An agent-based epidemic model REINA for COVID-19 to identify destructive policies

Jouni T Tuomisto; Juha Yrjölä; Mikko Kolehmainen; Juhani Bonsdorff; Jami Pekkanen; Tero Tikkanen

Empirical data is currently emerging about the effectiveness of actions of social isolation, closing of schools, and other mobility restrictions. The available estimates show a wide range of outcomes: Pepe et. al (2020) estimates based on mobile device location data only 19% reduction in potential encounters in Italy on 7-10 March during Italy's third week of travel restrictions, while Jarvis et al. (2020) estimates 73% reduction of average daily contacts in the UK under lock-down in late March 2020. Google (2020) estimates 52% decrease in mobility in retail and recreation sectors in Finland on 2020-03-29 from the baseline. It is difficult to make a practical interpretation about what relative reduction is achieved by which particular action. Yet, the findings of Jarvis et al. suggests that realistic but not complete lock-down measures can indeed lead to very significant relative reduction of mobility. Bayesian and other methods could be tested in updating parameter estimates by using data from countries with different strategies.