New Data-Driven Approaches Would Improve How FDA Monitors Safety of Approved Drugs

Drug Safety

New research in the INFORMS journal Management Science finds that the U.S. Food and Drug Administration (FDA) could enhance public safety by employing new data-driven approaches to identify adverse effects of drugs faster and more reliably, which would improve the regulatory decision-making and oversight processes. This approach would substantially contrast with the current FDA approaches for monitoring drug safety, which suffers major drawbacks (e.g., they are based on voluntary reports that can result in inaccurate or untimely decisions).

The study applies this approach to a highly publicized FDA black box warning for a drug used to treat diabetes. In May 2007, the FDA issued a safety alert for the anti-diabetes medication rosiglitazone, followed by a black box warning soon thereafter. The alert was issued on the same day the New England Journal of Medicine (NEJM) published an article reporting that rosiglitazone use is associated with a 64% increase in the risk of cardiovascular (CV) mortality and 43% increase in the risk of heart attack. However, subsequent studies were skeptical of the FDA decision to issue a black box warning and questioned the validity of methods employed in the NEJM study.

“Using comprehensive data from the Veterans Health Administration on over 300,000 diabetes patients, we find that rosiglitazone was not associated with CV mortality or an increased risk of heart attack,” says Vishal Ahuja, one of the study authors, from Southern Methodist University. “This is a conclusion that the FDA also reached as it retracted the [black box warning] six years after issuing it.”

Ahuja, alongside fellow researchers Carlos Alvarez of Texas Tech University and John Birge and Chad Syverson of the University of Chicago, emphasizes that this paper demonstrates the importance of using real-world data to support regulatory decision-making. This is consistent with the mandate of the 21st Century Cures Act, passed by U.S. Congress in 2016.

“The use of robust analytical methods that rely on large, reliable, and relevant observational databases is critical to improving the timeliness and accuracy of FDA’s decisions and allow it to perform a proactive, holistic assessment of drugs,” concludes Ahuja.

Edited by Gary Cramer