Using Routine Surveillance Data to Estimate the Epidemic Potential of Emerging Zoonoses: Application to the Emergence of US Swine Origin Influenza A H3N2v Virus

Cauchemez, Simon; Epperson, Scott; Biggerstaff, Matthew; Swerdlow, David; Finelli, Lyn; Ferguson, Neil M.

Up to this point, in the scenario where the detection of a case did not affect detection of other cases from the same cluster (scenario 1), we assumed that cluster information was not available, and thus F could not be calculated. Let's now assume that such information is available and therefore both G and F can be estimated; this allows us to estimate both R and the case detection rate r, as illustrated in the simulation study shown in Figure 6 . First, 12G gives a point estimate for R. Second, once R is estimated, it is possible to infer the case detection rate by determining by how much proportion F differs from proportion G. If the overdispersion parameter k is known (e.g., from detailed contact tracing data), it is possible to accurately estimate the case detection rate; otherwise we can derive informative bounds ( Figure 6 ).