|Year : 2020 | Volume
| Issue : 6 | Page : 192-200
Facemasks for prevention of viral respiratory infections in community settings: A systematic review and meta-analysis
Nishant Aggarwal1, Vignesh Dwarakanathan2, Nitesh Gautam1, Animesh Ray3
1 Research Assistant, All India Institute of Medical Sciences, New Delhi, India
2 Senior Resident, Department of Community Medicine, All India Institute of Medical Sciences, New Delhi, India
3 Assistant Professor, Department of Medicine, All India Institute of Medical Sciences, New Delhi, India
|Date of Submission||04-May-2020|
|Date of Decision||09-May-2020|
|Date of Acceptance||10-May-2020|
|Date of Web Publication||2-Jun-2020|
Department of Medicine, All India Institute of Medical Sciences, New Delhi
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Background: There is paucity of evidence on the effectiveness of facemask use in COVID-19 in community settings. Objectives: We aimed to estimate the effectiveness of facemask use alone or along with hand hygiene in community settings in reducing the transmission of viral respiratory illness. Methods: We searched PubMed and Embase for randomized controlled trials on facemask use in community settings to prevent viral respiratory illnesses published up to April 25, 2020. Two independent reviewers were involved in synthesis of data. Data extraction and risk-of-bias assessment were done in a standard format from the selected studies. Outcome data for clinically diagnosed or self-reported influenza-like illness (ILI) was recorded from individual studies. Pooled effect size was estimated by random-effects model for “facemask only versus control” and “facemask plus hand hygiene versus control.” Results: Of the 465 studies from PubMed and 437 studies from Embase identified from our search, 9 studies were included in qualitative synthesis and 8 studies in quantitative synthesis. Risk of bias was assessed as low (n = 4), medium (n = 3), or high (n = 1) risk. Interventions included using a triple-layered mask alone or in combination with hand hygiene. Publication bias was not significant. There was no significant reduction in ILI either with facemask alone (n = 5, pooled effect size: −0.17; 95% confidence interval [CI]: −0.43–0.10; P = 0.23; I2 = 10.9%) or facemask with handwash (n = 6, pooled effect size: (n=6, pooled effect size: −0.09; 95% CI: -0.58 to 0.40; P = 0.71, I2 = 69.4%). Conclusion: Existing data pooled from randomized controlled trials do not reveal a reduction in occurrence of ILI with the use of facemask alone in community settings.
Keywords: Coronavirus, COVID-19, hand hygiene, masks, pandemics, severe acute respiratory syndrome coronavirus 2
|How to cite this article:|
Aggarwal N, Dwarakanathan V, Gautam N, Ray A. Facemasks for prevention of viral respiratory infections in community settings: A systematic review and meta-analysis. Indian J Public Health 2020;64, Suppl S2:192-200
|How to cite this URL:|
Aggarwal N, Dwarakanathan V, Gautam N, Ray A. Facemasks for prevention of viral respiratory infections in community settings: A systematic review and meta-analysis. Indian J Public Health [serial online] 2020 [cited 2021 Sep 20];64, Suppl S2:192-200. Available from: https://www.ijph.in/text.asp?2020/64/6/192/285600
| Introduction|| |
COVID-19, like other respiratory viral illnesses, is thought to spread primarily by droplets, leading to contact of droplets with mucous membranes, either directly or through fomites. While social distancing has been recommended as the most effective means for spreading SARS-CoV-2 infection, wearing a facemask has remained a popular choice in the community. Countries such as Austria, Denmark, South Korea, Singapore, and Japan and a few Indian states/union territories such as Delhi, Uttar Pradesh, Maharashtra, Odisha, and Chandigarh have recommended using facemask in public. The World Health Organization maintains that there is insufficient evidence for the use of medical masks by healthy individuals at home or even in mass gatherings. However, the United States Centers for Disease Control and Prevention (CDC) updated their recommendations in early April to advise the use of a cloth face cover for the general public in settings where social distancing may not be achievable.
Till date, there is no direct evidence on the use of facemasks in preventing transmission of infection in the current COVID-19 pandemic, as suggested by a preprint systematic review. However, different studies including randomized controlled trials (RCTs) have reported (though variably) on the effectiveness of masks with or without other measures (such as hand hygiene) in preventing spread of respiratory viral illnesses. The results of such studies may be extrapolated to the current COVID-19 pandemic considering the paucity of evidence. Influenza-like illnesses (ILIs), which are usually a seasonal phenomenon around the globe, are commonly caused by influenza viruses (influenza A, B, and C), adenovirus, respiratory syncytial virus, human rhinovirus, parainfluenza viruses 1–4, and human metapneumovirus (HMPV) along with human coronaviruses.,,, We aimed to conduct a systematic review and meta-analysis to estimate the effectiveness of medical mask usage for prevention of ILI in the community settings.
| Materials and Methods|| |
This study was conducted according to the preferred reporting items for systematic reviews and meta-analysis guidelines (PRISMA).
- Facemask: Standard triple-layered medical procedure mask or surgical masks which can be fitted over the face covering the nose and mouth
- Hand hygiene: Intervention consisting of alcohol-based handrub, liquid soap, or alcohol-based gel
- Influenza like-illness: Self-reported or clinically diagnosed fever of 100°F or greater, oral or equivalent, and cough and/or sore throat as defined by the CDC.
On April 25, 2020, we searched the PubMed and Embase databases for articles meeting our inclusion and exclusion criteria without any time limits. The search terms of our study were as follows:
- PubMed – (Community OR general public OR general population OR house*OR slum* OR population based) AND (mask OR facemask OR N95 OR surgical mask OR PPE OR Personal protective equipment) AND (respiratory virus* OR COVID OR COVID-19 OR SARS-CoV-2 OR Influenza OR Flu OR SARS OR MERS OR Ebola OR H1N1). Filters: Case Reports, Classical Article, Clinical Study, Clinical Trial, Controlled Clinical Trial, Journal Article, Multicenter Study, Observational Study, Randomized Controlled Trial
- Embase – (Community OR general public OR general population OR house*OR slum* OR population based) AND (mask OR facemask OR N95 OR surgical mask OR PPE OR Personal protective equipment) AND (respiratory virus* OR COVID OR COVID-19 OR SARS-CoV-2 OR Influenza OR Flu OR SARS OR MERS OR Ebola OR H1N1). Filters: Article, Article in press.
Review articles were excluded. The reference lists of all relevant articles were screened for identifying additional studies. Our search was limited to articles in English language.
The detailed criteria were as follows:
Inclusion criteria: (1) study type: randomized controlled trials; (2) setting: community; (3) participants: humans with influenza or other viral respiratory illnesses; (4) intervention: the use of facemask with or without other strategies; and (5) outcome: clinically diagnosed influenza or ILIs. Exclusion criteria: (1) trials conducted in the hospital and health-care settings; (2) nonhuman studies; and (3) studies involving mass gatherings, such as Hajj pilgrimage.
The reviewers (NA and VD) independently extracted the data on a predesigned Excel sheet, and any discrepancy was resolved after discussion. The following study and patient characteristics were extracted from each study: sample size of each arm, number of clusters in each arm, intra-cluster correlation coefficient, type of intervention(s), number of events and nonevents among control and intervention groups, given adjusted odds ratio, and 95% confidence interval.
The reviewers (NA and VD) independently screened the articles based on the titles and abstracts for potential inclusion in our study. The full-text version of selected articles was reviewed next. If a consensus about inclusion of a particular study could not be reached, the disagreements were resolved with collective discussion. From the studies that were identified, we extracted study type, study settings, sample size, details of intervention and outcome, number of cases and noncases in both arms of the intervention, and odds ratio (if given). Our approach to article selection is outlined in [Figure 1].
Risk of bias was assessed by two reviewers (NA and VD) for the method of random sequence generation and allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome reporting (attrition bias), and selective reporting (reporting bias). For each item, risk was either “high”, “low” or “unclear”. We used the Cochrane risk-of-bias tool (RoB v2.0), along with the additional considerations for cluster-randomized trials wherever it was applicable. Any disagreements were resolved by mutual discussion. A summary of risk of bias is outlined in [Figure 2].
|Figure 2: Summary of the risk of bias according to Cochrane risk-of-bias tool.|
Click here to view
Data were extracted by NA and VD cross-checked by VD before being entered into the Excel sheet. It was imported to STATA, and the final analysis was done using STATA version 12 (StataCorp, College Station, Texas, USA). Self-reported or clinically diagnosed ILI events were considered as the desired outcome. The sample sizes were adjusted using intra-cluster correlation provided, if the studies were cluster-randomized trial, using general variance inflation method as described in Cochrane handbook. Outcome at the end of period or at 4 weeks was considered. In one study by Larson et al., ILI rate per 1000 person-weeks was reported. Events were calculated for 4 weeks using the provided ILI rate, sample size of each arm, and the assumption that ILI rate would be constant throughout the time period. Odds ratio was the principal summary measure, and wherever the adjusted odds ratio/risk ratio and 95% CI was given, we preferred using them for effect estimate. This was cross-checked by calculating odds ratio with the given number of events and nonevents in intervention and control arm. For meta-analysis, we converted odds ratio into log (odds ratio) and extracted standard error of log (odds ratio). Publication bias was assessed using the Begg's test and visually represented with a funnel plot [Figure 3]. Outcomes were then pooled, and two meta-analyses were done: first, meta-analysis of the use of mask versus control, and second, meta-analysis of the use of mask along with hand hygiene versus control. Heterogeneity was assessed by visual inspection and statistical methods (I square). Random-effects model was planned for calculating the effect size. In sensitivity analysis, we planned to perform the same meta-analysis with fixed-effects model for I2< 25% for “mask and hand hygiene” versus “hand hygiene only” to look into the effectiveness of mask in community setting.
|Figure 3: Results of the meta-analysis to determine the effectiveness of: (a) Mask use versus controls (forest plot) (P = 0.226). (b) Mask use versus controls (funnel plot) (P = 0.806). (c) Masks + hand hygiene versus controls (forest plot) (P = 0.714). (d) Masks + hand hygiene versus controls (funnel plot) (P-value = 0.46).|
Click here to view
| Results|| |
After a search through the PubMed and Embase databases, we identified a total of 902 records. We went through 83 full-text articles, of which 74 were excluded by the inclusion and exclusion criteria [Figure 1]. A systematic review was conducted for 9 cluster RCTs, the results of which are summarized in [Table 1].
Of the nine RCTs, six were done in household setting,,,,,, and two were done with university students as participants in university residence halls., Six of these studies compared three arms of intervention – hand hygiene alone, both hand hygiene and facial masks versus no intervention. In all but one RCT, facemasks were used for respiratory protection of contacts of the index case or the individuals still at risk of getting the infection with,,,,, or without, the use of mask by the index case. On the other hand, Canini et al (2010) used face masks as source control only for the index case but not the contacts. Therefore, we could not include this study in our meta-analysis. Only one study used N95 in one of their intervention arms. This study compared the efficacy of nonfit P2 masks (similar to N95), surgical masks versus no intervention, and found no significant difference in the rates of ILI among the three arms. Nonfit P2 masks were used because the fit testing was not feasible for the general community in the setting of an ongoing pandemic. There were a few notable findings recorded in the trials. Among the two RCTs conducted by Aiello et al. in 2010 and 2012, a significant reduction in the weekly ILI rate in the medical masks and hand hygiene arm during week 4–6 and week 3–6 was noted, respectively. Similarly, two trials conducted by Cowling et al., found a significant difference in the rate of ILI in the hand hygiene and medical mask arm if the interventions were applied within the first 36 h of illness onset in the index case. An RCT conducted by Simmerman et al. discovered a significant increase in the odds ratio in the hand hygiene arm and hand hygiene plus facemask arm, suggesting an increased susceptibility to influenza cases with the interventions, which contradicts the perceived protective effect.
No significant publication bias was noted for studies included in mask only versus controls [Figure 3]. We found no significant association between mask use and decrease in events of ILI (pooled effect size: −0.17; 95% confidence interval [CI]: −0.43–0.10; P = 0.23). Heterogeneity was low by I2 statistics (I2 = 10.9%).
Pooled effect of the usage of mask and hand hygiene against controls was calculated [Figure 3]. Publication bias was not significant for studies included in mask and hand hygiene versus controls (P value by Begg's test = 0.45). There was no significant decrease in ILI events with mask and hand hygiene compared to controls (pooled effect size: −0.09; 95% CI: -0.58 to 0.40; P = 0.71, I2 = 69.4%).
We conducted the following sensitivity analyses [Figure 4]:
|Figure 4: Sensitivity analysis results of the meta-analyses to determine the effectiveness of: (a) Masks versus controls using fixed-effects model. (b) Masks + hand hygiene versus controls excluding the Simmerman et al. study (P value = 0.005). (c) Masks + hand hygiene versus hand hygiene only. (d) Masks + hand hygiene versus mask use only.|
Click here to view
- Using fixed-effects model for mask only versus control: Effect size remained insignificant even after using fixed-effects model (n = 5, pooled effect size: −0.19; 95% CI: −0.424–0.043; P = 0.11, I2 = 10.9%)
- Using random-effects model for mask and hand hygiene versus control after excluding study by Simmerman et al.: This study was excluded for the following reasons. Of the six studies used, effect size of a study by Simmerman et al. was outside the funnel. Furthermore, Simmerman et al. noted that the adherence to intervention was particularly poor and the social customs in Bangkok lead to low use of mask. In addition, due to H1N1 outbreak, mask use and hand hygiene in the control group substantially increased. The meta-analysis conducted after excluding this study showed that mask plus hand hygiene was associated with lower events of ILI (n = 5, pooled effect size: −0.344; 95% CI: -0.583 to -0.105; P = 0.005, I2 = 0%)
- Random-effects model between mask and hand hygiene (intervention group) versus mask only (control): Effect size was insignificant between both the interventions for preventing events of ILI (n = 3, pooled effect size: −0.136; 95% CI: −0.406–0.133; P = 0.321, I2 = 0%)
- Random-effects model between mask and hand hygiene (intervention) versus hand hygiene only (control): No significant difference was found in ILI events (n = 3, pooled effect size: 0.102; 95% CI: −0.485–0.689; P = 0.73, I2 = 0%).
| Discussion|| |
The pooled estimate of the randomized control trials did not show any significant reduction of ILI by the use of facemasks with or without hand hygiene in community settings. However, sensitivity analysis showed a significant protective role of facemasks and hand hygiene after the study by Simmerman et al. was excluded due to the high risk of bias and probable contamination. Thus, the use of facemasks along with proper hand hygiene methods, but not facemasks alone, showed statistically significant benefits vis-à-vis no interventions. It may be reasoned that addressing fomite-borne transmission in addition to prevention of droplet infections by masks is important to derive protective benefits, if any. However, in a pandemic situation, the exact degree of protection offered by facemask use, over and above other practices such as hand hygiene, still remains obscure.
Two main factors, namely correct fitting of facemask and adherence to facemask usage, are important in preventing droplet infections. Adherence to facemask use was low in the intervention groups in all the studies which reported it [Table 1]. Correct fit of facemask was not reported in any of the studies included in this meta-analysis. These may be representative of the real-life scenario where both long-term adherence and proper usage of facemasks are considered to be major hurdles for widespread implementation and derived benefits. In addition, the use of medical mask may provide a false sense of security and may lead to neglect of other essential practices such as social distancing and hand hygiene, while also diverting mask supply away from the health-care workers.
Our study had a few limitations. Only a small number of studies were found eligible in our meta-analyses. For calculation of pooled estimate for mask use against control, larger weightage was given to one study. In addition, it may be assumed that certain practices such as hand hygiene may already be in practice by all arms in the studies, which would lead to regression toward null value. Furthermore, in the studies in which facemask was used by the index case and contacts, there could have been a time lag between infection inoculation and start of intervention. Subjects of the study could have also acquired infection outside the study area, which may not have been accounted in the individual studies. Another limitation of our study was the nonuniformity in the definition of ILI among the studies. Most of our included studies,,,, used the standard CDC case definition of ILI as one of the outcomes, but some studies,, used a different definition as described in [Table 1].
Thus, using facemask alone may not be an effective strategy in preventing transmission of viral respiratory diseases in community settings. Existing data pooled from the RCTs do not reveal a reduction in the incidence of ILI with the use of facemask alone in community settings. Well-conducted clinical studies are warranted to determine whether these results can be applied to the COVID-19 pandemic.
| Conclusion|| |
Available evidence does not confirm a protective effect of face mask usage alone in a community setting against influenza-like illnesses (and potentially, the COVID-19). For maximum benefit, mask use should be combined with other essential non-pharmaceutical interventions like hand hygiene.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Lewnard JA, Lo NC. Scientific and ethical basis for social-distancing interventions against COVID-19. Lancet Infect Dis 2020. pii: S1473-3099(20)30190-0.
Concerns around Public Health Recommendations on Facemask use among Individuals who are not Medically Diagnosed with COVID-19 Supported by a Systematic Review Search for Evidence. 2020; published online April, 10. Available from: https://doi.org/10.21203/rs. 3.rs-16701/v3
. [Last accessed on 2020 Apr 30].
Nguyen DNT, Mai LQ, Bryant JE, Hang NLK, Hoa LNM, Nadjm B, et al. Epidemiology and etiology of influenza-like-illness in households in Vietnam; it's not all about the kids! J Clin Virol 2016;82:126-32.
Puzelli S, Valdarchi C, Ciotti M, Dorrucci M, Farchi F, Babakir-Mina M, et al.
Viral causes of influenza-like illness: Insight from a study during the winters 2004-2007. J Med Virol 2009;81:2066-71.
Razuri H, Malecki M, Tinoco Y, Ortiz E, Guezala MC, Romero C, et al
. Human coronavirus-associated influenza-like illness in the community setting in peru. Am J Trop Med Hyg 2015;93:1038-40.
Fu Y, Pan L, Sun Q, Zhu W, Zhu L, Ye C,et al
. The clinical and etiological characteristics of influenza-like illness (ILI) in outpatients in Shanghai, China, 2011 to 2013. PLoS One 2015;10:e0119513.
Moher D. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Ann Intern Med 2009;151:264.
Higgins JPT, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD, et al
. The cochrane collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011;343:d5928.
StataCorp. Stata Statistical Software: Release 12. College Station, TX: StataCorp LP; 2011.
Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA. Cochrane Handbook for Systematic Reviews of Interventions version 6.0 (updated July 2019). Cochrane, 2019. Available from www.training.cochrane.org/handbook. [Last accessed on 2020 May 03].
Larson EL, Ferng YH, Wong-McLoughlin J, Wang S, Haber M, Morse SS. Impact of non-pharmaceutical interventions on URIs and influenza in crowded, urban households. Public Health Rep 2010;125:178-91.
Bland JM, Altman DG. Statistics notes. The odds ratio. BMJ 2000;320:1468.
Aiello AE, Murray GF, Perez V, Coulborn RM, Davis BM, Uddin M, et al. Mask use, hand hygiene, and seasonal influenza-like illness among young adults: A randomized intervention trial. J Infect Dis 2010;201:491-8.
Aiello AE, Perez V, Coulborn RM, Davis BM, Uddin M, Monto AS. Facemasks, hand hygiene, and influenza among young adults: A randomized intervention trial. PLoS One 2012;7:e29744.
Canini L, Andréoletti L, Ferrari P, D'Angelo R, Blanchon T, Lemaitre M, et al. Surgical mask to prevent influenza transmission in households: A cluster randomized trial. PLoS One 2010;5:e13998.
Cowling BJ, Fung RO, Cheng CK, Fang VJ, Chan KH, Seto WH, et al. Preliminary findings of a randomized trial of non-pharmaceutical interventions to prevent influenza transmission in households. PLoS One 2008;3:e2101.
Cowling BJ, Chan KH, Fang VJ, Cheng CK, Fung RO, Wai W, et al. Facemasks and hand hygiene to prevent influenza transmission in households: A cluster randomized trial. Ann Intern Med 2009;151:437-46.
MacIntyre CR, Cauchemez S, Dwyer DE, Seale H, Cheung P, Browne G, et al. Facemask use and control of respiratory virus transmission in households. Emerg Infect Dis 2009;15:233-41.
Simmerman JM, Suntarattiwong P, Levy J, Jarman RG, Kaewchana S, Gibbons RV,et al. Findings from a household randomized controlled trial of hand washing and facemasks to reduce influenza transmission in Bangkok, Thailand. Influenza Other Respir Viruses 2011;5:256-67.
Suess T, Remschmidt C, Schink SB, Schweiger B, Nitsche A, Schroeder K, et al. The role of facemasks and hand hygiene in the prevention of influenza transmission in households: Results from a cluster randomised trial; Berlin, Germany, 2009-2011. BMC Infect Dis 2012;12:26.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
|This article has been cited by|
||SARS-CoV-2, surgeons and surgical masks
| ||Mohammad Ibrahim Khalil,Gouri Rani Banik,Sarab Mansoor,Amani S Alqahtani,Harunor Rashid |
| ||World Journal of Clinical Cases. 2021; 9(10): 2170 |
|[Pubmed] | [DOI]|
||Nonpharmacological interventions to reduce respiratory viral transmission: an evidence-based review with recommendations
| ||E. Yuen,J. Fried,C. Salvador,D.A. Gudis,R.J. Schlosser,S.A. Nguyen,E.A. Brennan,N.R. Rowan |
| ||Rhinology journal. 2021; 0(0): 0 |
|[Pubmed] | [DOI]|
||Effectiveness of community face mask use on COVID-19 epidemiological trends and patterns in Italy: evidence from a “translational” study
| ||Nicola Luigi Bragazzi,Naim Mahroum,Giovanni Damiani,Jude Dzevela Kong,Jianhong Wu |
| ||Infectious Diseases. 2021; 53(4): 252 |
|[Pubmed] | [DOI]|
||PM2.5, NO2, wildfires, and other environmental exposures are linked to higher Covid 19 incidence, severity, and death rates
| ||Luke Curtis |
| ||Environmental Science and Pollution Research. 2021; |
|[Pubmed] | [DOI]|
||Is risk compensation threatening public health in the covid-19 pandemic?
| ||Eleni Mantzari,G James Rubin,Theresa M Marteau |
| ||BMJ. 2020; : m2913 |
|[Pubmed] | [DOI]|
||Face masks and COVID-19: don’t let perfect be the enemy of good
| ||Benjamin J Cowling,Gabriel M Leung |
| ||Eurosurveillance. 2020; 25(49) |
|[Pubmed] | [DOI]|