Add proportions for NHSN (changed to just total reporting hospitals)

[aysim319]

Due to time constraint the main metric that would be available once nhsn is deployed is raw counts. However, due to the reporting mandate not being consistent; the number of reporting hospital should be considered.

This was mentioned in previous iterations, but at the time was not a factor as normalizing the data would result in noise and therefore would not add consider number of reporting hospitals.

At the final stages of implementing the indicator, it was decided that the proprotions should be added, but in a future release in order prevent more delays in initial release of nhsn

[aysim319]

potential method was dicussed

We could report [totalnumberreportinghospital]/max_per_geo([totalnumberreportinghospital]) (fraction of hospitals currently reporting compared to max ever seen in a location) as the signal and [totalnumberreportinghospital] as the sample size

but led to the conclusion of just introducing the total reporting hospital

One concern about using the concept of "max ever seen in a location" is that this value can change over time. If you meant the max over the specific period when reporting wasn't mandatory (May 1 through Oct 31), then it feels like an ad-hoc, one-off kind of thing -- what do you do if there is another period of non-mandatory? (and also, even during mandatory periods, the number of reporting hospitals can change, and should therefore be reported.
I think it would be cleaner to just report the total number of reporting hospitals (not only during the non-mandatory period), and let users derive reporting proportions on their own if they would like.