Association Rule Mining in the U.S. Vaccine Adverse Event Reporting System (VAERS)

• Activity Number: 176
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Computing
• Author(s): Lai Wei*+ and John Scott
• Companies: FDA and FDA/CBER/OBE
• Keywords: association rule discovery ; spontaneous reporting system ; safety ; signal detection ; vaccine

Abstract

Spontaneous adverse event reporting systems are critical tools for monitoring the safety of licensed medical products. We apply association rule mining (ARM) in the US Vaccine Adverse Event Reporting System (VAERS) for the detection of potential vaccine-adverse event combinations. ARM finds associations between one vaccine and one or more adverse events (i.e. a syndrome). This is in contrast to currently used signal detection algorithms, which are designed to find bivariate associations between a vaccine and a single adverse event. By detecting vaccine-syndrome associations in addition to vaccine-adverse event associations, our approach is sensitive to potentially complex signals, which may be particularly important when monitoring novel medical countermeasures products, such as pandemic influenza vaccines. We find that ARM can detect a variety of potential vaccine-adverse event signals efficiently. To reduce the number of redundant association rules found based on our algorithm, we propose a method for post-processing of association rules so that they can be used as a screening method to identify patterns that may need further investigation in the vaccine safety surveillance.