I remember attending an interview where Barry Marshall jokingly claimed that he was the only person who'd ever earned a Nobel prize with a study size of n=1. At the time it seemed like a reasonable claim, not just a joke, and it's now interesting to me to see how wrong he was with that joke.
The team that discovered ibuprofen routinely tested candidates on themselves.[1] It wasn't uncommon. My dad did it at Merck in the '70s. In one case, he and his lab partner took a candidate appetite suppressant (serotonergic) and then went out for Chinese. They ate like kings and concluded it didn't work.
Bit of experience here: in the USA the FDA gets very upset about unsanctioned experimentation. There is a quasi-exception. The one drug I worked on that made it into trials had previously been tried on one human -- me. I already knew it wouldn't have tox implications -- we were more interested in the injection procedure itself. Plus we did some diffusion tests on my tissue which showed results consistent with the guinea pigs. The quasi-exception is that I was the CEO -- even the chief scientist couldn't do it since it could later be claimed I coerced her. But supposedly the CEO couldn't be coerced.
It's only a quasi-exception since you still have to report it in the back of your NDA (New Drug Application) and you're reporting a violation of procedure. But the NDA is enormous (used to be truckloads of documents, though now electronic), and by that time you have hundreds of human experiences with the drug. Obviously with an n of 1 it can't count as trial data, you're just writing up that once you were naughty.
(the whole process is weird -- technically, when you apply to test a new drug you're not actually asking permission, you are actually asking the Agency is not to prosecute you for using an unlicensed drug because you promise to be careful and that your plans fit into one of the exceptions written up in section 505(b) of the food and drug act.)
I remember attending an interview where Barry Marshall jokingly claimed that he was the only person who'd ever earned a Nobel prize with a study size of n=1. At the time it seemed like a reasonable claim, not just a joke, and it's now interesting to me to see how wrong he was with that joke.
Yeah, I was surprised not to see his case mentioned until I learned from the article itself how common it is.
The team that discovered ibuprofen routinely tested candidates on themselves.[1] It wasn't uncommon. My dad did it at Merck in the '70s. In one case, he and his lab partner took a candidate appetite suppressant (serotonergic) and then went out for Chinese. They ate like kings and concluded it didn't work.
[1] http://www.bbc.com/news/health-34798438
Bit of experience here: in the USA the FDA gets very upset about unsanctioned experimentation. There is a quasi-exception. The one drug I worked on that made it into trials had previously been tried on one human -- me. I already knew it wouldn't have tox implications -- we were more interested in the injection procedure itself. Plus we did some diffusion tests on my tissue which showed results consistent with the guinea pigs. The quasi-exception is that I was the CEO -- even the chief scientist couldn't do it since it could later be claimed I coerced her. But supposedly the CEO couldn't be coerced.
It's only a quasi-exception since you still have to report it in the back of your NDA (New Drug Application) and you're reporting a violation of procedure. But the NDA is enormous (used to be truckloads of documents, though now electronic), and by that time you have hundreds of human experiences with the drug. Obviously with an n of 1 it can't count as trial data, you're just writing up that once you were naughty.
(the whole process is weird -- technically, when you apply to test a new drug you're not actually asking permission, you are actually asking the Agency is not to prosecute you for using an unlicensed drug because you promise to be careful and that your plans fit into one of the exceptions written up in section 505(b) of the food and drug act.)