Seems we need to apply the distributions (they often include prisoners etc) to the higher number (4-6M) or do estimates of usage off the lower number (2.8-3.2M).
Most of those in groups not counted in the lower estimates will never get treated - certainly not with very expensive treatments.
zip
PS - I did join the longs on GILD friday, deeming the selloff overdone even with my negative perception of the nature of the HCV market opportunity.