key: cord-0684410-cch158vk authors: MacIntyre, C. Raina; Nguyen, Phi-Yen; Chughtai, Abrar Ahmad; Trent, Mallory; Gerber, Brian; Steinhofel, Kathleen; Seale, Holly title: Mask use, risk mitigation behaviours and pandemic fatigue during the COVID-19 pandemic in five cities in Australia, the UK and USA: a cross-sectional survey date: 2021-03-23 journal: Int J Infect Dis DOI: 10.1016/j.ijid.2021.03.056 sha: 0321669b32904bcc9d3c3891ab13974c071d5caa doc_id: 684410 cord_uid: cch158vk Objectives To determine patterns of mask wearing and other infection prevention behaviours in cities where mask wearing is not a cultural norm, over two time periods of the pandemic. Methods A cross-sectional survey of masks and other preventive behaviours in adults ≥18 years was conducted in five cities (Sydney, Melbourne, London, Phoenix and New York). Data was analysed according to the epidemiology of COVID-19, mask mandates and a range of predictors of mask wearing. Results The most common measures used were avoiding public areas (80.4%), hand hygiene (76.4%), masks (71.8%) and distancing (67.6%). Over 40% of people avoided medical facilities. These measures decreased from March-July 2020. Pandemic fatigue was associated with younger age, low perceived severity of COVID-19 and declining COVID-19 prevalence. Predictors of mask wearing were location (US, UK), mandates, age <50 years, education, having symptoms and knowing someone with COVID-19. Negative experiences with mask wearing and low perceived severity of COVID-19 reduced mask wearing. Most respondents (98%) believed that hand washing and distancing were necessary, and 80% reported no change or stricter adherence to these measures when wearing masks. Conclusion Pandemic mitigation measures were widely reported across all cities, but decreased between March and July 2020. Pandemic fatigue was more common in younger people. Cities with mandates had higher rates of mask wearing. Promotion of mask use for older people may be useful. Masks did not result in reduction of other hygiene measures. In the absence of a vaccine, non-pharmaceutical measures such as physical distancing, masks and hand hygiene have been key to controlling COVID-19 during the first year of the pandemic (Seale et al., 2020) . Whilst Asian countries have a longer history of mask-wearing for both infection and pollution, especially after SARS (2003) J o u r n a l P r e -p r o o f (Burgess and Horii, 2012; Sin, 2016) , the use of masks in the community is not a cultural norm in Western countries (MacIntyre and Chughtai, 2015) , and mask wearing was initially discouraged, whilst handwashing was promoted. As the pandemic unfolded, many countries mandated mask wearing during different stages of the pandemic. In New York City, an executive order was introduced on April 15 th 2020 (close to the peak of the first wave) mandating face coverings in public settings where physical distancing is not possible (Government of New York State, 2020). Likewise, Phoenix declared mandatory mask use from June 20 th , 2020 (before the epidemic peak) (Phoenix City . In Sydney, masks were not mandated, but by June 2020, guidance was updated to recommend mask use when physical distancing was not possible, for example on crowded public transport (New South Wales Department of Health, 2020). London initially had no mask wearing recommendation, but a mandate was issued on July 24 th for all public indoor spaces (London City Hall, 2020) . In Melbourne, Australia, a second wave in July resulted in a mask mandate on July 23 rd 2020 (Department of Health and Human Services Victoria, 2020a). There has also been changing guidance over time about community use of face masks from health agencies such as the World Health Organization (WHO) and, the Department of Health Australia, the US CDC and Public Health England.(Australian Government Department of Health, 2020; Centers for Disease Control and Prevention, 2020; Chan et al., 2020 ; Government of the United Kingdoms, 2020a) Initial reluctance to recommend masks, as well as active advice to not wear masks may have been influenced by global shortages of masks (MacIntyre and . The WHO also expressed concern that the use of facemasks may give people a false sense of security and lead to reduction of other infection control measures such as hand washing and physical distancing (World Health Organization, 2020). As the pandemic progressed, evidence accumulated to show that around 40 to 45% of infections were asymptomatic (Oran and Topol, 2020) . This means infections cannot be readily identified and infected people may be unaware they are infected, which increases the utility of universal masking. The available evidence on mask use in the community showed a net protective effect of masks (Chu et al., 2020; MacIntyre and Chughtai, 2020) . There may also be fatigue with social restrictions, mask mandates and other risk mitigation measures over a long period of a pandemic, yet no data to better understand this. It is therefore important to gather evidence about community understanding, experiences and practices around the use of masks and other risk mitigation measures during the pandemic, in settings with different disease incidence and different policies. It is also important to understand whether mask use affects other risk mitigation behaviours. This study aimed to provide insights into mask wearing and other infection prevention behaviours in five cities where mask wearing is not a cultural norm, over two time periods of the pandemic. A cross-sectional survey was conducted in five cities from three countries without a culture of mask wearing. . These cities were selected to represent a spectrum of low, medium and high incidence of COVID-19, and a range of policies towards community mask use (ranging from a mask mandate to no mandate), with a population of at least 5 million. The cities were classified as low incidence (<1 case per 100,000), medium incidence (1-10 cases per 100,000) and high incidence (>10 cases per 100,000), based on COVID-19 incidence at the start of the survey in July 2020. Adults ≥18 years old, of any gender, living in the selected cities in 2020 and willing to consent were included in the study, with sampling proportionate to population size. A market research company, Dynata (Dynata, 2020), randomly distributed the survey link by email to a representative sample of their panel members in Sydney, London, Melbourne, Phoenix and New York City. An algorithm using geolocation was used to screen and identify eligibility to take part in the survey. Dynata's worldwide consumer research panel includes over 60 million people from over 94 countries, including Australia, the United Kingdom (UK), and United States (US), that have been profiled on demographic and health attributes (Dynata, 2020). Panel members undergo a verification process to ensure reliability and accuracy of responses and to avoid duplicate participants. Data provided from panel members undergoes regular quality checks, such as participation limits, screening questions, digital fingerprinting, and capturing and removing participants that provide illogical responses or do not spend sufficient time on surveys. No identifying information is provided or collected. Panel members that chose to open the survey link were screened for inclusion based on age and location. Participants that met the inclusion criteria were directed to a participant information and consent page that J o u r n a l P r e -p r o o f provided details about the research study and had to provide consent to proceed with the online survey. Failure to complete the entire survey was considered withdrawal of consent. The survey was launched on July 10 th 2020 and closed on July 27 th 2020. The survey collected data on socio-demographics, risk factors for COVID-19, attitudes towards nonpharmaceutical interventions, adoption of and experiences with mask use, infection control behaviours and attitudes (Appendix Table 1 ), in people aged 18 years or over in Australia, the UK and the US. It took about 10-15 minutes to complete the survey. Questions on risk mitigation behaviours and mask wearing were asked about the early period of the pandemic (March-April 2020) and at the time of survey (July 2020). Data were collected using a web-based survey platform, Redcap. We powered the study a priori to identify a 20% difference in the rate of mask use between cities with (New York and Phoenix) and without mask mandates (Sydney, Melbourne and London) with 95% confidence and 80% power. We assumed a mask use prevalence of 60% among cities without mask mandate and 80% among cities with mask mandate (Babalola et al, 2020; Jones S et al, 2020 ) and a sampling ratio of 0.3, yielding a minimum required sample size of 194. Hence, we aimed to recruit a total of 2,150 participants, from which the samples were selected proportionate to population size, age and gender distribution of the sampled cities. i.e. 200 people from Sydney, 150 from Melbourne, 300 from London, 1200 from New York City, and 300 from Phoenix. Post-hoc power analysis was conducted via large sample approximation using G*Power 3.1.9.7 (Faul et al, 2009 ). In a two-sided test with α = 0.05, the recruited sample size (N=2,343) enables detection of at least 20% difference in mask use (i.e. odds ratio of 1.2), yielding a power of 94.1%. For socio-demographic and health related multiple-choice questions, answers were coded as ordinal variables if one option is allowed, or separate binary variables for each option if multiple options were allowed. Participants were asked to rank their confidence in the government and their trust in COVID-19 information provided by the government using a Likert scale of 0-4, where 4 represented the highest level of trust or confidence and 0 was none at all. In logistic regression, ratings of 3-4 were coded as 1 (high) and ratings lower than 3 were coded as 0 (low). Pandemic fatigue was coded as present if a participant reported a net decline in number of protective measures in July 2020 compared to March-April 2020. Mask mandate was coded as present for cities with a mandatory mask policy implemented before July 27 th , 2020 and absent for others. Cities were classified as either experiencing a declining in cases in July 2020 compared to March 2020 (Sydney, London, New York) or not (Melbourne, Phoenix). Descriptive statistics were performed for variables relating to mask use, perception and experience of masks, and other infection control behaviours such as hand-washing and physical distancing. In order to interpret Phoenix City Hall, 2020) were compared relative to the epidemic curves. Statistical analyses were performed to detect significant inter-city differences in health status, mask use, experience of mask shortages, and other health behaviours during the COVID-19 pandemic. Chi-square test was used for binary variables, one-way ANOVA test for continuous variables, and Kruskal-Wallis H test for ordinal variables. Two-sample proportional test was used to compare percentage mask use between March-April and July 2020 and between age groups. Statistical significance was defined at α-level of 0.05. Univariate and multivariate logistic regression was used to identify predictors of mask use and pandemic fatigue. All data were cleaned prior to analysis. Analysis was completed using Stata version 14 (Stata Corporation, College Station, TX, USA). In total, 2,343 people from the greater metropolitan areas of each city were surveyed, comprising 200 from Sydney, Australia, 148 from Melbourne, Australia, 291 from London, UK, 1,204 from New York City, NY, USA and 500 from Phoenix, AZ, USA. Trust in state or local government was higher than trust in national government (Table 1) , and generally trust in government was higher in Australia than the UK or US (Appendix Table A4 ). The mean age was 50.8±17.8. Participants were younger in Sydney (45.3±15.8), London (45.2±15.6) and Melbourne (46.7±17.9). More than half of the participants (n=1,288, 55.0%) had at least one co-morbidity (Appendix Table A2 ). Mask use between March and July 2020 relative to the epidemic curve in each city is shown in Figure 1 . The average incidence per 100,000 population between March-April 2020 is 0.33 for Melbourne, 0.61 for Sydney, 1.79 for Phoenix, 5.10 for London and 25.98 for New York. Participants reported adopting a wide range of measures in March-April 2020 (early stage of the pandemic) and later (July 2020) to reduce risk of COVID-19 ( Figure 2) . The most common measures early in the pandemic were: avoiding crowded areas, public transport and physical shops (80.4%), practising hand hygiene (washing hands, using hand sanitizers, not touching face) (76.4%), wearing masks (71.8%), physical distancing (67.6%), restricting visitors (60.9%), reducing visits to medical facilities (42.9%) and avoiding contact with sick people (31.5%). There was a consistent decrease reported in almost all risk mitigation behaviours between March-April 2020 and July across the cities ( Figure 2 ). However, mask use increased in all cities except Sydney, where it decreased. Melbourne, which was in the midst of a second wave at the time of the survey, had an increase in masks, avoiding contact with sick individuals and medical facilities, and physical distancing. Table 1 shows a total of 1,683 participants (71.8%) reported wearing a mask of any type during the COVID-19 pandemic. Prevalence of mask-wearing was significantly lower in Sydney (45.5%) and Melbourne (51.4%), and higher in London (70.8%), Phoenix (75.6%) and New York (77.4%) (p<0.001). Overall, mask use among participants 50 years old was significantly lower than that of people <50 years old (p=0.0150), but the difference is not significant in New York (p=0.0945), Phoenix (p=0.0955) and Melbourne (p=0.2667) (Appendix Table 3 ). Compared to the overall prevalence (55.3%), cloth mask use was higher in Phoenix (75.4%) and much lower in Sydney (only 16.8%). (Appendix Table 4 ). Table 1 also shows that most participants did not report any negative issues while wearing a mask (n=1318, 77.5%). Reported problems included receiving negative or racist remarks (n=144, 8.5%), embarrassment (n=141, 8.3%), being stared or laughed at (n=133, 7.8%), or being mistaken for being sick (n=51, 3.0%). Almost all participants believed that when wearing a mask, they also need to wash hands (n=1647, 97.9%) and adhere to physical distancing (n=1641, 97.5%); more than half (n=991, 58.9%) reported no change in other risk reduction measures when wearing a mask, 20% (n=335) reported stricter adherence to these measures and 21.0% (n=353) reported less adherence to one or more measure. Table 1 shows that in the period of March-June 2020, 12.5% participants (n=293) reported symptoms of a chest infection or cold or flu-like illnesses. Approximately a quarter of participants (n=609, 26.0%) reporting being tested for COVID-19, most commonly because they had symptoms (6.3%) or they were contacts of a confirmed case (6.4%). A total of 123 participants tested positive for COVID-19, representing a positive rate of 20.2% (5.2% of total population). Melbourne has a higher proportion of the population who were tested (n=67, 45.3%), but a significantly lower proportion of positive tests (n=1, 0.7%). More than one-third of participants (n=874, 37.3%) had a family member, work colleague or friend who had confirmed COVID-19. The proportion was lowest in Sydney and Melbourne, and highest in New York. The mean perceived severity of COVID-19 was lowest in London (57.2±25.2) and highest in Melbourne (63.9±25.1). The mean perceived risk of getting COVID-19 was lowest in Phoenix (52.2±23.8) and highest in New York (56.4±23.6). The mean perceived effectiveness rating for N95/P2, surgical and cloth masks was higher in New York and Phoenix than other cities. Results for comparison between cities are presented in Appendix Table 4 . In univariate regression analysis (Table 2) , respondents in the United Kingdoms and United States were more likely to wear masks during the COVID-19 pandemic compared to respondents in Australia (p<0.001). There was significant relationship between mask use and age (p=0.014) and having negative issues while wearing masks (p<0.001), but not gender (p=0.953). Mask mandate was associated with mask use (p<0.001), but was excluded from multivariable regression due to collinearity with country of residence. After adjusting for effects of other covariates, multivariable analysis showed that experiencing negative issues (p<0.001) and low perceived severity of COVID-19 (p=0.033) remain significantly associated with lower mask use. Factors associated with higher mask use were: age <50 (p<0.001), tertiary education (p=0.001), wearing a mask before the pandemic (p<0.001), knowing a family member, friend or colleague who was diagnosed with COVID-19 (p=0.001), having a chest infection, cold or flu-like symptoms in March-June 2020 (p=0.002) and self-reported adherence to local mask guideline (p=0.006) ( Table 2) . Multivariable regression (Table 3) shows that younger age (<50) (p=0.002) and low perceived severity of COVID-19 (p=0.01) are associated with pandemic fatigue. People living in cities that experience declining COVID-19 incidence between March 2020 and July 2020 (p=0.002) are also more likely to show pandemic fatigue. Mask mandates have a strong effect on mask use, and masks do not reduce compliance with other control measures. A range of social distancing and hygiene measures were used, with hand hygiene being the most common and more prevalent than physical distancing or mask use. This may reflect the strong focus on hand hygiene in early pandemic messaging. We confirmed that over 40% of people avoided medical facilities and healthcare during the pandemic, which is cause for concern. Pandemic fatigue was seen in all cities except Melbourne, which was experiencing a resurgence of COVID-19 during the survey period. Phoenix also was experiencing a resurgence at the time, but participants reported reduced preventive measures by July. Younger age and male gender predicted pandemic fatigue, pointing to these being key groups for health promotion messaging in protracted epidemics. Some studies suggest that males may be less compliant with mask guidance due to perceived association with a lack of 'masculinity' (Seale et al., 2020) , but we did not find any difference in mask use by gender. The surveyed cities represented a wide spectrum of disease incidence and varied community mask policies. Age restriction was stricter in London (compulsory for all 3 years old), Phoenix and New York (2 years old) than Melbourne (12 years old). Fines were more severe in Melbourne (A$250), London (£200) and Phoenix (US$250) than New York ($50 for non-compliance on public transit) (Bowling, 2020; Department of Health and Human Services Victoria, 2020a; London City Hall, 2020; Office, 2020) . In three cities (New York, Phoenix and Melbourne), mask mandates occurred close to the epidemic peak. In London, masks were mandated about three months after the epidemic peak, after substantial advocacy (Greenhalgh et al., 2020) . Only Sydney had no mask mandate (New South Wales Department of Health, 2020), and also had the lowest incidence of COVID-19 and the lowest rate of mask use. Except for Melbourne, all cities with mask mandates reported mask usage in excess of 70%, with the highest rates in New York and Phoenix. Our findings are consistent with a global survey on knowledge, attitudes and practices (KAP) for COVID-19 prevention measures (Johns Hopkins Bloomberg School of Public Health, 2020), which individuals in Australia reported lower prevalence of mask use (47%) than those in the United Kingdom (66%) and United States (87%) in July 2020. Our data shows that mask wearing decreased after the age of 50, with a significant difference by age in Australia and the UK, but not in the US. This is a concern given older people are at highest risk for serious complications and death (Leung, 2020) , and barriers among older people should be investigated. In contrast, in Asian countries, higher mask usage has been reported among older people (Lee et al., 2020; Seale et al., 2020) . We found that negative issues experienced while wearing masks reduced the likelihood of mask-wearing. Early in the pandemic, people of Asian ethnicity have reported racism and harassment, while others reported being suspected of criminal intent while wearing masks (Zine J, 2020) . The negative connotations of disease and identity concealment associated with masks in Western countries prior to the COVID-19 pandemic, may have been further catalysed by rising geopolitical tension (Ma and Zhan, 2020) and early public health messages which actively discouraged mask use . Mandating masks may eliminate this stigma by making it a mainstream behaviour (Betsch et al., 2020; Feng et al., 2020) . Our findings were consistent with a recent report, which shows higher levels of trust in the government among Australians (54%) than those in UK (41%) or US (34%) (Evans et al., 2020). However, trust was not a significant predictor of mask wearing. This study was not without limitations. The method of recruitment from an online panel and non-response from panelists may potentially introduce response bias (Baker et al., 2010) . However, Dynata panels have over 60 million people and are widely used in research. In addition, the survey was only administered online and in English. As a result, we may have excluded non-English speakers or people with limited access to the Internet, who may be different to their English-speaking or Internet-using counterparts. Although we surveyed mask use by ethnicity, we did not incorporate it into the multivariable regression model because of the complexity associated with multiple-response questions. Ethnicity is an important factor influencing mask adoption because of its link to culture, socioeconomic status and family/community norms (Sim et al., 2014) . Last but not least, our survey only provides a cross-sectional description of mask use, which relied on recall for reporting behaviours in the early pandemic period and may introduce recall bias. The pandemic has seen a rise in mask use in cities with no previous mask culture, as well as adoption of a wide range of other preventive behaviours, with hand hygiene being the most common. Mask use was widespread in the studied cities, especially where mandates were issued, but usage was lower in older adults. This age effect was not seen in the US, and may reflect cultural differences. Masks did not result in a net change in other risk mitigation behaviours such as hand washing and distancing. The reduction of risk mitigation behaviours between March and July 2020 may reflect changing epidemiology of the local pandemic and a corresponding change in risk perception. Both Melbourne and Phoenix were experiencing a resurgence during the survey, but only Melbourne showed an increase in most risk mitigation behaviours. The reduction in these behaviours in all other cities may reflect pandemic fatigue, which was more common in young people and males. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. CR MacIntyre reports being on advisory board for development of masks at Ascend Technologies, and having consulted for development of masks or Detmold and Atmos, outside the submitted work. AA Chughtai reports testing of filtration of masks by 3M for his Ph.D. more than 10 years ago. 3M products were not used in his research. He also has worked with CleanSpace Technology on research on fit testing of respirators (no funding was involved). PN Nguyen, M Trent, B Gerber, K Steinhofel and H Seale declare no conflict of interests. The study was funded by the Medical Research Future Fund by the Australian Government [Grant number APP1201320]. The funding organisation had no role in the design of the study; the collection, analysis, and interpretation of the data; nor the decision to approve publication of the finished manuscript. The study was approved by the Human Research Ethics Committee of the University of New South Wales (project number HC200460) and conform to the principles embodied in the Declaration of Helsinki. Australian Government Department of Health Massachusetts Institute of Technology, Global Outbreak Alert and Response Network, Facebook Data for Good Research synthesis: AAPOR report on online panels Social and behavioral consequences of mask policies during the COVID-19 pandemic Phoenix will require mask or other face covering starting Saturday. AZ Cent Japan's 'safety blanket' of surgical face maskwearing Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses Rational use of face masks in the COVID-19 pandemic 17: Continuing Temporary Suspension and Modification of Laws Relating to the Disaster Emergency Face coverings: when to wear one, exemptions, and how to make your own 2020a Government of the United Kingdoms. Coronavirus (COVID-19) in the UK 2020b Face masks for the public during the covid-19 crisis Johns Hopkins Bloomberg School of Public Health. KAP COVID: Exploring Knowledge, Attitudes and Practices for COVID-19 Prevention Imperial College London Big Data Analytical Unit and YouGov Plc Practice and technique of using face mask amongst adults in the community: A cross-sectional descriptive study Risk factors for predicting mortality in elderly patients with COVID-19: A review of clinical data in China Face covering guidance 2020 To mask or not to mask amid the COVID-19 pandemic: how Chinese students in America experience and cope with stigma A rapid systematic review of the efficacy of face masks and respirators against coronaviruses and other respiratory transmissible viruses for the community, healthcare workers and sick patients Facemasks for the prevention of infection in healthcare and community settings Physical distancing, face masks, and eye protection for prevention of COVID-19 NYS-COVID19-Tracker 2020 Governor Cuomo Issues Executive Order Directing MTA to Develop Plan to Bolster Mask Compliance On Public Transit System 2020 Prevalence of Asymptomatic SARS-CoV-2 Infection : A Narrative Review Phoenix Requires Face Coverings; Frequently Asked Questions 2020 Improving the impact of nonpharmaceutical interventions during COVID-19: examining the factors that influence engagement and the impact on individuals The use of facemasks to prevent respiratory infection: A literature review in the context of the Health Belief Model Masking fears: SARS and the politics of public health in China Unmasking the racial politics of the coronavirus pandemic Mask use during COVID-19: A risk adjusted strategy Advice on the use of masks in the community, during home care and in healthcare settings in the context of the novel coronavirus (COVID-19) outbreak Having COVID-19-like symptoms in The author would like to acknowledge the support of the staff at the Biosecurity Program (The Kirby Institute) who assisted with the administration of the survey.