Study population and setting
The study included 124 households in Beijing, China with an index case of laboratory-confirmed COVID-19 diagnosed prior to February 21, 2020. Surveys of index cases and their family members were conducted from February 28, to March 7, 2020.
Summary of Main Findings
41 out of 124 households (33%) experienced secondary transmission within 2 weeks of index case diagnosis, with a total of 77 secondary transmissions among 335 family members of index cases (23%). The proportion of children younger than 18 years old with secondary infection was lower than that of adults (36% vs. 70%, p<0.01). In univariate analyses of mask use, lower odds of any secondary transmission were observed in households with index cases who reported wearing a mask “all the time” after symptom onset versus those reporting “never” wearing a mask (OR: 0.30, 95% CI: 0.11 to 0.82); households in which at least one family member wore a mask “all the time” before index case symptom onset vs. households with no mask use (OR: 0.22, 95% CI: 0.07 to 0.69); and households in which all family members wore masks after index case symptom onset vs. none (OR: 0.20, 95% CI: 0.07 to 0.60). In multivariable logistic regression, higher odds of any secondary transmission were associated with the index case having diarrhea (OR: 4.10, 95% CI: 1.08 to 15.60) and having close contacts (<1 meter) with family members. Lower odds of any secondary transmission were associated with one or more family members, including the index case, wearing a mask before primary case symptom onset (OR: 0.21, 95% CI: 0.06 to 0.79) and household cleaning with chlorine- or ethanol-based disinfectant once a day or more, relative to less frequently (OR: 0.23, 95% CI: 0.07, 0.84).
Survey questions addressed behavior of both index cases and family members, and distinguished mask use before and after symptom onset.
Important methods, including the study inclusion criteria (40 households were excluded from analysis for unclear reasons) and the definition of secondary household transmission, were poorly described. It is not clear how secondary cases were ascertained; for example, if there was testing of all household members regardless of symptoms, or when tests were administered. The methods also do not make clear how secondary cases were distinguished from cases that may have arisen from the same exposure that gave rise to the index case in a household. Given the known heterogeneity in incubation periods, this is a possible source of bias. The timing of secondary transmission relative to index case symptom onset is not reported in detail. Asymptomatic and mild cases are likely under-represented among index cases. Responses are subject to recall bias (i.e., respondents may have misremembered) and social acceptability bias (i.e., respondents may have given answers that are perceived to be more acceptable). Multivariable analysis did not distinguish between the index case wearing a mask vs. household members wearing a mask. The type of masks used were not described.
This is one of the only pieces of available evidence regarding the effectiveness of mask use in preventing transmission of SARS-CoV-2 outside of healthcare settings.
This review was posted on: 12 June 2020