netterie / Undiagnosed

0 stars 1 forks source link

Explore KC 2012 vs 2015 estimates #15

Open netterie opened 7 years ago

netterie commented 7 years ago

In the 2016 Epi report, KC did something else:

Percent undiagnosed was calculated as 6% among MSM for King County3 ; prior estimate of 15% was used for non-MSM (based on CDC and Washington State estimates) resulting in an estimate of 7.3% overall, rounded up to 8% for a slightly more conservative estimate (this may be the most uncertain bar in the continuum). Estimated people living with HIV/AIDS is calculated by dividing “diagnosed and presumed living in King County” residents by .92. (p33)

The last run I did for Susan was in September, and we got Base Case 6.6% undiagnosed for MSM in 2015 (see Dropbox Undiagnosed Fraction/Users/KC).

The difference arises from longer infection windows in the newer dataset --> longer mean TID. Is this a real change or an artifact? KC folks seem to have believed it was an artifact and went ahead with their own "adjustment."