It seems pretty safe to guess that the inequality is reversed. But probabilities are not enough to answer the question. One must weigh expected costs and benefits.

A crude method, not of estimating any probabilities, but directly assessing the final decision, is to look at the existence of countries with different systems. That they have different waiting periods shows that the decision is not blindingly obvious. Moreover, since governments tend to be more sensitive to the publicity of concentrated harm, I would expect them to err in the direction of exclusion. That I have not heard of a recent Brazilian getting HIV from blood transfusion suggests that not only is their policy better, but that P(D|N) is unmeasureably low.

That just addresses the issue of HIV, which is what these policies that mention 1977 are about. One might worry that future diseases will appear first among gays. For this, it is much harder to determine the probabilities.