New Meta-Analysis out
SUMMARY AND CONCLUSION
This review was commissioned partly because of controversy around a Cochrane review which was interpreted by some people as providing definitive evidence that masks don’t work (
9). Our extensive review of multiple streams of evidence from different disciplines and study designs builds on previous cross-disciplinary narrative reviews (
233,
412) and aligns with the recent call from philosophers of science to shift from a “measurement framework” (which draws solely or mainly on RCTs) to an “argument framework” (which systematically synthesizes evidence from multiple designs including mechanistic and real-world evidence) (
19). Using this approach, we have demonstrated a more nuanced set of conclusions, summarized below, and have revealed why certain inaccurate assumptions and defective reasoning about the science of masks and masking seem to have become widely accepted among certain groups.
We began by reviewing basic science evidence on the transmission of SARS-CoV-2 and other respiratory pathogens and showed that there is strong and consistent evidence that they spread predominantly by the airborne route. We also showed that masks are effective, and well-fitting respirators are highly effective, in reducing transmission of respiratory pathogens, and that these devices demonstrate a dose-response effect (the level of protection increases as adherence to masking increases).
We then provided a methodological critique of clinical trials of masks in the control of respiratory disease epidemics and outbreaks, including listing common design flaws. We summarized evidence from RCTs, including repeating methodologically flawed meta-analyses, and showed that respirators are significantly more effective than medical or cloth masks, especially (and to the extent that) they are actually worn in all potentially hazardous circumstances.
We also reviewed an extensive body of observational and modeling evidence which showed that, overall, masking and mask mandates are effective in reducing community transmission of respiratory diseases during periods of high community transmission. The observational findings are particularly striking since various inherent limitations of such designs are likely to bias findings toward the null.
Our review of adverse effects and harms of masks found strong evidence to refute claims by anti-mask groups that masks are dangerous to the general population. We also found that masking may be relatively contraindicated in individuals with certain medical conditions and that certain groups (notably D/deaf people) are disadvantaged when others are masked.
We summarized evidence from multiple countries and cultures which shows that masks are important sociocultural symbols about which people care deeply (positively or negatively). We also showed that adherence (and non-adherence) to masking is sometimes linked to political and ideological beliefs and to widely circulated mis- or disinformation, and hence hard to change.
In a section on mask policy, we described how governments and organizations need explicit policies on using masks for prevention and control of respiratory infections, covering personal protection of at-risk groups; protection in specific settings, including workplaces and healthcare facilities; seasonal respiratory infections; and pandemics. These policies need to be based on sound risk assessment, risk management, and implementation principles.
Finally, we reviewed environmental impacts from single-use masks and respirators and highlighted novel materials and designs with improved performance and less environmental risk.
We believe this evidence supports several important conclusions and implications for further research.
First, the claim that masks don’t work is demonstrably incorrect, and appears to be based on a combination of flawed assumptions, flawed meta-analysis methods, errors of reasoning, failure to understand (or refusal to acknowledge) mechanistic evidence, and limitations in critical appraisal and evidence synthesis. Masks and respirators work if and to the extent that they are well-designed (e.g., made of high-filtration materials), well-fitting and actually worn. The heterogeneity of available mask RCTs does not appear to have been fully understood by some researchers who have conducted high-profile meta-analyses of the same. It is time for the research community to move on from addressing the binary question “do masks work?” through unidisciplinary and epistemologically exclusionary study designs and pursue more nuanced and multi-faceted questions via
interdisciplinary designs.
A fruitful avenue for future research, for example, would be the combination of experimental, observational and modeling data to refine our understanding of
when universal masking should be introduced during respiratory epidemics and
how best to promote and support masking policies in different situations and settings, and especially for groups at increased risk, during such outbreaks. Research on ventilation, filtration and other measures to improve indoor air quality was beyond the scope of this review [it has been covered elsewhere (
330,
395)], but there is scope for cross-disciplinary modeling to bring the science of indoor air into more direct dialogue with that of infectious disease transmission and masking to address the question of when and in what circumstances indoor masking can be deemed unnecessary (or, alternatively, advised or mandated) based on air quality. As noted in Modeling Masking, some research groups have begun to contribute to this interdisciplinary knowledge base.
Second, given that masking is an effective (though not perfect) intervention for controlling the spread of respiratory infections, and that it may be particularly important in the early stages of pandemics (when the pathogen may be unknown and drugs and vaccines are not yet available), improving understanding among scientists, clinicians, policymakers and the public about the effectiveness of masks and respirators is an urgent priority. The continuing recalcitrance of many (though not all) in the infection prevention and control community on this issue could prove a major threat to public health in future pandemics, particularly since such individuals often hold influential positions on global and national public health decision-making bodies.