Case-control study of behavioural and societal risk factors for sporadic SARS-CoV-2 infections, Germany, 2020-2021 (CoViRiS study)

Abstract

During the COVID-19 pandemic in Germany, a variety of societal activities were restricted to minimize direct personal interactions and, consequently, reduce SARS-CoV-2 transmission. The aim of the CoViRiS study was to investigate whether certain behaviours and societal factors were associated with the risk of sporadic symptomatic SARS-CoV-2 infections. Adult COVID-19 cases and frequency-matched population controls were interviewed by telephone regarding activities that involved contact with other people during the 10 days before illness onset (cases) or before the interview (controls). Associations between activities and symptomatic SARS-CoV-2 infection were analysed using logistic regression models adjusted for potential confounding variables. Data of 859 cases and 1 971 controls were available for analysis. The risk of symptomatic SARS-CoV-2 infection was lower for individuals who worked from home (adjusted odds ratio (aOR) 0.5; 95% confidence interval (CI) 0.3-0.6). Working in a health care setting was associated with a higher risk (aOR: 1.5; 95% CI: 1.1-2.1) as were private indoor contacts, personal contacts that involved shaking hands or hugging, and overnight travelling within Germany. Our results are in line with some of the public health recommendations aimed at reducing interpersonal contacts during the COVID-19 pandemic.

Keywords: COVID-19; SARS-CoV-2; infectious disease epidemiology; public health; virus infection.

Figure 1.

Adjusted odds ratios with 95% confidence intervals of associations of household characteristics with symptomatic SARS-CoV-2 infections. Each exposure (household size (three categories); household with child/children of a certain age group (two categories: yes/no); household with a risk group person (two categories: yes/no)) was analysed in a logistic regression model with five adjusting covariables (see the ‘Methods’ section for details). Analysis of the association with children’s age was restricted to participants living with at least one child <18 years of age (259 cases; 687 controls). Analysis of the association with living with a ‘risk group person’ was restricted to participants not living alone (708 cases; 1 623 controls). Examples of ‘risk group person’: person>60 years of age or/and with chronic disease. Observations with missing values in the variable of interest or any of the five adjusting covariables were excluded from analysis.

Figure 2.

Adjusted odds ratios with 95% confidence intervals of workplace-related associations with symptomatic SARS-CoV-2 infections. Each exposure was analysed in a logistic regression model with five adjusting covariables (see the ‘Methods’ section for details). Analysis of the association with face masks when in direct contact with colleagues and with room ventilation was restricted to participants who reported direct contact indoors with at least one colleague (332 cases; 522 controls). Observations with missing values in the variable of interest or any of the five adjusting covariables were excluded from analysis.

Figure 3.

Adjusted odds ratios with 95% confidence intervals of private contacts-related associations with symptomatic SARS-CoV-2 infections. Each exposure (two categories: yes/no) was analysed in a logistic regression model with five adjusting covariables (see the ‘Methods’ section for details). Analysis of outdoor versus indoor contacts was restricted to participants who reported only indoor contacts or only outdoor contacts (324 case-patients; 821 controls). Observations with missing values in the variable of interest or any of the five adjusting covariables were excluded from analysis.

Figure 4.

Adjusted odds ratios with 95% confidence intervals of associations of exposures (social and other activities) with symptomatic SARS-CoV-2 infections. Each single exposure was analysed in a logistic regression model with five adjusting covariables (see the ‘Methods’ section for details). Analysis of indoor eating was restricted to participants who reported eating at a restaurant (95 cases; 333 controls). Analysis of visiting a bar/pub indoors was restricted to participants who reported visiting a bar/pub (45 cases; 177 controls). Observations with missing values in the variable of interest or any of the five adjusting covariables were excluded from analysis. HH: household.

Figure 5.

Adjusted odds ratios with 95% confidence intervals of associations of exposures related to transport and travelling with symptomatic SARS-CoV-2 infections. Each single exposure was analysed in a logistic regression model with five adjusting covariables (see the ‘Methods’ section for details). Analysis of details of travelling was restricted to participants who reported any domestic travelling (148 cases; 440 controls). Observations with missing values in the variable of interest or any of the five adjusting covariables were excluded from analysis. HH: household.

Figure 6.

Adjusted odds ratios with 95% confidence intervals of associations of pre-existing medical conditions and personal characteristics with symptomatic SARS-CoV-2 infections. Each single variable was analysed in a logistic regression model with five adjusting covariables (see the ‘Methods’ section for details). The variable ‘frequency of smoking’ had four categories. Other variables shown in Figure 6 had two categories (yes/no). Observations with missing values in the variable of interest or any of the five adjusting covariables were excluded from analysis. BMI: body mass index; Ref: reference category; Y/N: yes versus no.