Skip to main content

Using High-Dimensional Propensity Scores to Automate Confounding Control in a Distributed Medical Product Safety Surveillance System

    Basic Details

    This article describes how confounding control is one aspect of the safety evaluation process that can be effectively automated to achieve the necessary speed and scale for the safety evaluation process at reasonable cost without sacrificing validity. The article discusses use of the high-dimensional propensity score (hd-PS) algorithm for automating confounding control in sequential database cohort studies.


    Jeremy A. Rassen ScD; Sebastian Schneeweiss MD, ScD 

    Corresponding Author

    J. A. Rassen, Department of Medicine, Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women's Hospital and Harvard Medical School, Boston, MA, USA. E-mail: