Lockdown impact on age-specific contact patterns and behaviours in France

In the first trimester 2020, a significant number of countries implemented general lockdowns of their populations to contain the quickly expanding SARS-CoV-2 epidemic and avoid major saturation of health care capacity. Understanding how these unprecedented measures impacted population behaviour and contact patterns is key to predict more accurately the health, social and economic impacts of such extreme actions if they were to be applied to future outbreaks. We set up an online survey to measure how the lockdown affected social contact patterns in France, and collected information from 42,036 participants aged 18 years and over between April 10 and April 28, 2020. Among the participants who normally worked outside home prior to the lockdown (72% of the survey population), 68% reported that they had moved to working from home and 17% reported being unemployed during the lockdown.Only 2% of participants used public transport during lockdown, as opposed to 37% before it. Participants reported increased frequency of washing hands, switch in greeting behaviour, but generally limited use of masks outside home. 138,934 contacts were reported, with an average 3.3 contacts per individual per day (1.7 for individuals aged >65 years old compared to 3.6 for younger age-groups). This represented a 70% reduction compared with previous surveys, consistent with reductions in transmission rates measured during the lockdown. Contacts in workplaces, shops, and transports on the previous day were respectively reported in only 11%, 31% and 0.5% of the participants. For those who maintained a professional activity outside home, the frequency of contacts at work dropped by 79%. This study shows that the lockdown dramatically affected population’s behavior, work, risk perception and contact patterns. Both frequency and heterogeneity of contacts were affected, impacting potential important features of virus dissemination. Such surveys are essential to evaluate more accurately the impact of past or future lockdowns and anticipate epidemic dynamics in these conditions.