61 Phase 2 Clinical trial success: Association with statistical decision-making in preceding rodent experiments

Objectives/Goals: In neurological animal research, statistical misapplication may lead to overoptimism in a therapy’s potential for successful translation. This pilot study investigated whether human clinical trials that fail have higher prevalence of statistical misapplication in preceding animal experiments, compared to human trials that succeed. Methods/Study Population: Phase 2 clinical trials for 3 neurological conditions were identified on ClinicalTrials.gov and classified as successful or failed based on advancement to Phase 3 and/or preplanned efficacy test results. PRISMA guideline methods were used to systematically search for preclinical animal experiments (same indication and intervention) preceding the start of the human trial. Data were gathered from animal articles by collectors blinded to human trial outcome and included items describing reporting transparency, experimental design and sample sizes, and statistical tests applied. Statistical mistakes were coded based on mismatch between test and design. Rates of mistakes were compared between articles preceding successful and non-successful human trials using weighted point estimates and 95% confidence interval. Results/Anticipated Results: The final sample included 24 trials (16 successful) and 70 associated animal studies. Transparency was poor, with infrequent reporting of group allocation method (39%), sample sizes adequate to evaluate attrition ( Discussion/Significance of Impact: Statistical misapplication is common in animal research, and this pilot study has demonstrated that preclinical statistical mistakes may indeed occur more frequently prior to failed human trials. Mistakes and lack of transparency may lead to overoptimism in preclinical experimental findings, with consequences for subsequent human translation.