There's a New York Times Room for Debate feature has an excellent discussion of the ethics of trials for Ebola treatments and vaccines. Here's part of the essay by Nancy Kass and Steven Goodman:
Ethics is not just figuring out which side poses better arguments; often it’s best to find a third way. Given the breadth and deadly nature of the current Ebola outbreak, and unknowns about treatments, an "adaptive approach" seems most appropriate. Adaptive approaches allow researchers to plan a sequence of studies, or modify a single study in almost real time, as they learn more about a drug. In West Africa, for example, the first 40 Ebola patients in a trial could all get an experimental treatment, and nobody would take a placebo. If nearly all patients survived, in settings where most others were dying with the same supportive care, then it is possible that placebo testing could be avoided, and subsequent trials could randomize to different doses or treatments.
But if the results of the first trial, without placebos, revealed anything less than an almost certain cure, a design with proper controls would have to be initiated, and explained to those participating in the trial. Patients must be told that the drug is not a guaranteed life-saver, so they can see the point of the control group. (And given the multiple beliefs about Ebola among West Africans, creative approaches to promoting understanding and consent are important as well.) These placebo-controlled trials could themselves be adaptive in design, randomizing more patients to whichever therapy appears most effective, until the verdict is clear. If we are to design trials to minimize suffering and death in a whole population, we must temper our compassion with humility about what we think we know.