Don't we want to include the modeling of sero-prevalence as a function of age and time at the same time that we developed [here](https://choisy.github.io/hfmd/analysis.html)?