Last week at a meeting on clinical trials for central nervous system treatments, researchers convened to grapple with what they see as a serious and underrecognized problem in drug development: the fact that many people in clinical trials stop taking their experimental medications or don't take them as prescribed. These nonadherent participants can diminish the apparent effects of a drug compared with placebo and can potentially mask a drug's side effects or safety risks. As researchers recognize the effect of the problem on their trial data, some are taking greater pains to measure adherence through blood tests or electronic monitoring, and some are considering ways to design more adherence-friendly trials.