Dynamic Visualization Tools to Simplify the Analysis of Complex Clinical Drug Safety Data

• Activity Number: 440
• Type: Invited
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Ana Szarfman*+
• Companies: U.S. Food and Drug Administration
• Keywords: interactive graphic display ; drug safety ; automated analyses ; patient profiles ; drill down to source data ; data auditing

Abstract

Interactive, automated data visualization tools provide disciplined approaches to investigate and document potential drug safety problems within different data resources. These tools display the sequence of events for multiple patients and within single patients allowing the user to quickly switch from one modality to the other. They reveal the data peculiarities and missing data. To understand the global picture, rows of multi-patient data are aligned by start of treatment or other event and sorted by duration of treatment or time to an outcome. To understand individual patients, multiple data types are put on a common timeline, aligned by start time of a drug or an event and linked to narrative text. This display also shows the data points and the start and end dates of medications, events, indications, and procedures. The ability to "drill down" from a global display or from any result to the individual patient profile and source data increases the certainty about the results. With automation, multiple reviewers can use the same datasets and outputs, so that the analytical processes can be easily understood and reused, fostering communication between all parties.