ABSTRACT
Background: Ecologic studies support non-pharmacologic measures to protect against SARS-CoV-2, but few studies have evaluated their effects at the individual level in the community. We aimed to assess the magnitude of the associations of social distancing and mask use with laboratory-confirmed infection by SARS-CoV-2 in adults living in Porto Alegre, Brazil, a city with 1,483,771 inhabitants.Methods: We conducted a population-based case-control study from late April to June 2020. Cases came from a list of all cases mandatorily notified to municipal authorities; controls were antibody-negative participants of three representative household surveys conducted at the same time-period.Findings: In logistic regression analyses of 271 cases and 1396 controls adjusted for age, sex, race, educational attainment, income, household size and pandemic moment, those reporting moderate to greatest adherence to social distancing had between 59% (OR=0·41; 95%CI 0·24-0·70) and 75% (OR=0·25; 0·15-0·42) lower odds of becoming infected, compared to those reporting very little adherence. Lesser out-of-household exposure reduced odds between 52% (OR=0·48; 0·29-0·77) and 75% (OR=0·25; 0·18-0·36), compared to going out every day all day. In a subsample of controls with data on mask use and cases of equivalent pandemic moment (198 cases and 420 controls), mask use reduced odds of infection by 87% (OR=0·13; 0·04-0·36).Interpretation: Greater social distancing and always using masks while away from home provided major protection against SARS-CoV-2. These simple measures can be of great benefit during the upcoming phases of the pandemic.Funding: IATS/FAPERGS, UNIMED Porto Alegre, Instituto Cultural Floresta, Instituto Serrapilheira, Ministry of Health.Declaration of Interests: The authors have no conflicts of interest to declare.Ethics Approval Statement: The ethics committee of the Hospital de Clínicas de Porto Alegre approved our study (No. 31499420·5·0000·5327) and the Brazilian National Ethics Committee (No. 30415520·2·0000·5313) approved the accompanying seroprevalence surveys. All participants gave prior informed consent, in written form by the controls and verbally for cases.