An additional point on the cost of HIT via infection: if the infection prevalence is high, the total number infected with be quite a bit higher than the HIT - HIT overshoot. HIT causes the rate of new infections to stop increasing, and with more infections, to decrease. But it doesn't just suddenly stop all new infections (as the author, of course, already know).

So those arguing for HIT via infection typically substantially underestimate the cost.