Apply IHR at time of infection, not time of observation

[damonbayer]

          Not for this PR, but i think we can go ahead and change this in subsequent versions.

Originally posted by @dylanhmorris in #17 (comment)

see also CDCgov/ww-inference-model#43

[dylanhmorris]

Bumping this.

[SamuelBrand1]

This is a good point