Meta-analysis of Placebo Response in Randomized Clinical Trials of Antipsychotic Drugs Using PANSS Focusing on Different Approaches to the Handling of Missing Data.

ABSTRACT

BACKGROUND AND OBJECTIVE: Schizophrenia treatment has been shifting to resocialization by efficacious antipsychotic drugs. However, even some of the pivotal studies of approved new antipsychotic drugs with proven efficacy had failed due to high placebo response. The aim of this study was to identify the potential factors affecting placebo response by meta-analysis for randomized clinical trials for antipsychotic drugs using Positive and Negative Syndrome Scale (PANSS) focusing on the current methodological change in the handling of missing data [from last observation carried forward (LOCF) to mixed-effect models for repeated measures (MMRM)] for successful future clinical trials. METHODS: Recent trends in the degree of placebo response were investigated between publication year (1993 to 2016) and the mean change of PANSS total score in the placebo arm. The potential factors affecting the degree of placebo response, such as study design and operational factors, were analyzed separately by meta-regression for LOCF- and MMRM-based data. RESULTS: There was no correlation between publication year and the mean change of PANSS score in the placebo arm in schizophrenia studies of 10 years applying MMRM. The number of countries and treatment setting in MMRM-based data and study duration in LOCF-based data were significantly associated with placebo response. CONCLUSION: Placebo response in schizophrenia clinical trials published between 2007 and 2016 has not increased over time. Differences in the healthcare environment among countries were suggested to affect the evaluation of antipsychotic drugs. Further analyses on the potential factors of placebo response for MMRM-based data are required.