Placebo-controlled (PC) Phase 3 trials are critical for the registration of new atypical antipsychotic medications (AAP) for schizophrenia but use of placebo when efficacious treatments exist has been questioned.

To investigate evidence for the use of placebo in clinical trials of schizophrenia via a meta-analysis of large, PC trials of new AAPs.

Using the FDA Summary Basis of Approval reports, we examined outcome data from all Phase 3 clinical trials that evaluated investigational AAPs. Publications from peer-reviewed literature were also identified. The main outcome variables were: symptom improvement in individual treatment arms, clinical response, therapeutic failure.

Meta-regression indicated a highly significant difference between improvement in the placebo and the active arm (p<0.0001). Effect size (ES) estimate for the placebo arm revealed that patients in this arm obtained a statistically significant but clinically negligible symptom reduction (Cohen d: ~0.15; p<0.004) while active-treated patients displayed a substantial symptom reduction (Cohen d: ~0.70; p<0.0001). Active treatments showed a highly significant (p<0.001) superiority vs. placebo in clinical response and therapeutic failure, with failure rates often exceeding the rate of clinical response. ESs for change varied substantially across trials, with an ES range of d=0.8 for the placebo and the active arms, respectively.

Variable ESs across studies support the view that placebo control has major importance in trials of new AAPs. However, efforts should focus on finding design alternatives and to minimize the risks of PC trials so that they may be conducted in ethically acceptable manner.