Reference: Ann Intern Med. 2022 Dec;175(12):1629-1638
Practice Point: N-95 masks offer the best protection for healthcare workers against COVID-19, even if this study tries to make us feel better about wearing surgical masks.
EBM Pearl: For noninferiority trials, and we can’t stress this enough, a per-protocol analysis is the preferred method to analyze data to minimize risk of inappropriately concluding the intervention is just as good as the comparator, when in fact it is inferior.
Masks have been mandated for healthcare workers since the beginning of the COVID-19 pandemic, the consensus being that a well-fitted N-95 mask or equivalent is best. Unfortunately, we all know that N-95 masks are often unavailable and always uncomfortable. In addition, if not well-fitted, the effectiveness decreases substantially. The main alternative in healthcare settings has been surgical masks, which are less protective in laboratory settings but with unknown effectiveness compared to N-95 masks in the real world. A recent randomized trial led by researchers in Canada tried to answer this question.
This noninferiority trial began in May of 2020 in Canada, but sites in Pakistan, Israel, and Egypt were subsequently added. There were numerous other changes to the protocol, mostly out of necessity as the pandemic evolved. The trial was designed to see if surgical masks were as good as (noninferior) N-95 masks, with a noninferiority margin originally set at 5%, meaning that if rates of COVID infection were no more than 5% worse with surgical masks compared to N-95s, they would be considered noninferior. However, as the pandemic evolved, there was a need to account for people who got COVID or who got vaccinated (the vaccine was not available at the start of the trial) and the noninferiority margin was changed to a [very large] hazard ratio of less than 2. The reason for choosing such a large hazard ratio was not explained. At the end of the trial, the researchers found that the hazard ratio for surgical masks compared to N-95 masks was 1.14 with a 95% CI of 0.77 to 1.69, which was considered noninferior.
While it was nice to see a randomized trial about something practical and important, this one has so many limitations that it’s hard to draw any meaningful conclusions. Some of the limitations were unavoidable due to the changes in the pandemic over time, such as the varying virulence levels of different strains of COVID-19 and the development of multiple vaccines which led to significant heterogeneity in the outcomes in the four countries. However, some of the limitations of this trial were the result of unexplained and unfortunate choices, such as the extremely wide inferiority margin and the lack of a per-protocol analysis. Even the authors do not conclude that surgical masks are noninferior, only that the rates of infection over time with surgical masks will probably be less than twice as much as with N-95s. This is hardly a ringing endorsement. Overall, these results are not reliable enough to warrant changing our current understanding that N-95 masks are more effective than surgical masks for preventing aerosol transmission, even with all of the challenges of the healthcare work environment.
For more information, see the topic COVID-19 Infection Control and Prevention in DynaMed.
DynaMed EBM Focus Editorial Team
This EBM Focus was written by Alan Ehrlich, MD, Executive Editor at DynaMed and Associate Professor in Family Medicine at the University of Massachusetts Medical School. Edited by Katharine DeGeorge, MD, MS, Deputy Editor at DynaMed and Associate Professor of Family Medicine at the University of Virginia; Dan Randall, MD, Deputy Editor at DynaMed; Nicole Jensen, MD, Family Physician at WholeHealth Medical; Vincent Lemaitre, PhD, Senior Medical Writer at DynaMed; and Sarah Hill, MSc, Associate Editor at DynaMed.