Are models useful? Reflections on simple epidemic projection models and the Covid-19 pandemic (Q2019318)

This paper concerns with epidemic projection models. First, the exponential model on total number of infections is introduced. In reality, the size of an epidemic cannot exceed the size of the entire population. To account for this and the fact that those infected run out of susceptibles to infect, then the logistic model is reviewed. To overcome the drawback of symmetry of the logistic function, the so-called Exo-r model, \(\dot{I}=r(t)I\), is introduced, where \(I(t)\) is the total number of infected at time \(t\). The focus is on the doubling time, maximum incidence, and the time to reach the maximum incidence. The total number of infections \(I_{\infty}=\lim_{t\to\infty}I(t)\) is also exploited. The model is fitted to data on infections in (mainland) China and deaths in the United States for the Covid-19 pandemic. In conclusion, simple models like this Exo-r model can plausibly capture a patchwork of complex epidemic unfolding heterogeneously in time and space.