Quality assurance for artificial intelligence algorithms applied to pharmacovigilance

CERSI Collaborators: Taziarchis Botsis, MS, MPS, PhD; Hao Wang, PhD; Kory Kreimeyer, MS

FDA Collaborators: Robert Ball, MD, MPH, ScM; Oanh Dang, PharmD, BCPS; Suranjan De, MS, MBA

Project Start Date: June 18, 2024

Regulatory Science Challenge:

Pharmacovigilance increasingly uses Artificial Intelligence to automate specific tasks that humans cannot complete easily. However, most applications do not reach the level of performance necessary for using Artificial Intelligence without human intervention. It is, therefore, essential to accurately measure this performance using optimal quality assurance methods. This project will review relevant literature and collect information about methods that may sufficiently support the evaluation of Artificial Intelligence in Pharmacovigilance.

Project Description and Aims/Goals:

The primary objective of this projects is to conduct a literature review on methods that could support the evaluation of Artificial Intelligence used in Pharmacovigilance. Investigators will collect and analyze several publications to collect all relevant information and include it in a technical report. This report will document the methods and findings from this review.

Anticipated Outcomes/Impact:

Investigators will collect information about strategies that may assist Pharmacovigilance experts in evaluating the use of Artificial Intelligence in this area. It will disseminate this scientific knowledge to the United States Food and Drug Administration by delivering a final report on the project findings.