analysis of existing

studies to demonstrate their protective ability in a health

care setting and the eectiveness of cloth masks in eective

ﬁltering of the transmission of the virus (i.e., protecting

others). Second, we present an international comparison of

the timing and introduction of face mask policies in relation

to COVID-19. Third, existing reviews on this topic have

largely focussed on medical or transmission aspects related

to face masks, with a lack of attention to the behavioural

factors underlying perception and adherence to usage.

This relates to the fourth contribution, which is providing

a more systematic literature review, also beyond medical

and clinical literature

to broader sources and databases.

This allows us to evaluate comprehensive themes about

face mask and covering behaviour with interrelated non-

pharmaceutical interventions such as social distancing and

hygiene measures and embeds the face mask literature

across a wider range of topics and more extensive period of

time. In turn, it likewise allows us to learn from behavioural

knowledge reaped from previous respiratory infections and

pandemics.

2. Eectiveness of cloth face coverings

2.1 Background and existing knowledge: health care

settings of surgical masks and respirators

Current knowledge on the eectiveness of face masks to

prevent virus transmission from COVID-19, SARS, MERS and

H1N1 is mostly limited to studies of surgical masks and N95

respirators. The majority of existing studies are conducted

in health care settings and focus on protection of the mask

wearer as opposed to wearing a mask for the protection

of others. This distinction is vital since mask wearing for

the general public occurs in non-clinical situations (home,

public transport, shops, restaurants) and involves both

protection of oneself but also others. Surgical masks and

N95 respirators were included in the most recent systematic

review and meta-analysis published in the Lancet

. Based

on 29 studies, the authors concluded that the use of both

N95 respirators and surgical masks (including similar re-

usable masks) were associated with large reductions in

virus transmission. In this meta-analysis, they also found that

mask wearing in non-health care settings is protective and

statistically signiﬁcant (RR=0.56, CI 0.40-0.79). There were,

however, some concerns about this study including diculty

in separating eects of dierent types of PPE (masks, eye

protection), potential confounders and the transferability of

results to community settings.Another meta-analysis found

that medical masks provided similar protection to N95

respirators in protecting against viral respiratory infections

in healthcare settings

. We emphasise that the majority

of studies have been conducted in health care settings

and there are therefore caveats in the ability to transfer

results directly to community settings (see Appendix 5,

GRADE recommendations). Protective equipment in health

care settings may be more eective because of training,

knowledge and the environment. As we note in relation

to ‘package policies’ (Section 5.3), masks are generally

introduced as one of many policies such as hand hygiene

and distancing and thus dicult to examine in exclusion.

Both distance but also duration of contact are likewise vital

(but rarely examined), which may dier across settings.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 5

2.2 Evidence of eectiveness of public wearing of masks

and coverings in community-based studies

A repeated concern raised by some is that there are few

randomised control trials (RCTs) with conclusive results

examining the eectiveness of face masksconducted in

community settings. As we discuss in relation to our

GRADE recommendations in Appendix 5, and as others

have noted, RCTs are challenging for evaluating face masks

in a public setting given both the ethical and practical

considerations

8, 9

. This echoes experts in the ﬁeld who have

urgently called for the implementation of face masks and

coverings for the general public

. We note that there have

also been no clinical trials of coughing into your elbow,

social distancing and quarantine, yet these measures have

been widely adopted and are considered as eective.

A recent study identiﬁed 10 RCTs that examined the

eectiveness of facemasks on reducing inﬂuenza virus

infection in the community from 1946-2018

. The study

did not distinguish estimates by the type of mask but did

examine masks in combination with hand hygiene. The RCTs

were heterogeneous across community settings ranging

from Hajj pilgrims, university and households settings. In

a pooled meta-analysis, the authors conclude that there

was no signiﬁcant reduction in inﬂuenza transmission with

the use of face masks (RR 0.78, 95% CI 0.51-1.20, p=0.25).

But notably, the authors state that: “most studies were

underpowered because of limited sample size, and some

studies also reported suboptimal adherence in the face

mask group.”

BOX 1

Face mask versus face coverings

Masks often refer to surgical or respiratory masks that medical sta use whereas coverings encompass broader types

and materials such as homemade cloth masks. Not all masks and coverings are equal, with ﬁltration comparisons

discussed later in this report.

Respirators

N95, FFP1/2/3 and other forms are

seal-tested respirator masks that

protect health care workers. These

masks seal around the nose and

mouth, have tangled ﬁbres that

contain ﬁlters. We note that there

are also dierences amongst these

such as those with and without

valve protection, which we do

not discuss here.

Surgical masks

This is a form of personal protective

equipment (PPE) worn by health

workers that ﬁts loosely over the

nose and mouth, often blue squares

that hook over the ears.

Cloth face coverings

These are face masks that can be

purchased or made in the home

using a variety of fabrics. Research

on a variety of fabrics and patterns of

face coverings has shown that tightly

woven fabrics such as cotton, denim

or tea cloths ﬁlter the best and that a

combination of multiple layers is the

most eective. Loosely woven fabrics

like a scarf have been shown to be

the least eective. Attention must

also be placed on how well it ﬁts on

the face; it should loop around the

ears or around the back of the neck

for better coverage.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 6

They concluded that “In theory, transmission should be

reduced the most if both infected members and other

contacts wear masks, but compliance in uninfected close

contacts could be a problem.”

A non-peer reviewed medRxiv pre-print meta-analysis of

around 20 studies* (cluster-RCTs, cohort studies, case-

control, cross-sectional) conducted in community settings

concluded that wearing a face mask slightly reduces

the odds of infection by the wearer by around 6%

Observational studies found greater eectiveness.

They concluded that RCTs likely underestimated ecacy

due to poor compliance and that observational studies

likely over-estimated ecacy because of self-reported

symptoms and confounding. They also found that face

mask wearing was consistently protective across settings

including the general community, schools and universities,

and visits to health care clinics. They concluded that face

mask wearing was probably not protective during mass

gatherings, but they note this should be judged with caution

since they are drawn only from Hajj pilgrimage studies.

Pilgrimages would have dierent multiple transmission

pathways and a longer duration of recurrent contact. A

major limitation noted by the authors was that the type of

face mask was rarely explicitly stated in the studies and

they had to infer masks were of surgical grade, leaving no

indication of the eectiveness of cloth or non-surgical masks

or coverings for the general public.

Although there are few community-based RCTs, there is

evidence from mask wearing that occurred within the public

in Beijing

, which examined the transmission of COVID-19

within families and close contacts of 335 people in 124

families from 28 February to 27 March 2020. They found

that face mask use before the family member developed

symptoms was 79% eective but that wearing a mask after

the onset of illness was not signiﬁcantly protective. The

risk of transmission in the household was 18 times higher

with those who had frequent daily close contact with the

infected family member, compared to those who did not. A

combined study in Hong Kong examining hospital workers

and household members of SARS patients (N=1,192) found

that frequent mask use in public venues, frequent hand

washing and disinfecting living quarters were signiﬁcantly

protective factors

. A case-control study in Beijing of 94

unlinked and 281 community-based controls found that case

patients (i.e., became infected) were more likely to have

chronic medical conditions, eaten outside the home or taken

taxis frequently and that the use of masks was strongly

protective

. Another study was conducted in a community

setting in Vietnam of nine persons with serological evidence

of SARS from a sample of 212 close contacts but does not

have a speciﬁc focus on face masks, but does conﬁrm a

higher risk for direct carers of those who are infected

As noted previously largely due to the experience of

previous respiratory infections, the wearing of masks in

the community is strongly recommended in many Asian

countries, with high to almost universal uptake.

Our literature review revealed that no systematic review and

meta-analysis had yet been conducted on the eectiveness

of other types of cloth masks beyond surgical masks and

N95 respirators. As noted above, reviews that did exist

focussed on surgical and R95 respirators in health care

settings

or did not distinguish between the type of mask

A meta-analysis scrutinizing the eectiveness of these

alternative mask types is therefore a contribution, given

that cotton and paper masks are being recommended by

some governments

and that there is diculty in sourcing

surgical masks for the general public. As noted in the

GRADE evaluation of our work in Appendix 5, and above,

there are two strong caveats to our meta-analysis. First, it

has been conducted in a health care setting and second,

that all studies focus on source protection (i.e., protecting

the wearer). In a community setting dierent circumstances

would be at play and protection would be both of oneself

but also in blocking transmission to protect others.

A strength of this analysis is that the studies are in a

relatively homogenous setting.

* We note ‘around 20’ since dierent numbers are reported in the text and ﬁgures. In the results section, 31 studies are listed of which 28 were reported as

suitable for meta-analysis yet 21 studies are listed in the meta-analysis results (Figure 2).

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 7

2.3 Meta-analysis of cloth and paper masks of protection of

wearers in health care settings

We identiﬁed four potential studies and further cross-

reference checks and analysis of existing systematic

reviews on face masks revealed one additional non-

English language study (see Table 1, Yin et al. 2004). One

of the studies did not pass our eligibility criteria due to the

absence of a clear control group (i.e., they lacked a ‘no

mask’ control group and compared only medical versus

cloth masks)

. The excluded study was conducted in a

healthcare setting in Vietnam and is notably the only RCT

study on cloth masks. Another important eligibility criterion

was the presence of separate estimates on cotton and

paper masks. Our meta-analysis is thus based on four

quantitatively comparable studies which provide ﬁve

estimates from healthcare settings in China (Table 1). One

article

was in Chinese, which we able to have translated

and obtain the necessary information.

2.4 Mask types and outcome as risk of infection

The primary outcome of interest was risk of infection,

with three studies providing statistics for SARS cases

and one for inﬂuenza A H1N1. Cotton masks include cloth

masks and ≥12-layer gauze masks, following speciﬁcations

provided in the studies. Estimates of gauze masks are

included due to the fact that some Chinese healthcare

workers made their own masks from layers of gauze during

the SARS outbreak

. The second mask type of interest are

alternative masks made from paper, also used during the

SARS epidemic.

TABLE 1

Characteristics of comparative studies included in mask type meta-analysis.

Study Country Setting Virus Type of

study

Comparison groups Sample

size

Main ﬁndings

Control

group

Intervention

group

Zhang et

al. 2012

150

Beijing,

China

Health

care

H1N1 Case-

control

No face

mask; 16.6%

infected

Cloth face

mask;

20.5%

infected

56 Cloth mask use did

not signiﬁcantly

decrease the risk of

infection in health

care setting

Liu et al.

2009

151

Beijing,

China

Health

care

SARS Case-

control

No face

mask; 12.1%

infected

≥12-layer

gauze;

6.5%

infected

477 Healthcare

workers who wore

cotton masks had

signiﬁcantly lower

risks of infection

Seto

et al.

2003

152

Hong Kong,

China

Health

care

SARS Case-

control

No face

mask; 13.3%

infected

Paper mask;

7.1% infected

111 Healthcare workers

who wore paper

masks had lower

risk of infection

Yin et al.

2004

153

Guangdong

Province,

China

Health

care

SARS Case-

control

No face

mask; 81.8%

infected

a. ≥12-layer

gauze;

22.8%

infected

b. Paper

mask; 50%

infected

a. 213

b. 55

Healthcare workers

who wore both

types of masks

had lower risk of

infection

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 8

2.5 Data analysis

We quantify associations of mask use with incidence of

infection by employing standard random eects meta-

analysis. We assess risk ratios (RRs) with 95% conﬁdence

intervals (CIs) and performed subgroup analysis by face

mask type. Due to the small number of studies, we were

unable to perform additional sensitivity analyses (e.g.,

dierences by country or dierent settings). Since our

sample is relatively homogenous with respect to research

design, country, and health care setting, we do not expect

inter-study heterogeneity to seriously bias our results.

As noted in the limitations, the low number of studies is,

however, one concern. We also performed Begg’s and

Egger’s tests which did not reveal presence of signiﬁcant

publication bias (P>0.05).

2.6 Meta-analysis results: Cotton masks associated with

infection reduction

The results of meta-analysis are shown in Figure 1 with

pseudo 95% CIs (see also Appendix 3, Fig A3.2). For SARS

and H1N1 infections, the use of both cloth/≥12-layer gauze

and paper masks is associated with a statistically signiﬁcant

reduction of the infection risk (pooled RR=0.49 with 95%

CI: 0.30 to 0.78, N=888). However, there is considerable

heterogeneity in the ﬁndings (I

=59.3%* and signiﬁcant

Q-test with P=0.03). The use of cotton masks is associated

with a 54% lower relative odds of infection in comparison

to the no mask groups (RR=0.46; 95% CI: 0.22-0.97; N=746)

with a coecient heterogeneity I

of 66.6% (Q-test P=0.05).

For paper masks, the relative odds of infection were 39%

lower than in the no mask group (RR=0.61; 95% CI: 0.41-

0.90; N=166; I

=0.0%). On average, we can conclude that

cotton masks exhibit a greater protective potential than

paper masks. The results on paper masks should be

interpreted with caution since there are only two estimates

that emanate from small samples in comparison to the

cotton masks studies and particularly the comparatively

larger sample sizes in previous meta-analyses on surgical

and N95 respirators

. Once again, we note that this is

about the protection of the wearer and not about reducing

spread, which we cover in the next section. We also note

that these are case-control studies and do not show causal

relationships.

* We use I² statistics to quantify between-study heterogeneity, where I²>50% representing a potential for substantial heterogeneity. Notably, it is not

uncommon in medical meta-analyses to have a I² of 80%.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 9

2.7 Face masks for protection of others: Eectiveness of

cloth face masks varies by fabric type, mask construction

and gaps

As the previous meta-analysis demonstrates, the type of

face covering is essential, but it focussed on the protection

of the wearer. Another persistent question is: Are certain

types of cloth face coverings that can be easily made in the

home or purchased more protective to shield transmission

of infection to others? Although there is currently limited

evidence on this in relation to COVID-19, a recent study

published April 24 2020 examined the performance of

various commonly available fabrics used in cloth masks and

coverings. Note that the COVID-19 virus produced by an

infected person is in the respiratory mucus and distributed

out in larger particles, making the type of fabric and ability to

penetrate this fabric important. A variety of common fabrics

and their combinations were used including cotton, silk,

chion, ﬂannel and various synthetics.

Figure 2 takes their original results and plots them on a

graph to visually demonstrate the dierences in mask

ﬁltration ability by several types of masks. The central

conclusions are that: (i) ﬁltration for various fabrics when

a single layer is used performs relatively worse, ranging

from 5-80% and 5-95% for particle sizes of <300 nm and

>300 nm respectively. (ii) cotton (particularly high grade

thread counts) is particularly good at ﬁltration, and this was

(iii) particularly with more layers cotton (600 TPI, 2 layers)

(99.5 ±0.1 error) whereas (iv) a hybrid material of (cotton-

chion, cotton-silk, cotton-ﬂannel) performed the best at

>80% (particles <300 nm) and >90% (for particles >300 nm).

They concluded that this enhanced performance of hybrids

was likely related to the combination of mechanical and

electrostatic-based ﬁltration. Finally (v), the eectiveness

of all masks, including N95, surgical and cloth masks were

seriously reduced when a gap was introduced, suggesting

the importance of proper ﬁt and usage.

FIGURE 1

Forest plot of risk ratios of the association of cotton and paper mask use with viral infection causing

SARS and inﬂuenza A H1N1.

0.05 0.25 1 5

Risk Ratio

Liu et al., 2009

Yin et al., 2004.1

Zhang et al., 2012

Yin et al., 2004.2

Seto et al., 2003

China

SARS

H1N1

SARS

8/123

46/202

9/44

22/44

2/26

43/354

9/11

2/12

9/11

11/72

0.54 [0.26, 1.11]

0.28 [0.19, 0.41]

1.23 [0.30, 4.94]

0.61 [0.41, 0.92]

0.54 [0.13, 2.28]

0.49 [0.30, 0.78]

Study Risk Ratio [95% CI]

Random overall (I

= 59.3%)Random overall (I

= 59.3%)

Cotton mask (e.g. cloth, 12−16 layer gauze) vs no mask

Paper mask vs no mask

Country

Virus

Events, Events,

masks (n/N) no masks (n/N)

0.46 [0.22, 0.97]

0.61 [0.41, 0.90]

Random subtotal (I

= 66.6%)

Random subtotal (I

= 0.0%)

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 10

The authors therefore conclude from this study that

combinations of various commonly available fabrics used

in cloth masks and face coverings can provide signiﬁcant

protection against the transmission of aerosol particles.

3. International face mask and covering policies

3.1 Varied and changing nature of policy information

International face mask policies, often in the form of

recommendations, have been introduced across many

countries in relation to COVID-19. Within these policies

there has been a distinction between recommendations for

respiratory (e.g., N95) and surgical masks for medical and

health care workers, versus face coverings (e.g., homemade

of fabric) for the general community. A deﬁnition of dierent

types of masks and coverings and terminology can be

found in Box 1, with a more detailed explanation on the

eectiveness of fabric and material dierences between

cloth masks described later in this report.

Notably, policies on face masks and coverings in relation

to COVID-19 have changed over time and been varied and

inconsistent between large supranational organisations

such as the World Health Organization (WHO) and advice

provided by various countries and regions (e.g., states,

provinces) within them. As of March 14 2020, 67 countries

had introduced policies, with many more implementing

policies between April and May 2020 (Figure 3A)

. The

majority of the facemask policies were inaugurated on

March 14, three days after the WHO declaration of the

coronavirus outbreak as a pandemic. On April 6 2020,

the World Health Organization (WHO) recommended that

healthy people in the community did not need to wear a

mask, and that they should be worn only by those who

are feeling unwell and are coughing and sneezing, as well

as caring for someone who is infected

. Advice from the

European Centre for Disease Prevention and Control (ECDC)

issued a similar statement on April 8 2020. Here they stated

that there was “no evidence that non-medical face masks

or other face covers are an eective means of respiratory

protection” and that there is “limited indirect evidence

showing that non-medical face masks made from various

materials may decrease the release to the environment of

respiratory droplet produced by coughing.”

FIGURE 2

Filtration eciencies of various fabric type test specimens (error).

Source: Adapted from Table 1

154

. Note: The ﬁgure shows ﬁltration eciencies at a ﬂow rate of 1.2 CFM.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 11

Conversely, there was almost universal mask wearing

without any policy in certain Asian countries that had

previous experience with SARS, which we explore in the

next section.

Large countries that had no previous history of face

coverings also adopted new measures. On April 3 2020, in

light of knowledge that a signiﬁcant portion of individuals

with COVID-19 are asymptomatic and can still transmit the

virus

, the United States Centre for Disease Control and

Prevention (CDC) recommended wearing non-surgical cloth

face coverings in public settings where social distancing is

hard to maintain (e.g., grocery stores). They also speciﬁcally

noted that they did not advocate the use of surgical masks

for the general public

. We note that in the same week,

the CDC recommended the use of non-medical face masks

while the ECDC stated that non-medical face masks are not

eective.

Then on June 5 2020, WHO published a correction of

their statement in early April with updated guidance

recommending that governments across the world should

recommend that the public should wear face masks in

public areas to help reduce the spread of COVID-19

This included encouraging mask wearing where there is

widespread transmission and physical distancing is dicult,

such as on public transport, in shops or in other conﬁned or

crowded environments. The WHO stressed that face masks

were one of a range of tools to reduce the risk of viral

transmission and that face masks should not give a false

sense of protection.

Across the United Kingdom recommendations and

mandates have varied considerably. On April 28 2020,

Scotland provided recommendations (not mandatory) that

the public should wear face masks in enclosed spaces

where social distancing is dicult to achieve. This was

followed by mandatory wearing on public transport as of

June 22 2020

. On June 9 2020, Wales recommended

that face coverings could be used where it might be dicult

to stay 2 meters away from others and advised using

three-layer non-medical face coverings

. England formally

introduced its ﬁrst face mask policy in early June, mandating

that as of June 15 2020, face coverings were mandatory

on public transport

. Although there has been some public

messaging about wearing face coverings in England, which

we explore in a later section, this has not been in the form of

clear and consistent formal advice.

3.2 Face mask requirements and recommendations:

An international comparison

As shown in Figure 3 (panel A), as of June 15 2020, most

countries (121 of 188 where data is available) required

face masks to be worn in the entire country, 19 in parts

of the country only, 28 did not require mask wearing, 14

recommended masks or covering, but did not require

mask-wearing and 6 Asian countries had no requirements,

but experienced virtually universal usage**. In Asian nations

such as China, Taiwan or Hong Kong, masks were already

common even before the coronavirus pandemic, credited

to populations accustomed to wearing coverings due

previous experience with the SARS and H1N1 outbreaks, or

pollution

. The use of face masks is also not new in Latin

America and were mandatory during H1N1 for instance in

Brazil

and Mexico

3.3 Location of mask wearing policy

Policies also vary by the location of where facemasks are

mandated. We illustrate measures of as of June 15 2020 in

Figure 3B with detailed policy categories described in Table

A1.1. Policies can be largely grouped into:

• mask wearing required for everyone in public places (71

countries).

• only indoor places (e.g., in relation to social distancing,

type of indoor space (15 countries),

• public transport and crowded places (in relation to

number of people, social distancing and venue)

(12 countries),

• certain public places (major cities only, in relation to social

distancing) (9 countries),

• public transport only (7 countries)

• universal mask usage but no formal policy (6 countries)

** China, Hong Kong, India, Japan, Malaysia, South Korea.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 12

FIGURE 3

Source: Masks4all data

155

data as of June 15 2020. This source is regularly updated and corrected but is not an ocial governmental or supranational

source. The authors used the information as provided in good faith, and note that each source for national policy is linked to an external source. Notes:

Panel A, data available for 188 countries. Panel A. Full country (121); Parts of country (19), No (28); No, but recommended (14); No, but universal usage (6).

Panel B see Appendix 1, Table A.1. Panel B, data available for 120 countries.

Face mask policies across the world.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 13

4. Behavioural factors related to face mask adherence:

Systematic literature review

4.1 Data collection systematic review

Following the PRISMA

and MOOSE

reporting guidelines

(see Appendix 2 Methods), we developed several

customized Python functions to undertake a systematic

review across the three databases of PUBMED, Scopus and

Web of Science (WoS), building on previous scientometric

work in genetics

38, 39

. Given the comparatively slower

publication pace of the social and behavioural sciences,

also noted elsewhere

, we also included some pre-

print non-peer reviewed articles and note this fact where

mentioned and in our GRADE recommendations

of the

quality of evidence (see Appendix 5). This extended search

allowed the inclusion of non-medical literature including

materials (e.g., face mask materials) and socio-behavioural

literature. We expanded the queries to include search terms

related to multiple derivations of face masks and coverings

(e.g., facemask, face mask, N95 respirator, surgical mask,

FFP3, cloth mask, face covering, all search queries available

upon request). Given the relatively recent nature of the

COVID-19 pandemic and to obtain a richer body of literature

and lessons learned, in addition to research on COVID-19

and coronavirus, we also included previous respiratory

pandemics such as the 1918 Spanish ﬂu, severe acute

respiratory syndrome (SARS), H1N1 inﬂuenza, Middle East

respiratory syndrome (MERS), H5N1 inﬂuenza and ﬂags for

additional policies on social distancing and isolation. More

detailed information on study selection can be found in

Methods (Appendix 2, Figure A2.1).

The aim of this rapid review was to focus on behavioural

factors related to compliance, with ﬁve central themes

that emerged: i) public understanding of the virus, ii) risk

perception, iii) previous national experience with pandemics,

socio-political systems, and trust in government and science,

iv) individual characteristics; and, v) perceived barriers. Most

of these themes have been previously identiﬁed such as

for example, in a review of qualitative research of SARS and

H1N1 in 17 studies

and a systematic review of 9 bioevent

studies in the United States

, but themes also dier due to

the wider breadth of the literature reviewed here, additional

focus on COVID-19 and international literature.

4.2 Public understanding of virus transmission

A central theme that emerged from the literature on public

adherence to face mask and coverings is the importance

of personal and cultural beliefs and understanding of

how respiratory viruses are spread. Core factors are: (i)

understanding how it is spread and, importantly, whether

asymptomatic individuals can transmit the virus, (ii) whether

mask wearing is for one’s own individual protection or to

protect others, (iii) clarity on diagnosis of COVID-19 and

inability and reticence to self-diagnose; and, (iv) ecacy

to adopt the required behaviour of face mask or covering

usage to counter the threat.

A systematic literature review of previous respiratory

pandemics (SARS, H1N1) found that the general perception

of how respiratory viruses are transmitted is that it is by air,

only within a particular proximity, by symptomatic others only

and more likely in cold ambient and water temperatures

For COVID-19 as with other respiratory viruses, droplets

are produced when an individual coughs, sneezes, talks

or breaths, which then convert to aerosols and become

airborne. Droplets can land on surfaces and can remain

viable. Aerosols are much smaller than droplets and thus

can more easily penetrate dierent types of material.

Knowledge rapidly changed about COVID-19 transmission,

particularly in the early phases, with a growing number of

studies demonstrating sizeable levels of asymptomatic

transmission

A related issue is the understanding of whether face masks

are used for individual protection against contracting the

virus versus wearing one to protect others. An international

poll of face mask wearing during March 12 to April 12

across 15 countries examined this (N=29,000, ~2,000

per country)

. In the UK (41%), Australia (47%), Russia and

Canada (35%) a sizeable proportion did not see the value

in wearing a face mask if they were not sick, suggesting

that they were not aware of asymptomatic transmission.

This is not entirely surprising, also in light of WHO and

other national advice that had initially focussed on

individual mask wearing for only infected individuals to

protect others in early April 2020.(19) This is compared to

comparatively lower levels reporting the same in Vietnam

(7%), China (9%), Japan (11%), but also Spain (8%) and Italy

(9%), all nations that had adopted or continued to have high

levels of face mask wearing by mid-April 2020. As noted

previously, this could be related to the dierent phases of

the outbreak. Respondents in the UK showed the lowest

levels of understanding that face masks can be worn to

protect others with 15% reporting ‘I expect people around

me to wear a face mask so I don’t get sick’ compared to for

instance Japan (58%) or Vietnam (55%).

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 14

During both SARS, H1N1 and repeated again with COVID-19,

there was also high diagnostic uncertainty particularly at the

start of the pandemic, which challenges individuals’ ability

to self-diagnose. Particularly during H1N1, but also COVID-19,

individuals expressed doubts about their ability to identify

symptoms and whether and how they could distinguish

between the pandemic and seasonal ﬂu symptoms. Studies

in the UK

47, 48

and New Zealand

of the H1N1 outbreak found

that individuals had strong fears and concerns about their

own judgement and ability to self-diagnose, which in turn

inﬂuenced their behaviour in relation to self-isolation and

use of remote healthcare. During H1N1, the ‘vagueness’

of the symptoms and dierentiating them was listed as a

central challenge for individuals. Indeed alternative data

collection of ‘real-time’ tracking crowdsourced a wide variety

of COVID-19 symptoms across a spectrum of mild to serious

systems ranging from loss of smell and taste to breathing

diculties, which at that time had not been included as

symptoms in many countries

Finally, the manner in which individuals in the community

respond to the threat of a respiratory infection is inﬂuenced

by their beliefs about the ecacy of the intervention and

perceived costs of protective behaviours

. Ecacy refers

to the beliefs about an individuals’ ability to successfully

adopt behaviours and the eectiveness of adopting

behaviours in eliminating the health threat. Literature from

the SARS coronavirus outbreak and H1N1 2009 pandemic

found that perceptions of risk, anxiety about the infection

and the ecacy of the intervention are pivotal

52, 53, 54, 55, 56,

57, 58, 59, 60

. Behavioural change is highly contingent on the

communication of risk, individual appraisal of risk and the

perceived ability to make the change

. A literature review of

over 65 studies examining over 20 public health issues for

instance, concluded that the key factors driving behavioural

change are increases in threat severity, threat vulnerability,

response ecacy and self-ecacy facilitated adapted

intentions or behaviours

. We return to this topic later in

this section when we discuss barriers, conﬂicting policy

advice and confusion about the eectiveness of face masks

and coverings, which in turn impacts the ecacy individuals’

would place on adopting face mask and covering

interventions.

4.3 Risk perception: perceived likelihood of infection

and perceived beneﬁts

A clear theme that emerged in the literature was the

importance of individual risk perceptions, the notion of

‘othering’ and belief that ‘it won’t happen to me’. Core

factors related to this theme are an: (i) overly optimistic

risk assessment of not contracting or transmitting the

virus, (ii) incorrect judgements about the role of proximity;

and, (iii) denying personal risk via ‘othering’ by blaming or

dierentiating oneself from vulnerable groups perceived to

be at a higher risk.

A central challenge isolated in the literature is that many

individuals view themselves as less vulnerable and more

capable than others, generally underestimating health risks,

ﬁnding it unnatural to respect strict isolation to protect others

and have only a limited awareness of actions that pose

a health risk

. A number of studies focus on individuals’

incorrect assessment of risk and overly optimistic sense

they will not contract the virus. A study of SARS in Canada,

for instance, found that a common aspect of risk perception

was the denial about their risk of contracting SARS because

they did not feel sick

. Individuals’ optimism can in turn lead

to an underestimation of contracting COVID-19 and thus

ignoring public health messages A study of 1,591 US-based

individuals in the ﬁrst week of the COVID-19 pandemic from

March 11-16 2020 examined individual’s perception of risk

Within ﬁve days, as they gained awareness about the virus,

perceptions of risk increased yet they still underestimated

their personal risk of infection. This, however, substantially

varied amongst individuals, isolating a subgroup of those

who persistently remained disengaged, unaware and did

not practice any protective behaviour. A pre-print non-

peer reviewed study on medRxiv surveyed individuals

across eight countries between mid-March to April 19 2020

(N=66,266)

. They found that the perception of individual

threat of COVID-19 was the highest in Italy, followed by the

UK, Spain and others with Germany being the lowest. The

authors note that the perceived threat was also in relation to

the phase of the outbreak with Italy and UK, two of the most

aected countries in Europe. They also found that the level

of threat was related to the trust in government and health

care systems, which were high in Germany, we return to

socio-political systems later.

Perceived proximity to the outbreak also played a role.

During the H1N1 2009 pandemic, individuals assumed they

had lower risk if they had a higher perceived health status

or that they perceived the outbreak was outside of their

proximity

. Proximity is often evaluated by individuals in

terms of geographical distance, but also own perceived

dierences in their own living environment. Another

common belief found in studies in the UK, US, New Zealand

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 15

and Australia in relation to previous pandemics was that

geographical proximity was protective and that respiratory

viruses were unlikely in a ‘modern, developed country’ and

hence a perceived lack of urgent risk and lag in adopting

public health messages

68, 69, 70

. A study in the UK about

public attitudes surrounding H1N1 found that individuals

believed that respiratory viruses were only more likely to

emerge in ‘other’ living environments such as those with low

hygiene levels, high population density, poor border control

and health systems

. Some, however, related the high ability

and propensity for viruses to spread worldwide more rapidly

due to air travel.

Distancing is a typical way of dealing with the negative

impact of health risks by using what is characterized in

the literature as ‘othering’

. Othering refers to blaming or

dierentiating oneself from ‘the other’, which in turn denies

personal risk. During H1N1, but also repeated with COVID-19,

was the designation of a vulnerable group of ‘others’ that

needed to be shielded and were perceived to be at a higher

risk of infection. This included those with chronic health

problems, impaired immune systems (e.g., undergoing

cancer treatment) or in frontline occupations (e.g., teachers,

health workers). Although the literature on this point is

largely from the H1N1 pandemic, it appears to echo similar

experiences in the COVID-19 pandemic. A general narrative

in this literature is the public belief that ‘it won’t happen to

me’ and have an inability to rationally evaluate the individual

risk of infection while also actively distancing themselves

from the threat by clarifying their dierence to ‘other’ groups

and circumstances that would have a higher risk of infection.

We discuss the link of othering with potential discrimination

later in Section 3.5 on individual and group dierences.

4.4 Previous national experience with pandemics, socio-

political systems and trust in government and science

There is a strong national variation in the acceptance and

usage of face masks, which has been attributed to several

key factors. These are: (i) previous experience with viral

infections (e.g., SARS, H1N1), (ii) normalisation and history

of mask wearing for other reasons (e.g., pollution) and

rapid adoption for those without a face mask history; and,

(iii) socio-political systems, linked to individualistic versus

socially cohesive structures, political polarization and trust in

government.

Previous experience with viral infections such as SARS and

H1N1 is linked to more universal and early mask use and

acceptance during COVID-19

. In addition to face mask

policies, these countries also simultaneously introduced

a battery of other interrelated non-pharmaceutical

interventions. After SARS, most governments from the

nations that were deeply inﬂuenced had already set up rapid

responses and policies that would allow them to react swiftly

in the event of another respiratory pandemic

. In many

Asian countries impacted by SARS, broad communications

had already previously been tested and put in place, such

as media messaging and billboards showing how to wash

hands and wear masks properly

. At the start of SARS 65%

of respondents in Hong Kong said they wore a mask

Singapore, for instance had previously distributed over 1

million ‘SARS toolkits’ which included a digital thermometer,

two surgical masks and instructions in four languages

They had also previously developed random and electronic

monitoring of compliance. After the ﬁrst COVID-19 case

was reported in Singapore on January 22, 2020, the

country introduced deep and swift measures by February

7 2020. All non-essential gatherings were cancelled; daily

temperature and health checks were performed in schools

and workplaces, and face mask wearing and physical

distancing were quickly advised in public places. So deep

was the intervention that an unintended consequence

was that inﬂuenza levels declined steeply from a mean of

57.3% (ﬁrst 6 weeks of 2020) to 3.5% (week 14), lower than

any inﬂuenza levels for the past 3 years

. The authors

attributed this to the introduction of wide and deep non-

pharmaceutical interventions, which included face masks

in addition to the suspension of mass gatherings, social

distancing and public promotion of the social responsibility

to stay at home.

Second, in nations where individuals have previously worn

face masks for other reasons such as pollution (e.g., India,

China), there is higher and more normalized compliance

(see previous section on international policies). There is

limited knowledge of face mask wearing in countries without

a previous history, but an initial study demonstrated that face

mask adoption could happen rapidly

, which we discuss in

more detail later in this report.

A third more heterogeneous strand of literature focuses

on socio-political systems, linked to individualistic versus

socially cohesive structures, political polarization and trust

in government. One body of research draws from research

on socio-political systems that compare more community-

based social cohesion versus those from more individualistic

based structures

. This has been posited in the form of

‘tight versus loose cultures’, with countries such as the US,

Italy and Brazil positioned as having the weakest social

norms and being more permissive

. These socio-political

structures have in turn been linked to the ability to engage

in coordinated action.

This strand of literature currently exists largely in pre-print

non-peer reviewed format due to the slower publication

processes within the social sciences, noted previously. It

examines the inability for coordinated action due to political

polarization – which is notable in countries such as the

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 16

US, UK and Brazil, which in turn results in distrust of the

opposing party and beliefs in false information that can

undermine public health messages. Using geolocation data

of daily movements from January until April 23 2020 from

around 35 million unique devices and debit card transaction

data of over 12 million cards in the same period in the US,

one study found that residents in Republican voting counties

were less likely to stay at home after a state order

Conversely, those from Democrat counties were more likely

to switch to online e-commerce spending after stay at home

state orders were implemented. Adherence was also related

to the political aliation of the governor and suggested

that bipartisan support was essential to maximize the

eectiveness of policies. We note some concerns with these

ecological studies using political variables as there are very

likely many other confounders that play a role.

Another medRxiv pre-print examining the level of conﬁdence

in local and national health care systems and governments

and the WHO during COVID-19 found individuals reported

the lowest levels in the UK and US across all institutions

compared to other countries such as Spain, the Netherlands,

Italy and Germany

. Another study drawn from two MTurk

studies of US respondents (N=1153) found that political

conservatism predicted lower compliance with social

distancing

. They found that the politicization of COVID-19

prompted conservatives to discount mainstream media

reports of the severity of the virus, leading them to downplay

the health risks and adhere less to social distancing

protocols, also when controlling for key demographic and

psychological variables. As demonstrated in more general

research, the result can be echo chambers and less cross-

group information sharing

. These more narrowing lines of

information can be ampliﬁed by social media streams.

Another pre-print non-peer reviewed study examined trust

in science and government during the COVID-19 and the

relationship with compliance with public health measures

Using digital trace data from Twitter and survey data

collected online via Telegram and Facebook, they examined

the evolution of trust in science in Italy during early phases

of COVID-19. They found that there was an initial reliance

on information seeking from scientists and public-health

authorities. Trust in science and institutions (local or

national government) emerged as a consistent predictor

of both knowledge about COVID-19 and adherence to

non-pharmaceutical measures. As the epidemic peaked in

Italy, however, they found a reverse in information seeking

and trust in science and health authorities, interpreted

by the authors as an erosion of trust. Interestingly, using

an experiment they found that those who held incorrect

information about COVID-19 gave no or even lower

importance to scientiﬁc information about the virus. Many

disadvantaged communities and particularly racial and

ethnic minorities hold a low level of trust in public institutions

due to persistent experiences of discrimination

, with

particularly low levels of trust in the health care system

4.5 Individual characteristics: Vulnerability, compliance

and discrimination

There has also been attention to the study of dierences

in the vulnerability and adoption of non-pharmaceutical

interventions across groups. This work can largely be

divided into topics examining: (i) face mask usage in

relation to vulnerable groups more likely to die if infected,

(ii) dierence between demographic groups in relation

to asymptomatic infection and compliance to non-

pharmaceutical interventions such as face masks; (iii)

additional traits such as personality or physical traits (e.g.,

eye-glass wearing, activity level) that make it more dicult

to wear a mask; and, (iv) discrimination or social reactions to

mask wearing.

Throughout previous pandemics and with COVID-19 there

has been a focus on vulnerable groups more prone to

infection and death from the virus. A considerable amount

of COVID-19 research has focussed on dierences in death

rates due to individual characteristics such as age

other vulnerabilities related to co-morbidities (hypertension,

obesity, diabetes)

, being male, from certain ethnic groups

(African, Asian) or working in front-line occupations

. Some

of these factors are related to deprivation such as some

groups are more likely to become infected since they

are more often key workers (e.g., bus drivers, care), rely

on public transport or have poor or no internet contact,

putting them at a higher risk of leaving the household and

higher risks of contact

. Here policies have been eective

in promoting the wearing of face masks for protection

around those who are ill, with the majority within health care

settings. Several studies conducted in relation to H1N1 and

during SARS, concluded that the perceived need or wish

to care for sick (isolated) loved ones overrode concerns

about self-protection and personal distancing. A Canadian

study of SARS reported that although compliance with the

quarantine order was high, within house protocols such as

mask wearing were uneven and often ignored when a family

member was sick and required care

Another emerging area of COVID-19 research is the study of

how compliance to non-pharmaceutical interventions such

as wearing face masks varies across socio-demographic

groups and ﬁrst particularly age groups. Representative

data measuring health behaviour across eight countries

(Belgium, France, Germany, Italy, the Netherlands, Spain,

UK, US) published in a medRxiv non-peer reviewed pre-

print discussed previously, found that younger people

perceived the COVID-19 threat lower to themselves than

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 17

older groups, also related to lower adherence to non-

pharmaceutical behaviours

. Mask wearing was the highest

in older 65+ age groups in all nations except the UK and

the Netherlands where mask wearing was both low but also

more evenly enacted across age groups. In the UK, the only

clear dierences in these behaviours across age groups

were higher levels of hand sanitizer use by the youngest

groups (18-24) and the avoidance of public transportation

by the oldest groups (65+). Conversely, a medRxiv non-

peer reviewed pre-print study by Goldstein and Lipsitch

examining weekly COVID-19 cases in Germany found that

the incidence of infection increased the strongest amongst

younger age groups between the starts and end of March.

These relative increases were found for all individuals under

35, but were especially large amongst those between 20

and 25 (RR=1.4; 95% CI: 1.27-1.55). The authors suggest that

increased mixing and lower adherence of social distancing

practices amongst these age groups could be responsible

for this relative increase.

This representative data also found particularly important

dierences by sex

. Women exhibited substantially higher

perceptions of threat compared to men, which translated

into the adoption of a wider-range of preventive behaviours.

They conclude that one of the most protective factors for

women during COVID-19 has been their adoption of multiple

protective interventions. Women were more likely to wear

face masks across the nations that were examined with

the exception of the UK and the Netherlands where face

mask wearing was similar across the sexes. In spite of the

fact that the case fatality rate for COVID-19 has been shown

as substantially higher for men

, this does not translate

into higher perceptions of threat or related behavioural

protection, suggesting a need to target this and other

groups for future communication campaigns.

Some early psychological literature published mainly

in pre-print non-peer reviewed literature has examined

psychological factors related to face mask wearing. Given

that these are non peer-reviewed pre-prints, the selectivity

of respondents, small sample size and often-artiﬁcial

experimental circumstances in which the research was

carried out, replication would be necessary before using

this as evidence to draw ﬁrm conclusions (see Appendix

5). A pre-print study asked 86 participants to assess how

they felt wearing a mask while being exposed to groups

with dierent levels of mask wearing

. Exposure to social

groups where more wore a mask reduced the strange

feeling of wearing a mask, suggesting that as mask-wearing

is easier when a larger majority of society also wears

a mask. Another non-peer reviewed PsyArXiv preprint

engaged in an internet-based study (N=457) using ‘human-

like’ characters wearing a mask or exhibiting dierent

facial expressions (neutral, happy, angry). They found that

reports of interpersonal distancing were reduced when the

character was wearing a face-mask as they were perceived

as more trustworthy.

Historical reports from the bubonic plague through to

HIV and COVID-19 have reported a rise in violence and

discrimination against stigmatized groups that carry a

disease

. Early in the COVID-19 period there is anecdotal

evidence of anti-Asian discrimination in some areas

100

with a PsyArXiv preprint reporting stigma faced by the

Chinese community outside of China due to wearing face

masks, particularly in countries such as the United Kingdom

that did not adopt early face mask policies

101

. During the

SARS pandemic, chronic kidney disease patients in Hong

Kong were perceived by the public as high risk ‘super

spreaders’ of infection. The study concluded that this

potentially stigmatized group wore masks as a symbol of

socially responsible action, but also to protect themselves

during the pandemic

102

. Another strand of the literature has

found that interventions such as frequent hand-washing

or mask-wearing has the potential to attract social stigma,

embarrassment or discrimination. Face mask wearing was

found in some studies as problematic since some were

concerned that it would make others see them as indicating

infection. This was the case in a study of face mask wearing

within a Hispanic community in the United States

103

4.6 Perceived barriers: supply concerns, resource

constraints and comfort

Another segment of the literature from this review looked

at the key barriers to face mask usage by the general

public, which were identiﬁed primarily as: (i) supply concerns

and inability to source them, (ii) resource constraints, (iii)

concerns about comfort, appropriate ﬁt and incorrect usage;

and, (iv) environmental waste.

A recurrent theme in the literature were shortages of

protective equipment, including face masks. A considerable

amount of the face mask literature in this systematic review

related to the shortage or lack of supply of face masks and

PPE, particularly for health workers

104, 105, 106

. During SARS, the

lack of protective equipment and especially masks led to a

‘state of panic’ in Taiwan

107

. So great was the shortage during

SARS that the Japanese government donated thousands

of masks and other protective equipment to Vietnam and

in Toronto, Canadian doctors sought to purchase their own

supplies

108

. The extreme shortage of

PPE was a strong theme from previous epidemics such as

SARS, where lack of worldwide protective masks, gloves

and respirators was positioned as the one of the key

‘lessons learned’

109

. This worldwide shortage of masks

and other PPE, however, repeated itself once again with

COVID-19 in many countries.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 18

The absence of clear agreed standards for making and

manufacturing face coverings is also lacking. Since

manufacturing issues have shown to play a role in the

provision of PPE, clear guidelines for manufacturers and

those making masks at home is essential. Information needs

to be provided to produce coverings that reach the proper

standards, such as the higher ecacy multi-layer hybrid

masks discussed in Section 2.7. The CDC in the US for

instance, provides a very clear tutorial on how to make a

face mask at home and information about washing, re-use

and how to wear face coverings.

A shortage of face masks and protective equipment for

medical sta also produced a feeling amongst the public

that face masks were unavailable or that wearing face

masks would unduly compete with medical resources. There

were also cases of face masks and other PPE that had been

imported from other countries as being ‘recalled’ which

may have impacted public trust. An international poll of

face mask wearing during COVID-19 from March 12 to April

12 across 15 countries, reported that the largest perceived

drawback in wearing face masks in most countries was

getting access to one

110

. Speciﬁcally they responded to the

question ‘even if I wanted to wear a face mask, I have heard

they are not available anywhere or are too expensive’. This

was the highest in Japan (57%), France (49%), Germany

and Spain (45%) and the UK (42%). A systematic review

demonstrated that negative emotions such as fear can lead

to a change in behaviour only if people feel that they are

able to control the threat. If they cannot – such as facemask

regulations without a clear supply of access to facemasks –

reactions will be defensive due to feelings of helplessness

111

Fear is thus only eective when individuals feel a strong

ability or level of ecacy, otherwise it will elicit a defensive

and negative response.

One of the greatest risks of virus spread is the inability of

some individuals to adopt policy recommendations due

to resource constraints such as money to buy or ability to

make their own face covering. The Italian government for

instance, set the price of surgical masks to a maximum of 50

cents to assuage this problem

112

. Some community members

may lack the resources to purchase or make face masks but

also live in areas with high neighbourhood density or lack

the ability to social distance. A variety of studies conducted

during previous pandemics in the UK and US, found that

perceived obstacles for following isolation, distancing and

other measures were related to the economic pressures

to continue to work and fulﬁlling familial and workplace

commitments

113

. A study of H1N1 in 2009 in the UK reported

incidences of individuals breaking compliance such as

isolation due to boredom, job security and economic

strain

114

. A repeated fear in the US literature was shutting

down the economy and ability to ﬁnancially survive in the

household when schools were closed or individuals were

unable to work

115

. Translating these ﬁndings to face masks,

it seems paramount to ensure broader access for those

unable to purchase or make face coverings, such as cutting

costs or free distribution to certain groups.

Multiple studies concluded that compliance with wearing

masks for a longer period of time was hindered due to the

fact that they were uncomfortable

116

. A dearth of literature

has examined this in the health care setting, including sores

and headaches due to long periods of wearing

117

, which

we do not cover in this review. A block randomized study

of mask-wearing in the United States isolated three main

issues related to the comfort of mask wearing in the home

118

First, the level of intensity or physical exertion impacted

comfort and compliance. Second, environmental factors

were key such as temperature, ventilation and apartment

size. Finally, the mask ﬁt was key which was variable

depending on whether they had eye glasses and the facial

structure of respondents namely high/low cheekbones,

bridge of the nose and shape of the face. Those with low

cheekbones and small nose bridges had diculties in mask

wearing due to the need to constantly adjust the mask.

Others noted diculties due to heat and dampness after

wearing it or having eyeglasses. The authors also noted

that young children persistently touched and grabbed their

mother’s mask. It should be noted that this study examined

surgical masks only (which have less ventilation) and

not cloth masks (see Box 1). Although there are fewer

community studies of face mask wearing, appropriate ﬁt

appears to be key to avoid adjusting or removing the mask.

Finally, there has been the additional concern largely voiced

outside of academic publications that particularly if disposal

masks were to be used there would be considerable

environmental waste.

5. Public adherence to face mask and coverings,

relationship with other interventions and importance

of clear and consistent public messaging

A central question asked by many governments without

a previous history of face mask or covering wearing is:

(i) whether their populations would actually adopt this

more invasive behaviour; and, (ii) how face mask wearing

compares to and is related to other non-pharmaceutical

interventions. After exploring this, we then describe the

importance of clear and consistent public messaging,

focusing on an example in the UK.

5.1 Adherence to face mask wearing: a cross-national

comparison

We were able to locate results from two cross-national

representative surveys, which at the time of writing are

pre-prints and not peer-reviewed publications but provide

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 19

a cross-national comparison of face mask and covering

wearing from mid-March to mid-April 2020. It is vital to

note that respondents in these studies were asked about

wearing a face mask due to COVID-19 between mid-

March to mid-April 2020 and that countries were all on

dierent disease trajectories, with varying and changing

policy recommendations during this period. The most

comprehensive study is a representative survey measuring

health behaviour carried out by the Max Planck Institute

for Demographic Research between March 13 to April 19

2020 (N=66,266) across eight countries (Belgium, France,

Germany, Italy, the Netherlands, Spain, UK, US). It is

published as a medRxiv non-peer reviewed pre-print that

examines attitudes and behaviours surrounding COVID-19

including face mask wearing and other non-pharmaceutical

measures

119

. They found that the wearing of a face mask

substantially increased over time (all p <0.001, exception

the Netherlands, Belgium) (see Figure 4 for results of

ﬁnal week of study). Sharp increases of mask wearing of

the general population were observed in Italy (to above

80%), Spain (above 65%), and the US (above 70%). Levels

of face mask usage in France and Belgium in that period

rose to around 40%. The UK and the Netherlands had the

lowest reported levels of face mask wearing, though the

UK still had increases in reported face mask wearing to

over 20% by April 19. This was in addition to hand hygiene

increases particularly in the UK, Germany, Italy and the US.

Focussing on the UK, the study found that behaviours that

were the most adopted in this period was the reduced use

of transportation (p < 0.001) and a sharp increase in social

distancing (p <0.001).

Another survey poll conducted by IPSOS in 15 countries

between March 12 to April 12 of around 2,000 persons per

county (N=29,000) also found cross-national variation in

the wearing of face masks to protect against COVID-19

120

Wearing face masks for protection was at very high levels

in Vietnam (91%), China (83%), Japan (77%) and India (76%),

which were relatively stable over the one month period. Like

the previous study, Italy (81%) and Spain (62%) showed the

highest levels of adopting mask wearing, likely also reﬂected

by the relatively earlier onset and outbreak in those nations.

Lower levels were reported in that period in France (34%),

Canada (28%), Australia (21%), Germany 20%) and the United

Kingdom (16%). The United States had a sharp increase from

11% March 12-14 to 50% within one month by April 9-12, likely

related to the policy change discussed previously of face

coverings in public by the CDC on April 3 2020

121

. Notably,

a PsyArXiv non-peer reviewed pre-print was able to use this

policy introduction as a natural experiment to demonstrate

the impact of the CDC’s recommendation on reported

mask wearing and buying

122

. Using a large nationally-

representative survey across the US (N=3,933) that was in

the ﬁeld, they found signiﬁcant increases in mask wearing

(+12 percentage points) and buying (+7 points) which they

concluded demonstrates the speed at which this behaviour

can be adopted by the population and the importance of

national leadership and clear communication.

5.2 Adherence to other non-pharmaceutical interventions:

face masks and coverings in perspective

We also examined variation in the adherence to a variety

of non-pharmaceutical interventions to put face mask and

covering wearing in perspective. Here we can draw from

the recent cross-national COVID-19 study on reported

behaviours mentioned in the previous section

123

. Here a

hierarchy of adherence to non-pharmaceutical interventions

emerges during COVID-19, which is shown in Figure 4.

A strong caveat, however, is that interventions are rarely

introduced or judged alone but rather as packages,

discussed in more detail in the conclusion. For comparison,

we also examined nine studies mostly of nationally

representative samples covering a range of outbreaks in the

US (see Appendix 3, Additional Results, Figure A3.1)

There was virtually universal enactment of avoiding public

transportation and social distancing, which was the highest

across all countries. This was followed by very high levels

of hand-washing, the highest in the UK which reﬂects a very

strong and clear campaign. This COVID-19 behaviour is in

line with previous research examining the H1N1 pandemic

and SARS who found individuals were very familiar with

hand and respiratory hygiene behaviour (e.g., hand

washing, cough/sneeze etiquette). This in turn meant that

they perceived them as acceptable and common-sense

behaviour they could easily adopt to reduce infection

transmission

124

. The next type of behaviour was the use

of alcohol-based hand sanitizers, which has considerable

cross national variation which may be related to lack of

supply. Wearing face masks was also at very high levels in

many countries by the third week in April 2020. There was

scepticism in the media and by some governments that

public without previous experience of wearing face masks

and coverings would not comply, yet as Figure 4 shows,

self-reported uptake relatively high in most countries,

reaching particularly high levels in Italy (83.8%), the US

(65.8%) and Spain (63.8%). It remained very low, however, in

the UK at 25.9% and the Netherlands at just 7.2%. Appendix

3 also shows compliance to other interventions such as

avoidance of suspected infected people, avoiding touching

eyes, nose and mouth, stopping close contact (shaking

hands, hugging) and avoiding public events, crowds and

cancelling social plans.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 20

5.3 Package policies: face mask usage, physical distancing

rules and hand sanitizer

As noted previously, when countries introduce non-

pharmaceutical interventions, they often work concurrently

to reduce infection transmission. Measures are rarely

introduced as individual interventions, but rather as a

‘package’ including distance, masks and hand hygiene. Due

to the introduction of often multiple interventions at one time

and varying levels and progression of the virus outbreak

across regions, it is dicult – or arguably impossible – to

evaluate the eectiveness of one sole intervention.

A persistent question has been the relationship of face mask

wearing to other behaviours such as social distancing or

disinfection, which is dicult to deﬁnitely determine for the

reasons argued above. A large meta-analysis published

on June 1 2020 in the Lancet linked various interventions

including face masks, eye protection and social distancing

125

They concluded that the risk of being infected was 13%

within one metre and 3% beyond that distance. They

reported that for every extra metre of distance of up to

three meters, the risk is further reduced by half. Wearing a

face mask and eye protection were found to signiﬁcantly

reduce risk of infection, with the duration and intensity

of contact likewise key factors. We note, however, that

this study had several problems and cannot be directly

translated to mask wearing in the general public and was

therefore, for instance, not included in previous advice for

the general public provided by DELVE

126

. Reasons include

the fact that the bulk of the data focussed on MERS and

SARS (and not COVID-19), that most studies were conducted

in healthcare settings (i.e., not wearing in the public), and

that the focus was thus more on protection of the wearer

than source protection (i.e., preventing transmission from

the person wearing the mask). Some of the studies that had

no infections in either the masked or unmasked groups

FIGURE 4

Percentage of individuals who reported having adopted speciﬁc behaviours in response to COVID-19,

3rd week in April 2020 by country.

Source:Perrotta et al. (2020)156, we are grateful that they shared raw data & detailed information. Bar charts show median values and 95% CI as errors.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 21

were not scored, which biased towards the beneﬁt of mask

wearing. There are also concerns about the study of the

relationship between distance and risk, which is based on

extrapolations and the assumption of a linear relationship

between distance and risk (whereas it is likely exponential).

We note therefore, concerns about the conclusions of this

study and particularly those related to distancing.

At the time of writing in late June 2020, governments across

the world had dierent rules regarding social distancing

and were examining whether other non-pharmaceutical

interventions would be eective in reducing those levels. At

the time of writing, countries with the largest distance of 2

metres (6.5 feet) are Canada, Spain and the UK. The US has

1.8 metres, Germany and Australia 1.5, whereas the WHO,

China and France all suggest a 1 metre (3.3 feet) distance.

In England, the government previously advised to: “Keep

your distance if you go out - 2 metres apartwhere possible”.

In the COVID-19 guidance for employers and employees,

the Department of Business notes: “Maintain two-metre

social distancing,where possible” and “Where possible, you

should maintain two metres between people…Where it’s

not possiblefor people to be two metres apart, you should

do everything practical to manage the transmission risk.” On

June 23 2020, this was updated with the announcement

that as of July 4 2020, the government suggests where it

is not possible to stay two metres apart, guidance will allow

people to keep a social distance of ‘one metre plus’, with

plus seeming to suggest some sort of additional mitigation

127

In response to easing the lockdown, countries such as

South Korea and Portugal maintained the 2 metre distance

where possible, but then clariﬁed the etiquette of combining

dierent types of interventions (hand sanitizer, mask use)

when the 2 metre rule was not possible. In South Korea,

the advice is to maintain the 2 metre guideline (or ‘two

arms lengths’) in general, which is relaxed to one metre

inside shops, restaurants and cafes. As of June 01 2020 in

Portugal, the 2 metre rule was suggested where possible.

In Japan, the ‘3C’ rule has been implemented, which

is to avoid crowded places, closed spaces with poor

ventilation and close contact, all with wearing a face mask

where possible. The emphasis has been less on rules and

more on understanding how to avoid the transmission of

droplets from one household to another, such as through

social bubbles and tight networks

128

. Distance is often not

addressed in exclusion, but as a policy package of almost

universal use of hand sanitizer when entering (shops,

restaurants) and demanding that face masks are worn inside

shops. There is also education about etiquette in places like

restaurants, with sanitizers at tables, all sta wearing masks,

instructions to take o masks when sitting and putting it

on when going to the washroom. At the moment Canada

is also moving towards a policy across dierent provinces

that combines face masks with a reduction of the 2 metre

‘hockey-stick’ rule. Health Canada now recommends that

people wear cloth masks when social distancing of 2 metres

is not possible, particularly in crowded public settings, such

as stores, shopping areas and public transportation.

5.4 The importance of communications, clear and consistent

public messaging

This review of the behavioural literature on face masks and

coverings also revealed the importance of communications

and public messaging during pandemic outbreaks on the

eective implementation of and adherence to face mask

and other non-pharmaceutical interventions (see e.g.,

Figure 5). A study of non-pharmaceutical interventions in

Canada during the SARS outbreak found that inconsistent

information from various sources prompted individuals

to question the credibility of available information

129

. This

inconsistent information resulted in fear and denial of the

pandemic. Many participants in the study expressed doubts

about the information from the public health department,

which in turn inﬂuenced their level of perceived risk. For

example, there were mixed messages about who needed to

quarantined, with some members of the household asked to

be quarantined whereas others were not.

Figure 5 also provides a recent example of mixed

messaging that might be confusing to the public and

contradicts some of the knowledge in this report. The top

panel shows some positive aspects of the messaging such

as clarifying that it is an altruistic behaviour to protect others

and face coverings can be worn. It however engages in

othering of a vulnerable group who is the least likely to

break rules (and be symptomatic), does not focus on self

protection and now as we explore in the next section shows

the least protective and non-recommended item - a scarf.

Literature examining the H1N1 2009 pandemic in the UK and

Spain concluded that the public became sceptical about the

way in which the communication about this new respiratory

infection was presented, particularly by the media. They

found the communications to be unreliable, premature,

inconsistent, sensationalist and unduly alarmist. Several

UK studies reported scepticism from individuals due their

perception of the media’s propensity to create hype and

panic in what they viewed as an attempt to scare people

130,

131

. A Spanish study of H1N1 likewise concluded that reporting

was at ﬁrst sensationalist, followed by contradictory advice

coming from their own Spanish politicians and ocials

versus other international leaders

132

. Doubts about the

trustworthiness of information and a general information

fatigue were related to people disregarding advice in New

Zealand and the UK in

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 22

FIGURE 5

Example of public messaging about face coverings, UK Government, 27 June 2020.

Note: Full text above reads ‘Wear a face covering when you go to the shop. It will help protect others from #coronavirus if you are infected but not

displaying symptoms."

Source:Twitter, UK Prime Minister @10Downing Street, June 17 (accessed 17 June 2020).

PROs

• Good – asymptomatic –

altruistic – protect others

• Good – clear covering is cloth

(not surgical mask, N95)

CONs

• Bad: othering of older, vulnerable group,

least likely to break rules, expression

unclear, ominous grey background

• Bad: wearing scarf – least eective

cloth covering

• Bad: Unclear – above 'when you go

to the shop' – text on photo suggests

in the shop

• Bad: focus only pn protecting others and

not self protection

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 23

relation to H1N1

133,

134

. Another reason for scepticism about

the information was related to mixed messages in the

media and feeling that it was dicult to sift between facts

and opinions.

As many governments that lacked experience from SARS

and previous outbreaks, the WHO, the most trusted global

supranational organisation for health advice, has changed

positions on face mask use and wearing for the general

public between early April 2020

135

and revised in June

2020

136

. This likely initiating a repeat of some of the public’s

previous reaction to changing information about SARS

and H1N1. As the initial section on international policies on

face masks has shown, nations and even states, countries

or regions within these have adopted dierent face mask

policies. As noted previously, this has been the case with

Scotland, Wales and England.

6. Conclusion

This report provides evidence on the main policy directives

of face masks in the international sphere in addition to

the core factors related to the adherence of wearing

a face mask and covering. We found many similarities

across previous pandemics and ‘lessons learned’ that

appeared to repeat themselves from the Spanish ﬂu

through to SARS, MERS, H1N1 to COVID-19. It was clear

that core socio-behavioural factors were pivotal such as

the public’s understanding of the virus, risk perception,

previous experience with mask wearing and socio-political

systems, individual characteristics and perceived barriers.

We likewise found importance in consistent and clear public

messaging; including adopting a package of policies and

that countries without a previous history of mask wearing did

indeed adopt this behaviour. Finally, we produced evidence

showing that cotton face coverings can provide signiﬁcant

protection against the transmission of aerosol particles.

The key points from this report are:

• Cloth face coverings are eective in reducing source

virus transmission, i.e., outward protection of others, when

they are of optimal material and construction (high grade

cotton, hybrid and multilayer) and ﬁtted correctly and for

source protection of the wearer

• Socio-behavioural factors are vital to understanding

public adherence to wearing face masks and coverings,

including public understanding of virus transmission, risk

perception, trust, altruism, individual traits, perceived

barriers

• Face masks and coverings cannot be seen in isolation but

are part of ‘policy packages’ and it is imperative to review

interrelated non-pharmaceutical interventions in tandem

including hand hygiene, sanitizers and social distancing

when maintaining the 2 metre or 1 metre+ distancing rule

is not possible

• Consistent and eective public messaging is vital to

public adherence of wearing face masks and coverings.

Conﬂicting policy advice generates confusion and lack

of compliance. Populations without a previous history

of mask wearing have rapidly adopted face coverings

during the COVID-19 period.

This report provides scientiﬁc evidence in which experts

and governments can inform their decision-making but

does not extend to direct policy directives. As noted in

the disclaimer and elsewhere

137

, but also in our GRADE

recommendations (Appendix 5), research and policy-making

in this area is ongoing and continuously under revision.

We do note however, that the current advice in England

regarding non-surgical face coverings for the general public,

employers and employees does not align with the broader

science evidence in this report. For example, COVID-19

secure guidance for businesses and sta issued by HM

Government on June 14 2020

138

states (see Appendix 4 for

entire text, emphasis added by authors):

“There are some circumstances when wearing a face

covering may be marginally beneﬁcial as a precautionary

measure. The evidence suggests that wearing a face

covering does not protect you, but it may protect others if

you are infected but have not developed symptoms...

It is important to know that the evidence of the beneﬁt of

using a face covering to protect others is weak and the

eect is likely to be small, therefore face coverings are

not a replacement for the other ways of managing risk,

including minimising time spent in contact, using ﬁxed teams

and partnering for close-up work, and increasing hand and

surface washing.”

“ Three main factors stand in the way of prevention: First,

public indierence. People do not appreciate the risks

they run. The second factor…..is the personal character of

the measures which must be employed…It does not lie in

human nature for a man who thinks he has only a slight

cold to shut himself up in rigid isolation... Third, the highly

infectious nature of the respiratory infections adds to the

diculty of their control.”

Major George A Soper, 1919, The Lessons of the Pandemic, Science

157

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 24

We note that although evidence is mentioned, there is no

clear reference to the speciﬁc material.

We also note that particularly in relation to face masks

and coverings, there has been a particular precaution in

some contexts such as England, that seems to override

the scientiﬁc evidence and lack transparency in decision-

making

139, 140

. It may be attributed to several factors. First, in

the face mask and covering sphere, there has been a focus

on highlighting very small fragmented pieces of knowledge

and assertion that evidence was not strong due to the

lack of clear RCT

141

. As noted previously, there have also

been no clinical trials of coughing into your elbow, social

distancing and quarantine, yet these measures are seen

by the public and policy-makers as common sense and

have been widely adopted and are considered as eective.

The heterogeneity of the research designs does not ﬁt the

standard RCT evidence-based medicine approach

142

, yet

there are still many high quality studies (see Appendix 5).

A non-peer reviewed medRxiv pre-print of a systematic

review of facemasks likewise concluded that RCTs “may

not be the best quality evidence to evaluate a population

behaviour like facemask use that is likely to be imperfectly

implemented”. They conclude that “compared to RCTs,

observational data (cohort and case-control studies) may

give superior quality evidence for ecacy of facemask

wearing to avoid inﬂuenza-like-illness, given they are trying

to relate actual behaviour to outcomes”

143

. Another non-

peer reviewed pre-print study released on 23 June 2020

linked the face mask wearing rate to country’s COVID-19

death rates

144

. As we note, however, throughout this report

although there may be a correlation, it is likely never

one policy in isolation and rather a combined layering or

package of policy eects. Second, the lack of decisive

measures and changing positions on face masks and

coverings for the general public by the WHO and some

governments has undoubtably fuelled this uncertainty

about their eectiveness. Third, an additional debate has

been about the applicability of results across multiple

settings (e.g., health care versus in the community),

pandemics (e.g., can SARS research be relevant for

COVID-19) and cross-national dierences. We note that

although there are core dierences, there can be many

standard ‘lessons learned’ from previous and other

experiences that appear to be repeating themselves. Finally,

recommendations and guidelines often either implicitly or

explicitly considered supply issues and concerns about

access and competition of the public taking away vital PPE

equipment of surgical masks. This is a logistical and supply

issue and not an issue about the eectiveness of face

masks and coverings.

We note various limitations of our work and provide an

attempt to scrutinize and GRADE

145

our report in relation to

the quality of the evidence (Appendix 5). The vast literature

review also covered non-COVID-19 studies and dierent

nations, which although had the strength of breadth,

ignored intricate dierences. We also note that although

we report cloth face coverings to be eective, the meta-

analysis was in a health care setting and the fabric tested

within a laboratory, but evidence was supplemented from

observational studies

146

. Further testing in community

settings would be desirable. Few RCTs have been

conducted to examine the eectiveness of dierent types

of face masks and coverings. But as noted throughout this

report, in addition to ethical concerns, this seems highly

unrealistic to devise such a study, particularly in current

circumstances.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 25

Appendix 1. International policies on facemask requirements and recommendations

TABLE A1.1

Type of requirement by number of countries (as of June 15 2020)

General category Total

countries

Detailed categories Total

countries

Indoor public places 15 All commercial establishments 1

All indoor public places 10

All indoor public places and outdoor within 20 meters of others 1

All indoor public places with multiple people 2

Supermarkets, banks and some indoor spaces 1

Everywhere in public 71 Everywhere in public 71

Certain public places 9 Certain public places 1

Everywhere in public (major cities) 1

Everywhere in public where social distancing isn’t possible 6

Public roads and business employees 1

Public transport only 7 Public transport only 7

Public transport and

crowded places

12 Public transport and schools 1

Public transport and shopping 1

Public transport and stores 2

Public transport + everywhere in public where social distancing isn’t possible 1

Public transport + everywhere in public with more than 10 people 1

Public transport + markets + most public places 1

Public transport + select states: everywhere 1

Public transport + shopping 2

Public transport + shops 1

Public transport, markets, supermarkets & crowded places 1

Universal mask usage 6 Universal mask usage 6

Total countries

information

available***

120 120

Source: Masks4all data158; general categories created by authors. Note: ***This information is only available for 120 countries.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 26

Appendix 2. Data and Methods

A2.1 Further information systematic review study selection

Figure A2.1 provides a detailed illustration of study selection

and the period of the aforementioned previous respiratory

outbreaks. After duplicates were removed and selection

criteria was enacted, we produced a harmonized ﬁle. The

majority of the studies (561) on aggregate are returned

from queries relating to H1N1, but for the year of 2020, it is

naturally COVID-19 (263 in 6-months to date). In this rapid

policy response brief that needed to be produced very

quickly, we do not provide details of all study selection

and exclusion, but will do so in a more detailed future

publication. Brieﬂy, we identiﬁed and excluded

duplicate articles and studies that did not include humans.

We included all studies that were returned from three

leading bibliographic databases (Scopus, PubMed and

Web of Science). Due to the rapid shifts in knowledge

surrounding COVID-19 and longer publication time for most

behavioural and social science journals, we also included

some pre-print non-peer reviewed articles from SocRxiv,

PsyRxiv, MedRxiv, bioRxiv and SSRN and indicate this when

evidence is provided. There was no selection on language

but the majority of articles are in English. We included all

research designs, with the exception of the meta-analysis,

discussed in detail in relation to that analysis.

FIGURE A2.1

Contents of the systematic review study of the face mask and related literature.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 27

Appendix 3. Additional results

Note: In these US studies, the majority of the samples are taken from randomly drawn representative national samples of US adults studying SARS

(N=1025)

159

, response to a hypothetical serious infectious outbreak (N=500)

160

, H1N1 (N=1290)

161

, national representative US opinion polls 2009-10 (N varies)

162

a random sample of adults in Arizona (H1N1, N=727)

163

and in two counties in North Carolina (H1N1, N=207)

164

. The remainder used convenience samples of

adult travellers at 4 international airports (H5N1, N=1301)

165

, parents of children in San Antonio, Texas after a H1N1 outbreak (N=727)

166

and an internet survey of

Stanford alumni and students (H1N1, N=6249)

167

. A comparative summary of further details about these studies can also be found in

168

FIGURE A3.1

Compliance of various non-pharmaceutical health interventions during previous outbreaks, US studies only.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 28

Appendix 4. UK government advice on face coverings

(14 June 2020)

Full excerpt from document: HM Government. Working

safely during the COVID-19 in construction and other

outdoor work. COVID-19 secure guidance for employers,

employees and the self-employed, Version 2.0 updated 14

June 2020

147

There are some circumstances when wearing a face

covering may be marginally beneﬁcial as a precautionary

measure. The evidence suggests that wearing a face

covering does not protect you, but it may protect others if

you are infected but have not developed symptoms.

A face covering can be very simple and may be worn in

enclosed spaces where social distancing isn’t possible. It

just needs to cover your mouth and nose. It is not the same

as a face mask, such as the surgical masks or respirators

used by health and care workers.

Similarly, face coverings are not the same as the PPE used

to manage risks like dust and spray in an industrial context.

Supplies of PPE, including face masks, must continue to be

reserved for those who need them to protect against risks in

their workplace, such as health and care workers, and those

in industrial settings like those exposed to dust hazards.

It is important to know that the evidence of the beneﬁt of

using a face covering to protect others is weak and the

eect is likely to be small, therefore face coverings are

not a replacement for the other ways of managing risk,

including minimising time spent in contact, using ﬁxed teams

and partnering for close-up work, and increasing hand and

surface washing. These other measures remain the best

ways of managing risk in the workplace and government

would therefore not expect to see employers relying on

face coverings as risk management for the purpose of their

health and safety assessments.

Wearing a face covering is optional and not required by law,

including in the workplace. If you choose to wear one, it is

important to use face coverings properly and wash your

hands before putting them on and before taking them o.

FIGURE A3.2

Funnel plot with pseudo 95% conﬁdence intervals, meta-analysis.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 29

Employers should support their workers in using face

coverings safely if they choose to wear one. This means

telling workers:

• Wash your hands thoroughly with soap and water for

20 seconds or use hand sanitiser before putting a face

covering on, and before and after removing it.

• When wearing a face covering, avoid touching your face

or face covering, as you could contaminate them with

germs from your hands.

• Change your face covering if it becomes damp or if

you’ve touched it.

• Continue to wash your hands regularly.

• Change and wash your face covering daily.

• If the material is washable, wash in line with

manufacturer’s instructions. If it’s not washable, dispose of

it carefully in your usual waste.

• Practise social distancing wherever possible.

You can make face-coverings at home and can ﬁnd

guidance on how to do this and use them safely on GOV.UK

Appendix 5. GRADE Recommendations

The authors apply GRADE (Grading of Recommendations,

Assessment, Development, and Evaluation)

recommendations

148

. The advantage of using GRADE is

that it ensures both a systematic process and transparency

of research and transparently note the quality of evidence

for each topic studied. The limitations of GRADE is that the

steps and recommendations are narrowly gauged towards

medical research. For instance, the ﬁrst step is an a-priori

ranking of ‘high’ quality to the yardstick of randomized

control trials and ‘low’ to observational studies, with the

underlying assumption that RCTs are less prone to bias.

Bias is related to lack of blinding, the trial being cut short,

etc., which does not cover common bias issues in socio-

behavioural research.

Although we ﬁnd this approach useful for transparency,

and explicitly noting the strengths, limitations and our

assessment of the quality of evidence, strictly applying

GRADE recommendations is problematic for the current

study for two reasons. First, our report covers a vast array of

social and behavioural research vital to our understanding

of face mask and covering wearing, which is by deﬁnition

almost always observational studies, many of which are

considered of very high quality within those disciplines.

Second, a core criticism of the face mask literature, and in

particular cloth or non-surgical face coverings has been the

lack of RCTs. As mentioned in the report and eloquently

argued elsewhere

149

, this is unrealistic for ethical but also

practical reasons. This may be another reason that ‘weak’

or ‘lack of evidence’ has been ascribed to face mask and

coverings for the general public.

We note there have also been no clinical trials of hand-

washing, coughing into your elbow, social distancing and

quarantine, yet these measures have been widely adopted

and are considered as eective.

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 30

TABLE A5.1

GRADE recommendations

Topic Section Type of

research

Strengths Limitations Level quality

of evidence

Eectiveness

of cloth face

coverings.

2.2-2.6 Meta-analysis

of infection

reduction of

cotton masks.

More systematic and

empirical examination

beyond only a narrative

review.

Relatively

homogeneous studies

in similar setting,

country.

No RCTs were able

to be included, with

estimates come from

observational research

designs only.

Studies all in a

healthcare setting in

one country.

Examined SARS and

H1N1 and not COVID-19

setting.

Small number of

studies.

Moderate-quality (to

translate to public

setting).

No broader community

settings (e.g., transport,

shops) conducted as it

is virtually impossible to

conduct RCTs

2.7 Eectiveness

of cloth masks

by fabric type,

construction.

Rigorous study carried

out with multiple types

of fabrics and hybrid

construction.

Attention to use by

adding measure of

‘gap’ as proxy for

incorrect ﬁt or usage.

Tests carried out in lab

and not community

setting.

May be other

measures beyond ﬁt

and gap related to

eectiveness.

Moderate- to high-

quality (need to

translate to community

setting; replication).

International

face mask

policy

comparative

data.

3 International

comparative

data is scarce

with few

comparative

measures

available. Data

is taken from

Masks4all

170

Contains data for 188

countries.

Direct links to the

source of each policy

are provided.

Does not divide into

dierent regions (e.g.,

UK, Scotland, Wales or

States in US).

Not an ocial or

supranational data

source.

Lack of other

comparative databases

to check validity.

Moderate-quality (at

least 1 primary source

with traceable links).

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 31

Topic Section Type of

research

Strengths Limitations Level quality

of evidence

Behavioural

literature

systematic

review (high

quality unless

section

otherwise

noted below)

4 Highly

heterogeneous

literature taken

from multiple

sources of

PUBMED, Web

of Science,

Scopus and

preprints

(PsyRArix,

SocRxiv, SSRN,

MedRxiv,

bioRxiv)

Highly

heterogeneous

literature with

multiple study

designs

Captures wide breadth

of interdisciplinary

research

Captures hard to

measure topics such

as risk perception and

public attitudes

Preprints capture most

recent knowledge on

the topic

Strength of breadth of

knowledge

Disciplines approach

topics in varied

manners, making

direct comparisons

sometimes challenging

Due to slower

publication process

of social sciences,

COVID-19 studies often

pre-prints without peer

review (when this is the

case it is indicated in

the review)

Dicult to empirically

or systematically

analyse as in for

example a meta-

analysis or RCT

High-quality (contains

multiple systematic

reviews with consistent

results, but we note

that in some cases the

quality of studies are

mixed)

Beyond

COVID-19 and

coronavirus,

included also

literature

on previous

pandemics

such as SARS,

H1N1, MERS

Larger body of

literature to draw

conclusions from

Ability to pick up

‘lessons learned’

that are not possible

since COVID-19 is

still ongoing in many

countries

Dierences in

pandemics (country,

virus)

Knowledge not always

directly applicable due

to national, medical or

societal dierences

(e.g., culture of mask

wearing, trust in

government)

As above

4.4 Topic of

socio-political

systems

and trust in

government

and science

Builds on larger body

of research, theories

COVID-19 research

available using large-

scale data and multiple

types of research

designs

COVID-19 research

based on several

pre-print non-peer-

reviewed studies

(indicated in section)

Moderate-quality

(smaller base of COVID

speciﬁc literature from

pre-prints)

4.5 Vulnerable

groups &

discrimination

in relation to

COVID-19

Builds on existing

literature of

discrimination of

groups during other

pandemics

Most COVID-19

research on this topic

are pre-prints and not

peer reviewed

Experiments of

discrimination, small n,

selective samples, may

not translate to real-

world settings

Low-quality to no

evidence (COVID-

speciﬁc research)

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 32

Topic Section Type of

research

Strengths Limitations Level quality

of evidence

4.6 Perceived

barriers

Extensive material and

studies, particularly on

COVID-19 since 2020

Fit and comfort

examined in detail

(virtually all in relation

to health care

professionals)

Mostly in relation to

public health workers,

less on general public

Fit and comfortable of

surgical & respirator

masks very dierent

from cloth coverings

High-quality for health

workers

Limited evidence for

general public

Public

adherence

to face mask

coverings and

relationship

to other

interventions

5.1 Cross national

comparisons

of face mask

wearing and

coverage during

COVID-19

Two nationally

representative surveys

from dierent sources of

8 countries (N=66,266)

and 15 countries

(N=29,000)

Both cover period of

mid-March to mid-April

2020 using self-reports

Countries at dierent

stages of disease

trajectories with varying

& changing policies

over that period

Limited country

coverage

Moderate- to high-

quality with consistent

results across studies

5.2 Policy packages

of non-

pharmaceutical

interventions

Multiple studies and

systematic review

across US and multiple

pandemics

Cross-national

COVID-19 study of

current interventions

Dicult to separate

dierent interventions,

often introduced in

tandem

Dicult to know

whether intervention

was mandatory or

recommended, plus

changes over time

High-quality, multiple

studies with consistent

results

5.3 Face mask

usage in relation

to physical

distancing

Meta-analyses that

include multiple

interventions in health-

care settings

COVID-19 information

on this topic available

from governmental

websites and news

sources

Dicult to translate

research from health-

care settings to general

public

Fast moving

government advice

Information from news

sources is not peer-

reviewed research

No to low-quality for

general public (note:

dicult to obtain

evidence due to

diculties in separating

the impacts of dierent

policies, current topic)

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 33

Appendix 6. Preparation of Report

Report prepared for the SET-C Group by

Professor Melinda Mills FBA, MBE, University of Oxford,

Leverhulme Centre for Demographic Science

Dr. Charles Rahal, University of Oxford, Leverhulme Centre

for Demographic Science

Evelina Akimova, University of Oxford, Leverhulme Centre

for Demographic Science

Members of The Royal Society SET-C who contributed

to and commented on this report

Professor Peter Bruce FRS (Chair), The Royal Society

Professor Sir Roy Anderson FMedSci FRS, Imperial

College London

Professor Charles Bangham FMedSci FRS, Imperial

College London

Professor Richard Catlow FRS, The Royal Society

Professor Christopher Dye FMedSci FRS, University

of Oxford

Professor Sir Marc Feldmann AC FAA FMedSci FRS,

University of Oxford

Professor Sir Colin Humphreys FREng FRS, Queen

Mary University of London

Professor Frank Kelly FRS, University of Cambridge

Professor Melinda Mills FBA, University of Oxford

Professor Sir John Skehel FMedSci FRS, The Francis

Crick Institute

Professor Georey Smith FMedSci FRS, University

of Cambridge

Professor Alain Townsend FRS, University of Oxford

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 34

References

1 DELVE. 2020 Face masks for the general public. See https://rs-delve.

github.io/reports/2020/05/04/face-masks-for-the-general-public.html

(accessed 26 June 2020).

2 Centres for Disease Control and Prevention (CDC). 2020

Considerations for wearing cloth face coverings. See https://www.

cdc.gov/coronavirus/2019-ncov/prevent-getting-sick/cloth-face-

cover-guidance.html?CDC_AA_refVal=https%3A%2F%2Fwww.cdc.

gov%2Fcoronavirus%2F2019-ncov%2Fprevent-getting-sick%2Fcloth-

face-cover.html (accessed 26 June 2020).

3 World Health Organisation (WHO). Advice on the use of masks in the

context of COVID-19: interim guidance, 5 June 2020. See https://apps.

who.int/iris/handle/10665/332293 (accessed 26 June 2020).

4 Greenhalgh T et al. 2020 Face masks for the public during the

COVID-19 crisis. BMJ. (doi:10.1136/bmj.m1435).

5 Chu DK et al. 2020 Physical distancing, face masks and eye

protection to prevent person-to-person transmission of SARS-CoV-2

and COVID-19: a systematic review and meta-analysis. The Lancet,

395, P1973-1987. (doi:10.1016/S0140-6736(20)31142-9).

6 Op. cit, note 5

7 Bartoszko JJ et al. 2020 Medical masks vs N95 respirators for

preventing COVID-19 in healthcare workers: A systematic review and

meta-analysis of randomized trials. Inﬂuenza and Other Respiratory

Viruses, 14, 365-373. (doi:10.1111/irv.12745).

8 Xiao J et al. 2020 Nonpharmaceutical Measures for Pandemic

Inﬂuenza in Nonhealthcare Settings—Personal Protective and

Environmental Measures. Emerg. Infect. Dis. 26, 967–975.

9 Brainard JS, Jones N, Lake I, Hooper L, Hunter P. 2020 Facemasks

and similar barriers to prevent respiratory illness such as COVID-19: A

rapid systematic review. medRxiv. (doi:10.1101/2020.04.01.20049528).

10 Op. cit, note 4

11 Xiao J et al. 2020 Nonpharmaceutical Measures for Pandemic

Inﬂuenza in Nonhealthcare Settings—Personal Protective and

Environmental Measures. Emerg. Infect. Dis. 26, 967–975.

12 Brainard JS, Jones N, Lake I, Hooper L, Hunter P. 2020 Facemasks

and similar barriers to prevent respiratory illness such as COVID-19: A

rapid systematic review. medRxiv. (doi:10.1101/2020.04.01.20049528).

13 Wang Y et al. 2020 Reduction of secondary transmission of SARS-

CoV-2 in households by face mask use, disinfection and social

distancing: a cohort study in Beijing, China. BMJ Global Health, 5,

e002794. See https://gh.bmj.com/content/5/5/e002794 (accessed 26

June 2020).

14 Lau JTF et al. 2004 SARS transmission, risk factors, and prevention in

Hong Kong. Emerging Infectious Diseases, 10, 587–592. (doi:10.3201/

eid1004.030628).

15 Wu J et al. 2004 Risk factors for SARS among persons without known

contact with SARS Patients, Beijing, China. Emerging Infectious

Diseases, 10, 210–216. (doi:10.3201/eid1002.030730).

16 Tuan PA et al. 2007 SARS transmission in Vietnam outside of the

health-care setting. Epidemiology and Infection, 135, 392–401.

(doi:10.1017/S0950268806006996).

17 Chu DK et al. 2020 Physical distancing, face masks, and eye

protection to prevent person-to-person transmission of SARS-CoV-2

and COVID-19: a systematic review and meta-analysis. Lancet.

(doi:10.1016/S0140-6736(20)31142-9).

18 Xiao J et al. 2020 Nonpharmaceutical Measures for Pandemic

Inﬂuenza in Nonhealthcare Settings—Personal Protective and

Environmental Measures. Emerg. Infect. Dis. 26, 967–975.

19 Op. cit, note 2

20 MacIntyre C R et al. 2015 A cluster randomised trial of cloth masks

compared with medical masks in healthcare workers. BMJ Open. 5,

e006577. (doi:10.1136/bmjopen-2014-006577).

21 Yin W et al. 2004. Eectiveness of personal protective measures in

prevention of nosocomial transmission of severe acute respiratory

syndrome. Zhonghua Liu Xing Bing Xue Za Zhi. 1, 18-22. See https://

pubmed.ncbi.nlm.nih.gov/15061941/ (accessed 26 June 2020).

22 Andersen B M. 2019 Prevention and Control of Infections in

Hospitals: Practice and Theory. Springer International Publishing.

(doi:10.1007/978-3-319-99921-0).

23 Op. cit note 5

24 #MASKS4ALL, 2020 What Countries Require Masks in Public or

Recommend Masks? See https://masks4all.co/what-countries-require-

masks-in-public/ (accessed 26 June 2020).

25 World Health Organisation (WHO). Advice on the use of masks in

the context of COVID-19: interim guidance, 6 April 2020. See https://

apps.who.int/iris/bitstream/handle/10665/331693/WHO-2019-nCov-

IPC_Masks-2020.3-eng.pdf?sequence=1&isAllowed=y (accessed 26

June 2020).

26 European Centre for Disease Prevention and Control (ECDC). Using

face masks in the community: reducing COVID-19 transmission from

potentially asymptomatic or pre-symptomatic people through the use

of face masks. Technical Report 8 April 2020. See https://www.ecdc.

europa.eu/sites/default/ﬁles/documents/COVID-19-use-face-masks-

community.pdf (accessed 26 June 2020).

27 Rothe C et al. 2020 Transmission of 2019-nCoV Infection from an

Asymptomatic Contact in Germany. NEJM, 382, 970–971. (doi:10.1056/

NEJMc2001468).

28 Op. cit note 2

29 Op. cit note 3

30 Scottish Government. Face coverings mandatory on public transport

from 22 June. See https://www.gov.scot/news/face-coverings-

mandatory-on-public-transport-from-22-june/ (accessed 26 June

2020).

31 Welsh Government. Face coverings: frequently asked questions.

See https://gov.wales/face-coverings-frequently-asked-questions

(accessed 26 June 2020).

32 Department for Transport. Face coverings to become mandatory on

public transport. See https://www.gov.uk/government/news/face-

coverings-to-become-mandatory-on-public-transport (accessed 26

June 2020).

33 Rothstein MA et al. 2003 Quarantine and Isolation: Lessons learned

from SARS. A Report to the Centers for Disease Control and

Prevention. See https://biotech.law.lsu.edu/blaw/cdc/SARS_REPORT.

pdf (accessed 26 June 2020).

34 Garcia LP. 2020 Use of facemasks to limit COVID-19 transmission.

Uso de máscara facial para limitar a transmissão da COVID-19.

Epidemiol Serv Saude. 29, e2020023. (doi:10.5123/S1679-

49742020000200021).

35 Ministry of Health Mexico. 2009 Actions to contain transmission of

inﬂuenza A H1N1. Acciones para contener la transmisión de Inﬂuenza

A HINI. en Secretaría de Salud, México. See http://www.issste-

cmn20n.gob.mx/Archivos%20PDF/Acciones_Basicas_inﬂuenza_

SS_20090511.pdf (accessed 26 June 2020).

36 Liberati A et al. 2009 The PRISMA statement for reporting systematic

reviews and meta-analyses of studies that evaluate healthcare

interventions: explanation and elaboration. BMJ. 339, b2700–b2700.

(doi:10.1136/bmj.b2700).

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 35

37 Stroup DF et al. 2000 Meta-analysis of Observational Studies in

Epidemiology: A Proposal for Reporting. JAMA, 83, 2008-12. (doi:

10.1001/jama.283.15.2008).

38 Mills MC, Rahal C. 2019 A scientometric review of genome-wide

association studies. Communications Biology, 2. (doi:10.1038/s42003-

018-0261-x).

39 Mills MC, Rahal C. 2020 The GWAS Diversity Monitor tracks diversity

by disease in real time. Nature Genetics, 52, 242–243. (doi:10.1038/

s41588-020-0580-y).

40 Van Bavel JJ et al. 2020 Using social and behavioural science to

support COVID-19 pandemic response. Nature Human Behaviour, 4,

460–471. (doi:10.1038/s41562-020-0884-z).

41 Guyatt GH et al. 2008 GRADE: an emerging consensus on rating

quality of evidence and strength of recommendations. BMJ. 336,

924–926. (doi:10.1136/bmj.39489.470347.AD).

42 Teasdale E et al. 2014 Public perceptions of non-pharmaceutical

interventions for reducing transmission of respiratory infection:

systematic review and synthesis of qualitative studies. BMC

Public Health, 14. See https://bmcpublichealth.biomedcentral.com/

articles/10.1186/1471-2458-14-589 (accessed 26 June 2020).

43 Gershon RR et al. 2018 Adherence to Emergency Public Health

Measures for Bioevents: Review of US Studies. Disaster Med. Public

Health Prep. 12, 528–535. (doi:10.1017/dmp.2017.96).

44 Op. cit note 42.

45 Op. cit note 26

46 Bricker D. 2020 More people say they’re wearing masks to protect

themselves from COVID-19 since March. IPSOS. See https://www.

ipsos.com/en/more-people-say-theyre-wearing-masks-protect-

themselves-covid-19-march (accessed 26 June 2020).

47 Caress AL et al. 2010 Exploring the needs, concerns and behaviours

of people with existing respiratory conditions in relation to the H1N1

‘swine inﬂuenza’ pandemic: a multicentre survey and qualitative study.

Health Technology Assessment, 14, 1-108. (doi:10.3310/hta14340-01).

48 Teasdale E, Yardley L. 2011 Understanding responses to government

health recommendations: public perceptions of government

advice for managing the H1N1 (swine ﬂu) inﬂuenza pandemic.

Patient Education and Counselling, 85, 413–418. (doi:10.1016/j.

pec.2010.12.026).

49 Gray L et al. 2012 Community responses to communication campaigns

for inﬂuenza A (H1N1): a focus group study. BMC Public Health. 12,

205. (doi:10.1186/1471-2458-12-205).

50 Menni C et al. 2020 Real-time tracking of self-reported symptoms to

predict potential COVID-19. Nature Medicine. (doi:10.1038/s41591-020-

0916-2).

51 Teasdale E et al. 2012 The importance of coping appraisal in

behavioural responses to pandemic ﬂu. British Journal of Health

Psychology, 17, 44–59. (doi:10.1111/j.2044-8287.2011.02017.x).

52 Bish A, Michie S. 2010 Demographic and attitudinal determinants of

protective behaviours during a pandemic: A review. British Journal of

Health Psychology, 15, 797–824. (doi:10.1348/135910710X485826).

53 Cowling BJ et al. 2010 Face masks to prevent transmission of

inﬂuenza virus: a systematic review. Epidemiology and Infection, 138,

449–456. (doi:10.1017/S0950268809991658).

54 Goodwin, R et al. 2009 Initial psychological responses to Inﬂuenza A,

H1N1 (“Swine ﬂu”). BMC Infectious Diseases, 9. (doi:10.1186/1471-2334-

9-166).

55 Lau JTF et al. 2011 Changes in knowledge, perceptions, preventive

behaviours and psychological responses in the pre-community

outbreak phase of the H1N1 epidemic. Epidemiology and Infection,

139, 80–90. (doi:10.1017/S0950268810001925).

56 Lau JTF et al. 2007 Anticipated and current preventive behaviours in

response to an anticipated human-to-human H5N1 epidemic in the

Hong Kong Chinese general population. BMC Infectious Diseases, 7.

(doi:10.1186/1471-2334-7-18).

57 Liao Q et al. 2010 Fielding, Situational Awareness and Health

Protective Responses to Pandemic Inﬂuenza A (H1N1) in Hong Kong:

A Cross-Sectional Study. PLoS One, 5, e13350. (doi:10.1371/journal.

pone.0013350).

58 Prati G et al. 2011 Compliance with recommendations for pandemic

inﬂuenza H1N1 2009: the role of trust and personal beliefs. Health

Education Research, 26, 761–769. (doi:10.1093/her/cyr035).

59 Rubin GJ et al. 2010 The impact of communications about swine

ﬂu (inﬂuenza A H1N1v) on public responses to the outbreak: results

from 36 national telephone surveys in the UK. Health Technology

Assessment, 14, 183-266. (doi:10.3310/hta14340-03).

60 Rubin GJ et al. 2009 Public perceptions, anxiety, and behaviour

change in relation to the swine ﬂu outbreak: cross sectional

telephone survey. BMJ, 339, b2651–b2651. (doi:10.1136/bmj.b2651).

61 Floyd DL et al. 2000 A Meta-Analysis of Research on Protection

Motivation Theory. Journal of Applied Psychology, 30, 407–429.

(doi:10.1111/j.1559-1816.2000.tb02323.x).

62 Op. cit note 61

63 Op. cit note 41

64 Cava MA et al. 2005 Risk Perception and Compliance with Quarantine

During the SARS Outbreak. Journal of Nursing Scholarship, 37,

343–347. (doi:10.1111/j.1547-5069.2005.00059.x).

65 Wise T et al. 2020 Changes in risk perception and protective

behaviour during the ﬁrst week of the COVID-19 pandemic in the

United States. PsyArXiv Preprints. (doi:10.31234/osf.io/dz428).

66 Perrotta D et al. 2020 Behaviours and attitudes in response to the

COVID-19 pandemic: insights from a cross-national Facebook survey.

medRxiv. (doi:10.1101/2020.05.09.20096388).

67 Op. cit note 42

68 Op. cit note 49

69 Hilton S, Smith E. 2010 Public views of the UK media and government

reaction to the 2009 swine ﬂu pandemic. BMC Public Health, 10.

(doi:10.1186/1471-2458-10-697).

70 Seale H et al. 2012 Examining the knowledge, attitudes and practices

of domestic and international university students towards seasonal

and pandemic inﬂuenza. BMC Public Health, 12. (doi:10.1186/1471-2458-

12-307).

71 Op. cit note 70

72 Joe H. 2003 Risk: From perception to social representation. British

Journal of Social Psychology, 42, 55–73. (doi:10.1348/0144666037632

76126).

73 Op. cit note 46

74 Op. cit note 32

75 Op. cit note 32

76 Op. cit note 32

77 Op. cit note 32

78 Chow A et al. 2020 Unintended consequence: inﬂuenza plunges

with public health response to COVID-19 in Singapore. The Journal of

Infection, S0163-4453, 30262-0. (doi:10.1016/j.jinf.2020.04.035).

79 Op. cit, note 46

80 Betsch C et al. 2017 On the beneﬁts of explaining herd immunity in

vaccine advocacy. Nature Human Behaviour, 1. (doi:10.1038/s41562-

017-0056).

81 Gelfand MJ et al. 2011 Dierences between tight and loose cultures: a

33-nation study. Science, 332, 1100–1104. (doi:10.1126/science.1197754).

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 36

82 Painter M, Qiu T. 2020 Political beliefs aect compliance with

COVID-19 social distancing orders. SSRN Preprint. (doi:10.2139/

ssrn.3569098).

83 Op. cit note 67

84 Rothgerber H et al. 2020 Politicizing the COVID-19 pandemic:

ideological dierences in adherence to social distancing. PsyArXiv

Preprint. (doi:10.31234/osf.io/k23cv).

85 Bakshy E et al. 2015 Exposure to ideologically diverse news and

opinion on Facebook. Science, 348, 1130–1132. (doi:10.1126/science.

aaa1160).

86 Battiston P et al. 2020 Trust in science and experts during the

COVID-19 outbreak in Italy. OSF Preprints. (doi:10.31219/osf.io/twuhj).

87 Smith SS. 2020 Race and trust. Annual Review of Sociology, 36, 453-

475. (doi:10.1146/annurev.soc.012809.102526).

88 Boulware, LE. 2003 Race and trust in the health care system. Public

Health Reports, 118, 358-365. (doi:10.1016/S0033-3549(04)50262-5).

89 Dowd JB et al. 2020 Demographic science aids in understanding

the spread and fatality rates of COVID-19. PNAS, 117, 9696–9698.

(doi:10.1073/pnas.2004911117).

90 Nepomuceno MR et al. 2020 Besides population age structure, health

and other demographic factors can contribute to understanding

the COVID-19 burden. PNAS, 117, 13881–13883. (doi:10.1073/

pnas.2008760117).

91 Richardson S et al. 2020 Presenting characteristics, comorbidities,

and outcomes among 5700 patients hospitalized with COVID-19

in the New York City area. JAMA. 323, 2052-2059. (doi:10.1001/

jama.2020.6775).

92 Verhagen, DM et al. 2020 Forecasting spatial, socioeconomic and

demographic variation in COVID-19 health care demand in England

and Wales. BMC Medicine, 18. (doi:10.1186/s12916-020-01646-2).

93 Op. cit note 64

94 Op. cit note 67

95 Goldstein E, Lipsitch M. 2020 Temporal rise in the proportion of

both younger adults and older adolescents among COVID-19 cases

in Germany: evidence of lesser adherence to social distancing

practices? medRxiv. (doi:10.1101/2020.04.08.20058719).

96 Op. cit note 67

97 Peckham H et al. 2020 Sex-Bias in COVID-19: A Meta-Analysis and

Review of Sex Dierences in Disease and Immunity. SSRN Preprint.

(doi:10.21203/rs.3.rs-23651/v1).

98 Carbon CC. 2020 The Psychology of Wearing Face Masks in Times of

the COVID-19 Pandemic. SSRN PrePrint. (doi:10.2139/ssrn.3584834).

99 Cohn SK. 2012 Pandemics: waves of disease, waves of hate from

the Plague of Athens to A.I.D.S.*. Historical Research, 85, 535–555.

(doi:10.1111/j.1468-2281.2012.00603.x).

100 Russell A. 2020 The rise of coronavirus hate crimes. The New Yorker.

See https://www.newyorker.com/news/letter-from-the-uk/the-rise-of-

coronavirus-hate-crimes (accessed 26 June 2020).

101 Xio Z et al. 2020 The face mask and the embodiment of stigma.

PsyArXiv Preprint. (doi:10.31234/osf.io/fp7z8).

102 Siu JYM. 2010 Another nightmare after SARS: knowledge perceptions

of and overcoming strategies for H1N1 inﬂuenza among chronic renal

disease patients in Hong Kong. Qualitative Health Research, 20,

893–904 (doi:10.1177/1049732310367501).

103 Ferng Y et al. 2011 Barriers to mask wearing for inﬂuenza-like illnesses

among urban hispanic households. Public Health Nursing, 28, 13–23.

(doi:10.1111/j.1525-1446.2010.00918.x).

104 Wu H et al. 2020 Facemask shortage and the novel coronavirus

disease (COVID-19) outbreak: Reﬂections on public health measures.

EClinicalMedicine. 21, 100329. (doi:10.1016/j.eclinm.2020.100329).

105 Yuan EJ et al. 2020 Where to buy face masks? Survey of applications

using Taiwan’s open data in the time of coronavirus disease 2019.

Journal of the Chinese Medical Association, 83, 557-560. (doi:10.1097/

JCMA.0000000000000325).

106 Wang MW et al. 2020 Mask crisis during the COVID-19 outbreak.

European Review for Medical and Pharmacological Sciences, 24,

3387–3399. (doi:10.26355/eurrev_202003_20707).

107 Op. cit note 32

108 Op. cit note 32

109 Op. cit note 32

110 Op. cit note 46

111 Witte K, Allen M. 2000 A meta-analysis of fear appeals: implications

for eective public health campaigns. Health, Education and

Behaviour, 27, 591–615. (doi:10.1177/109019810002700506).

112 Presidenxa de Consigilio die Ministri. Ordinaza n. 11/2020 Il

commissario straodrdinario per l’attuazione e il coordinamento delle

misure di contenimento e contrasto dell’emergenza epidemiologica

COVID-19. See http://www.governo.it/sites/new.governo.it/ﬁles/

CSCovid19_Ord_11-2020-txt.pdf (accessed 26 June 2020).

113 Adams-Prassl A et al. 2020 Inequality in the impact of the coronavirus

shock: new survey evidence for the UK. Cambridge Working Papers

in Economics. (doi:10.17863/CAM.52477).

114 Baum NM et al. 2009 “Listen to the People”: Public deliberation about

social distancing measures in a pandemic. The American Journal of

Bioethics, 9, 4–14. (doi:10.1080/15265160903197531).

115 Op. cit note 48

116 Op. cit note 64

117 Luo P et al. 2020 Topical rh-aFGF: An eective therapeutic agent for

facemask wearing-induced pressure sores. Dermatologic Therapy.

(doi:10.1111/dth.13745).

118 Op. cit note 105

119 Op. cit note 67

120 Op. cit note 46

121 Op. cit note 2

122 Goldberg M et al. 2020 Mask-wearing increased after a government

recommendation: a natural experiment in the U.S. during the

COVID-19 pandemic. PsyArXiv Preprint. (doi:10.31234/osf.io/uc8nz).

123 Op. cit note 67

124 Op. cit note 42

125 Op. cit note 5

126 Op. cit note 1

127 UK Prime Minister’s Oce. PM announces easing of lockdown

restrictions: 23 June 2020. See https://www.gov.uk/government/

news/pm-announces-easing-of-lockdown-restrictions-23-june-2020

(accessed 26 June2020).

128 Block P et al. 2020 Social network-based distancing strategies to

ﬂatten the COVID-19 curve in a post-lockdown world. Nature Human

Behaviour, 4, 588–596. (doi:10.1038/s41562-020-0898-6).

129 Op. cit note 64

130 Op. cit note 47

131 Op. cit note 70

132 Prieto Rodgriguez MA et al. 2012 La visión de la ciudadanía sobre la

epidemia de gripe H1N1 2009-2010: Un enfoque cualitativo. Index de

Enfermería. 21, 38–42. (doi:10.4321/S1132-12962012000100009).

133 Op. cit note 42

134 Op. cit note 43

135 Op. cit note 24

136 Op. cit note 3

137 Op. cit note 40

FACE MASKS AND COVERINGSUPERSCRIPT FOR THE GENERAL PUBLIC  26 JUNE 2020 37

138 HM Government. Working safely during the COVID-19 in

construction and other outdoor work. COVID-19 secure guidance

for employers, employees and the self-employed, Version 2.0

updated 14 June 2020. See https://assets.publishing.service.gov.uk/

media/5eb961bfe90e070834b6675f/working-safely-during-covid-19-

construction-outdoors-140620.pdf (accessed 26 June 2020).

139 Alwan NA et al. 2020 Evidence informing the UK’s COVID-19 public

health response must be transparent. The Lancet, 395, 1036-1037.

(doi:10.1016/S0140-6736(20)30667-X).

140 Op. cit note 4

141 Greenhalg T. 2020 Will COVID-19 be evidence-based medicine’s

nemesis? PLoS Med. 17. (doi:10.1371/journal.pmed.1003266).

142 Op. cit note 143

143 Brainard JS, Jones N, Lake I, Hooper L, Hunter P. 2020 Facemasks

and similar barriers to prevent respiratory illness such as COVID-19: A

rapid systematic review. medRxiv. (doi:10.1101/2020.04.01.20049528).

144 Miyazawa D, Kaneko G. 2020 Face mask wearing rate predicts

country’s COVID-19 death rates. medRxiv. (doi:10.1101/2020.06.22.2013

7745).

145 Op. cit note 41

146 Xiao J et al. 2020 Nonpharmaceutical Measures for Pandemic

Inﬂuenza in Nonhealthcare Settings—Personal Protective and

Environmental Measures. Emerg. Infect. Dis. 26, 967–975.Brainard

JS, Jones N, Lake I, Hooper L, Hunter P. 2020 Facemasks and similar

barriers to prevent respiratory illness such as COVID-19: A rapid

systematic review. medRxiv. (doi:10.1101/2020.04.01.20049528).

147 https://assets.publishing.service.gov.uk/

media/5eb961bfe90e070834b6675f/working-safely-during-covid-

19-construction-outdoors-140620.pdf (accessed June 23 2020)(95)

Section 6.1, p. 25

148 Op. cit note 41

149 Op. cit note 4

150 Zhang D et al. 2013 Factors Associated with Household Transmission

of Pandemic (H1N1) 2009 among Self-Quarantined Patients in Beijing,

China. PLoS One. 8, e77873. (doi:10.1371/journal.pone.0077873).

151 Liu W et al. 2009 Risk factors for SARS infection among hospital

healthcare workers in Beijing: a case control study. Tropical

Medicine and International Health, 14, 52–59. (doi:10.1111/j.1365-

3156.2009.02255.x).

152 Seto W et al. 2003 Eectiveness of precautions against droplets and

contact in prevention of nosocomial transmission of severe acute

respiratory syndrome (SARS). Lancet. 361, 1519–1520. (doi: 10.1016/

s0140-6736(03)13168-6).

153 Op. cit note 15

154 Konda A et al. 2020 Aerosol Filtration Eciency of Common Fabrics

Used in Respiratory Cloth Masks. ACS Nano. 14, 6339–6347.

(doi:10.1021/acsnano.0c03252).

155 Op. cit note 23

156 Op. cit note 67

157 Soper GA, 1919 The lessons of the pandemic. Science 49, 501-506.

(doi:10.1126/science.49.1274.501).

158 Op. cit note 23

159 Blendon RJ et al. 2004 The public’s response to severe acute

respiratory syndrome in Toronto and the United States. Clinical

Infectious Diseases, 38, 925–931 (doi:10.1086/382355).

160 Blendon RJ et al. 2006 Attitudes toward the use of quarantine in a

public health emergency in four countries. Health Aairs, 25, W15–

W25. (doi:10.1377/hltha.25.w15).

161 Ibuka Y et al. 2010 The dynamics of risk perceptions and

precautionary behaviour in response to 2009 (H1N1) pandemic

inﬂuenza. BMC Infectious Diseases, 10, 296. (doi:10.1186/1471-2334-10-

296).

162 SteelFisher, GK et al. 2010 The public’s response to the 2009 H1N1

inﬂuenza pandemic. NEJM, 362, e65. (doi:10.1056/NEJMp1005102).

163 Kim Y et al. 2015 Public risk perceptions and preventive behaviours

during the 2009 H1N1 inﬂuenza pandemic. Disaster Medicine and

Public Health Preparedness, 9, 145–154. (doi:10.1017/dmp.2014.87).

164 Horney JA et al. 2010 Intent to receive pandemic inﬂuenza A (H1N1)

vaccine, compliance with social distancing and sources of information

in NC, 2009. PLoS One. 5, e11226. (doi:10.1371/journal.pone.0011226).

165 Yanni EA et al. 2010 Knowledge, attitudes, and practices of US

travelers to Asia regarding seasonal inﬂuenza and H5N1 avian

inﬂuenza prevention measures. Journal of Travel Medicine, 17,

374–381. (doi:10.1111/j.1708-8305.2010.00458.x).

166 Loustalot F et al. 2011 Household transmission of 2009 pandemic

inﬂuenza A (H1N1) and nonpharmaceutical interventions among

households of high school students in San Antonio, Texas. Clinical

Infectious Diseases, 52, S146–S153. (doi:10.1093/cid/ciq057).

167 Jones JH, Salathé M. 2009 Early assessment of anxiety and

behavioural response to novel swine-origin inﬂuenza A(H1N1). PLoS

One, 4, e8032. (doi:10.1371/journal.pone.0008032).

168 Op. cit note 43

169 Op. cit note 23

170 Op. cit note 23

171 Op. cit note 23

DISCLAIMER

This paper has drawn on the most recent evidence as of 26 June 2020 and has not been subject to formal peer-review. Further evidence on this topic

is constantly published and the Royal Society and British Academy may return to this topic in the future. This independent overview of the science has

been provided in good faith by experts and the Royal Society and British Academy and paper authors who accept no legal liability for decisions made

based on this evidence.

THANKS

The Royal Society is grateful to the Leverhulme Trust for its support for the Society’s pandemic response work.

The text of this work is licensed under the terms of the Creative Commons Attribution License which permits unrestricted use, provided the original

author and source are credited. The license is available at: creativecommons.org/licenses/by/4.0