In treating diseases of the central nervous system (CNS), only about 9% of the drugs that enter Phase I testing survive to launch, a low percentage compared with drugs in other therapeutic areas. Also, about 50% of the failed trials fail because they cannot demonstrate efficacy in Phase II, a 15% increase from the previous decade. These combined with a 50% failure rate of CNS Phase III trials create a high hurdle for conducting a successful positive CNS trial.

When a trial fails to demonstrate the study drug’s efficacy, high placebo response is commonly observed. High placebo response, when present, makes it harder to show treatment effects. This is supported by a recent report on schizophrenia trials, where placebo response is increasing at an alarming rate that signal detection becomes difficult. Notably, recent trends in schizophrenia trials are actually similar to trends previously observed in trials of major depressive disorder.

To tackle the problem of high placebo response, this session will provide a platform for statisticians involved in the CNS drug area to share insights and to discuss potential solutions that can be implemented in future trials.

3 key questions to be addressed: 1)What is the negative impact of high placebo response? Has the phenomenon of high placebo response been observed in recent psychiatric clinical trials? If so, what are the possible contributing factors?

2)What are the study design elements that can be implemented to potentially reduce placebo effects and thereby increase signal detection?

3)Coupled with improved study design(s), are there suitable statistical methods available for analyzing psychiatric trial data? Can these methods deal with extensive missing data frequently observed in psychiatric clinical trials?