COVID-19 Epidemic in Switzerland: Growth Prediction and Containment Strategy Using Artificial Intelligence and Big Data

Abstract

AbstractUsing a previously developed agent-based artificial intelligence simulation platform (EnerPol) coupled with ‘Big Data,’ the evolution and containment of COVID-19 in Switzerland is examined. The EnerPol platform has been used in a broad range of case studies in different sectors in all of Europe, USA, Japan, South Korea and sub Saharan Africa over the last 10 years. In the present study, the entire Swiss population (8.57 million people), including cross-border commuters, and the entire Swiss public and private transport network that is simulated to assess transmission of the COVID-19 virus. The individual contacts within the population, and possible transmission pathways, are established from a simulation of daily activities that are calibrated with micro-census data. Various governmental interventions with regards to closures and social distancing are also implemented. The epidemiology of the COVID-19 virus is based on publicly available statistical data and adapted to Swiss demographics. The predictions estimate that between 22 February and 11 April 2020, there will be 720 deaths from 83’300 COVID-19 cases, and 73’300 will have recovered; our preliminary variability in these estimates is about 21% over the aforementioned period. In the absence of governmental intervention, 42.7% of the Swiss population would have been infected by 25 April 2020 compared to our prediction of a 1% infection over this time period, saving thousands of lives. It is argued that future scenarios regarding relaxation of the lockdown should be carefully simulated, as by 19 April 2020, there will still remain a substantial number of infected individuals, who could retrigger a second spread of COVID-19. Through the use of a digital tool, such as Enerpol, to evaluate in a data-driven manner the impacts of various policy scenarios, the most effective measures to mitigate a spread of COVID-19 can be devised while we await the deployment of large-scale vaccination for the population globally. By tailoring the spatio-temporal characteristics of the spread to match the capacity of local healthcare facilities, appropriate logistic needs can be determined, in order not to overwhelm the health care services across the country.