Nonparametric Method for Analyzing Recurrence of Adverse Events in Randomized Clinical Trials

• Activity Number: 657
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Masanori Ito and Misun Lee*
• Companies: Astellas Pharma Global Development and Astellas Pharma Global Development
• Keywords: Recurrent Adverse Events ; Mean Cumulative Function ; Kaplan-Meier ; Non-parametric ; Modified Log-rank Test

Abstract

Adverse events (AEs) from randomized clinical trials are often summarized as a crude incidence rate based on the percentage of patients who had at least one occurrence of AE or by an exposure-adjusted incidence rates. The AEs that are reported multiple times may be inadequately addressed by the crude summary. An approach with mean cumulative function (MCF) can assess multiple or reoccurring AEs to subjects. A method was proposed by taking the MCF of AEs which is a similar approach to a Cox model. Such approaches can assess the AE profiles between the treatment groups, however, the AE occurring days from either treatment groups are considered for both treatment groups based on the regression model. This may misrepresent the actual occurrence days of AEs. Furthermore, the assumption of proportionality between treatment groups is not often valid. Therefore, we propose a nonparametric MCF method as an additional approach in conjunction with the widely used crude incidence approach. We investigated the results from a randomized Phase 3 study and the results are presented. Throughout simulation, we further investigated s and graphical approaches are presented to compare the semi-parametric MCF method with our proposed nonparametric MCF method.