UC Develops New Approach to Web-Based Drug Withdrawal Warnings

CINCINNATI—The University of Cincinnati (UC) has developed a faster approach for informing consumers online when the Food and Drug Administration (FDA) withdraws a medication.

The new approach works with NetWellness.org, a commercial-free, consumer health Web site produced by Ohio's three medical research universities–UC, Case Western Reserve University and Ohio State University.

"Information about drug withdrawals may not reach patients quickly enough to prevent potentially dangerous side effects," explains Peter Embi, MD, lead author on the UC-based study published in the most recent online edition of the Journal of Medical Internet Research. "Given the public's growing use of the Web for health information, it's important that Web-based consumer health content is kept up to date, particularly that involving withdrawal of a potentially harmful medicine.

"However," says Embi, assistant professor of medicine and a researcher at UC's Institute for the Study of Health, "it's been shown that many sites don't update their content for days, or even weeks, following an FDA drug withdrawal.

"Our new approach," he says, "allows just one person to modify affected Web pages in less than an hour and within just hours of an FDA drug withdrawal announcement."

Embi says he hopes this new method will encourage those responsible for other Web sites carrying critical consumer health information to adopt a similar 24-hour response standard to drug withdrawals.

"There's evidence that patients continue to use medications for some time after their withdrawal, which occasionally causes harmful effects," Embi explains, "so it's important to inform the public as quickly as possible."

Before developing the new approach, Embi and his team evaluated NetWellness' previous update process following the FDA's withdrawal of the painkilling, anti-inflammatory drug Vioxx (rofecoxib) on Sept. 30, 2004.

Notified of the Vioxx withdrawal, NetWellness staff searched their site for the drug and related terms. References to Vioxx information were temporarily pulled and evaluated by medical experts. Content was either changed and reposted or permanently removed.

The researchers found that while this original process was ultimately effective, it took nearly three weeks and considerable human input to complete NetWellness content changes after the Vioxx withdrawal.

To improve on this, the research team developed their new approach and then tested it following FDA withdrawal of the anti-inflammatory drug Bextra (valdecoxib) on April 7, 2005. The results were significantly better than with the previous approach, they say. In just 18 hours after the drug recall, all instances of Bextra on NetWellness were updated and active, says Embi.

The new method, initiated by the FDA's automated MedWatch E-List alert and combined with modified technology and people processes, allows updates of all relevant NetWellness pages with minimal manual input and within 24 hours of a drug recall.

Now when the FDA announces a drug withdrawal, the NetWellness team receives an alert by MedWatch E-List and immediately checks the site's database for instances of the drug name, including trade and generic names. Instances are then hyperlinked using an automated find-and-replace function built into the site's content management system.

Simultaneously, pages containing these references receive a hyperlinked warning box indicating the availability of important information about the withdrawn drug.

All hyperlinks point directly to a new NetWellness "warning" page containing information about the FDA alert and to additional links to either the FDA or the drug manufacturer's Web site.

A recent study in the Journal of the American Medical Informatics Association reports that the majority of Web-based sources, including the popular consumer health site WebMD.com, took several days to update the Bextra drug information. A subscription-based health information Web site for health professionals, PDR.net, took 268 days to update Bextra content.

Embi plans to further automate this new update method and develop similar responses to other types of FDA health warnings relevant to consumers.

Study coauthors included Mark McCuistion, Charles Kishman, Doris Haag and Steven Marine, all from UC, and Prasad Acharya, from Wright State University.