FDA Developing New Drug Safety Tools: Dal Pan

Share

FDA says it is developing new processes to help identify drug postmarket safety issues. For example, the agency’s Sentinel Initiative is using the FDA Sentinel Innovation Center (IC), to test “machine learning and natural language processing technologies to more effectively access and integrate data in clinical narratives for safety studies,” CDER Office of Surveillance and Epidemiology director Gerald J. Dal Pan says in an online post.

 “Through a series of methodological and infrastructure development projects, FDA has developed capabilities to conduct signal identification analyses in the Sentinel System to complement data from FAERS [FDA Adverse Event Reporting System],” Dal Pan writes. “The analytic packages and results from these projects are available on the Sentinel Initiative Web site.” The Sentinel System uses real-world data from administrative claims databases and electronic health records on medication usage and medical diagnoses involving over 300 million people in the U.S. to obtain evidence about adverse drug reactions.

Additionally, Dal Pan highlights CDER’s Newly Identified Safety Signal process, a postmarket safety initiative, which “allows for a standardized, interdisciplinary approach to systematically identify, evaluate, and address safety signals. The agency is also implementing natural language processing and machine learning methods to improve review and analysis of FAERS data and is exploring applying these techniques to the medical literature.” He points to an Information Visualization Platform (InfoViP) as an example of a decision-support software tool that will allow safety reviewers to “focus on evaluating complex data and detecting safety concerns.”

Read more