Study population and setting
In April 2020, 6,024 Danish adults with no history of COVID-19 infection who reported spending on average 3 hours every day outside their households were randomized to one of two study arms. The intervention group received a supply of fifty surgical masks (including instructions for proper use) and were recommended to wear these masks in congregate settings outside their homes, while the control group received guidance to follow standard public health recommendations for preventing COVID-19 transmission (e.g. physical distancing, frequent handwashing, avoiding crowded public settings). At the time of study, mask use in community settings was low. All participants received materials and instructions for SARS-CoV-2 point-of-care antibody (IgM and IgG) testing in their homes at baseline and after 1-month follow-up as well as RT-PCR oropharyngeal testing at 1-month follow-up (or following COVID-19 symptom onset during the study period). Participants responded to weekly follow-up emails to report antibody test results, COVID-19 symptoms, and compliance with face mask recommendations (intervention group only), measured by self-reported “complete” adherence, “predominant” adherence, or non-adherence in the past week.
Summary of Main Findings
Fewer than half (46%) of participants in the intervention group reported complete adherence to mask use recommendations, and 47% reported predominant adherence. At the end of the study period, 1.8% of participants in the intervention group and 2.1% of participants in the control group had evidence of SARS-CoV-2 infection as determined through RT-PCR or antibody testing. Overall, the relative odds of COVID-19 infection was 18% lower among participants among participants in the intervention group (OR 0.54-1.23). In sensitivity analyses, intervention group participants who reported insufficient compliance (7%) with mask use recommendations were excluded, but results were similar to the primary intent-to-treat analysis.
he authors randomized participants to receive recommendations for face mask use in attempt to strengthen causal inferences of any observed associations between use of surgical masks and COVID-19 infections. The authors also simulated data using multiple imputations to measure the impact of lost-to-follow-up (19%) on observed outcomes.
Ascertainment of adherence to face mask recommendations in the intervention group through a single self-reported categorical measure (full compliance, predominant compliance, or non-adherence) is subject to misclassification, which likely introduced error into the observed associations. Low adherence to face mask recommendations in the intervention arm, with fewer than half of participants reporting complete compliance, suggests observed non-significant differences in COVID-19 infections between study arms is an artifact of poor intervention compliance and limited study power to detect significant difference between arms rather than face mask ineffectiveness in preventing COVID-19. The study also did not assess indirect benefits of face mask use as source control. Lastly, specimen self-collection among participants and self-reporting of antibody and antigen (i.e., PCR) testing results may have been prone to error.
This study is among the first to use a randomized design to assess whether surgical masks reduce COVID-19 susceptibility among wearers in a setting where mask use is limited.
This review was posted on: 18 November 2020