Face masks have been recommended as one of the first line of defense against SARS-CoV-2 infection for almost two years. So far, most of the evidence behind this recommendation – and my own belief that we should introduce universal masking in public places – has been based on measurements in scientific experiments under laboratory conditions or information from the investigation of special events. As such, some people have questioned whether face masks can be effective in real-world conditions. Well, for the first time a study from California shows that they are.
Face masks are a nuisance. You are uncomfortable. They prevent us from seeing each other’s facial expressions, disrupting the communication and interpersonal connections that are so important to well-being during these last two years. Also, we know face masks aren’t perfect. Still, they might be good enough to make the inconvenience worthwhile. In my opinion, this study shows that they are.
The new study, conducted by the California Department of Health and Human Services and published in the Journal of the CDC Morbidity and Mortality Weekly Report, was a case-control design. Individuals who tested positive for SARS-CoV-2 were demographically matched to other individuals who did not test positive. Both groups were asked about their masking behavior over the past two weeks.
Statistical analyzes showed that people who reported wearing a mask in public were about half as likely to be infected as those who did not. (The results of this study are presented in terms of “odds ratios,” which are related to, but not quite the same as, relative risk.) For people who “always” (rather than “sometimes” or “most of the time”) was the estimated effect even more significant.
A second part of the study attempted to distinguish between cloth masks, surgical masks and N95/KN95 respirators. Not unexpectedly, N95/KN95 were found to reduce the likelihood of infection compared to people not wearing a mask. The surprising thing to me is how effective they were, reducing the relative odds by 83%. Cloth masks and surgical masks were found to be less effective.
A worrying finding of the study is that the evidence for protection among people who reported wearing a mask only “sometimes” (compared to “most of the time” or “always”) is very weak. This could be due to the relatively small number of people who only wore their mask part of the time (less than 10% of the study participants). But it’s probably true that inconsistent or infrequent use just doesn’t provide as much protection.
The study has some weaknesses. The biggest is that the number of participants in the study was simply too small to make very accurate estimates of effectiveness. For example, although the best estimate of risk reduction from wearing an N95/KN95 is 0.17, the confidence interval ranges from 0.05 to 0.64. For this reason, statistical tests aimed at making a fine-grain distinction between the effectiveness of different mask types or usage patterns have been inconclusive.
What was inconclusive was the effectiveness of masks in general, even when not used consistently, and the pattern of increasing effectiveness with mask type and frequency of use. This is consistent with the CDCs current guidance that everyone wears the best possible protective mask that fits well and is worn consistently.