Basic prediction methodology for covid-19: estimation and sensitivity considerations

Tom Britton

For this reason we refrain from predicting the progress of the epidemic during this main phase but only give the rough picture. Once the early stage of the epidemic has passed the exponential growth rate r (or r E in case preventive measures have been introduced) starts to decline. This goes on up until the time t at which the cumulative fraction infected Λ I (t) exceeds 1 − 1/R 0 (or R E in case of preventive measures). This is the time when the expected number of infections no longer exceeds one due to reduced community susceptibility. After this time the number of new infections starts dropping with time thus eventuelly leading the epidemic to a halt. The effect of introducing preventive measures reduces the initial growth rate as described in earlier sections. It also delays the timing of the peak, and particular the height of the peak, closely related the maximal healthcare burdon. The amount of peak-reduction is model dependent, but for a simple model, reducing R 0 = 3 to R E = 2 reduces the height of the peak by around 50%.