Details
Description
With growing access to linked claims and EHR data, pharmacoepidemiology faces new opportunities—and ongoing challenges like unmeasured confounding in real-world studies. This presentation showcases methods from the FDA Sentinel Innovation Center to address confounding in observational research. Using data from over 25 million individuals in the Sentinel Real-World Evidence Data Enterprise (RWE-DE), it highlights targeted learning, doubly-robust estimation, plasmode bias analysis, and automated negative controls. It was presented at the 2025 ISPE Annual Meeting.
Additional Information
Contributors
Presenter(s)
Xu Shi