Data Science for COVID-19

COVID-19 is a novel, rapidly spreading zoonotic disease, which is characterized by fever, shortness of breath and coughing. In December 2019 the first case of this disease was reported in Wuhan, China. Since then, the virus has infected more than 30 million people and caused more than 929 thousand deaths (John Hopkins University, 2020). The virus is transmitted through droplets during close unprotected contact with in infected individual and is genetically similar to corona viruses that cause severe respiratory syndrome (SARS) and the Middle East respiratory syndrom (MERS).COVID-19 is highly infectious, as shown by the preliminary estimates of the basic reproduction number R0 which ranges from 2.8 to 5.5 in the absence of quarantines and social distancing measures (Nussbaumer-Streit et al., 2020). 

Already the disease has shown to have a major impact on utilization of Intensive Care Units (ICU). For instance in Grasselli et al. (2020) it was stated that an immediate sharp increase could be observed in ICU admissions in Lombardy, Italy. As of 7th March 2020, the current total number of patients with COVID-19 occupying an ICU bed (n = 359) represented 16% of the currently hospitalized patients with COVID-19 (n = 2217). Also in the United State of America COVID-19 has led to a dramatic increase in the demand for ICU beds .  

Aim of the project was to develop a simulation model in order to determine the impact of several measures, such as social distancing, on the utilization of the Intensive Care Units in the Dutch city of Rotterdam. Several approaches were applied such as agent-based modelling and discrete event simulation. Besides simulation approaches also time series analysis (ARIMA) was applied. 

The result of the projects were several models able to predict the number of infections with the virus and the consequent effect of the utilization of the ICU.  

Dr. Raymond Hoogendoorn
Senior Researcher