Effect of risk status for severe COVID-19 on individual contact behaviour during the SARS-CoV-2 pandemic in 2020/2021—an analysis based on the German COVIMOD study

Abstract

Background One of the primary aims of contact restriction measures during the SARS-CoV-2 pandemic has been to protect people at increased risk of severe disease from the virus. Knowledge about the uptake of contact restriction measures in this group is critical for public health decision-making. We analysed data from the German contact survey COVIMOD to assess differences in contact patterns based on risk status, and compared this to pre-pandemic data to establish whether there was a differential response to contact reduction measures. Methods We quantified differences in contact patterns according to risk status by fitting a generalised linear model accounting for within-participant clustering to contact data from 31 COVIMOD survey waves (April 2020-December 2021), and estimated the population-averaged ratio of mean contacts of persons with high risk for a severe COVID-19 outcome due to age or underlying health conditions, to those without. We then compared the results to pre-pandemic data from the contact surveys HaBIDS and POLYMOD. Results Averaged across all analysed waves, COVIMOD participants reported a mean of 3.21 (95% confidence interval (95%CI) 3.14,3.28) daily contacts (truncated at 100), compared to 18.10 (95%CI 17.12,19.06) in POLYMOD and 28.27 (95%CI 26.49,30.15) in HaBIDS. After adjusting for confounders, COVIMOD participants aged 65 or above had 0.83 times (95%CI 0.79,0.87) the number of contacts as younger age groups. In POLYMOD, this ratio was 0.36 (95%CI 0.30,0.43). There was no clear difference in contact patterns due to increased risk from underlying health conditions in either HaBIDS or COVIMOD. We also found that persons in COVIMOD at high risk due to old age increased their non-household contacts less than those not at such risk after strict restriction measures were lifted. Conclusions Over the course of the SARS-CoV-2 pandemic, there was a general reduction in contact numbers in the German population and also a differential response to contact restriction measures based on risk status for severe COVID-19. This differential response needs to be taken into account for parametrisations of mathematical models in a pandemic setting.

Contact behaviour before, during and after the COVID-19 pandemic in the Netherlands: evidence from contact surveys in 2016-2017 and 2020-2023

“…indication for influenza vaccination) in any of the survey rounds nor the baseline survey. This was confirmed by the COVIMOD study in Germany [14] that used the same medical risk definition. However, in some study rounds of the CONNECT study in Canada, it was found that participants with comorbidities had fewer contacts than participants without comorbidities [13].…”

“…The effect of education level on the difference in number of contacts during stringent and relaxed measures was also reported for a contact study in Hungary [33]. The absence of an association between number of contacts and medical risk group (defined as indication for influenza vaccination) was found in the COVIMOD study in Germany [15]. In one of the six rounds of the CONNECT study in Canada, it was found that persons with a higher medical risk have fewer contacts, but this was also the case for their baseline survey [14].…”

“…For instance, contacts in younger age groups were greatly affected by school closures [11], while contacts in older age groups decreased due to suspended social gatherings [8,12]. Persons with comorbidities may have reduced their contacts more than persons without comorbidities [8,13,14], given their higher probability of severe outcomes [15]. Also the socio-economic status of a person may have influenced contact behaviour, as persons with a low socio-economic status often have professions that preclude working from home [16].…”