Interpreting clinical trial outcomes marred by placebo response with an assessment of false-negative and true-negative clinical trials in depression using propensity-weighting

Abstract

Abstract The objective of this study was to evaluate the performances of the propensity score weighted (PSW) methodology in a post-hoc re-analysis of a failed and a negative RCTs in major depressive disorders. The conventional study designs, randomizations, and statistical approaches do not account for the baseline distribution of major non-specific prognostic and confounding factors such as the individual propensity to show a placebo effect (PE). Therefore, the conventional analysis approaches implicitly assume that the baseline PE is the same for all subjects in the trial even if this assumption is not supported by our knowledge on the impact of PE on the estimated treatment effect (TE). The consequence of this assumption is an inflation of false negative results (type II error) in presence of a high proportion of subjects with high PE and an inflation of false positive (type I error) in presence of a high proportion of subjects with low PE value. Differently from the conventional approach, the inverse of the estimated PE probability was used as weight in the mixed-effects analysis to assess TE in the PSW analysis. The results of this analysis indicated an enhanced signal of drug response in the failed trial (ClinicalTrials.gov NCT00584974) and confirmed the absence of drug effect in the negative trial (ClinicalTrials.gov NCT0058497). This approach can be considered as a reference approach that minimize the risk of inflating either type I or type II error in contrast to what happens in the analyses of RCT studies conducted with the conventional statistical methodology.