I bet it's control groups.

The FDA thinks similarly. In late 2021, they canceled trials of Paxlovid because it was immoral to have a control group. They did not then _waive the trial requirement_, meaning Paxlovid was both too proven to deny but too unproven to be administered.

Having a control group, but specifically in a study only given to terminally ill patients. Leave the control group to a subsequent study after effectiveness has been established in terminally ill patients.

I understand control groups in general, those are fine. But this is specifically in trials where people are pinning their last hopes out of desperation on some medicine and some random drop of the dice seals their fate.

I agree with you in principle, but where do you draw the line for terminally ill? 100% terminal? in what time frame? 99% terminal? 90%? 50%?

I'd imagine that somebody who has, say, a 10% chance to die in 3 month, is also quite desperate.

(a) some random drop of the dice already sealed their fate. If anything, being admitted to an experimental group and then clustered into control or not, is the DM giving a second roll of the dice.

(b) most therapies don't improve prognosis in terminal cases and some make it worse. We don't know which beforehand. In the cases where the therapy makes it worse, was it therefore immoral to not put more people in the control group?

I'd argue the lack of knowledge makes it a wash morally. If we knew beforehand whether the therapy would be effective, we wouldn't need the control group, but we also wouldn't need the study.