Smooth exponential fitting and prediction on COVID-19 transmission charateristics in Italy using SEIR model

Italy is the first country in Europe with more than 200,000 people infected with COVID-19. As a medium-sized country, it has an average mortality rate of COVID-19 about 14.2%, which is much higher than the rates of China, USA, and most of other countries, and has caused the extensive concern of the international community. It is heartening that the number of new deaths from COVID-19 in Italy has recently been limited to single digits per day. As a representative of European countries, Italian successful experience in fighting the epidemic is well worth learning from, and the transmission characteristics of COVID-19 in Italy should get its due attention. Here we conducted a statistical analysis of COVID-19 data from Italy, based on a discretized SERI epidemic model. Considering the potential infectivity of COVID-19 during incubation periods, we processed the real-time statistics data of COVID-19 between February 15 and August 13, 2020 in Italy, and calculated the values of daily transmission rates β and reproduction numbers R
₀
using the 7-day moving average method and gave a detailed error analysis. During the fitting time interval, the mean values of β and R
₀
for COVID-19 are calculated as 0.058-0.068 and 2.47-2.94, respectively, while the current value of R
₀
on August 13 is calculated to be ~1.24-1.47, indicating that the epidemic is getting better. We compared our results on COVID-19 in Italy with those in China, USA and the World, and discussed possible influence of other fitting methods on the results of this paper. Combined with Italian clinical medical data, our research will provide the theoretical and information support to Italian policy makers and help people all over the World understand the dynamics of transmission evolution of the virus and design effective heath intervention strategies to combat COVID-19 in the future.