Abstract:
|
Drug safety is a major concern in pre- and post-marketing settings. It is important to profile and predict adverse events to reduce attrition during development and post approval. The advent of digital clinical trials and electronic health records will lead to an increase in safety data, i.e., the big data realm (the 4 V's). Analyzing safety data based on frequency of AEs and less on quantitative approaches is neither sufficient nor optimal for characterization and prediction of patient safety in the big data setting. Big data concepts and machine learning have shown potential utility, e.g., these approaches can help identify clusters based on various features which facilitate discovery of associations between drugs and AEs. In this round table discussion, the utility of data science, big data ideas, and machine learning in drug safety will be considered in the pre- and post-marketing settings focusing on the following: experiences using these approaches in analyzing safety data, advantages they offer relative to other methods, their routine use, and those best suited to address safety questions. The discussion will also highlight relevant methodological approaches.
|