357 – Statistical Methodologies in Consulting
Identifying the Potential Risk Factors of Venous Thromboembolism Events in Clinical Trials
Kao-Tai Tsai
Analyzing safety data and trying to identify the potential risk factors of an adverse event is a both interesting and extremely challenging task. Due to the correlations and possible interactions of patient's medical history, medications received before and after the trial, as well as other factors, it can make this a complex multivariate data analysis problem and the simple summary can be very misleading. In addition, subjects with or without the safety event are usually not readily comparable in most of the randomized trials. In this article, as a case study, we show the exploratory data analysis performed in identifying the risk factors for Venous Thrombus Embolism using data analytical techniques such as matching, mediation, and regressions. Data from recent clinical trials are used to demonstrate the statistical procedures utilized.