Placebo-controlled trials are usually considered the standard (and most stringent) measure of a drug's efficacy. It's a surprisingly high hurdle to clear. All sorts of things that people swear by, and all sorts of new things that you'd be sure would work fail when they're up against a similarly sized and colored dose of sugar.

But you can't always run a placebo group, because it isn't always ethical to do so. For a life-threatening condition, the comparison group has to be the current best standard of care (which, after all, is what you're trying to beat.) For lesser diseases, a trial against a known therapy can also be appropriate, although it's usually done after a placebo-controlled one has already been run.

But there's one situation where you can run a placebo control for a deadly condition: when the best standard of care is nothing at all.

Several forms of cancer fall into that category. Pancreatic, renal, and hepatic cancers, for example, exhaust their best available treatments very quickly. Some of the patients in that situation then offer themselves as subjects for clinical research, for which we in the drug industry are extremely grateful. With any luck, we'll be able to find something that works well enough to unblind as quickly as possible.

And when that happens, the disease is no longer in the "placebable" category. There's now an active agent, a possible treatment, and thus a new standard of care. Several cancers have moved off the list in recent years, and here's hoping that the process continues.