Back to Agenda
Session 11: Signal Detection and Evaluation
Session Chair(s)
James Buchanan, PharmD
President
Covilance LLC, United States
New methods of detecting safety signals are evolving. This session will explore novel ways in which safety signals may be identified within social media postings as well as the use of large language models to detect adverse events within clinical notes from electronic health records. The results from a horizon scan will be summarized concerning the use of AI tools and natural machine learning for safety surveillance in the absence of a REMS as well as the use of real-world data (RWD) by Sentinel for signal detection.
Learning Objective : - Understand the application of disproportionate reporting rate algorithms to the evaluation of social media sources of adverse event information
- Describe how large language models can be applied to detecting adverse events from electronic health information
- Discuss how Sentinel utilizes real-world data for signal detection
- Explain how AI and natural machine learning tools can assist in safety surveillance in the absence of a REMS
Speaker(s)
Social media safety monitoring: Emergence of a new global tool for early signal detection? Experience from the COVID-19 vaccine
Patrick M. Caubel, MD, PhD, MBA
Pfizer Inc, United States
Chief Safety Officer
Automated Adverse Event Detection from Electronic Health Records using Large Language Models
Vivek Rudrapatna, MD, PhD
UC San Francisco, United States
Assistant Professor; Co-Director, Center for Real-World Evidence
A Horizon Scan Perspective on the use of AI and Real-World Data for Signal Detection
Ariela G Chick, MPH
Perspective PV, United States
Senior Strategic PV Advisor
Signal Detection for Clinical Trial Design: Finding the Right Questions
Peg Fletcher, MD, PhD
MedAssessment, Inc., United States
President
Have an account?