Nice paper (linked in the article) but I see one thing that looks like an error. When comparing different start dates for intervention the have 3 curves on a plot. Then for "under reporting" they have another plot which looks like the same 3 curves multiplied by a constant to compensate for the under reporting.
The problem is that the intervention consists of increasing the percent of patients getting treatment from 10 percent to 70 percent. But with under reporting by 2x that would be an actual increase from 5 percent to 35 percent, which would have a much less significant effect on the total number infected - possibly not even ending the epidemic. Perhaps they assume the intervention will include solving the under reporting problem.
The problem is that the intervention consists of increasing the percent of patients getting treatment from 10 percent to 70 percent. But with under reporting by 2x that would be an actual increase from 5 percent to 35 percent, which would have a much less significant effect on the total number infected - possibly not even ending the epidemic. Perhaps they assume the intervention will include solving the under reporting problem.
Otherwise a very nice report: http://www.cdc.gov/mmwr/preview/mmwrhtml/su63e0923a1.htm?s_c...