Masks During Pandemics Caused by Respiratory PathogensEvidence and Implications for Action | Infectious … – JAMA Network

Key Points

Question During the COVID-19 pandemic, what has been learned about whether face mask use is associated with lower transmission of SARS-CoV-2 in community settings, and how has it been learned?

Findings Literature review revealed many high-quality observational studies demonstrating the association of face mask use in the community and of mask mandates with reduced spread of SARS-CoV-2. Randomized clinical trials conducted during the pandemic provide limited information.

Meaning Robust available data support the use of face masks in community settings to reduce transmission of SARS-CoV-2 and should inform future responses to epidemics and pandemics caused by respiratory viruses.

Importance As demonstrated by the influenza virus and SARS-CoV-2, viruses spread by the respiratory route can cause deadly pandemics, and face masks can reduce the spread of these pathogens. The effectiveness of responses to future epidemics and pandemics will depend at least in part on whether evidence on masks, including from the COVID-19 pandemic, is utilized.

Observations Well-designed observational studies have demonstrated the association of mask use with reduced transmission of SARS-CoV-2 in community settings, and rigorous evaluations of mask mandates have found substantial protection. Disagreement about whether face masks reduce the spread of SARS-CoV-2 has been exacerbated by a focus on randomized trials, which are limited in number, scope, and statistical power. Many effective public health policies have never been assessed in randomized clinical trials; such trials are not the gold standard of evidence for the efficacy of all interventions. Masking in the community to reduce the spread of SARS-CoV-2 is supported by robust evidence from diverse settings and populations. Data on the epidemiologic, environmental, and mask design parameters that influence the effectiveness of masking provide insights on when and how masks should be used to prevent transmission.

Conclusions and Relevance During the next epidemic or pandemic caused by a respiratory pathogen, decision-makers will need to rely on existing evidence as they implement interventions. High-quality studies have shown that use of face masks in the community is associated with reduced transmission of SARS-CoV-2 and is likely to be an important component of an effective response to a future respiratory threat.

More than 3 years after the COVID-19 pandemic began, the use of face masks in the community remains controversial. Vaccination, treatment, population immunity, and other developments have enabled a return to a semblance of prepandemic life, but disagreement about what the evidence shows about masksand the implications for their usepersists. SARS-CoV-2 is still a disruptive and deadly presence, and future epidemics or pandemics caused by pathogens spread by the respiratory route are a near certainty.1 Failure to understand the evidence on the role of masks in preventing the spread of SARS-CoV-2 could undermine our ability to respond to epidemics and pandemics caused by respiratory pathogens.

Evidence on Community Masking at the Advent of the COVID-19 Pandemic

In early 2020, when SARS-CoV-2 was spreading globally and the World Health Organization declared a Public Health Emergency of International Concern, there were neither vaccines against nor treatments for COVID-19. Furthermore, we lacked understanding of the viruss routes of transmission, extent of presymptomatic and asymptomatic spread, and degree of transmissibility. As a result, prevention and control strategies were based on what was known about transmission of other respiratory pathogens, especially influenza viruses and previously characterized human betacoronaviruses, such as those that cause severe acute respiratory syndrome (SARS), Middle East respiratory syndrome (MERS), and the common cold.2-4 Public health officials needed to consider that available evidence came from studies on specific types of masks (particularly N95 respirators vs surgical masks), that there was variation in the aims of masking interventions (ie, wearer protection vs source control), and that epidemiologic and behavioral differences between study settings (especially in health care settings vs the community) might influence apparent mask effectiveness.

These considerations are essential to the interpretation of evidence on mask efficacy and effectiveness. When the COVID-19 pandemic began, there were limited data on masks to prevent community transmission of respiratory pathogens.2 In addition, there were legitimate concerns that studies on prevention of influenza might not be relevant to COVID-19, as the influenza and SARS-CoV-2 viruses differ in their degree of transmissibility, proportion of transmission attributable to asymptomatic shedding, and other factors. Studies on community masking and transmission of SARS-CoV-2 were eagerly awaited and, once available, widely cited. As the pandemic progressed, high-quality data demonstrated the benefits of masking as well as increased protection associated with certain mask types and patterns of use. However, the robustness and nuance of those data have at times been overshadowed by attention to one particular type of study: the randomized clinical trial (RCT).

RCTs on Community Masking During the COVID-19 Pandemic

Two RCTs5,6 on mask use in the community conducted during the COVID-19 pandemic have been published.7 The first, which randomized 3030 people in Denmark to receive surgical masks and a recommendation to mask outside the home vs no intervention, produced inconclusive results.5 The incidence of SARS-CoV-2 infection was 20% lower in the intervention group, but the sample size was insufficient for the difference in infection rates to achieve statistical significance. During the study period, it is possible that other interventions diluted measurable effects of masking or that detected infections were largely transmitted within households. Only 46% of participants in the intervention group reported masking as recommended. Rates of community masking were also low; any effects of the intervention would have been derived primarily from wearer protection alone.8

The second, much larger study6 showed that widespread community masking in Bangladesh was associated with a modest, statistically significant reduction in symptomatic SARS-CoV-2 infections. This was a cluster RCT in which 600 villages with more than 340000 residents were randomized to receive either cloth or surgical masks and promotion of masking in public spaces vs no intervention. Mask use was observed to be 3 times more common (42% vs 13%) in intervention than control villages, with the largest increases in mask use observed in mosques. In villages assigned to the intervention, the incidence of symptomatic SARS-CoV-2 infection was reduced 9.5% overall and 35% among people aged 60 years or older compared with controls. Surgical masks appeared to be more effective than cloth masks.

Why do we not have more RCTs on masks to prevent the spread of SARS-CoV-2? In a pandemic caused by a lethal respiratory virus, it is difficult to find a setting in which it is ethical and feasible to randomize people to masking vs no masking. The time needed for RCTs to be funded, designed, and implemented further limits their feasibility during public health emergencies.9 Other challenges to conducting RCTs of masking during a pandemic include adequately powering a study amid fluctuating community transmission levels and crossover between study groups (ie, inadequate adherence by participants in the intervention group or adoption of the intervention by those in the control group). Perfect adherence with the intervention (ie, wearing a mask correctly at all times of potential exposure) can be impractical, biasing results toward the null. Although imperfect adherence might mimic reality, it can complicate the interpretation of study results. It is not possible to conduct a blinded RCT on mask use vs no mask use, and crossover can be a particular problem if the intervention carries social meaning or if fear of the disease influences adherence with randomization assignment.

In the clinical world, choices between therapeutic regimens to treat various conditions are, whenever possible, guided by the results of carefully planned and executed RCTs. Although the RCT is often referred to as the gold standard, methodological considerations correct the misconception that RCTs are necessarily superior.10 RCTs are not the onlyor even the most importantway to assess the efficacy of health interventions. Although a well-designed RCT can provide valuable and internally valid information on the efficacy of a health intervention, in many areas of public health and medical practice, RCTs are impractical to conduct.11 Furthermore, extrapolation to populations and settings outside the study may be invalid. Many highly effective policies and recommendations that reduce illness and injury have never been assessed in either an individual-level or cluster RCT. Examples include speed limits on highways; seatbelt and motorcycle helmet laws; interventions such as taxation and smoke-free indoor areas to reduce tobacco use; and putting babies to sleep on their backs to reduce the risk of sudden infant death syndrome.11-13 In public health practice, evidence that does not come from RCTs is, appropriately, almost invariably assessed and used in support of policy making.

When RCT data are sparse, creating a summary measure of effect using meta-analysis can seem appealing.14 Meta-analysis was originally developed as a tool for combining results of multiple RCTs that assessed the efficacy of the same therapeutic agent against the same disease outcome, especially when available studies were of limited size and therefore limited statistical power. A meta-analysis does not create new data. Rather, it assesses previously conducted studies, assigning weights based on study size and quality, to improve understanding of an intervention. If the studies available for inclusion in a meta-analysis differed in their methods, populations, contexts, or measurements, combining them for the purposes of conducting a meta-analysis may yield invalid results and conclusions. Meta-analysis can work well for simple interventions expected to have consistent effects across populations (eg, the effect of a particular drug on a specific patient outcome). However, if masking is the intervention under investigation, the effectiveness of the intervention might vary greatly depending on the type of mask, masking behavior in the nonintervention group, force of infection in the community, whether the intervention is designed for individual protection or for source control, uptake of the intervention, and characteristics of the circulating pathogen. If any of these components vary among studies, different interventions are being tested. Lumping studies together because they include a mask can yield invalid conclusions and conceal important findings of individual studies. The results of a Cochrane review of RCTs and cluster RCTs on interventions to reduce the spread of respiratory viruses14 exemplified this pitfall; studies on masking to prevent influenza virus transmission and studies on masking to prevent SARS-CoV-2 virus transmission were analyzed together. When interpreting study findings, decision-makers should strive to understand the conditions under which specific interventions are likely to be useful.

Observational Studies on Community Masking During the COVID-19 Pandemic

Before the emergence of SARS-CoV-2, observational studies of measures to prevent transmission of the related betacoronaviruses that cause SARS and MERS indicated that use of respirators or masks was associated with a large reduction in the risk of infection in health care settings and in the community.2 Since the pandemic began, high-quality case-control, cohort, and ecologic studies support the effectiveness of masks in the community to prevent transmission of SARS-CoV-2 (Table 1).6,7,15-21 Proof of concept that masking can reduce transmission of SARS-CoV-2 comes from laboratory evaluations that have used simulated human respiration and other techniques to show that cloth masks, surgical masks, and N95 respirators reduce the spread of potentially infectious respiratory droplets and aerosols.15,22,23 Masks can offer effective source control and some wearer protection; reduction of droplet and aerosol spread is greatest when both the source and the exposed individual are masked.15,22,24 Studies show that mask type and fit influence efficacy.15,17,22,25,26 The higher filtration efficiency of N95 respirators compared with surgical masks provides further evidence of efficacyessentially, a dose-response association of the intervention with the outcome.

Observational studies have demonstrated the effectiveness of mask use to prevent transmission of SARS-CoV-2 on airplanes,27 in schools,28,29 and among household30 and community17,18,31 contacts of individuals with COVID-19. A COVID-19 outbreak on the USS Theodore Roosevelt aircraft carrier was particularly instructive. Ships are high-risk environments for respiratory disease outbreaks because they bring people together for prolonged periods in often poorly ventilated close quarters. The outbreak on the USS Theodore Roosevelt occurred early in the COVID-19 pandemic, before crew members would have had immunity to SARS-CoV-2. More than 80% of those who reported not masking were infected; the odds of infection were 30% lower among those on the ship who reported masking.32

In many COVID-19 outbreak settings, including the USS Theodore Roosevelt, masking was not the sole implemented public health measure (eg, hand hygiene, isolation of cases, physical distancing). Confounding can be a particular problem in observational studies. Studies demonstrating the effectiveness of masks to prevent COVID-19 have attempted to address this by controlling for factors, such as other interventions7,30,33-35 or demographic characteristics that might influence risk of infection.17,31 Others have used stratified analysis and shown that the impact of masking varies across patterns of use, populations, and settings. For example, masking in the household was protective if performed in the days prior to symptom onset in the index case,30 demonstrating the risk of presymptomatic transmission and the effectiveness of masking to prevent asymptomatic transmission. Masks were associated with a greater reduction in risk of infection among unvaccinated than vaccinated individuals and with reduced risk of infection among those exposed outside their households but not among those exposed within their households31 (possibly because low rates of masking within households precluded sufficient assessment). Several studies have shown increased protection with greater consistency of masking.7,17,18

Other Considerations: Mask Mandates, Risk-Benefit Calculations, and Areas of Uncertainty

Whether masks work is a different question from whether mask mandates work. The effectiveness of mandating an even partially effective intervention depends on many factors, and the impact of the intervention can be challenging to demonstrate. If adherence to a public health mandate is low, a mandate is unlikely to have an impact (seat belts reduce the risk of death, if they are worn). Higher rates of indoor masking in parts of Asia (eg, Hong Kong, Japan, Korea, and Singapore) may account for lower rates of infection and death, especially early in the pandemic.36,37 For a population in which use of the intervention is already common, a statistically significant reduction in infection rates will be more difficult to establish. Furthermore, assessment of the effectiveness of mask mandates requires either cluster randomized studies or ecologic studies in which the unit of observation is the group, not the individual. Such studies have been done: rigorous evaluations of mask mandates in several settings suggested substantial protective benefits. In Germany, an opportunity to generate high-quality data arose when different regions mandated masking at different times during the COVID-19 pandemic. Mask mandates were associated with a 45% reduction in SARS-CoV-2 infections.38 Variation in timing of mask mandates across the United States provided a similar study opportunity, and a matched cohort analysis of more than 400 US counties showed that enactment of a mask mandate was associated with a 25% reduction in COVID-19 incidence 4 weeks later.20 Although it is possible that cases might soon peak without intervention if masking is implemented when incidence is increasing, US communities with mask mandates had less transmission than those without mandates after controlling for potential confounders, including premandate incidence.19,20

The risk-benefit calculations that shape public health recommendations may differ by setting and may change over time (Table 2).15,17-20,22,24-26,28-33,35,39-46 When the COVID-19 pandemic began, scarcity of medical masks and respirators precluded their use outside of health care settings. There was concern that community members wearing masks might self-contaminate with SARS-CoV-2 or might fail to practice other public health measures due to a false sense of security. However, although respiratory viruses can contaminate the outside of masks when masks are worn for hours in high-exposure clinical settings,46 the relevance of this finding to community settings is unclear. There is no compelling evidence that masking is associated with neglect of other public health measures; in fact, studies have suggested the opposite.18

Other concerns raised about masks included possible impacts on respiratory function; although masking can be uncomfortable, especially in warm conditions, there is no compelling evidence of consequential deleterious effects on physiology, including during exercise.47 It can be difficult for young children to wear well-fitting masks, and the possibility that masking may impede cognitive and social development40,41 suggests that this risk should be considered, balancing with possible benefits of masking. There is abundant evidence that school closures are deleterious to childrens health and that masking in schools decreases transmission of SARS-CoV-2 within schools.28,29,39 Using measures including masking to protect high-risk people in the school community and to keep schools open is likely to result in better health and educational outcomes than school closures. Consideration of trade-offs should inform future decisions about masking in schools to prevent the spread of respiratory viruses, and frequent reassessments of the epidemiologic context and available evidence can help maximize benefits and reduce disruption and potential harms.

Although available evidence strongly suggests that masking in the community can reduce the spread of SARS-CoV-2, knowledge gaps persist. It is challenging to disentangle the impacts of masks from those of other interventions on transmission of SARS-CoV-2. The effectiveness of masks may differ between variants of SARS-CoV-2. Until recently, respirators such as N95s were not widely available outside health care settings. We lack precise estimates of the extent to which the community spread of SARS-CoV-2 is reduced at different levels of uptake of different mask types in different contexts. However, there is alignment between findings from laboratory models and limited available effectiveness data: a study on the use of masks or respirators in indoor public settings17 showed that respirators were more protective against SARS-CoV-2 infection than surgical masks, which were more protective than cloth masks.

Effectiveness depends on many factors. No public health intervention, even a highly efficacious vaccine, is 100% effective. Even the best masks will not provide complete protection, and benefits of masking are limited if masks are not worn everywhere transmission occurs (eg, health care workers who consistently wear masks while working with patients but not in break rooms with other health care workers or in the community can be infected in the latter settings). In any pandemic or epidemic, masking will be just one of a series of interventions. The most effective strategies to limit illness and death from SARS-CoV-2 and other respiratory pathogens involve a layered response, including vaccination when available, isolation of infectious people, and protection through risk reductionincluding use of high-quality masks in areas and at times and by vulnerable populations when the pathogen may be spreading. The COVID-19 pandemic and the global mpox outbreak are sobering reminders that we will confront new infectious disease threats in the future. Despite new approaches to developing and manufacturing vaccines (particularly mRNA technology) that can reduce the time between pathogen discovery and vaccine availability, that time frame will still be months at best and, for some pathogens, years or decades. Thus, decision-makers will again need to rely on existing and rapidly generated evidence as they implement interventions to mitigate disease spread. In these circumstances, RCTs and meta-analyses have important limitations and should not form the sole, or even primary, basis of public health decisions. Available evidence strongly suggests that masking in the community can reduce the spread of SARS-CoV-2 and that masking with the highest-quality masks that can be made widely available should play an important role in controlling whatever pandemic caused by a respiratory pathogen awaits us.

Accepted for Publication: September 6, 2023.

Published: October 31, 2023. doi:10.1001/jamanetworkopen.2023.39443

Open Access: This is an open access article distributed under the terms of the CC-BY License. 2023 Cash-Goldwasser S et al. JAMA Network Open.

Corresponding Author: Shama Cash-Goldwasser, MD, MPH, Resolve to Save Lives, 85 Broad St, Ste 1626, New York City, NY 10004 (shamacg@gmail.com).

Author Contributions: Dr Cash-Goldwasser had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Cash-Goldwasser, Reingold, Jackson, Frieden.

Acquisition, analysis, or interpretation of data: Cash-Goldwasser, Reingold, Luby, Frieden.

Drafting of the manuscript: Cash-Goldwasser, Reingold, Jackson, Frieden.

Critical review of the manuscript for important intellectual content: Cash-Goldwasser, Reingold, Luby, Frieden.

Administrative, technical, or material support: Cash-Goldwasser, Reingold, Frieden.

Supervision: Reingold, Frieden.

Conflict of Interest Disclosures: Dr Frieden reported that Resolve to Save Lives received funding from Bloomberg Philanthropies, #startsmall, Gates Philanthropy Partners, which is funded with support from the Chan Zuckerberg Foundation, and the Bill & Melinda Gates Foundation. No other disclosures were reported.

Funding/Support: This analysis was conducted with funding from Resolve to Save Lives. Resolve to Save Lives is funded by grants from Bloomberg Philanthropies; the Bill and Melinda Gates Foundation; and the Chan Zuckerberg Initiative DAF, an advised fund of Silicon Valley Community Foundation.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Read the original here:

Masks During Pandemics Caused by Respiratory PathogensEvidence and Implications for Action | Infectious ... - JAMA Network