Data pooling in clinical studies: Statistical model and methods as well as our experience
*Shun Zhenming, Sanofi-Aventis 

Keywords: pooling, safety evaluation

This presentation will be composed of 4 parts: (1) Discussion on statistical hypothesis and models for demonstrating evidence in safety and efficacy where integrated or pooled evaluation can be considered. Our conclusion is that a pooling strategy could be more valid for safety evaluation; a pooled evaluation could be meaningful for efficacy if the statistical hypothesis is formulated correctly (Shun et al, 2005, Stat in Med). (2) A brief review of the existing methods for meta-analysis (3) Two real life examples: An example (Wei-Xiang Qi et al: Clinical Drug Investigation, 2014, and Eur J Clin Pharmacol, 2013) based on the safety data from a cancer drug and another one (Angela Webster et al, 2010) based on the efficacy data from a drug of induction therapy for prophylaxis against acute rejection in kidney transplant will be presented together with our evaluation on the validity and interpretation of the outcome: The issues of the pooled data summary in the example are selection bias in study selection, the nature of the studies that provided the source data, and interpretation of the findings. (4) Conclusion: We summarize our conclusion and recommendation.