A Further Look Into Meta-Analysis of Rare Event Data for Drug Safety Assessment

• Activity Number: 32
• Type: Contributed
• Date/Time: Sunday, August 4, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yujun Wu*+ and Hui Quan and Peng-liang Zhao
• Companies: Sanofi US and Sanofi and Sanofi US
• Keywords: Meta analysis ; rare event ; safety assessment ; non-inferiority

Abstract

In drug development, meta-analysis has been increasingly used to evaluate the drug safety profile by combing results from multiple randomized trials, e.g., evaluation of cardiovascular risk for new antidiabetic therapies. For a binary outcome, when the event rate is low, zero events are often observed in either or both arms in some trials, and then the risk ratios or odds ratios for these trials can not be calculated. The common methods are to either exclude them from the meta-analysis or add a continuity correction factor to each cell of the corresponding 2x2 tables in the analysis. It is known that both approaches may result in biased results. We take a further look into the performances of different meta-analysis methods for rare event data via extensive simulations. Under various clinical trial settings and for different outcome measures (e.g., risk difference, risk ratio, and odds ratio), we investigate different types of continuity correction factors. We also examine methods for constructing confidence intervals and testing for non-inferiority for drug safety assessments and compared their performances.