The placebo effect is important because it might _increase_ rate of HIV transmission by resulting in more risky behavior. Or, it might reduce rate of HIV transmission by causing the person to focus on the possibility of infection more.
It would be unethical to give a placebo to someone for this kind of disease.
If you knew that it would be "unethical to give a placebo to someone" then you wouldn't have to do the trial!
Indeed - there have been situations in which the Trial turns out to be so positive - that the study is canceled, and all participants are given the treatment. I think that was actually the situation with the circumcision trials - circumcision turned out to be so ludicrously low risk, and so effective at reducing HIV infections in heterosexual male populations in Africa, that they, ethically, had to prematurely end the study and circumcise all the participants.