Why estimating population-based case fatality rates during epidemics may be misleading

Lucas Böttcher; Mingtao Xia; Tom Chou

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03. 26.20044693 doi: medRxiv preprint These solutions hold for the different regions shown in the phase plot of Fig. S2 and are equivalent to those for I(τ > t, t). Corresponding expressions forP d (t) andP r (t) can be found and used to construct M 1 p (t). Fig. S3 (a) shows the magnitude of I(τ, t) in the t − τ plane when we set S(t) = S constant (so that the first equation in Eq. (18) does not apply) such that β 1 SS ≈ 0.158/day. In this case, the epidemic continues to grow in time, but the mortality rates M 0,1 p (t) nonetheless converge as t → ∞. In Fig. S3(b) , we set β 1 S = 0 for t > t q to model strict quarantining after t q = 50 days. We observe no new infections after the onset of strict quarantine measures. In both cases (quarantine and no quarantine), we use ρ(τ ; n = 8, γ = 1.25) (see Eq. (7) in the main text) to describe the initial distribution of infection times τ . As time progresses, more of the distribution of τ moves towards smaller values until quarantine measures take effect (see Fig. S3 (c) and (d)).