A note on parameter estimation for variant Creutzfeldt-Jakob disease epidemic models.

A recent series of papers has raised issues regarding estimation of the key epidemiological parameters of variant Creutzfeldt-Jakob disease (vCJD) from fitting survival models to case data. In particular, it was stated that the scale of the epidemic cannot be estimated and must be fixed in any analysis. We show that this problem is an artefact of the approximate likelihood used in these papers to facilitate model-fitting, and is not a concern if estimation is based on the full likelihood. We also address comments in these papers regarding the choice of incubation period distribution.